ie dvenmt a) re i . "WEE My Fis ; park, A hae g ” Vie) l ; y Ae yA atihs va ! re i | NL i} ray ? hy Big ; uy tilt i @ ul ISSN 1347-0558 PGre 2 ee Pap . ads 2 Ps bad BW SE Cagek UR? EQS 22 > 7 5 4, Ij ¥ Ml BE ee ay Ok) ’ BYE 3 Se SA ‘, , ‘di i AU 3 ikl ORNITHOLOGICAL SCIENCE Vol.1 No.1 January 2002 _ The Ornithological Society of Japan ORNITHOLOGICAL SCIENCE Official journal of the Ornithological Society of Japan Editor-in-Chief Keisuke Ueda, Rikkyo University, Tokyo Associate Editors Teruaki Hino, Forestry and Forest Products Research Insti- tute, Kyoto Hidetsugu Sakai, Nihon University, Tokyo Editorial Board Masahiko Nakamura, Joetsu University of Education, Joetsu Isao Nishiumi, National Science Museum, Tokyo Kazuo Okanoya, Chiba University, Chiba Eiichiro Urano, Kyoto University, Kyoto Advisory Board Alexander V. Andreey, Institute of Biological Problems, Magadan Walter J. Bock, Columbia University, New York Jiro Kikkawa, The University of Queensland, Brisbane Woo-Shin Lee, Seoul National University, Suwon Bernd Leisler, Max-Planck-Gesellschaft, Radolfzell Anders P. Moller, Universite Pierre et Marie Curie, Paris Richard Noske, Northern Territory University, Casuarina Pilai Poonswad, Mahidol University, Bangkok Lucia Liu Severinghaus, Academia Sinica, Taipei Navjot S. Sodhi, National University of Singapore, Singapore Jeffrey R. Walters, Virginia Polytechnic Institute and State University, Blacksburg John C. Wingfield, University of Washington, Seattle Jeong-Chil Yoo, Kyung-Hee University, Seoul Editorial Policy Ornithological Science publishes reviews, original articles, short communications and comments covering all aspects of ornithology. Manuscripts are judged on the basis of their con- tribution of original data and ideas or interpretation. All arti- cles are peer-reviewed by at least two researchers expert in the field of the submitted paper. Manuscripts are edited where necessary for clarity and economy. Ornithological Science aims to publish as rapidly as is con- sistent with the requirements of peer-review and normal pub- lishing constraints. Submission Manuscripts and editorial correspondence should be ad- dressed to: Dr. Teruaki Hino, Associate Editor, Forestry and Forest Products Research Institute, Kyoto 612-0855, Japan. Tel: +81-75-611-1201, Fax: +81-75-611-1207, e-mail: tkpk@affre.go.jp For detailed instructions concerning the submission of man- uscripts, please refer to the Instructions to Authors at the in- side of the back cover of each issue of the journal or visit the OS] web page: http://wwwsoc.nii.ac.jp/osj/ Subscriptions Membership of the Ornithological Society of Japan (OSJ) is open to anyone interested in ornithology and the aims of the OSJ. The annual fee for an ordinary member is Japanese Yen 5,000. Members are entitled to receive two journals. Ornitho- logical Science (in English; two issues per year) and Japanese Journal of Ornithology (with English summary; two issues per year). To join the OSJ, please apply to: The OSJ Office, Laboratory of Biodiversity Science, School of Agricultural and Life Sciences, The University of Tokyo, Yayoi 1-1-1, Tokyo 113-8657, Japan. Tel: +81-3-5841-7541, Fax: +81-3-5841-8192, e-mail: osj@lagopus.com Copyright Submission of a manuscript implies: that the work de- scribed has not been published before; that is not under con- sideration for publication elsewhere; that its publication has been approved by all co- authors, if any, as well as by the re- sponsible authorities at the institute where the work has been carried out; that, if and when the manuscript is accepted for publication, the authors agree to automatic transfer of the copyright to the OSJ; and that the manuscript will not be pub- lished elsewhere in any language without the consent of the copyright holders. ISSN: 1347-0558. All articles published in the Ornithological Science are pro- tected by copyright, which covers the exclusive rights to re- produce and distribute the article (e.g. as offprints) as well as all translation rights. No material published in this journal may be reproduced photo-graphically or stored on microfilm, in electronic data bases, video disks, etc., without first obtain- ing written permission from the copyright holders. Special re- quests should be addressed to: Dr. Keisuke Ueda, Editor- in-Chief, Laboratory of Animal Ecology Faculty of Science, Rikkyo University, Ikebukuro, Tokyo 171- 0021, Japan, Tel: +81-3-3985-2596, Fax: +81-3- 3985-2596, e-mail: keisuke@rikkyo.ac.jp Printer: Kokusai Bunken Insatsusha Co., Ltd., Takada-no-baba 3-8-8, Shinjuku-ku, Tokyo 169-0075, Japan. Tel: +81-3-3362-9741, Fax: +81-3-3368-2822. Cover design: Eiichiro Urano PREFACE Lom From Japan through Asia to the world: ~ building bridges in ornithological science On behalf of the Ornithological Society of Japan, I would like to express my great pleasure at the publica- tion of this first edition of Ornithological Science. This is an international peer-reviewed journal that will be published twice yearly. The purposes of the journal are to promote ornithological studies in Japan and other Asian countries, and to facilitate the exchange of information between ornithological scientists working both inside and outside Asia. In Japan, China, Korea, and many other Asian countries, both the amount and importance of ornithological research are increasing. Research in Asia encompasses areas of ornithological science that have emerged rela- tively recently, including the behavioral ecology, conservation biology and molecular phylogenetics of birds, as well as more traditional fields such as bird distribution, population and community ecology and dynamics, and physiology. However, often the results of ornithological research carried out in Asia have been published in domestic journals solely in Asian languages. Although the work may be of high quality and great impor- tance, language barriers and difficulty acquiring published materials can prevent foreign scientists accessing and understanding the research done by their colleagues. We would like to help remove these difficulties with the publication of Ornithological Science. In Asia, there are so many interesting birds and subjects for research. Although we know that the amount of research being done in this region is increasing, we believe that many interesting and important studies of Asian avifauna remain unpublished. As a venue for the publication of original ornithological research, Or- nithological Science is open to all scientists, and there are advantages to publishing within its pages, as de- scribed on the reverse of the front cover and in the editorial of this edition. The publication and dissemination of interesting work encourages further study, for example, by scientists extending research to other species and groups of birds. We hope that Ornithological Science will facilitate the development of ornithological research in this way. As well as publishing original research, the editors of Ornithological Science would like to encourage scien- | tists to contribute review papers on particular topics, to promote the understanding of different fields of re- search. Examples of topics for review could include the origin and evolution of birds endemic to Asia, mate se- | lection and social structure of Asian birds e.g. babblers and pheasants, the present status and conservation of | endangered birds in Asia, the use of molecular data for ecological bird studies, the use of GIS (Geographic In- _ formation Systems) for habitat and distribution analyses, and modern technologies used in migration studies. | Producing reviews in a common language, based on previous publications in scientists’ native languages, , should give a better perspective on the state of research in various fields of Asian ornithology. The avifaunas of different countries are, or used to be, connected to each other. Bird species of one area | were differentiated from those occupying neighboring areas. Migratory birds move over a number of different | countries. Therefore, in order to study the origin, distribution, ecology, migration and conservation of birds, in- | ternational cooperation and collaboration are very important. During the last ten years, I have enjoyed collabo- | rating with Korean, Chinese, Indian, Russian, and American ornithologists to satellite-track the migration of | cranes, swans and storks. We have obtained an enormous amount of valuable data on the migration routes, mi- | gration patterns through time, relative importance of resting sites, and habitat characteristics of important sites. | Our studies have resulted in various conservation activities relevant to these birds and their habitats. Through my involvement with this research, I realized that international cooperation and collaboration are extremely important. My colleagues and I have already published many papers in various academic journals, but this at- | tractive new ornithological journal is appealing to us as a publication venue, and we are looking forward to | presenting our work within its pages. I would like to encourage collaborative research groups, e.g. Asian re- | search groups studying geese, swans, egrets and raptors, to publish their results here. Publishing the results of joint work in international journals like Ornithological Science should help broaden the scope of our research activities, and increase the depth of our friendships. This year, the Ornithological Society of Japan will celebrate its 90th anniversary. We now have about 1,200 members, and this number is increasing. Most members of the Society are Japanese. However, we encourage new members from many countries to join, particularly ornithological scientists working in Asian countries. We would like to broaden our research horizons, while forming as many friendships and collaborative working relationships as possible. With the publication of Ornithological Science, the Ornithological Society of Japan will enter into a new generation of its history. Ornithology in Japan, and ornithologists working there, will ben- efit from their area of interest and expertise becoming more open. This will facilitate an increasing level of connection with other ornithological scientists in Asia and around the world. Japan is a beautiful country with a diverse landscape including forests, mountains and islands, and has a lot of interesting areas for ornithologi- cal studies. It is my great pleasure to welcome visitors to our country to study birds, and also to enjoy birds through subscribing to Ornithological Science. I sincerely hope that Ornithological Science will help build bridges connecting ornithological scientists from different countries. January 4, 2002 Hiroyoshi HIGUCHI President of the Ornithological Society of Japan i) EDITORIAL “Ornithological Science”, the new English publication from Japan Ornithological Science has evolved from the Japanese Journal of Ornithology, which was composed of both English and Japanese papers and was published four issues; numbers 50 volumes last year. Twenty-five years ago, when I joined the society, many papers were written in Japanese. However, in recent years most of them are written in English. This marked a great change in our society and an outstanding growth in a new generation of Japanese ornithologists. We felt now was a good opportunity to divide the journal into two parts: English and Japanese. This is the first issue of the English addition. Ornithological Science will be published twice a year. It includes special features, original articles, reviews, short communications, and commentaries on all aspects of ornithology. All papers will be critically reviewed by at least two referees. If accepted, submissions will be published in a high quality format without a long delay. I hope that this new journal will have be of international caliber and will contain high quality papers of ornithology from not only Japan but also other Eastern and/or South-Eastern Asian nations. Of course, we would be glad to receive papers from other parts of the world. In keeping with our aims, we have invited powerful Advisory Board members from Asian, European, Australian, and North American countries listed on the back cover. We welcome your support as authors, readers, referees, and subscribers. Keisuke UEDA Editor-in-Chief i nih of aur fremieitinan eweor i hohe x sen of da aya will qalptuestie fee witty anyeiveynaty i mw have a : ny is Tye "89 ovte, Were ever, oF A i REN OH RT YATE Me = i Tin ur anon of Onuitholitgews Sipe Tee Chriitetegical 5 7 cin ene rite? iftiis é rAlhw ; PASsniry “sriniiiy} ywoay vi Deedes aa acs, dep dreeertnigu qey pare lenin, a nee cats AM tts Py but wad OVAL | Oi ae! pramueleee 2 cago): oatmeal) uo heateikquyy ne ip wey ak +9) OO That need oye uth io-eriepel josie ape Sei aoe yA elo Ohara; enrbAn ene mmcbre viteuoncnds 800 m) where winter con- ditions would not support the resident species. Hence the migratory habit allows segregation by habitat, both in summer and winter, despite a large overlap where there is aggressive competition for space in the _ breeding season. 10) Interspecific associations Many species form mixed-species flocks in winter and benefit in various ways such as early warning of predators and improved ability to find patchy re- sources (Bell 1985; Ford 1989). There may be more advantages in joining a mixed-species flock than a single-species flock, because ecological differences (as discussed above) will tend to reduce the total level of competition. There may also be direct bene- fits. Four examples can be cited from south-eastern Australian forests, based on well known but mainly anecdotal information. Firstly, Superb Lyrebirds Menura novaehollandiae forage by scratching vigor- ously at the forest floor, helping maintain understorey structure and regeneration (Ashton & Bassett 1997). In the process they displace more arthropods than they can catch or consume themselves. Lyrebirds are often followed by loose groups of other insectivores such as Eastern Yellow Robins Eopsaltria australis, White-browed Scrubwrens Sericornis frontalis and Pilotbirds Pycnoptilus floccosus that take advantage of this revealed resource (Higgins et al. 2001). Sec- ondly, many Grey Fantails migrate from foothill forests for the winter when flying insects are scarce, but those that remain spend substantial amounts of time following bark-foraging birds and catching winged insects displaced while those species forage behind loose bark. The main bark foragers concerned are White-throated Treecreepers Cormobates leu- cophaea, Red-browed Treecreepers Climacteris ery- throps, Crested Shriketits Falcunculus frontatus and flocks of Varied Sittellas Daphoenositta varia. Both Grey Fantails and White-throated Treecreepers join mixed feeding flocks in winter (Bell 1985; Ford 1989). Thirdly, Willie Wagtails often concentrate their winter foraging round large mammals (kanga- roos Macropus spp. or domestic stock), presumably benefiting from concentrations of insects near fresh warm dung. Fourthly, honeyeaters and other birds may congregate at fresh wounds in trees where mar- supial possums or gliders (especially Yellow-bellied Gliders Petaurus australis) have made scars to ex- tract exudates (Russell 1981; Loyn 1985b and unpub- lished; Chapman et al. 1999). These examples show how birds may benefit from the activities of unrelated birds and mammals in their environment. DISCUSSION The patterns of segregation revealed in this review resemble those described by Lack (1971) for conti- nental avifaunas in various parts of the world. He identified habitat differences as much the commonest means of ecological isolation in continental passer- ines, and this is confirmed for the distinct group of birds inhabiting forests of south-eastern Australia. Differences of this sort allow multiple species to co- exist broadly in an area, but not to share habitats at the fine scale. Differences in foraging stratum or sub- strate were found to be important in allowing some species pairs to share the same habitat at the same time, increasing the diversity of those habitats. The species diversity of a given habitat is expected to be a function of its structural and floristic complexity, and those characteristics will set limits on the extent to which bird species can co-exist. Tilman (1982) pre- sented a set of theoretical models for predicting out- comes of competitive exclusion and co-existence among organisms that may be useful in further inter- pretation of continental data on Australian birds. The patterns identified differ in one respect from those described elsewhere, and that relates to the ag- R. H. LOYN gressive honeyeater species that form interspecific territories (Dow 1977; Loyn et al. 1983; Loyn 1987a, b; Clarke 1995; Clarke & Schedvin 1997; Grey et al. 1997, 1998; Catterall et al. 1998 and in press). Ford (1989) discussed these species in relation to their communal breeding behaviour, and suggested that predator avoidance may have been a key driver for the evolution of communal breeding in these birds, many of which inhabit open and exposed woodland environments. A further reason is suggested by the experiments where Bell Miners were removed and other birds decimated their previously protected food supply: this showed that the level of resources needed to maintain the colony would not exist unless it was protected by an adequate number of birds within the group. Conversely, young birds would face great dif- ficulties in establishing new territories without being part of a large enough group to maintain food sup- plies through territorial defence. A parallel situation has been reported for White-winged Choughs Corco- rax melanoleucos, where young birds are encouraged to remain within the group (by kidnapping if neces- sary) in order to deter destructive attacks on the nest by other groups, in this case by conspecifics (Hein- sohn 1987). Many factors contribute to evolution of communal breeding, but the ecological advantages (or necessities) of group living should be given due weight among them. Elsewhere in the world, various bird species appear to live as aggressive groups in temperate woodland habitats, and further work may show that communal defence of resources is an im- portant factor in evolution of such systems. This paper has focused mainly on ecological segre- gation between congeneric species, but the cases of interspecific aggression highlight the fact that compe- tition occurs between all species, and can help deter- mine the nature and health of the ecosystem. Wood- land birds have declined in southern Australia (Robinson 1993; Barrett et al. 1994), and are vulnera- ble to competition from Noisy Miners in small grazed patches of forest (Loyn 1987a; Grey et al. 1997, 1998; Bennett 1999): active habitat management is needed to reverse such declines. A full analysis of competitive forces and the way they shape communi- ties (Cody 1974; Kikkawa & Anderson 1984) must consider the full range of competing species. Species- based approaches will continue to be useful to con- servation managers (e.g. Lambeck 1997; Loyn et al. 2001) but further understanding of ecological interac- tions will help progress to holistic ecosystem man- agement. ACKNOWLEDGMENTS This paper is based on experience from a large num- ber of studies, and many colleagues have assisted in those studies and contributed ideas at various times. Specific information for this paper was provided by An- drew Bennett, Carla Catterall, Mike Clarke, Rohan Clarke, Richard Hill, Ed McNabb and Ralph MacNally. Kym Saunders, Phoebe Macak and Carol Harris helped compile data and references. David Choquenot, Mike Clarke, Jiro Kikkawa and an anonymous referee pro- vided useful comments on a draft. I am grateful to all, and to Teruaki Hino for inviting me to prepare this paper. REFERENCES Ashton DH & Bassett OD (1997) The effects of forag- ing by the superb lyrebird (Menura novaehollandiae) in Eucalyptus regnans forests at Beenak, Victoria. Aust J Ecol 22: 383-394. Aumann T (1988) The diet of the brown goshawk, Ac- cipiter fasciatus, 1n south-eastern Australia. Aust Wildl Res 15: 587-594. Baker-Gabb DJ (1984) The feeding behaviour and ecol- ogy of seven species of raptor overwintering in coastal Victoria. Aust Wildl Res 11: 145-160. Barrett GW, Ford HA & Recher HF (1994) Conserva- tion of woodland birds in a fragmented rural land- scape. Pac Conserv Biol 1: 245-256. Bell HL (1985) The social organisation and foraging be- haviour of three syntopic thornbills (Acanthiza spp.). In: Keast A, Recher HF, Ford HA & Saunders D (eds) Birds of the eucalypt forests and woodlands: ecology, conservation and management. pp 151-163. Surrey- Beatty, Sydney. Bennett AF (1999) Linkages in the landscape: the role of corridors and connectivity in wildlife conservation. IUCN, Gland. Bennett AF, Brown G, Lumsden L, Hespe D, Krasna S & Silins J (1998) Fragments for the future. Depart- ment of Natural Resources & Environment, Mel- bourne. Blakers M, Davies SJJF & Reilly PN (1984) The atlas of Australian birds. Melbourne University Press, Melbourne. Catterall CP, Piper S & Goodall K (in press) Noisy miner irruptions associated with land use by humans in south east Queensland: causes, effects and manage- ment implications. Proc R Soc Queens]. Catterall CP, Kingston MB & Park K (1997) Use of remnant forest habitat by birds during winter in sub- tropical Australia: patterns and processes. Pac Con- Ecological segregation in SE Australia serv Biol 3: 262-74. Catterall CP, Kingston MB, Park K & Sewell S (1998) Effects of clearing lowland eucalypt forests on a re- gional bird assemblage. Biol Conserv 84: 65-81. Chapman A, Bradford MG & Hoskin CJ (1999) Sap suckers: a novel bird ‘guild’ in wet sclerophyll forests of tropical north Queensland. Emu 99: 69-72. Chesterfield EA, Loyn RH & Macfarlane MA (1984) Flora and fauna of Barmah state forest and their management. Forests Commission Victoria, Research Branch Report 240, Melbourne. Choquenot D, Krebs CJ, Sinclair ARE, Boonstra R & Boutin S (2001) Vertebrate community structure in the boreal forest. In: Krebs CJ, Boutin S & Boonstra R (eds) Ecosystem dynamics of the boreal forest. pp 437-462. Oxford University Press, Oxford. Christidis L & Boles WE (1994) The taxonomy and species of birds of Australia and its territories. Royal Australasian Ornithologists Union Monograph 2. RAOU, Melbourne. Clarke MF (1995) Co-operative breeding in Aus- tralasian birds: a review of hypotheses and evidence. Corella 19: 73-90. Clarke MF (1997) A review of studies of the breeding biology of Australian birds from 1986-95: biases and consequences. Emu 97: 283-89. | Clarke RH & Clarke MF (1999) The social organisation | of a sexually dimorphic honeyeater: the crescent honeyeater Phylidonyris pyrrhoptera, at Wilsons _ Promontory, Victoria. Aust J Ecol 24: 644-654. | Clarke MF & Schedvin N (1997) An experimental study _ of the translocation of noisy miners Manorina melanocephala and difficulties associated with dis- persal. Biol Conserv 80: 161-67. | Clarke MF, Griffioen P & Loyn RH (1999) Where do all | _ the bush birds go? Wingspan Suppl 17: 1-16. / Cody ML (1974) Competition and the structure of bird / communities. Princeton University Press, New Jersey. Collins BG & Briffa P (1982) Seasonal variation of abundance and foraging of three species of Australian honeyeaters. Aust Wildl Res 9: 557-569. Cooper RM & McAllan IAW (1995) The Birds of West- | ern New South Wales: A Preliminary Atlas. New South Wales Bird Atlassers Inc., Albury, Australia. Czechura GV (1985) The raptors of the Blackall- Conondale Ranges and adjoining lowlands, south- eastern Queensland. Corella 9: 37-44. Darwin C (1859) On the origin of species by means of natural selection, or the preservation of favoured races in the struggle for life. J. Murray, London. Reprint edition: Avenel Books, New York. ' Diamond JM (1981) The continuum of insularity: the relevance of equilibrium theory to the conservation of ecological islands. Afr J Ecol 19: 209-212. Dow DD (1977) Indiscriminate interspecific aggression leading to almost sole occupancy of space by a singe species of bird. Emu 77: 115-21. Dow DD (1978) Breeding biology and development of the young of Manorina melanocephala, a commu- nally breeding honeyeater. Emu 78: 207-222. Egan KH (1997) Seasonal changes in the pollen sam- pled from nectarivorous birds visiting an open forest at Menai, New South Wales. Corella 21: 83-87. Emison WB, Beardsell CM, Norman FI & Loyn RH (1987) Atlas of Victorian birds. CFL/RAOU, Mel- bourne. Ford HA (1985) A synthesis of the foraging ecology and behaviour of birds in eucalypt forests and wood- lands. In: Keast A., Recher HF, Ford HA & Saunders D (eds) Birds of the eucalypt forests and woodlands: ecology, conservation and management. pp 249-254. Surrey-Beatty, Sydney. Ford HA (1989) Ecology of Birds: An Australian Per- spective. Surrey Beatty & Sons, Sydney. Ford HA & Paton DC (1976) Resource partitioning and competition in honeyeaters of the genus Meliphaga. Aust J Ecol 1: 281-287. Ford HA & Paton DC (1977) The comparative ecology of ten species of honeyeaters in South Australia. Aust J Ecol 2: 399-407. Ford HA & Paton DC (1982) Partitioning of nectar sources in an Australian honeyeater community. Aust J Ecol 7: 149-159. Garnett ST & Crowley GM (2000) The action plan for Australian birds. Environment Australia, Canberra. Government of Victoria (1997) Victoria’s Biodiversity. Department of Natural Resources & Environment, Melbourne. Grey MJ, Clarke MF & Loyn RH (1997) Initial changes in the avian communities of remnant eucalypt wood- lands following a reduction in the abundance of noisy miners Manorina melanocephala. Wildl Res 24: 631-648. Grey MJ, Clarke MF & Loyn RH (1998) Influence of the noisy miner Manorina melanocephala on avian diversity and abundance in remnant Grey Box wood- land. Pac Conserv Biol 4: 55-69. Griffioen P (2001) Temporal changes in the distribution of birds in eastern Australia. PhD Thesis, La Trobe University, Melbourne. Groves RH (1981) Australian Vegetation. Cambridge University Press, Cambridge. Heinsohn RG (1987) Age-dependent vigilance in winter congregations of co-operatively breeding white- winged choughs Corcorax melanorhamphus. Behav Ecol Sociobiol 20: 303-306. R. H. LOYN Higgins PJ (1999) Handbook of Australian, New Zealand and Antarctic birds. Vol 4. Oxford Univer- sity Press, Melbourne. Higgins PJ & Davies SJJF (1996) Handbook of Aus- tralian, New Zealand and Antarctic Birds. Vol 3. Ox- ford University Press, Melbourne. Higgins PJ, Peter JM and Steele WK (eds) (2001). Handbook of Australian, New Zealand and Antarctic birds. Vol 5. Oxford University Press, Melbourne. Hindwood KA (1956) The migration of white-naped and yellow-faced Honeyeaters. Emu 56: 421-S. Keast A (1957) Variation and speciation in the genus Climacteris Temminck (Aves: Sittidae) Aust J Zool 5: 474-495. Keast A (1968) Seasonal movements in the Australian honeyeaters (Meliphagidae) and their ecological sig- nificance. Emu 67: 159-210. Keast A (1985) Bird community structure in southern forests and northern woodlands: A comparison. In: Keast A, Recher HF, Ford HA & Saunders D (eds) Birds of the eucalypt forests and woodlands: ecology, conservation and management. pp 97-116. Surrey- Beatty, Sydney. Keast A, Recher HF, Ford H & Saunders D (1985) Birds of eucalypt forests and woodlands: ecology, conser- vation and management. Surrey Beatty & Sons Pty Ltd, Sydney. Kikkawa J (1968) Ecological association of bird species and habitats in eastern Australia, similarity analysis. J Anim Ecol 36: 143-166 Kikkawa J & Anderson DJ (1984) Community ecology: pattern and process. Blackwell Scientific Publica- tions, Oxford. Kikkawa J & Pearse K (1969) Geographical distribution of land birds in Australia—a numerical analysis. Aust J Zool 17: 821-840. Lack D (1971) Ecological isolation in birds. Blackwell Scientific Publications, Oxford and Edinburgh. Lambeck RJ (1997) Focal species: a multi-species um- brella for conservation. Conserv Biol 11: 849-856. Low T (1994) Invasion of the savage honeyeaters. Aust Nat Hist 24: 26-33. Loyn RH (1980) Bird populations in a mixed eucalypt forest used for production of wood in Gippsland, Vic- toria. Emu 80: 145-56 Loyn RH (1985a) Ecology distribution and density of birds in Victorian forests. In: Keast A, Recher HF, Ford HA & Saunders D (eds) Birds of the eucalypt forests and woodlands: ecology, conservation and management. pp 33-46. Surrey-Beatty, Sydney. Loyn RH (1985b) Bird populations in successional forests of mountain ash Eucalyptus regnans. Emu 85: 213-230. 20 Loyn RH (1987a) Effects of patch area and habitat on bird abundances, species numbers and tree health in fragmented Victorian forests. In: Saunders DA, Arnold GW, Burbidge AA & Hopkins AJM (eds) Na- ture conservation: the role of remnants of native veg- etation. pp 65-77. Surrey Beatty & Sons, Sydney. Loyn RH (1987b) The bird that farms the dell. Nat Hist 96: 54-60. Loyn RH (1995) Bell Miners and the farming hypothe- sis—a comment. Emu 95: 145-146. Loyn RH (1998) Birds in patches of old-growth ash for- est, in a matrix of younger forest. Pac Conserv Biol 4: 111-121. Loyn RH, McNabb EG, Volodina L & Willig R (2001) Modelling landscape distributions of large forest owls as applied to managing forests in north-east Victoria, Australia. Biol Conserv 97: 361-376. Loyn RH, Runnalls RG, Forward GY & Tyers J (1983) Territorial bell miners and other birds affecting popu- lations of insect prey. Science 221: 1411-13. MacNally RC (1995) On large-scale dynamics and com- munity structure in forest birds: lessons from eucalypt forests in south-eastern Australia. Philos Trans R Soc London B 350: 369-379. MacNally RC (1996) A winter’s tale: among-year varia- tion in bird community structure in a south-eastern Australian forest. Aust J Ecol 21: 280-91. MacNally R & McGoldrick JM (1997) Landscape dy- namics of bird communities in relation to mass flow- ering in some eucalypt forests of central Victoria, Australia. J Avian Biol 28: 171-183 Marchant S & Higgins PJ (1990) Handbook of Aus- tralian, New Zealand and Antarctic birds. Vol 1. Ox- ford University Press, Melbourne. Marchant S & Higgins PJ (1993) Handbook of Aus- tralian, New Zealand and Antarctic birds. Vol 2. Ox- ford University Press, Melbourne. May RM (1978) The evolution of ecological systems. Sci Am 239: 160-175. McFarland DC (1986) The organisation of a honeyeater community in an unpredictable environment. Aust J Ecol 11: 107-120. McFarland D (1988) The composition, microhabitat use and response to fire of the avifauna of subtropical heathlands in Cooloola National Park, Queensland. Emu 88: 249-257. McLaughlin J (1990) Surveys and observations of the black-eared miner Manorina melanotis in Victoria 1989-90. RAOU Report 71. Menkhorst PW, Smales I & Quin B (1999) Helmeted honeyeater recovery plan 1999-2003. Dept. Natural Resources & Environment, Melbourne. Menkhorst PW & Davies JN (1983) The status and dis- Ecological segregation in SE Australia tribution in Victoria of the grey-fronted honeyeater Lichenostomus plumulus. Aust Bird Watch 10: 107-— Ui: Munro U & Munro JA (1998) Migratory restlessness in the yellow-faced honeyeater Lichenostomus chrysops (Meliphagidae), an Australian diurnal migrant. Ibis 140: 599-604. Munro U & Wiltschko W (1992) Orientation studies on yellow-faced honeyeaters Lichenostomus chrysops (Meliphagidea) during autumn migration. Emu 92: 181-5. Munro U, Wiltschko W & Ford HA (1993) Changes in the migratory direction of yellow-faced honeyeaters Lichenostomus chrysops (Meliphagidae) during au- tumn migration. Emu 93: 59-61. Nix HA (1976) Environmental control of breeding, post-breeding dispersal and migration of birds in the Australian region. Proc XVI Int Ornithol Congr: 272— 305. Olsen PD, Debus SJS, Czechura GV & Mooney NJ (1990) Comparative feeding ecology of the grey goshawk Accipiter novaehollandiae and brown goshawk Accipiter fasciatus. Aust Bird Watch 13: 178-192. Osborne WS & Green K (1992) Seasonal changes in composition, abundance and foraging behaviour of birds in the Snowy Mountains. Emu 92: 93-105. Paton DC (1980) The importance of manna, honeydew and lerp in the diet of honeyeaters. Emu 80: 213-26. Paton DC (1988) Migratory flocks of honeyeaters at Cape Jervis, South Australia. South Aust Ornithol 30: 133-4. Paton DC & Ford HA (1977) Pollination by birds of na- tive plants in South Australia. Emu 77: 73-85. Pearce J, Menkhorst PW & Burgman MA (1995) Niche overlap and competition for habitat between the hel- meted honeyeater and the bell miner. Wildl Res 22: 633-646. | Poiani A (1993) Bell miners: What kind of ‘farmers’ are they? Emu 93: 188-194. Pyke GH (1983) Seasonal patterns of abundance of hon- eyeaters and their resources in heathland areas near Sydney. Aust J Ecol 8: 217-213. Pyke GH (1985) The relationships between abundance of honeyeaters and their food resources in open forest areas near Sydney. In: Keast A, Recher HF, Ford HA & Saunders D (eds) Birds of the eucalypt forests and woodlands: ecology, conservation and management. pp 65-77. Surrey-Beatty, Sydney. Quammen D (1996) The song of the dodo: Island bio- geography in an age of extinctions. Random House Australia Pty Ltd, Sydney. Recher HF (1985) Synthesis: A model of forest and 21 woodland bird communities. In: Keast A, Recher HF, Ford HA & Saunders D (eds) Birds of the eucalypt forests and woodlands: ecology, conservation and management. pp129-35. Surrey-Beatty, Sydney. Recher HF & Holmes RT (1985) Foraging ecology and seasonal patterns of abundance in a forest avifauna. In: Keast A, Recher HF, Ford HA & Saunders D (eds) Birds of the eucalypt forests and woodlands: ecology, conservation and management. pp 79-96. Surrey- Beatty, Sydney. Recher HF, Gowing G, Kavanagh R, Shields J & Rohan-Jones W (1983) Birds, resources and time in a tablelands forest. Proc Ecol Soc Aust 12: 101-123. Recher HF, Holmes RT, Schultz M, Shields J & Ka- vanagh R (1985) Foraging patterns of birds in euca- lypt forest and woodland on the tablelands on south- eastern Australia. Aust J Ecol 10: 399-419. Robinson D (1990) The social organisation of the scar- let robin Petroica multicolor and flame robin P. phoenicea in southeastern Australia: a comparison between sedentary and migratory flycatchers. Ibis 132: 78-94. Robinson D (1992) Why do flame robins Petroica phoenicea migrate? A comparison between the social and feeding ecologies of the flame robin and the scar- let robin P. multicolor. Corella 16: 1-14. Robinson D (1993) Vale Toolern Vale: the loss of our woodland birds. Wingspan 9: 1-3. Rowley I (1974) Bird life. The Australian Naturalist Li- brary: William Collins Ltd., Sydney. Russell R (1981) How Fluffy Gliders led me to the tree of life. Habitat Aust 9: 7-8. Schodde R & Mason IJ (1999) The directory of Aus- tralian birds—Passerines. CSIRO Publishing, Mel- bourne. Serventy DL & Whittell HM (1976) Birds of Western Australia. University of Western Australia Press, Perth. Sibley CG & Alquist JE (1985) The phylogeny and classification of the Australo-Papuan passerine birds. Emu 85: 1-14. Silveira CE, Yen AL, Bennett AF, Brown GW, Hinkley SD, Loyn RH, Lumsden LF & Smith W (1997) Fauna of the box-ironbark study area. Report to Land Conservation Council, Melbourne. Sinclair ARE (1995) Population regulation of resident herbivores. In: Sinclar ARE & Arcese P (eds) Serengeti 2. Dynamics, management and conserva- tion of an ecosystem. University of Chicago Press, Chicago. Starks J (1987) The status and distribution of the black- eared miner (Manorina melanotis) in Victoria. Arthur Rylah Institute Technical Report 49, Melbourne. R. H. LOYN Stone C (1996) The role of psyllids (Hemiptera: Psylli- dae) and bell miners (Manorina melanophrys) in canopy dieback of Sydney blue gum (Eucalyptus saligna Sm.). Aust J Ecol 21: 450-458. Thackway R & Cresswell ID (1995) An interim biogeo- graphic regionalisation for Australia: A framework for setting priorities in the national reserves system cooperative program, Version 4.0. Australian Nature Conservation Agency, Canberra. Tilman D (1982) Resource competition and community structure. Princeton University Press, New Jersey. Traill BJ, Collins E, Peake P & Jessup S (1996) Current and past status of the birds of Chiltern—a box-iron- bark forest in north-eastern Victoria. Aust Bird Watch 16: 309-26. Wiens JA (1981) Single-sample surveys of communi- ties; are the revealed patterns real? Am Nat 117: 90- nN tN 98. Woinarski J (1985) Foliage-gleaners of the treetops, the pardalotes. In: Keast A, Recher HF, Ford HA & Saunders D (eds) Birds of the eucalypt forests and woodlands: ecology, conservation and management. pp 165-75. Surrey-Beatty, Sydney. Woinarski JCZ (1988) The vertebrate fauna of broom- bush Melaleuca uncinata vegetation in north-western Victoria, with reference to effects of broombush har- vesting. Aust Wildl Res 16: 217-238. Wykes BJ (1985) The helmeted honeyeater and related honeyeaters of Victorian woodlands. In: Keast A, Recher HF, Ford HA and Saunders D (eds) Birds of the Eucalypt Forests and Woodlands: ecology, con- servation and management. pp 205-17. Surrey- Beatty, Sydney. Ecological segregation in SE Australia Appendix 1. Pairs of congeneric native bird species inhabiting forests, woodlands and shrublands of mainland south-eastern Australia (excluding species found mainly in warm temperate rainforest), showing the assessed primary mechanisms for ecologi- cal segregation along with other potentially important mechanisms, degree of overlap in total range (Australia) and local range (study region) and degree of habitat overlap in the study region. Other potentially Gees Primary : p Overlap : 2 F : : important ; ; Habitat Family Species | Species 2 segregation p in total in local 4 mechanisms! SCETe ESE range’ range® erealap mechanisms Phasianidae Coturnix pectoralis C. ypsilophora h 2 2 1 Coturnix ypsilophora C. chinensis h SZ 2 2 1 Coturnix pectoralis C. chinensis h SZ 2 2 1 Accipitridae Accipiter fasciatus A. novaehollandiae fg fd 3 3 2, Accipiter novaehollandiae __ A. cirrhocephalus SZ fg, fd 3 3 1 Accipiter fasciatus A. cirrhocephalus SZ fe, fd S 5 3 Columbidae Phaps chalcoptera Ph. elegans h 3 3 3 Cacatuidae Calyptorhynchus banksii C. lathami r sz, fd 2, 0 y, Calyptorhynchus banksii C. funereus h 2 3 2D Calyptorhynchus lathami C. funereus fd 3 3 3 Psittacidae Trichoglossus haematodus _ T. chlorolepidotus SZ fd 3 3 4 Glossopsitta concinna G. pusilla SZ 4 4 3 Glossopsitta pusilla G. porphyrocephala h 2 2 1 Glossopsitta concinna G. porphyrocephala h 2 2 1 Polytelis swainsonii P. anthopeplus ir 0 0 2 Platycercus elegans P eximius h s 2 2 1 Neophema chrysostoma N. pulchella h m 1 1 0 Neophema chrysostoma N. chrysogaster h m 7) 2 l# Neophema chrysostoma N. splendida h m 2 3 O# Neophema chrysostoma N. petrophila h n,m 1 0 O# Neophema chrysostoma N. elegans h 2 2 l# Neophema elegans N. chrysogaster h m 1 1 O# Neophema elegans N. splendida h m 2 1 1 Neophema elegans N. pulchella r 0 0 0 Neophema elegans N. petrophila h n,m 2 1 l# Neophema petrophila N. pulchella r 0 0 0 Neophema petrophila N. splendida ir 0 0 0 Neophema petrophila N. chrysogaster h n,m 1 0 l# Neophema chrysogaster N. pulchella r 0 0 0 Neophema chrysogaster N. splendida r 0 0 0 Neophema pulchella N. splendida r 0 0 0 _ Cuculidae Cacomantis variolosus C. flabelliformis S 2 4 3 Chrysococcyx osculans Ch. basalis S 4 3 4 Chrysococcyx basalis Ch. lucidus h 3 3 2 Chrysococcyx osculans Ch. lucidus h 2 2 2) | Strigidae Ninox strenua N. connivens SZ fg, fd 3 2 1 Ninox connivens N. novaeseelandiae SZ fd 3 3 4 Ninox strenua N. novaeseelandiae SZ fd 3 3 4 | Tytonidae Tyto tenebricosa T. novaehollandiae h fd, fg 3 3 1 Tyto novaehollandiae T. alba SZ fd 3 3 2 Tyto tenebricosa T. alba h sz, fd, s 3 3 0 | Caprimulgidae Eurostopodus mystacalis E. argus h 1 1 1 Halcyonidae Todiramphus pyrrhopygia _ T. sanctus h SZ 2 3 i | Climacteridae Climacteris affinis C. erythrops r fg, s 0 0 0 Climacteris erythrops C. picumnus h S 3 2 0 Climacteris affinis C. picumnus s 2 3 2 23 Appendix 1. (Continued) Family Species | Maluridae Malurus cyaneus Pardalotidae Acanthizidae Meliphagidae Malurus cyaneus Malurus cyaneus Malurus splendens Malurus splendens Malurus Stipiturus malachurus Pardalotus punctatus Dasyornis brachypterus Sericornis frontalis Hylacola pyrrhopygia Calamanthus fuliginosis Gerygone mouki Gerygone fusca Gerygone olivacea Acanthiza pusilla Acanthiza pusilla Acanthiza pusilla Acanthiza pusilla Acanthiza pusilla Acanthiza pusilla Acanthiza pusilla Acanthiza apicalis Acanthiza apicalis Acanthiza apicalis Acanthiza apicalis Acanthiza apicalis Acanthiza uropygialis Acanthiza uropygialis Acanthiza uropygialis Acanthiza uropygialis Acanthiza uropygialis Acanthiza reguloides Acanthiza reguloides Acanthiza reguloides Acanthiza reguloides Acanthiza iredalei Acanthiza iredalei Acanthiza iredalei Acanthiza chrysorrhoa Acanthiza chrysorrhoa Acanthiza nana Anthochaera carunculata Philemon corniculatus Manorina melanophrys Manorina melanophrys Manorina melanophrys Manorina melanocephala Manorina melanocephala Manorina flavigula R. H. LOYN Species 2 M. splendens M. leucopterus M. lamberti M. leucopterus M. lamberti M. leucopterus S. mallee P. striatus D. broadbenti S. magnirostris H. cauta C. campestris G. fusca G. olivacea G. mouki . chrysorrhoa . iredalei nana apicalis uropygialis reguloides lineata iredalei lineata reguloides chrysorrhoa uropygialis iredalei . reguloides . chrysorrhoa lineata nana . iredalei . chrysorrhoa nana lineata nana lineata . chrysorrhoa lineata nana . lineata chrysoptera Ph. citreogularis M. melanocephala M. flavigula M. melanotis M. flavigula M. melanotis M. melanotis ALAR RRA RRARARRARRAARAARAAR ARE Primary segregation mechanisms! aun = Sao aah aa S SS ain ay Sa aa SS Sf Sf Se Sw eS Ss aS aS, = SS Other potentially important segregating mechanisms” SZ, S SZ SZ SZ SZ Overlap Overla : : P P Habitat in total in local 4 3 3 overlap range range WWN DO OWWWWN WW KH HhPWKE NK WN WHRW KH KH HhRWN HK WK WNN CORK WOWONN WN 24 WNN CONN WWHhRWWKH WP HRWWWKE WNW RWNHKE WU NK WK HNN COCK WOWOWWNN NY NOTTTOWNWNODOK OK BPWK DWH KEP NK WNNCKERWNNNKH TOW OF CH WrR Or WH NOS Ecological segregation in SE Australia Appendix 1. (Continued) Other potentially @xeths Primary p Overlap Habitat Family Species | Species 2 segregation Ten HanD in total in local 4 mechanisms! Seema range*> range* oneHlap mechanisms Lichenostomus chrysops L. leucotis S fd, m ?) 3 3 Lichenostomus chrysops L. melanops h fd, s 3 3) 1 Lichenostomus chrysops L. virescens h S 2, A) 0 Lichenostomus chrysops L. penicillatus h s 2 2 1 Lichenostomus chrysops L. cratitius r S 1 1 0 Lichenostomus chrysops L. ornatus r s 1 1 0 Lichenostomus chrysops L. plumulus r s 1 0 0 Lichenostomus chrysops L. fuscus h s,m 5 3 1 Lichenostomus virescens L. fuscus h s 1 1 0 Lichenostomus virescens L. leucotis h S 2 2 0 Lichenostomus virescens L. melanops h s 2 2 0 Lichenostomus virescens L. cratitius h s 2 2 0 Lichenostomus virescens L. penicillatus h s 2 2 0 Lichenostomus virescens L. ornatus h S 3 3 1 Lichenostomus virescens L. plumulus h S 4 3 0 Lichenostomus leucotis L. melanops h fd, s 3 2 1 Lichenostomus leucotis L. penicillatus s sz, fd 2 2 1 Lichenostomus leucotis L. ornatus s sz, fd 3 3 2 Lichenostomus leucotis L. fuscus S sz, fd 3 2 1 Lichenostomus leucotis L. plumulus s sz, fd 2 3 1 Lichenostomus leucotis L. cratitius Ss sz, fd 3 3 1 Lichenostomus melanops L. fuscus SZ s 4 4 4 Lichenostomus melanops L. ornatus h SZ, S 1 1 0 Lichenostomus melanops L. penicillatus h SZ, S 2 3 2 Lichenostomus melanops L. plumulus r SZ, S 0 0 0 Lichenostomus melanops L. cratitius r SZ, S 1 1 0 Lichenostomus cratitius L. penicillatus h s 7) 3 0 Lichenostomus cratitius L. fuscus G S 1 1 0 Lichenostomus cratitius L. plumulus h s 2, 3 3 Lichenostomus cratitius L. ornatus ?h s 4 4 4 Lichenostomus ornatus L. fuscus h 1 1 0 Lichenostomus ornatus L. penicillatus h 2 2 0 Lichenostomus ornatus L. plumulus ?h S 2 3 3 Lichenostomus plumulus L. penicillatus h 2, 2 0 Lichenostomus plumulus L. fuscus r 1 0 0 Lichenostomus fuscus L. penicillatus h fd, s 2 3 1 Melithreptus gularis M. brevirostris S SZ 2 3 3 Melithreptus brevirostris M. lunatus s 2 3 3 Melithreptus lunatus M. gularis s sz, mM 2 3 3 Phylidonyris pyrrhoptera Ph. novaehollandiae h fd, s 3 3 2 Phylidonyris pyrrhoptera Ph. melanops h fd, s 3 2 0 Phylidonyris pyrrhoptera Ph. nigra r fd, s 2 1 1 Phylidonyris pyrrhoptera Ph. albifrons ie 1 1 0 Phylidonyris novaehollandiae Ph. melanops h fd 3 3 1 Phylidonyris novaehollandiae Ph. nigra c S 2 2 3 Phylidonyris novaehollandiae Ph. albifrons r 1 1 0 Phylidonyris nigra Ph. melanops h fd, s 2 2 l Phylidonyris nigra Ph. albifrons r S 1 0 0 Phylidonyris albifrons Ph. melanops h fd, s 2 2 1 Ephthianura tricolor E. albifrons h 2 3 3 25 Appendix 1. (Continued) Family Species | Petroicidae Pomatostomidae Cinclosomatidae Pachycephalidae Dicruridae Campephagidae Artamidae Ephthianura tricolor Ephthianura aurifrons Petroica boodang Petroica boodang Petroica boodang Petroica boodang Petroica goodenovii Petroica goodenovii Petroica goodenovii Petroica phoenicea Petroica phoenicea Petroica rosea Pomatostomus temporalis Pomatostomus superciliosus Pomatostomus temporalis Psophodes olivaceus Cinclosoma punctatum Pachycephala olivacea Pachycephala olivacea Pachycephala olivacea Pachycephala olivacea Pachycephala rufogularis Pachycephala rufogularis Pachycephala rufogularis Pachycephala inornata Pachycephala inornata Pachycephala pectoralis Myiagra rubecula Myiagra cyanoleuca Myiagra rubecula Rhipidura rufifrons Rhipidura fuliginosa Rhipidura rufifrons Coracina novaehollandiae Coracina novaehollandiae Coracina novaehollandiae Coracina papuensis Coracina papuensis Coracina tenuirostris Artamus leucorynchus Artamus leucorynchus Artamus leucorynchus Artamus leucorynchus Artamus personatus Artamus personatus Artamus personatus Artamus superciliosus Artamus superciliosus Artamus cinereus Cracticus torquatus R. H. LOYN Other potentially Oueae ; Eniaaty, important ; P Ove Habitat Species 2 segregation : in total in local 4 mechanisms pa scent range’ range* overlap mechanisms E. aurifrons h 4 4 3 E. albifrons h 2 3 4 P. phoenicea h m 3 3 2$ P. rosea h s 2 3 1 P. goodenovii h 2 2 P. rodinogaster h 3 3 0 P. phoenicea h m 2 2 0 P. rosea h S 7) 2 0 P. rodinogaster h 2 2 0 P rodinogaster h m 3 3 1 P. rosea h S 2 3 1 P. rodinogaster s m 2 4 3 P. superciliosus h 2 A) 2 P. ruficeps h 2 3 1 P ruficeps h S 2 2 Il P. nigrogularis r 0 0 0 C. castanotus r 0 0 0 P. rufiventris s m 2 3 yD P. pectoralis s m 3 3} 3 P inornata h 0 0 0 P. rufogularis if 0 0 0 P inornata SZ fd 3 3} 4 P. pectoralis s SZ 3 3 2 P. rufiventris S SZ 3 3 4 P. rufiventris S m 3 3 4 P._ pectoralis h 2 3 3 P. rufiventris S m 2 3 2 M. cyanoleuca h 3 3 1 M. inquieta S sz, mM D) 3 1 M. inquieta S sz, mM 2, 2 2 M. fuliginosa S m 3 3 4 R. leucophrys h S 3) 3 1 R. leucophrys h S 3 3 0 C. maxima h S 3 3 3 C. papuensis h SZ 3 3 4 C. tenuirostris h SZ 3 3) 3 C. maxima h S 2 ] 1 C. tenuirostris h 3) 2 1 C. maxima h S 2 1 0 A. cyanopterus h 2 3 ] A. personatus h 4 3 1 A. superciliosus h 4 3 ] A. cinereus h 4 4 0 A. cyanopterus h 3 4 2 A. cinereus h 5 3 1 A. superciliosus none! 5 5 5) A. cyanopterus h 3 4 2 A. cinereus h 5 3 1 A. cyanopterus h m 2 3 0 C. nigrogularis h SZ 2 2 2 26 Ecological segregation in SE Australia Appendix 1. (Continued) Other potentially Oped Primary ; p Overlap ; ; ; ; : important : : Habitat Family Species | Species 2 segregation : in total in local 4 mechanisms! _S°8™°8ating range range® erenae mechanisms~ Strepera graculina S. versicolor S m 2 3 3 Corvidae Corvus coronoides C. tasmanicus h 2 2 1 Corvus coronoides C. bennetti h 2 3 Corvus coronoides C. mellori h 3 4 2 Corvus tasmanicus C. mellori h 1 2 1 Corvus tasmanicus C. bennetti r 0 0 0 Corvus mellori C. bennetti r 2 2 1 Passeridae Taeniopygia guttata T. bichenovii h 7) 2 0 Stagonopleura guttata S. bella h 2 2 0 Hirundinidae Hirundo neoxena H. nigricans h fg, s, m,n 2 4 wy) Hirundo nigricans H. ariel h n 4 4 1 Hirundo ariel H. neoxena n fg, s,m 2) 4 3 Sylviidae Cincloramphus mathewsi C. cruralis h 4 4 0 ' Ecological segregation mechanisms: ?=unsure; c=complex; fd=food; fg=foraging technique; h=habitat; m=migration; n=nest site; none! =none; r=range; s=stra- tum/substrate; sz=size. Range is given precedence to habitat as a prmary mechanism, unless the two species occupy distinctly different habitats where their ranges join. > Codes as above. Almost all species pairs show some differences in habitat and range, so these are not listed when considered only as secondary or potential mechanisms for ecological segregation. Degrees of overlap in range and habitat are described in the next three columns. The list of secondary mechanisms is not exhaustive. 3 Range overlap categories: 0=no contact; 1=ranges join, but little overlap; 2=substantial overlap; 3=range of one almost embraced by that of other, but lat- ter covers much greater area; 4=ranges mostly overlapping, with range of one usually embraced by that of other; S=ranges virtu- ally identical. 4 Habitat overlap categories: 0=no overlap; 1=mainly separate habitats, but some overlap; 2=lots of separate habitat; substantial overlap too; 3=mainly over- lapping but each species has separate habitats too; 4=one species has separate habitats but also completely overlaps habitat of the other species; 5=complete overlap. #=little or no overlap in breeding season, but may use similar habitats at other times. $=less overlap in winter than in breeding season 27 Appearliy |. . (Cowtiunsal) a mide ia! toed ello ho ted, uaniliies Terai: ots bo Nygard eg! yay \ntensctodane eS xpahrove sie) tt gt a gar” orion sealed — Wend) ag tewragis hiw — —— — une 2 wevO yininerr | in apaiee saerpeidematided sarnarntintioal os ae hroveecamten fad shariind gente yloime St be toade! opel ieo yionohgmamels ted aintihlll ming ens! evioage orgie h sou! setital atone se re n= 2 cownit radhweheaiegt giieadeess yma a memngtiard oh site ar ie vinkelerqah vo) ataignoe= > ‘ f l — 4 iyi [ : U j 0 ti + 1 ‘om ‘Crna Me flioatinil: p bere teres 1 ied jeterO qihav we Scented enunayt eS + wer whoo ot dished al wa ast ‘not mtpegine 7 Z as ; no ae + Dat Het pa! Sa ene 8 487 bie pabida exons aren, oi se bom i ae bap/estaw aanannet lo steer th le vigtihee ime acleian A : ~ e Rt £ pitfrons t 4 Spadyenard ayelinee x adiijroa . S158 hewiciimy Likielice |= > VAN beating ry . Myr oars he i +P pienedt .~ v ; rhe df hemmlong P vege icapetier = nici s slates = JP hewn ; ¥ [ ~ Ais Pints on red i, bw) S FetvrSiin prx.trwcrt) f° coblmnenvier s) Prohee pli 2) * Vibfierraie — 4 wien oh on P eth po & Meine rersam a4 = P pathtaiigrsi ideo feb femmes) y P jtpen\ilioves - Ae nd ee teen ae Beater es ’ ’ 5] U 7 Vimiatiovel wizaivrelis P ndligey wittheiiosen ndbepor “gates shoo isnon Sole daan = it (i iivpieren( didal=if ception Se shore ; VQNS90 lien enon spin eeicun. Vieiunisee Yeentigen 26 a we Ht oyna? = Ay l" Feviventear re / the toni eh - envainuaboae: lattnsten w cul i ead rem an menor mitt pMecihy tygleafia/epras n — =e J * ¥ a “ - : Y= f 2 > re! mms) “a : — yt ay — = RAeaip | . t J , > > pe ’ q ae ret es i tess Ser | ip " = ~ ~ ~ ry ; : i ro4@ . i * f s a i*- 7 , j aan Bi ° we ly — 8 a as — ‘ ay t 7 A * — <_< “ Ornithol. Sci. 1: 29-40 (2002) SPECIAL FEATURE Interspecific segregation and attraction in forest birds Comparative foraging ecology of five species of ground- pouncing birds in western Australian woodlands with comments on species decline Harry F. RECHER!”, William E. DAVIS, Jr.* and Mike C. CALVER? ' Centre for Ecosystem Management, School of Natural Sciences, Edith Cowan University, Joondalup, W.A. 6027, Australia ? College of General Studies, Boston University, Boston, MA 02215, USA 3 Biological Sciences, Murdoch University, Murdoch, W.A. 6150, Australia Abstract In this paper, we compare the foraging ecology of five Australian robins (Petroica multicolor, P. goodenovi, Eopsaltria griseogularis, Microeca fascinans, and Melanodryas cucullata) in woodlands of Western Australia. Australian robins are in- sectivorous and obtain the greatest proportion of their prey by pouncing from a perch to the ground. Data were collected at three different sites in eucalypt (Eucalyptus) woodland and two sites in acacia (Acacia) woodland. The species differed in habitat, structure of the ground substrates where prey were taken, proportion of foraging ma- noeuvres used, height of foraging perches and prey-attack distances, though there were broad overlaps in all foraging dimensions. Within a site, species were more sim- ilar to each other in their foraging behaviour and selection of foraging substrates than they were to conspecific individuals occurring elsewhere. This indicates that potential foraging behaviours were very broad, and their expression is determined by the char- acteristics of the habitat and available prey. At all sites, robins took prey from ground substrates characterised by a mosaic of bare soil, low ground vegetation, and litter. The smallest species, P. goodenovi, used lower perches than the other robins and probably searched for small prey which it located at short distances. P. goodenovi had the widest distribution and was the most abundant of the species studied. The impli- cations of these findings for the conservation of ground-foraging birds in Australia are discussed. ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2002 Key words Foraging ecology, Ground-pouncing birds, Petroicidae, Threatened species conservation, Woodlands Ground-foraging birds are common in Australian woodlands (Recher et al. 1985; Ford et al. 1986; Recher & Davis 1997, 1998). Among them are species which search for prey from a perch and pounce on ground-dwelling invertebrates and small vertebrates. This pouncing guild is comprised of such different birds as kingfishers (Alcedinidae), cuckoos (Cuculidae), Australian robins (Petroicidae), and butcherbirds (Artamidae), all of which are important components of bird communities in Australian wood- lands and forests (Recher et al. 1985; Holmes and Recher 1986). In this paper, we compare the foraging ecology of (Received 8 May 2001; Accepted 25 September 2001) * Corresponding author, E-mail: hjrecher @pacific.net.au 29 five Australian robins (Scarlet Robin Petroica multi- color, Red-capped Robin P. goodenovi, Western Yel- low Robin Eopsaltria griseogularis, Jacky Winter Microeca fascinans, and Hooded Robin Melanodryas cucullata) in eucalypt (Eucalyptus) and acacia (Aca- cia) woodlands of Western Australia. All are insectiv- orous and only rarely take small vertebrates and seeds (Barker & Vestjens 1990). Our objective is to describe the foraging behaviour of each species and the structure of the ground habitats from which they obtain prey. Previous studies of ground-foraging birds in Australia have considered the ground as a single substrate (e.g., Recher & Davis 1997, 1998), but the ground surface is a mosaic of vegetation, lit- ter, bare ground and coarse woody debris. There is therefore the potential for ground-foraging birds to H. F, RECHER et al. partition different parts of this mosaic and minimize competition for food resources on what is otherwise a single, horizontal plane. Thus, in considering the par- titioning of foraging resources among ground-pounc- ing birds, our emphasis is on micro-habitat selection of ground substrates. Because ground-foraging birds are among the most threatened on the Australian con- tinent (Recher 1999; Garnett & Crowley 2000), we conclude with comments on the reasons for their de- cline. METHODS 1) Study Sites Data were collected in Western Australia from June to October 1997 in Wandoo Eucalyptus wandoo and Powderbark Wandoo E. accedens woodlands in Dryandra State Forest (centred on 32°45’S, 116°55’E; 400 m asl) near the town of Narrogin; in mallee and Wheatbelt Wandoo/York Gum (E. capillosa/E. lox- ophleba) woodlands in Durrakoppin Nature Reserve near Kellerberrin (31°07'S, 117°13'E; 350 m asl); and in Salmon Gum/Gimlet/Morrel (E. salmonophloia/E. salubris/E. longicornis) woodlands at Yellowdine (31°22'S, 119°09'E; 400-450 m asl) east of Southern Cross. During July and August 1999, we collected data on ground-pouncers in woodlands dominated by Acacia spp. near Gascoyne Junction (25°03'S, 115°12’E; 150m asl) and Mt Magnet (28°04’S, 117°20'E; 375-400 m asl). Sites were selected where birds, including ground-pouncers, were abundant. None of the eucalypt woodlands in which we worked was grazed by domestic stock, nor had any been recently burnt. Winter rainfall in 1997 was aver- age with good herbaceous and shrub growth on all sites. In contrast, the Acacia woodlands at Gascoyne Junction and Mt Magnet were grazed heavily by sheep, goats and cattle and had significant soil degra- dation and loss. This was particularly so for the Mt Magnet area. There was no evidence of recent fires. Both sites had had two seasons of above average rainfall. Where it survived grazing by domestic ani- mals, ground vegetation in the acacia woodlands had an abundance of green foliage, flowers and seed. 2) Birds Although we sought out as many different pairs and individuals as possible, we inevitably recorded data on the same individuals on more than one occa- sion. Appendix | presents the number of observed foraging manoeuvres for each species at each site and 30 the estimated minimum number of pairs for which data were obtained. In 1997, only Red-capped (RCR), Scarlet (SR), and Western Yellow Robins (WYR), and Jacky Winter (JW) were present at Dryandra. For comparison of foraging behaviour, we therefore include observations of Hooded Robins (HR) at Dryandra made in 1995 (Recher & Davis 1998; unpubl.), but these were not used in analyses as conditions between the years differed and the ground substrates around perches were not recorded in 1995. Jacky Winter and Red-capped Robin occurred at Durrakoppin and Yellowdine, but only Red-capped Robins were found at Gascoyne Junction. At Mt Magnet, Hooded and Red-capped Robins were pres- ent, but despite extensive searching only two pairs of Hooded Robins were found (Appendix 1). All data were collected during the breeding season (August-October) and individuals were nesting or feeding fledglings at the time of observation. None of the data is from birds in their year of hatching. Mor- phological data were obtained from Baker et al. (1997; see Appendix 2). 3) Foraging Data For each individual encountered, we recorded up to five consecutive foraging manoeuvres (prey-at- tacks) following the procedures and terminology of Recher et al. (1985). Following Recher and Gebski (1990), the first manoevre observed was not recorded; records commenced with the second ma- noeuvre observed. For each observation, we recorded species, sex (if known), foraging manoeuvre, perch height, height and substrate of prey, and horizontal distance along the ground to prey from perch (dis- tance of attack). For some observations, perch height and/or the distance of attack was not recorded, usu- ally because they were not seen clearly. For this rea- son, some sample sizes differ between tables. 4) Foraging Habitat The habitats in which robins occurred were noted, but we did not quantify habitat attributes other than the ground substrates where prey were taken. We measured ground substrates used for foraging in the following way. As we followed foraging robins, perches from which a bird pounced to the ground were flagged with numbered, coloured tape so that they could located later. Within a three metre radius of each perch, we esti- mated percent cover of logs, coarse woody debris (e.g., branches, fallen dead shrubs), litter (e.g., dead Foraging ecology of ground-pouncing birds leaves, shed bark), bare ground, ground vegetation (grass, ferns and herbs), trees, and shrubs, and the number of trees and shrubs. The three metre radius was selected following earlier work (Recher and Davis unpubl.) which had shown that the average dis- tance from perch to prey for ground-pouncers was within three metres for all species. In the work re- ported here, we assumed that the entire three metre circle around a perch was searched equally for prey. As the measurements made were of superimposed layers of vegetation, debris and litter, total cover can exceed one hundred percent. Birds often moved short distances (<3m) between perches, or pounced re- peatedly from the same perch. We only recorded habitat data for successive pounces if the perches used were at least 6 m apart (1.e., no overlap of the 3 m radius circles around perches). Substrate measure- ments are therefore fewer than the recorded number of pounces. 5) Analysis We used all observations in the comparison of dif- ferences in foraging behaviour and set the accepted level of significance at P<0.01 to compensate for the lack of independence of some data (see Recher and Gebski 1989 for a justification of these procedures). Some rare behaviours (<1% of observations) were grouped with more common behaviours: hovering manoeuvres were combined with hawk and probe was combined with glean. Some infrequent behav- iours (1-5% of observations) (e.g., pounce bark, glean foliage) were grouped together as others and not included in analyses because of the large number of zero (nil) observations among species and sites. As we could not always confirm the sex of individ- uals or for species which are not sexually colour di- morphic, data for males, females and individuals of unknown sex were combined. All statistical tests were carried out using ‘Statis- tica for Windows (Statsoft Inc. 1999). Loglinear analyses were used to compare foraging manoeuvres between plots and species. Differences between species and site in the use of substrates were not tested as inspection of the data showed a high corre- lation between foraging manoeuvre and _ substrate. Correlations were also calculated between perch height and distance of attack. Correlations were cal- culated separately for species and sites because of | differences in the structure, height and floristic com- | position of the vegetation between sites. Only the Red-capped Robin was present at all sites, so the 3-way table showing foraging behaviour for all species at all sites was incomplete. However, it was possible to construct a complete 3-way table (species X site X foraging manoeuvre) for Red-capped Robin and Jacky Winter at Dryandra, Durrakoppin and Yellowdine. For Hooded and Red-capped Robins at Mt Magnet, and for Jacky Winter, Red-capped, Scarlet and Western Yellow Robins at Dryandra, 2- way tables (species X manoeuvre) were constructed. MANOVA was used to test for differences between species and sites in two foraging attributes (perch height and attack distance) and four habitat attributes (% shrub cover, % bare ground, % ground vegetation and % litter, including coarse woody debris and logs). Prey height was excluded because of the high fre- quency of 0 height (i.e., the prey were on the ground). Not every species was present at each site, so we first tested for differences between each species at each site. Jacky Winter and Red-capped Robin oc- curred at more than one site, so subsequent analyses tested for differences across sites for each of these. Dependent variables in MANOVA were screened for conformity to assumptions and transformed if neces- sary; perch height and attack distance were log-trans- formed to meet MANOVA assumptions and all per- cent data were arcsine transformed (Tabachnik & Fiddell 1996). Initial multivariate tests used a signifi- cance level of P<0.01, but univariate tests after an initial multivariate test was significant used P<0.05. The habitat attributes (% shrub cover, % bare ground, % ground plants and % litter) of each indi- vidual attack in which the bird pounced on ground- dwelling prey were also assessed using MANOVA (see Table 5 which includes omitted variables for completeness). The pattern of analyses follows that for foraging attributes. The means of the habitat variables and the means of the foraging variables for each species at each lo- cation were standardised before being subjected to cluster analysis based on complete linkage and Eu- clidean distances. The distance matrix from the clus- tering exercise was then used in multi-dimensional scaling (MDS) to determine if distinct groups of species or locations could be classified on the basis of habitat and foraging variables. In MDS, the axes do not have a numerical value and are expressed simply as dimensions without units (i.e., a visual picture) and are not proportional to the variances described. H. F. RECHER et al. Table 1. Foraging behaviour of ground-pouncing robins at five sites in Eucalyptus and Acacia woodlands in Western Australia: Red-capped Robin (RCR); Scarlet Robin (SR); Jacky Winter (JW); Western Yellow Robin (WYR); Hooded Robin (HR). Num- bers are percent of combined manoeuvre/substrate prey attacks. Sample size is given in parenthesis. See Appendix 1 for the num- ber of pairs for which data were recorded. Manoeuvre Pounce Glean Substrate Ground Ground:::::- A. EUCALYPTUS Woodlands Dryandra RCR (158) 74 6 SR (189) 80 1 JW (206) 64 0 WYR (198) 93 0 HR (41)! 80 0 Durrakoppin RCR (75) 83 1 JW (86) 52 0 Yellowdine RCR (55) 38 5 JW (73) 29 0 B. ACACIA Woodlands Gascoyne RCR (69) 59 0 Mt Magnet RCR (92) 79 3 HR (66) 89 0 Foraging behaviour Hawk Snatch Others” Bark Air Foliage--:::- Bark 1 6 4 6 3 3 5 7 3 l 0 15 10 10 1 0 2 2 1 2 15 0 5 0 0 1 7 5 2 7 25 1 15 0 11 35 0 9 1 55) 4 11 0 6 20 10 1 4 9 4 1 1 3 0 D 3 0 6 ' 1995 data adapted from Recher and Davis (1998); * includes glean foliage, pounce bark RESULTS 1) Foraging manoeuvres and substrates Table 1 combines the substrate of the prey with the foraging manoeuvre of the bird for the most common foraging behaviours. Except at Yellowdine, ground- pouncing was the most frequent foraging behaviour recorded and ground was the most common foraging substrate (Table 1). Hawking insects from the air and snatching prey from foliage and bark were the next most common behaviours and were the most frequent behaviours at Yellowdine. Red-capped Robins often gleaned prey, usually from the ground or bark, as did Hooded Robins at Dryandra in 1995 (Table 1). Red- capped robins sometimes hopped along the ground and gleaned prey from the soil surface, litter and low (<2cm high) vegetation. At Yellowdine, 5% of prey taken by Red-capped Robins were gleaned from the foliage of ground vegetation. Birds hopping on the ground also snatched prey from low vegetation or hawked it from the air. Loglinear analysis of the 3-way table (species x siteXforaging manoeuvre) for Red-capped Robins and Jacky Winters fitted a model involving significant 2-way interactions between behaviour and species ( 13=49.29, P<0.01) and between behaviour and site (y2=201.16, P<0.01). A higher proportion of Jacky Winter foraging behaviour was spent in hawking in- sects from the air and less pouncing to the ground than for Red-capped Robin. At Yellowdine, both hawked more frequently and pounced less often than elsewhere (Table 1). At Dryandra, there was a significant difference in foraging manoeuvres between bird species (Y= 18.05, P<0.01). Western Yellow Robins differed from other robins by taking almost all prey by pounc- ing (93%) and rarely snatching, hawking or gleaning (Table 1). Jacky Winters hawked and snatched prey more frequently than Scarlet and Red-capped Robins. In a 2-way comparison, there was no difference be- tween Red-capped and Scarlet Robins (73=3.52, P30:25)): At Mt Magnet, there was no significant difference in foraging manoeuvres between species (y¢= 3.94, 32 Foraging ecology of ground-pouncing birds Table 2. Mean perch height and attack distance in meters for five species of ground-pouncing robins at five sites in Western Australian woodlands: Red-capped Robin (RCR), Scarlet Robin (SR), Jacky Winter (JW) and Western Yellow Robin (WYR). Standard deviation shown in parenthesis. Hooded Robin (HR) data for Dryandra were obtained in 1995 ( Recher and Davis un- publ.). Heights and distances are for ground-pouncing foraging manoeuvres only. Perch height and/or attack distance were not al- ways recorded, while dashes indicate the species was absent from that site. Site Species ; Gascoyne Mt Dryandra Durrakoppin Yellowdine ee henet RCR No. observations 117 62 22 36 34 Perch height 1.2 (1.0) 1.3 (0.8) 1.5 (1.3) 1.3 (1.1) 0.8 (0.5) Attack distance 1.9 (1.7) 1.5 (1.3) 3.6 (3.0) 1.8 (1.9) 1.5 (1.3) SR No. observations 53 - - = = Perch height 1.5 (1.0) ~ - = = Attack distance 2.6 (2.1) - - - = JW No. observations 129 45 22 = = Perch height 1.5 (1.0) AE (AI) 1.9 (1.6) = = Attack distance 2.3 (1.8) 2.8 (1.8) 3.6 (3.7) _ = WYR No. observations 83 ~ - ~ = Perch height 1.5 (0.9) - - - = Attack distance DESi(22D)) - ~ _ - HR No. observations 15 - — = SW Perch height 2.1 (1.1) ~ = — 1.1 (0.8) Attack distance Ba (len) - _ - 79-3) (P25) p=0.03). However, gleaning comprised 12% of for- aging manoeuvres for the Red-capped Robin, but the Hooded Robin did not glean (Table 1). Red-capped Robins used similar proportions of _ foraging manoeuvres and substrates at Dryandra and Durrakoppin (73=1.68, P>0.5), but they pounced less, and hawked and gleaned more at Yellowdine than at Dryandra and Durrakoppin (y;=46.37, P< 0.01) (Table 1). Red-capped Robins at Gascoyne Junction pounced less and hawked and snatched more than at Mt Magnet (y;=7.49, P<0.01) (pounce vs all other behaviours combined). There was no difference in the behaviour of Jacky Winters between Dryandra and Durrakoppin (y7= 3.37, P>0.1) (pounce vs all other behaviours com- bined), but birds at Yellowdine pounced less and _hawked more than at the other sites (y;=42.47, _ P<0.01) (glean and other’ categories deleted) (Table 1). The behaviour of Hooded Robins at Dryandra in 1995 was similar to that of Hooded Robins at Mt Magnet in 1999, but the Dryandra birds took 15% of their prey by gleaning bark. At Mt Magnet, Hooded Robins took 6% of prey by gleaning foliage (Table iy 2) Perch height and distance of attack Although robins occasionally took prey at dis- tances exceeding six metres, most prey were taken within three metres of the perch from which it was sighted (Table 2). Except for Red-capped Robin at Yellowdine, 10 of the 11 species/site comparisons of perch height and attack distance were significantly correlated: the higher the perch, the greater the dis- tance at which prey were attacked (Table 3). For six of the seven comparisons possible, robins used higher perches when hawking than for ground pouncing (Table 4). However, sample sizes were small and the differences were not always significant (Table 4). Initial MANOVA at Dryandra found that species differed significantly (Wilks lambda,5; ;494)=0.97, P<0.001). Univariate tests revealed that these differ- H. F. RECHER et al. Table 3. Correlation coefficients between perch height and distance of attack for Red-capped Robin (RCR), Scarlet Robin (SR), Jacky Winter (JW), Western Yellow Robin (WYR), and Hooded Robin (HR) at five sites in eucalypt and acacia woodlands in Western Australia during 1997 and 1999. Data includes all available foraging behaviours including ground-pouncing and hawking (sally). Sample size is in parenthesis. (** P<0.001, * P<0.01, NS — not significant, P>0.01). Dashes indicate species was absent from that site in 1997. Site Bypeies Gascoyne Dryandra Durrakoppin Yellowdine Dee Mt Magnet RCR 0.39**(130) 0.62**(62) 0.32 (NS)(41) 0.71**(41) 0.46**(56) SR 0.30**(165) - - ~ JW 0.31**(163) 0.66**(45) 0.82**(21) - = WYR 0.45**(188) - = ~ — HR - = = — 0.70**(53) Table 4. Mean perch heights of ground pouncing robins compared to the perch heights of robins hawking and snatching. Species and sites without perch height data for hawking and snatching robins have been deleted. All P-values are significant using P=0.0014, after Bonferroni correction from an initial P of 0.01. Site Species GP Ht Hawk Ht t P Dryandra RCR 1.15 (106) 1.08 (4) tiga) 0.13 0.89 SR 1.63 (137) 2.52 (4) ti139)= 1.62 0.11 JW 1.51 (134) 3.26 (34) t166)= 4-92 <0.0001 Durrakoppin JW 2.10 (45) 3.18 (20) t3) 3.32 0.001 Yellowdine RCR 1.47 (290) 5.13 (18) tig5)= 7.34 <0.0001 JW 1.90 (22) 4.82 (41) t(61y=3.04 0.003 Gascoyne RCR 1.26 (36) 2.61 (7) tig: 3.19 <0.003 ences were related to differences in perch height (F3 649=4.53, P<0.01) and attack distance (F; ¢4)= 3.23, P=0.02). Red-capped Robins used lower perches and attacked prey closer to the perch than co-occurring robins (Table 2). Scarlet Robins, Jacky Winters and Western Yellow Robins used the same height perches and attacked prey at similar distances (Table 2). At 2.1m, the perch height of Hooded Robins at Dryandra in 1995 was greater than for other robins at Dryandra in 1997 (Table 2). Initial MANOVA at Durrakoppin found strong differences between the species (Wilks lambda, ,4= 0.82, P<0.001), while univariate tests revealed that these were caused by variation in perch height (F, jps=15.19, P<0.001) and distance of attack (F; \95=21.24, P<0.001). Red-capped Robins perched lower and attacked prey at closer distances than Jacky Winters (Table 2). No significant interspecific differences in foraging attributes occurred at either Yellowdine or Mt Magnet at the 0.01 level (Wilks lambda, ;.=0.90, P=0.045; 34 Wilks lambda, ,,,=0.98, P=0.39, respectively). Initial MANOVA found significant differences be- tween locations for Red-capped Robin (Wilks lambdag ¢g,=0.81, P<0.001), which univariate tests attributed to perch height (Fy ,.;=15.48, P<0.001) and distance of attack (Fy 35;=3.77, P=0.005). Perch height at Yellowdine was markedly higher than other sites (Table 2). Attack distances varied considerably, with the greatest values at Yellowdine (Table 2). There were significant differences in the foraging attributes of Jacky Winter between Durrakoppin, Dryandra and Yellowdine (Wilks lambda, ,;)=0.94, P=0.005), which univariate tests attributed to differ- ences in perch heights (F, 5,,=7.26, P<0.001). Perch heights at Durrakoppin and Yellowdine were similar, but higher than those at Dryandra (Table 2). 3) Characteristics of foraging habitat Initial MANOVA at Dryandra found that species differed significantly (Wilks lambda,, ,397=8.60, P<0.001). Univariate tests revealed that these differ- Foraging ecology of ground-pouncing birds Table 5. Habitat characteristics of ground substrates of Western Australia eucalypt and acacia woodlands on which Red-capped Robin (RCR), Scarlet Robin (SR), Jacky Winter (JW), Western Yellow Robin (WYR), and Hooded Robin (HR) hunted for prey. Sample size (N) is shown in parentheses. Measurements are based on a three metre radius around the perch from which the bird attacked prey. Values are means+standard deviation. % Cover Location Coarse Trees Shrubs peeoud Logs woody Litter Bate vegetation : ground debris Dryandra RCR (122) 1S =I) 9+12 24+25 1+3 Sae7/ 42+27 30+29 SR (140) 14+12 8+12 31426 344 5+6 46+26 18+19 JW (111) 16+10 4+7 22+23 S225 4+3 44+25 29+26 WYR (156) 17+16 10+14 Dee 22 4+5 BEES 57£28 17+19 Durrakoppin RCR (63) 15—20! 8+13 47+24 3+6 no data” 34+24 16+16 JW (48) 15—20! 10+13 30+19 4+6 no data* 41+21 Pea 28 Yellowdine RCR (20) IZ=14 17+16 jie 2 Dass} 8+6 35+29 44+22 JW (27) Paes) 6+10 Dae 4+4 Sae5) 48+26 43+25 Gascoyne junction RCR (21) 10+20 Sys 18) 13411 eee. 6+5 21+19 59+21 Mt Magnet RCR (55) 15+20 14+14 14+16 0 243 ABE A| 59+19 HR (43) 21+22 5+6 16+12 0 34 18+ 62+19 "Range of projected canopy cover for study site as measured by HFR in 1986. * At this site, coarse woody debris was included with litter as a single measure. ences were caused by differences in % shrub cover (F3 497=6.75, P<0.001), % bare ground (F; 49,= 13.45, P<0.001), % cover of ground plants (F; 45,= 5.99; ‘P=<01001), and % litter cover (F; 49;=7.75, P<0.001). Jacky Winters used ground substrates with less shrub cover than other species (Table 5). Red- capped Robins and Jacky Winters selected substrates with more bare soil than either Scarlet or Western Yellow Robins (Table 5). The cover of ground vege- tation was higher on the substrates used by Scarlet Robins, while Western Yellow Robins selected sites with a high proportion of litter (Table 5). Initial MANOVA at Durrakoppin found strong dif- ferences between the species (Wilks lambda, ,),= 0.86, P=0.003). Univariate tests revealed that these were caused by variation in % cover of ground vegetation (F, j9,=14.65, P<0.001) and % of bare ground F, ,),=7.03, P<0.001). Red-capped Robins foraged on substrates with more ground vegetation than those used by Jacky Winters (Table 5). Jacky Winter substrates had proportionately more litter and bare soil. Initial MANOVA at Yellowdine found no sig- 35 nificant differences between the species (Wilks lambda, ;,=0.75, P=0.054), but the foraging sites frequented by Red-capped Robins had more shrubs and less litter than where Jacky Winters foraged (Table 5). Initial MANOVA at Mt Magnet found strong dif- ferences between the species (Wilks lambda, = 0.78, P<0.001). This was the result of higher shrub cover at the foraging locations used by Red-capped Robins compared with those used by Hooded Robins (F, 94=20.78, P<0.001) (Table 5). 4) Intraspecific differences between sites Initial MANOVA found significant differences between locations for Red-capped Robin (Wilks lambdaj, g3;=0.42, P<0.001). Percentage shrub cover (Fy597=5-96, P<0.001), % bare ground (F4577= 34.10, P<0.001), % litter cover (F,,,,=33.17, P< 0.001), and % ground vegetation (F,5,,=10.05, P< 0.001) were significant univariate variables. Ground foraging substrates used by Red-capped Robins at Durrakoppin had more ground vegetation than all other sites, and there was more ground vege- H. F. RECHER et al. tation at Dryandra than Yellowdine (Table 5). The differences in ground vegetation were not significant between Dryandra and the Acacia woodland sites, but Yellowdine had less ground vegetation than Gas- coyne Junction and Mt Magnet. Shrub cover and the amount of bare soil were greater at Gascoyne Junc- tion and Mt Magnet than at Dryandra and Durrakop- pin (Table 5). An initial MANOVA found significant differences in the attack characteristics of Jacky Winter be- tween Durrakoppin, Dryandra and Yellowdine (Wilks lambdag ;,,=0.70, P<0.001). Percentage shrub cover (F, 16;=9-02, P<0.001) and % ground vegetation (F, }65=23.14, P<0.001) differed across sites. The ground foraging substrates used by Jacky Win- ters at Durrakoppin had greater shrub cover than those at Dryandra and Yellowdine (Table 5). For Jacky Winter foraging sites, ground cover was greater at Durrakoppin and less at Yellowdine than other sites. Correspondingly, the cover of litter and area of bare of soil was greater at Yellowdine (Table 5). 5) Foraging and Habitat Profiles With the Red-capped Robin at Durrakoppin form- ing an outlier, there are three distinct groups of wood- lands and species; Mt Magnet and Gascoyne Junction (Acacia woodlands), Yellowdine (Salmon Gum/Gim- let/Morrel eucalypt woodlands), and Dryandra/Dur- rakoppin (wandoo eucalypt woodlands) (Fig. 1). Indi- vidual species fit more neatly into location groupings than species groupings. That is, different species were more similar in their foraging and habitat pro- files at the same location than to conspecifics at other locations. Thus, woodland type is more important than species in describing the foraging behaviour of ground-pouncing robins. DISCUSSION As with all studies which compare co-existing species within a foraging guild, there were broad overlaps between species in the frequency of foraging manoeuvres, perch heights and foraging substrates (e.g., MacArthur 1958; Recher 1989). However, as shown by Calver and Wooller (1981) and Wooller and Calver (1981a), overlap can be high in some re- source dimensions, but when taken together the over- lap falls. For this reason, Cody (1974) emphasised the need to consider all resource dimensions together when investigating the competitive relationships among co-existing species. The results presented here considered only resource use during the breeding sea- son and at times when food was likely to be most abundant. Under these conditions, co-existing species may have very similar foraging ecology without nec- essarily competing for resources (Recher 1989; Wiens 1989). As food becomes less abundant, species become increasingly different in their use of re- sources (e.g., Recher 1989), a factor not considered in this study. Despite this limitation, there were significant dif- ferences in foraging behaviour between co-occurring species at all sites. Jacky Winters tended to take more aerial prey than other robins, while Western Yellow 1.2 DURRRCR O 0.8 MAGHR @) 0.4 DRYSR MAGRCR N re) O 8 Sis lichus GASRCR 2 0 O O E a DRYWYR -0.4 YELLRCR O -0.8 YELLJW 1.2 -1.2 -0.8 -0.4 0 0.4 0.8 1.2 1.6 Dimension 1 Fig. 1. Multi-dimensional scaling of foraging and habitat attributes for robins at six sites (four in eucalypt woodland and two in acacia woodland). The points are labelled with site and robin species. Sites: YELL (Yellow- dine), GAS (Gascoyne Junction), MAG (Mt Magnet), DRY (Dryandra), DURR (Durrakoppin); Species: RCR (Red-capped Robin), SR (Scarlet Robin), JW (Jacky Winter), WYR (Western Yellow Robin), HR (Hooded Robin). Foraging ecology of ground-pouncing birds Robins pounced to the ground more frequently than others. Significantly, there was no difference in forag- ing behaviours between Red-capped and Scarlet Robins at Dryandra, which was the only site where these two species co-occurred. Of the species studied, Red-capped and Scarlet Robins are the most similar in appearance (both are red, black and white), al- though the Scarlet Robin is half again as large as the Red-capped Robin (Appendix 2). Despite differences in foraging behaviour, the species studied relied primarily on prey taken from the ground or ground vegetation either by pouncing or snatching. It is therefore necessary to consider other aspects of their foraging ecology to fully appre- ciate how resources are allocated within this guild. 1) Habitat While present at the same sites, there were broad habitat differences among the species studied. At Dryandra, Jacky Winters were most common in open Wandoo woodlands with widely spaced large trees and few shrubs. Western Yellow Robins were most often associated with Wandoo and Powderbark Wan- doo woodlands on lateritic slopes with a patchy shrub layer to 1.5m, although they also foraged in open Wandoo woodlands and Mallet E. astringens planta- tions. Red-capped Robins were most common in habitats dominated by sheoak Allocasuarina spp. and Jam Wattle Acacia acuminata, but also occurred in eucalypt woodlands. Scarlet Robins used Powderbark Wandoo and Wandoo woodlands and frequently oc- curred in association with Jarrah E. marginata and Marri E. calophylla which the other species avoided. Most Scarlet Robins were found in habitats with a moderate to dense shrub layer, but they also used Mallet plantations where there is no shrub layer. The two most similar species at Dryandra, in terms of size, appearance and foraging behaviour, Red- capped and Scarlet Robins, have substantially differ- ent geographical distributions. Dryandra is one of the few places where they come into contact (Blakers et al. 1984; Saunders & Ingram 1995; Serventy & Whit- tell 1951). However, where they came into contact at Dryandra, Scarlet and Red-capped Robins were inter- specifically territorial with Scarlet Robins being dom- inant (H. Recher unpubl. obs.). At Durrakoppin, Red-capped Robins were most frequent in mallee (woodlands dominated by multi- stemmed eucalypts) and along the edges of Kwongan shrublands (floristically rich, sandy heathlands). Jacky Winters were restricted to Wheatbelt Wan- doo/York Gum woodland with a patchy shrub layer. At Yellowdine, Jacky Winters were only found in tall Salmon Gum woodland with an extensive, low (to 1 m) shrub layer and broad open areas. Red-capped Robins were most abundant in Salmon Gum/Gim- let/Morrel woodlands with a shrub layer dominated by Melaleuca and Acacia species. At Mt Magnet, Red-capped Robins were restricted to dense Acacia shrublands, while Hooded Robins occurred in open woodland with widely spaced trees, shrubs and small patches of shrubby woodland. Red-capped Robins were restricted to dense Acacia shrublands at Gas- coyne Junction, avoiding more open habitats and edges. Although the three meter radius around foraging perches was too small to fully describe the habitats used by robins, at Dryandra where all species were present, there were significant differences between species in the percent cover of shrubs, ground vegeta- tion, bare ground and litter. These differences largely reflected the different habitat types each species se- lected. Although less pronounced, similar differences occurred between Red-capped Robin and Jacky Win- ter at Durrakoppin and between Red-capped and Hooded Robins at Mt Magnet. In each instance, Jacky Winters and Hooded Robins frequented more open habitats than Red-capped Robins. This was also the case at Yellowdine where Jacky Winters and Red- capped Robins co-occurred, but the differences were not significant, possibly because of small sample sizes and the more uniform abundance of shrubs be- tween habitat types. 2) Resource sharing of ground substrates Within a site, species were more similar to each other in foraging attributes and the foraging sub- strates selected then they were to conspecifics at other sites. This indicates that potential foraging be- haviours were very broad, but their expression was largely determined by the characteristics of the re- source, which in this case are the attributes (structure) of the ground surface. We conclude that co-existence by apportioning ground substrates is not possible within this group of birds. None of the species studied specialised in tak- ing prey from ground vegetation, litter or bare ground. Instead, as described above, species are pri- marily segregated by habitat and secondarily by for- aging behaviour and substrates. This suggests that the availability (but not neces- sarily the abundance) of ground-dwelling prey within H. F. RECHER et al. the size range taken by these species is limited and likely to be easily depleted in the short-term. For this reason alone, ground-pouncing birds should segre- gate by habitat, use different foraging behaviours or differ in the size and type of prey when co-habiting. The interspecific territoriality between the Red- capped and Scarlet Robins at Dryandra, and between Scarlet and Flame Robins P. phoenica in southeastern Australia (Loyn 1980) is confirmation that very marked differences in size are required between ground-pouncing birds before they can co-exist. Other ground-pouncers, such as butcherbirds and kingfishers, which co-exist with the species studied, tend to be much larger and take many large prey in- cluding small vertebrates, use very different foraging behaviours (e.g., sweeping by woodswallows), or are prey specialists (i.e., cuckoos) (Baker & Vestjens 1990; Serventy & Whittell 1951; pers. obs.). 3) Sit and Search Robins are sit and search predators, but move con- tinually between perches often over long distances. Prey are visually located by a perched bird which then flies or drops to where the prey was seen. The area around a perch in which prey are located is a function of the height of the perch, the openness of the immediate habitat, and the behaviour of the bird. Some of the highest perches and attack distances were at Yellowdine where Jacky Winters and Red- capped Robins obtained a large proportion of their prey by hawking. The greater frequency of hawking also explains the higher perches and longer attack distances for Jacky Winters at Durrakoppin. At Dryandra, where most prey were taken by pouncing to the ground or by snatching prey from ground vege- tation or the lower part of tree trunks, perch heights were lower and attack distances for Jacky Winter were smaller. The correlation between perch height and attack distance suggests that high perching birds may search for prey at longer distances, while low perching birds search close to the perch. The frequent changing of perches, even when prey was taken, supports the suggestion that the availabil- ity of ground-dwelling prey is limited and that local resources (those around a perch or set of perches) are quickly depleted. 4) Perch height and prey size There were differences in the height of perches se- lected by ground-pouncing robins and the distances at which they located and attacked prey. In particular, 38 the Red-capped Robin used lower perches and took prey at closer distances than the other robins. Pre- sumably, as the smallest of the species studied, the Red-capped Robin takes the smallest sized prey (for examples of prey size choice, see Calver & Wooller 1981; Hespenheide 1971; Wooller & Calver 1981b) and so must be close to the ground to see them. Fur- thermore, they do not need to scan a large distance because small arthropods are proportionately more abundant than large ones (e.g., Janzen 1973; Majer & Recher unpubl. data). Larger robins presumably take larger prey and hence need to perch higher and scan larger distances to locate them. By selecting small prey, Red-capped Robins may be able to exploit closed, shrub habitats more effi- ciently than the larger species. In turn, this may ex- plain their relative abundance and wide distribution throughout Australia (Blakers et al. 1984). Robins using high perches in the denser habitats where Red- capped Robins are most frequent would have limited lines of sight and thereby be restricted in the area of substrate that could be searched for large prey either on the ground or in the air. 5) Implications for conservation Many studies in Australia and overseas indicate that bird species assemblages co-exist by partitioning the food resource on the basis of one or more of prey type, foraging height, foraging substrate, or foraging behaviour (e.g., Cody 1974; Ford et al. 1986; MacArthur 1958; Recher 1989; Recher & Davis 1998; Wheeler & Calver 1996; Wiens 1989). If the basis of this partitioning is disrupted (e.g., by changes in available foraging substrates) then the pat- tern of partitioning could be expected to change, with possible loss of one or more species from the assem- blage. Australian studies suggest that such changes have occurred and continue to occur, with past (Bur- bidge & McKenzie 1989; Recher & Lim 1990) and on-going implications for conservation (Recher 1999; Ford et al. 2001). Although there were differences in the structure of ground substrates on which different species of ground-pouncers foraged, within a site all species selected foraging sites best described as a mosaic of bare ground, litter and ground vegetation. Distur- bances, such as nutrient enrichment, which increase the extent and density of ground and shrub vegeta- tion; frequent burning, which creates an open habitat with increased areas of bare ground and reduced amounts of litter; and, grazing by domestic animals, Foraging ecology of ground-pouncing birds which reduces the amount of ground and shrub vegetation (Arnold & Weeldenburg 1998), compacts the soil and increases the amount of bare ground (Abensperg-Traun et al. 2000) will disadvantage ground-pouncers. Among the consequences of these disturbances are decreased abundances and possible changes in the size distribution of terrestrial arthro- pods favouring smaller species (see Abensperg-Traun et al. 2000; Wooller & Calver 1988). Such environmental changes appear to be responsi- ble for the decline of ground-foraging birds through- out Australia (Garnett & Crowley 2000; Recher 1999). Post hoc, we predict that the impact should be greatest on the largest species of robins; those requir- ing large prey and relatively large areas of suitable substrate around each perch. This appears to be the case with Jacky Winter and Scarlet and Hooded Robins declining significantly in abundance and distribution throughout their range, while Red-capped Robins persist in relative abundance. Ultimately, however, as degradation of ground substrates in Aus- tralia intensifies with continued land clearing, weed invasion and over grazing, all species will be ad- versely affected. Conservation of this group of birds in Australia requires landscape scale changes in graz- ing and fire management practices to preserve the foraging resources and ground substrates which this study demonstrates are required by each species. ACKNOWLEDGMENTS This work was conducted while WED was a visiting SASTEC Fellow at Edith Cowan University during 1997 and a recipient of a Faculty of Science Research Fellowship during 1999. Helpful comments on the man- uscript were received from H. A. Ford, J. A. Keast and R. Major, the editor and an anonymous referee. REFERENCES Abensperg-Traun M, Smith GT & Main BY (2000) Ter- restrial arthropods in a fragmented landscape: a re- view of ecological research in the Western Australian central wheatbelt. Pac Conserv Biol 6: 102-119. Arnold G & Weeldenburg J (1998) The effects of isola- tion, habitat fragmentation and degradation by live- stock grazing on the use by birds of patches of gimlet Eucalyptus salubris woodland in the wheatbelt of Western Australia. Pac Conserv Biol 4: 155-63. Baker G, Dettmann E, Scutney B, Hardy L & Drynan D (1997) Report on the Australian bird and bat banding scheme, 1995—96. Environment Australia, Canberra. Barker RD & Vestjens WJM (1990) The food of Aus- tralian birds. Vol 2. Passerines. CSIRO Publ, Mel- bourne. Blakers M, Davies SJJF & Reilly PN (1984) The atlas of Australian birds. Melbourne University Press, Melbourne. Burbidge AA & McKenzie NL (1989) Patterns in the modern decline of Western Australia’s vertebrate fauna: causes and conservation implications. Biol Cons 50: 143-198. Calver MC & Wooller RD (1981) Seasonal differences in the diets of small birds in the Karri forest under- storey. Aust Wildl Res 8: 653-657. Cody, ML (1974) Competition and the structure of bird communities. Princeton University Press, Princeton. Ford HA, Noske S & Bridges L (1986) Foraging of birds in eucalypt woodland in north-eastern New South Wales. Emu 86: 168-79. Ford HA, Barrett G, Saunders DA & Recher HF (2001) Why have birds in the woodlands of southern Aus- tralia declined? Biol Conserv 97: 71-88. Garnett S & Crowley G (2000) The Action Plan for Aus- tralian Birds. Environment Australia, Canberra. Hespenheide HA (1971) Food preference and the extent of overlap in some insectivorous birds, with special reference to the Tyrannidae. Ibis 113: 59-72. Holmes RT & Recher HF (1986) Search tactics of insec- tivorous birds foraging in an Australian eucalypt for- est. Auk 103: 515-30. Janzen DH (1973) Sweep samples of tropical foliage in- sects: effects of seasons, vegetation types, elevation, time of day, and insularity. Ecology 54: 677-708. Loyn RH (1980) Bird populations in a mixed eucalypt forest used for production of wood in Gippsland, Vic- toria. Emu 80: 145-156. MacArthur RH (1958) Population ecology of some war- blers in northeastern coniferous forests. Ecology 39: 599-619. Recher HF (1989) Foraging segregation of Australian warblers (Acanthizidae) in open-forest near Sydney, New South Wales. Emu 89: 204-215. Recher HF (1999) The state of Australia’s avifauna: a personal opinion and prediction for the new millen- nium. Aust Zool 31: 11-27. Recher HF & Davis WE (1997) Observations on the foraging ecology of a mulga bird community. Wildl Res 24: 27-43. Recher HF & Davis WE (1998) Foraging profile of a wandoo woodland avifauna during spring. Aust J Ecol 23: 514-28. Recher HF & Gebski V (1990) Analysis of the foraging ecology of eucalypt forest birds: sequential versus H. F. RECHER et al. single-point observations. In: Morrison ML, Ralph CJ, Verner J & Jehl JR Jr (eds) Avian Foraging: theory, morphology, and applications. pp 174-180. Allen Press, Lawrence. Recher HF & Lim L (1990) A review of the status of Australia’s terrestrial vertebrate fauna. Proc Ecol Soc Aust 16: 287-301. Recher HF, Holmes RT, Schulz M, Shields J & Ka- vanagh R (1985) Foraging patterns of breeding birds in eucalypt forest and woodland of south-eastern Australia. Aust J Ecol 10: 399-420. Saunders D & Ingram J (1995) Birds of Southwestern Australia. An atlas of changes in the distribution and abundance of the Wheatbelt Avifauna. Surrey Beatty & Sons, Chipping Norton, NSW. Serventy DL & Whittell HM (1951) A handbook of the birds of Western Australia. Paterson Brokensha, Perth, WA. Statsoft Inc (1999) STATISTICA for Windows. Statsoft Appendix 1. Number of observations of ground-pouncing birds at Dryandra, Durrakoppin and Yellowdine during 1997 and at Gascoyne and Mt Magnet during 1999. Data for Hooded Robin at Dryandra were collected in 1995. The esti- mated number of different pairs for which we obtained data at each site is shown in parenthesis. Dashes indicate the species was absent from that site. Inc, Tulsa, Okalahoma. Tabachnik B & Fiddell L (1996) Using multivariate statistics. Harper Collins Publ, New York, NY. Wheeler AG & Calver MC (1996) Resource partitioning in a island community of insectivorous birds during winter. Emu 96: 23-31. Wiens JA (1989) The ecology of bird communities. Vol 1. Cambridge University Press, Cambridge. Wooller RD & Calver MC (1981a) Feeding segregation within an assemblage of small birds in the Karri for- est understorey. Aust Wildl Res 8: 401-410. Wooller RD & Calver MC (1981b) The diet of three in- sectivorous birds on Barrow Island, Western Aus- tralia. Emu 81: 48-50. Wooller RD & Calver MC (1988) Changes in an assem- blage of small birds in the understorey of dry sclero- phyll forest in south-western Australia after fire. Aust Wildl Res 15: 331-338. Appendix 2. Comparative size of Red-capped Robin, Scar- let Robin, Western Yellow Robin, Jacky Winter, and Hooded Robin. Males and females are shown separately for sexually colour dimorphic species. Data are from the Australian Bird and Bat Banding Scheme (Appendix H in Baker et al. 1997). Data are means and standard deviation with sample size in parenthesis. Bird species Location RCR SR JW WYR HR Dryandra 158 (10) 189(8) 206(6) 198 (10) 41 (2)! Durrakoppin 75 (3) ~ 86 (3) = = Yellowdine 55 (8) - 75 (2) - - Gascoyne 69 (12) ~ - - Mt Magnet 92 (15) ~ — - 66 (2) ' From Recher and Davis (1998). Measurement SECIIES Weight Wing chord Tail length (g) (mm) (mm) RCR Male 8.6-0.7 (62) 62.9-2.1(67) 47.7-3.3 (35) Female 8.7-0.6(65) 61.5-2.5(67) 46.5-2.2 (38) SR Male 13.1-1.1(185) 75-2.2 (108) 55.3-3.3 (31) Female 13.6-1.5(99) 73.6-2.2(78) 55.2-5.2 (19) JW 15.1-2.5(30) 86.8-7.2(29) 60.6-8.5 (19) WYR 19.0-1.9(21) 89.3-3.7(19) 60.6-8.5 (19) HR Male 23.6-2.3 (23) 96.4-3.7(14) 71.2-1.8 (6) Female 22.5-1.6(18) 89.8-2.2(19) 67.7-2.9 (6) Omnithol. Sci. 1: 41-51 (2002) SPECIAL FEATURE Interspecific segregation and attraction in forest birds Heterospecific attraction among forest birds: a review Mikko MONKKONEN?* and Jukka T. FORSMAN* Department of Biology, University of Oulu, POB 3000, 90014 Oulu, Finland Abstract In this paper we review the evidence for a habitat selection process where ORNITHOLOGICAL colonizing individuals use other species presence as cues to profitable breeding sites. SCIENCE Our experimental studies in Fennoscandia and North America have shown that den- sity and species richness of migrant birds breeding in the forests respond positively to experimentally augmented titmice densities. We used analytical modeling to analyze ecological conditions, which may favor a habitat selection process where later arriv- ing individuals (colonists) use the presence of earlier established species (residents) as a cue to profitable breeding sites. We compared the fitness of two colonist strate- gies: colonists could either directly sample the relative quality of the patches (termed samplers) or, alternatively, they could also use residents as a cue of patch quality (cue-users). Model results suggested that cue-using strategy is more beneficial in most ecological conditions and that this may result in heterospecific attraction. Fur- ther field experiments showed that migrant individuals selected nest sites at close vicinity of nesting titmice, and bred earlier and reproduced better. We conclude that heterospecific attraction may be a common and widespread process among forest © The Ornithological Society of Japan 2002 birds particularly in seasonal environments. Key words Experimental studies, Habitat selection, Migrant vs. resident birds, Reproductive output, Species richness Heterospecific attraction was coined by Monk- konen et al. (1990) to describe habitat selection | process where individuals prefer selecting habitat patches already occupied by individuals of another species. It was hypothesized that particularly migrant birds may use residents as cues to profitable breeding | sites in conditions where direct and accurate assess- ing of the quality of available patches is difficult. Short available breeding time and large year-to-year | variation in conditions, both characteristics of north- em environments, would presumably render het- erospecific attraction a profitable habitat selection | strategy. Temperate and boreal passerine bird communities are comprised of resident and migrant species. Resi- | dents have to cope with local conditions on a year _round basis but migrants are able to avoid the low phases in food availability by migrating elsewhere for | the temperate winter. Individuals of many species (Received 28 May 2001; Accepted 30 August 2001) * Corresponding author, E-mail: mikko.monkkonen@oulu.fi * Present address: Montana Cooperative Wildlife Research Unit, Avian Studies Program, University of Montana, Missoula, MT 59812, USA show high fidelity to their previous year’s breeding territory (Hildén 1965). However, because of high adult mortality approximately half of the individuals in passerine breeding populations are first time breed- ers who need to make a selection among potential breeding sites. Therefore breeding habitat selection may only be made once in the life of a bird (Cody 1985) stressing the importance of this operation. Birds obviously use a multiple of cues when se- lecting breeding sites (Hildén 1965). These include habitat cues, habitat structure, floristics, food re- sources etc., as well as information on population density (Wiens 1989). Also interactions with other species may influence the occupancy of an area (Cody 1985). The effect of other species on local community diversity may either be positive, such as through heterospecific (see also Slagsvold 1980, Elm- berg et al. 1997) or conspecific (Alatalo et al. 1982; Doligez et al. 1999) attraction, or negative through intra- (e.g., Krebs 1971; Fretwell 1972; Monkkonen 1990) or interspecific (e.g., Reed 1982; Garcia 1983; Gustafsson 1987; Martin & Martin 2001a, b) compe- tition. The presence and density of predators may also decrease the desirability of habitat patches (Mar- M. MONKKONEN and J. T. FORSMAN tin & Roper 1988; Jarvinen 1990; Suhonen et al. 1994; Hogstad 1995), if settling individuals are able to actively avoid areas with high risk, as appears to be the case (Norrdahl & Korpimaki 1998). The idea that individuals prefer to settle close to conspecific individuals at least in some species can be traced back many decades (Lack 1948; Svardson 1949; Kalela 1952). Stamps (1988) reviewed hy- potheses to account for such attraction to conspecific individuals. Aggregated distribution may 1) provide protection against predators by means of communal defence or information delivered by neighbours, 2) benefit an individual if settled neighbours reflect habitat quality, 3) provide social stimulus and hence accelerate or improve breeding performance (e.g. pairing), or 4) improve ability to defend against in- truders or competitors. In principle, the three first hy- potheses may apply to heterospecific attraction as well. First, it has often been shown that individuals of many species may aggregate to breed in colonies or clumps for protection against nest and other predators (Slagsvold 1980). Second, birds are able to recognize vocalizations of heterospecific individuals, e.g. song (Monkkonen et al. 1996) and warning calls (Forsman & Monkkonen 2001; Gunn et al. 2000), and therefore birds can make use of information delivered by other individuals. Aggregations of heterospecific individu- als may also provide social stimulus if, for example, females are attracted to such ‘hot-spots’ of singing males, improving or accelerating pairing. For heterospecific attraction to operate certain eco- logical conditions must prevail. First, residents must honestly signal the quality of habitat. This seems a rather robust assumption. Resident birds are very likely less time constraint than migrants, and hence, can invest more time and energy in direct assessing of the relative quality of available habitat patches. Migrant birds are more time constrained in their breeding because in their northern breeding areas habitat selection, pairing, nest building, reproduction, and in some species also moulting has to be fulfilled within just few weeks. Even short delays in the onset of reproduction may accrue severe fitness costs (von Haartman 1967; Alatalo & Lundberg 1984; Harvey et al. 1985; Barba et al. 1995). The presence of resident individuals is likely an honest signal of predation risk in a patch, because residents have had time to evalu- ate the risk of predation or because residents in risky patches have already fallen victims of predation. Habitat selection and predation result in spatial varia- tion in the density of resident individuals in a land- scape. Only if such variation existed, would there be potential for migrants to gain information from resi- dent abundance. Second, migrant bird responding to this variation should be capable of making a choice between differ- ent habitat patches. In other words, heterospecific at- traction is more likely applied by a species that have to process much information from the landscape be- fore habitat selection, such as habitat generalists ap- parently have to do (see Dall & Cuthill 1997). Strict habitat specialists more likely respond directly to structural or floristic composition of habitats. Third, for fitness benefits from heterospecific attraction, res- idents must not cause serious resource depletion for migrants. If severe competition occurs between resi- dents and migrants, migrants should avoid settling in a patch together with residents (Cody 1985). There- fore, heterospecific attraction is more likely among species that do not compete over food and in condi- tions where food depletion is not a risk (food limita- tion less severe). In this paper we first provide some background in- formation on abundances of birds along biogeo- graphic gradients to exemplify the range of condi- tions where temperate breeding bird communities de- velop. In particular, we focus on the relative abun- dance of migrant and resident species in their breed- ing assemblages, which form the starting point for our later studies on species interactions in forest bird communities. Secondly, we review the evidence, both theoretical and empirical, for heterospecific attraction among forest birds. We consider the effects of this at- traction on both breeding numbers and community structure as well as on fitness components of breed- ing individuals. Finally, we provide a discussion about the generality, importance and conservation im- plications of heterospecific attraction. BIOGEOGRAPHICAL PATTERNS IN THE ABUNDANCE OF RESIDENT AND MIGRANT BIRDS AND IMPLICATIONS FOR HETEROSPECIFIC ATTRACTION A well-known geographic pattern in northern breeding bird communities is the increase in the pro- portion of migrant birds of the total species and pair numbers with increasing seasonality (MacArthur 1959; Wilson 1976; Herrera 1978; Morse 1989). Usu- ally migrants comprise a larger share of species and individuals in the north than in the south. This geo- graphical pattern is clear, for example, in western Eu- Heterospecific attraction rope where the proportion of tropical (trans-Saharan) migrants increases from an average less than 10% in the Mediterranean region to about 50% in Fennoscan- dia (Herrera 1978; for North American pattern, see e.g. Wilson 1976; Morse 1989). According to Herrera (1978) such a pattern is be- cause Carrying capacity of the environment during the severe season (winter) regulates the size of resident species populations below the levels of summer time carrying capacity, which in turn affects the number of migrants that may enter the habitats. In other words, migrants fit in the breeding assemblages in high num- bers only where resident populations are regulated to a low level (see also Morse 1989). Resident birds are usually considered superior competitors over mi- grants and this interspecific competition would keep migrant numbers low in areas and habitats where res- ident densities are high (Herrera 1978; O’Connor 1981). The pattern of increasing proportion of mi- grants toward north may also be because of geo- graphical variation in migrant abundance without any linkage to resident abundance. Helle and Fuller (1988), however, showed that total migrant densities do not vary very much from south to north in Europe indicating that the increasing proportion of migrants toward north is mainly due to decreasing resident abundance. More detailed new analysis by Forsman and MonkkG6nen (see Forsman 2000) showed that while resident densities generally declined with latitude mi- grant densities tended to peak at mid-latitudes (in central Europe) and were lower both further south and north (Fig. 1). The decrease in resident densities was not linear either, and densities north from 60°N were invariably low whereas further south highly variable. The unimodal density pattern of migrants was common to many genera (Phylloscopus, Fringilla, and Turdus) and suggests that migrant den- sities at geographical scale vary independently of res- ident numbers. The current evidence, therefore, does | not support the conclusion that competition with resi- dents would cause the geographical pattern in mi- grant proportions, and calls for alternative explana- tions. Forsman and Monkkonen (see Forsman 2000) also analysed the covariation between titmice (resident) _ and migrant (several genera) densities in Europe after removing the geographical trends in abundances to find out geographic areas where negative (competi- tion) and positive (e.g. heterospecific attraction) asso- _ clations between residents and migrants are more 20 — 18 4 16 4 Northern Europe ZZZ) Central Europe 2a Westem Europe Mediterranean —* A | ara SO Wa sone ss 7 2°, SRARRARAIY 0, 00,000.00, POLO PS Density (pairs/10 ha) <> ware OX rarer SOK SZ So LLL D> y SA OR =—_ — ON FD ODO OO N | | ! I | Gs Fringilla Phylloscopus Hole LES 2 Titmice Turdus Species groups Fig. 1. Density of different forest bird taxa and species groups in different parts of Europe based on published census results (see Forsman 2000). Northern Europe refers to areas north from latitude 60° and southern Europe to areas south from 45°. The area between these latitudes was divided into central Europe (east from 2°E) and western Europe (west from 2°E). Titmice include all European Parus spp. and the Long- tailed Tit (Aegithalos caudatus). Generic groups refer to all species in the genus, and Hole denotes hole nesting passerine birds other than Parus spp. (Ficedula hypoleuca, F albicollis, and Phoenicurus phoenicurus). likely than elsewhere. There were many significant positive associations but no significant negative asso- ciations. For example, the density of Fringilla spp. was positively and significantly associated with tit- mice densities in northern and central Europe, and that of Phylloscopus-warblers in central Europe. This result indicates that positive rather than competitive interactions may prevail between residents and mi- grants almost irrespectively of the geographic loca- tion. EVIDENCE FOR HETEROSPECIFIC ATTRACTION: NUMERICAL RESPONSES 1) Observational evidence That birds may be attracted to nest close to other species is a well-known pattern (Durango 1947; Koskimies 1957; Hildén 1964, 1965). Slagsvold (1980), for example, found that Bramblings (Fringilla montifringilla) and Redwings (Turdus ilia- cus) preferred to nest in Fieldfare (Turdus pilaris) colonies. These species obviously benefited from breeding within Fieldfare colonies because of com- munal defense against nest predators. M. MONKKONEN and J. T. FORSMAN First implications of heterospecific attraction among putative competitor species were found by Reed (1982). He demonstrated that Chaffinch (Fringilla coelebs) territories on the mainland of Scotland tended to co-occur with Great Tit (Parus major) territories more often than expected by chance alone. In island conditions, however, these two species seemed to compete with each other and occu- pied non-overlapping territories (Reed 1982). Reed (1982) concluded that the environment on the main- land is richer (more food) than on islands. These re- sults suggest that interspecific interactions may vary considerably according to environmental conditions from competition in situations were resources are limiting (on islands in Reeds case) to positive interac- tions in others (mainland). Positive association between territories of two species may also stem from overlapping habitat re- quirements or from concomitant settling in rich re- source. In a playback experiment, where habitat re- quirements of the species were controlled for, Timo- nen et al. (1994) demonstrated that two migrants species, the Chaffinch and the Willow Warbler (Phyl- loscopus trochilus), did not avoid settlement near or in resident (Parus spp.) territories. Also in this work, there was a tendency that migrants aggregated more than expected at the vicinity of residents. In this ex- periment, however, food availability was not con- trolled for and a possibility remained that actually both residents and migrants preferred settling in high quality food patches resulting in positive associations among species. 2) Experimental evidence To test for numerical response of migrants to resi- dent abundance in their breeding bird assemblages of forest birds we have conducted three rather similar experiments (M6nkkonen et al. 1990, 1997; Forsman et al. 1998). In these studies, we manipulated the oc- currence of resident titmice (Parus spp.). These ma- nipulations involved winter-feeding and putting up nest boxes to attract titmice on part of the study plots, and removal of titmice from some others. As a result, at the onset of breeding season migrants could make a choice between plots devoid of titmice and plots where titmice densities were augmented. Migrant re- sponses to manipulations were measured by census- ing their abundance on experimental plots at the height of the breeding season when breeding pairs possess territories. The experimental design where all plots received both treatments in alternate years ef- 44 fectively removed any year, food or site related ef- fects on the results (for detailed methodology, see original publications). Heterospecific attraction hy- pothesis predicts that migrant densities would be higher in plots with titmice than in empty plots. Our results, encompassing two different continents (Europe and North America) and a variety of condi- tions from north (Lapland) to south boreal forest zones (Minnesota), were consistent and provided sup- port for the heterospecific attraction hypothesis. The general pattern turned out to be a positive response by migrants to augmented titmice abundance. In all three locations migrant species richness tended to be higher when titmice were present than in absence of titmice (Fig. 2) but this trend was statistically signifi- cant only in Lapland. In Lapland and central Finland total migrant abundance responded significantly and positively to augmented titmice abundances, and in central Finland and in Minnesota, foliage gleaners showed a significant positive response. In each area there were one or two individual species showing positive, and none showing negative, response to tit- mice presence. The only other study, in addition to our experi- ments, where the effects of heterospecific attraction on species abundance and community assembly has been addressed is the work by Elmberg et al. (1998) on dabbling ducks. In line with our results, Elmberg et al. (1998) concluded that heterospecific attraction rather than competition affects species co-occurrence in dabbling ducks. 3) Theoretical considerations and a test Our experiments clearly showed that forest bird species might use each other’s presence as cues in breeding habitat selection in a wide variety of envi- ronmental conditions. Results from local experimen- tal work, however, do not lend themselves to make far-reaching conclusions about the importance and generality of the heterospecific attraction as a process. Therefore, we used analytical modeling to analyze ecological conditions, which may favor a habitat selection process where later arriving individ- uals (colonists) use the presence of earlier established species (residents) as a cue to profitable breeding sites (heterospecific attraction) (M6nkk6nen et al. 1999). In this model, colonists assessing potential breeding patches could select between high-quality source and low-quality sink patches. Residents occu- pied a proportion of the source patches. One patch can only foster one pair of colonists. Colonists could Heterospecific attraction Central Finland Minnesota 9 — 12:< 9 | 6 B 4 0 Total abundance Species richness 20 18 + 10 4 94 0 0 a) 30 = 25 = c 20 - © al 15 Dd) 2, 10 105 a 5 To) LL 0 0 Redwing Red-eyed Vireo 2 * 8 8 - * g ic c 6 - 6 < re | A\ ZA 5 al < 24 2 - 0 0 0H ADD REM ADD REM ADD REM Treatment Treatment Treatment Fig. 2. Mean species richness, mean number of pairs of migrant passerine birds and of foliage gleaning birds in three experimental studies (Lapland, Forsman et al. 1998; Central Finland, Monkkonen et al. 1990; Minnesota, Monkk6nen et al. 1997). Mean number of pairs for the most strongly responding species is also given (Redwing Turdus iliacus, Chaffinch Fringilla coelebs, Red-eyed Vireo Vireo olivaceus). Error bars denote +1 SD. ADD refers to augmented titmice density and REM to removal of titmice from the study plots. Asterisk refers to statisti- cally significant (P<0.05) difference between the treatments. either directly sample the relative quality of the We assumed that colonizing individuals use sequen- patches (termed samplers) or, alternatively, they tial-comparison tactic (SCT) when choosing among could also use residents as a cue of patch quality patches. It follows from SCT that colonists sample (cue-users). Cue-users gained benefit from lowered only a limited number of patches (maximum 5 costs when assessing occupied source patches. We patches). Our model does not result in an ideal distri- , compared the fitness between cue-users and sampler bution because colonists do not necessarily end up in different ecological conditions and varied, for ex- selecting the best available patch but the best of the ample, the proportions of sink, empty source and oc- _ evaluated ones. We used a variant of natural decision cupied source patches, as well as intensity of compe- theory where sequences can be depicted with deci- tition vs. benefits gained from social aggregations. sion tree diagrams (for details, see Monkkénen et al. 45 M. MONKKONEN and J. T. FORSMAN 1999). The results of the model indicated that the cue- using strategy is an efficient way to choose the best possible patch both when benefits from social aggre- gation exceeded the effects of competition (interspe- cific competition is not strong) but also when inter- specific competition is stronger than the benefits (re- sults in avoidance of occupied patches). Samplers can achieve higher fitness than cue-users only if the dif- ference in quality between occupied and unoccupied source patches is low (interspecific interactions weak). This was because the relatively more compli- cated patch selection procedure of cue-users creates costs, which override the benefits of avoiding the di- rect assessment of the patch quality, when gains are low. Consequently, cue-using strategy can be used both to avoid competition and to aggregate with het- erospecific individuals. Heterospecific attraction would occur whenever colonists gained some benefit from aggregating with residents, which exceeded the effects of competition. The model also predicted that the strongest attrac- tion to heterospecifics occurs when residents occupy approximately half of high-quality source patches. This is because in such conditions colonists can truly make a choice between empty and occupied source patches. If only few patches are occupied (or empty) chances of finding one are low and choices between empty and occupied source patches are infrequent. In other words, the response of colonists to resident abundance would not necessarily be linear along a whole gradient of resident abundance. To test this idea we conducted a further experiment in central Finland where resident densities in nine study plot were manipulated create a spectrum of res- ident densities, relative to previous year’s unmanipu- lated densities (Thomson et al. unpublished). In the first study year titmice were allowed to breed on plots at natural densities, but before the second breeding season their densities were manipulated by feeding, providing nest boxes and removals. Relative changes in migrant densities were analyzed against the differ- ence in titmice densities between years. The hypothe- sis was that intermediate change in titmice densities would be associated with the highest migrant densi- ties. The results were only partly consistent with the prediction. There was only very little evidence for a non-linear response. In general the result was a linear response of migrants as shown in Fig. 3 for foliage gleaning guild. This pattern matches well with our 46 Change in foliage gleaner density T Te -2 0 Change in titmice density ee a 2 K4 Fig. 3. The change in density of foliage gleaning birds be- tween two consecutive breeding seasons as a response to ma- nipulated titmice density (pairs/10 ha). Titmice densities were decreased (negative change), kept constant, or augmented (positive change) between the years. Foliage gleaners’ density significantly increases with increasing titmice density (regres- sion slope =1.16, df=7, P=0.032). earlier experiments and provide further support for heterospecific attraction. The prediction from our an- alytical model was not, however, confirmed. This is very likely because we did not manage to create high enough titmice densities in our experiment for non- linear responses to emerge. For example, fitting a quadratic curve to density response of the chaffinch in relation to titmice density suggests that the peak in chaffinch density would be achieved when titmice density is about 10 pairs/10 ha. Only after that point would chaffinch densities start to drop with an in- crease in titmice densities. Maximum densities in our area were about 6 pairs/10ha, which is rather high density compared to natural densities in the area, but, however, far below the threshold point for non-linear response. COLONISATION RATES AND FITNESS CONSEQUENCES OF HETEROSPECIFIC ATTRACTION ON DIFFERENT SPATIAL SCALES Habitat selection across scales Habitat selection of birds is regarded as a hierar- chical procedure during which factors affecting deci- sions of colonizing birds vary considerably (Hutto I) Heterospecific attraction 1985). On larger scales the influence of individual choice on selection is probably negligible (Hutto 1985) because the choice of geographical area or macrohabitat (e.g., forest type) is quite likely strongly genetically determined (Partridge 1978). However, the smaller the scale is the larger is the number of the characters and cues of the environment that can be taken into account by an assessing individual. The multitude of different biotic cues affecting habitat selection suggests that perceptual ranges of birds can be very wide and following decisions show considerable behavioural plasticity. Considering the importance of a certain factor in habitat selection, however, it is crucial to take into account the scale where the process takes place and whether or not the selection decisions are adaptive. Cues that are used during the selection process may be either nested across different scales or their effect may take place on one level only. Our earlier work on heterospecific attraction emphasize its effect on habitat patch selec- tion but Timonen’s et al. (1994) study suggested that it may operate on smaller scales as well, such as on territory level. To examine the importance of het- erospecific attraction in more detail we conducted a set of experiments, which involved two intersecting spatial scales: habitat patches in the landscape and territories within a patch (Seppanen et al. unpub- lished). In the landscape level we manipulated the densities of resident birds (Parus spp.) inhabiting iso- lated forest patches embedded on an agricultural landscape. In the territory level experiment we stud- ied the settlement of migrant birds in relation to nest of Parus spp. In both experiment we used the Pied Flycatcher as a model species. 2) Colonisation rates If heterospecific attraction is used by Pied Fly- catchers (Ficedula hypoleuca) when selecting breed- ing habitat patches, they should first select patches of high titmice density. There was a tendency that Pied Flycatchers preferred high tit density patches to zero density patches. Both the first male flycatcher of each study plot and the average arrival day of males tended to be slightly earlier on high tit density plots than on zero plots, though not statistcally significantly (Seppanen et al. unpublished). Female flycatchers had no response whatsoever to the treatments. On the level of territories Pied Flycatchers were let choose between two nest boxes: the other was close (25m) and the other one was farther away (100m) from an active tit nest. If heterospecific attraction is 47 used at this scale, nest boxes adjacent to a tit nest would be preferred to more distantly located nest boxes. On the territory level, males clearly preferred nest boxes closer to the tit nest to the nest boxes far- ther away. In 25 set-ups, out of the total 36 cases, males selected first the box closer to the tit nest first. The observed distribution differed significantly from random pattern (1-tailed P=0.014 based on resam- pling). Likewise, females preferred settling to close by nest-boxes to those farther away (23 vs. 12 boxes selected first, respectively; 1-tailed P=0.043). It is known that female flycatchers select the nesting site according to the quality of the site and not according to the quality of the male (Alatalo et al. 1986). There- fore, we can consider female preference as independ- ent of male selection even though males arrive first from migration. In this experiment we cannot com- pletely rule out the possibility that both titmice and flycatcher, independently, chose higher quality sites. 3) Fitness consequences At least equally as important as the scale issue is whether the habitat selection is adaptive or not. Even though heterospecific attraction results in positive as- sociation between titmice and migrant densities, it does not necessarily indicate that using heterospecific cues is beneficial in terms of reproductive success (see Pulliam 1988; Martin 1998). We tested if het- erospecific attraction results in an increased fitness in the Pied Flycatcher by comparing reproductive suc- cess in patches where tits nested with patches devoid of tits using the abovementioned experimental arrangement. In general, the presence of titmice had a positive effect on the reproductive success of the Pied Fly- catcher (Seppanen et al. unpublished). Flycatchers were able to start egg-laying earlier, and the time delay from the female arrival to the first egg was on average 1.7 days shorter in patches where tits were present than in patches devoid of tits. Moreover, nestlings hatched 1.7 days earlier, and there were on average 0.6 more nestlings in broods in patches with tits than in patches without tits. Fledglings growing up in the neighborhood of titmice were larger than in the nests farther away. The results of these two experiments clearly indi- cate two things. First, heterospecific attraction seems to be an adaptive habitat selection strategy in terms of reproductive success in the Pied Flycatcher. Second, the effect of heterospecific attraction on habitat selec- tion and fitness is potentially working on two over- M. MONKKONEN and J. T. FORSMAN lapping scales; the presence of tits is first used in roughly comparing the quality of habitat patches at the landscape level followed by a more fine-tuned small scale nest site selection with preference to neighborhood of tit nest sites (see also Timonen et al. 1994). The results of the experiments also provide an example of nested habitat selection across scales with possibly cascading effects from titmice presence in the landscape in terms of higher occupation rates and reproductive success. DISCUSSION To summarize, boreal forest environment provide an example of a system where using resident species as cues is a profitable strategy in the breeding habitat selection of migrant birds. The experiments con- ducted on two continents indicate that heterospecific attraction of migrants to titmice increase the diversity and total abundance in local breeding communities. Migrant birds apparently use titmice abundance in comparing the relative quality (food and/or predators) of habitat patches. Our work at the biogeographic scale suggests that heterospecific attraction might not be restricted to boreal conditions but may be a wide spread process in forests bird communities. Analyti- cal modeling approach suggested that this sort of cue- ing from residents in most cases creates fitness bene- fits and is therefore selected for. This was further shown in experiments on titmice and Pied Flycatch- ers. We observed that flycatchers preferred areas of high tit density in their settlement and, moreover, their reproductive success was higher in patches with tits than without them. Recent theoretical study has suggested that posi- tive interspecific interactions are plausible, common and intensive in a wide variety of environmental con- ditions. For example, Dodds (1988) showed that in highly seasonal (“boom and bust’) environments, positive interactions, such as facilitation and mutual- ism, are selected for. Similarly, Bertness and Call- away (1994) suggested that positive interactions should be particularly common in communities under a severe physical stress (e.g., in highly variable or seasonal environments) and/or experiencing high consumer (predation) pressure. Bird communities in temperate and particularly in boreal settings occur in conditions that very likely meet these conditions: sea- sonality is pronounced producing a large difference in resource levels between summer and winter (Blake et al. 1994) and predation pressure on adults, nests and 48 young birds is heavy (Hanski et al. 1996; Solonen 1997): We were able to show fitness benefits for Pied Fly- catchers from settling in patches with titmice, but many earlier studies, conducted further south in tem- perate forests, have also shown competitive interac- tions between flycatchers (Pied Flycatcher or Col- lared Flycatcher, F. albicollis) and tits (Slagsvold 1975; Sasvari et al. 1987; Gustafsson 1987, 1988; Merila & Wiggins 1995). For example, Gustafsson (1987) showed that tits affected negatively the fitness of collared flycatchers. Potential reason for these seemingly contradictory results is in the difference in densities of residents among studies. In these south- ern studies titmice densities have usually been 2-5 times higher than the highest densities in our study areas (4—5 pairs/10 ha). Contrasting results in our and Gustafssonson’s (1987) study suggests that interspe- cific interactions may change along with different densities of potential competitors. This matches well with the results of our analytical model, which pre- dicted stronger attraction to residents at intermediate abundance (see above). The results of the experimental studies indicate that birds’ readiness to follow heterospecific cues varies among places and species. Not all species re- sponded positively to increasing resident densities in our experiments, and there obviously is much varia- tion in within-species responses to resident densities according to local conditions. In what conditions are species more apt to using heterospecific cues? Young birds selecting their first nesting sites are very likely more susceptible to use heterospecific cues than older individuals, which usually return to their previous year’s breeding site. Young birds might also be better off by using heterospecific residents than conspecific as cues because the presence of conspecifics may not reflect relative quality of the breeding sites in the cur- rent year but rather conditions in the past when site selection was made. Given the extensive between- year variation in conditions taking one’s cue from residents is quite likely more beneficial. It follows that the intensity of heterospecific attraction should vary according to the proportion of young individuals in the breeding population. We earlier referred to habitat generalists as being a species group apt to het- ereospecific attraction because they are not very tightly dependent on any particular habitat feature. In two Finnish experiments (Mo6nkk6nen et al. 1990; Thomson et al. unpublished), we found that the Chaffinch, an acknowledged habitat generalist, Heterospecific attraction showed the strongest response to increased titmice density. In the northernmost experiment (in this area the Chaffinch is relatively few in number) the Bram- bling had also a positive response to augmented tit densities (Forsman et al. 1998). Brambling is also a habitat generalist and, in addition, does not show site- fidelity to previous year’s breeding sites (Enemar et al. 1984; Mikkonen 1983). Brambling is therefore free to use external cues in order to find as good breeding habitat as possible. The results of the experiments on Pied Flycacher provide also some evidence about the processes be- hind the heterospecific attraction. We have earlier suggested that the presence and density of tits is used to make quick assessment of relative quality among habitat patches in the landscape. Our results showed, that indeed, high tit density patches and nest boxes closer to the nest of tit were colonized earlier indicat- ing that tits were used as a measure of the patch and site quality resulting in increased fitness. Female fly- catchers in patches where tits were present showed shorter time lags between arrival and the onset of egg-laying. Pied Flycatchers, as many other birds, are time constrained in their breeding and an early start of the breeding has a positive effect on the reproduc- tive output (e.g., von Haartman 1967; Lundberg & Alatalo 1992). Our results also suggest that flycatch- ers may also benefit from the tits through enhanced feeding efficiency or predator vigilance. At the landscape level, heterospecific attraction re- sults in a clumped distribution of individuals and species, a common pattern in nature (Hanski et al. 1993). This would explain the common observation that some seemingly suitable habitat patches remain empty. If colonization of patches is more generally dependent on the presence of individuals of other species, this would further complicate population dy- namics in patchy landscapes. For example, metapop- ulation models, for the sake of realism, should incor- porate interspecific interactions such as heterospecific attraction. Given the increasing fragmentation of landscapes taking interspecific interactions into ac- count when assessing individual dispersal and popu- lation viability is becoming increasingly important. For example, many old-forest associated resident species in Fennoscandia show declining population trends because of habitat loss and fragmentation (e.g. Haila & Jaérvinen 1990). This may have negative ef- fects on migrant species as well, if colonization rates in remaining patches depend critically on heterospe- cific cues. 49 Our experiments focused only on migrant birds’ habitat selection and fitness, and provided no evi- dence of whether heterospecific attraction results in a mutualistic relationship where also residents benefit from migrants’ presence or benefits are asymmetric accruing only to migrants. This remains as a chal- lenge for future studies. The experiments in temper- ate forests have so far encompassed only the high end of the resident density gradient; resident densities in nest-box studies may be unnaturally high compared to natural densities (e.g. Wesolowski et al. 1987). It would be interesting to see results from an experi- ment similar to our flycatcher work conducted in temperate settings where numerical response and fit- ness effects were studied over the whole gradient of resident densities. These results would further test for the importance and intensity of interspecific interac- tions (both negative and positive) and even reveal threshold conditions where originally positive inter- actions turn into negative ones. ACKNOWLEDGMENTS The studies reviewed in this paper were financially supported by the Academy of Finland (project #44085), Emil Aaltonen Foundation, and Faculty of Sciences, Univ. Oulu. REFERENCES Alatalo RV, Lundberg A & Bjérklund M (1982) Can the song of male birds attract other males? An experi- ment with the pied flycatchers Ficedula hypoleuca. Bird Beh 4: 42-45. Alatalo RV & Lundberg A (1984) Density-dependence in breeding success of the pied flycatcher (Ficedula hypoleuca). J Anim Ecol 53: 969-978. Alatalo RV, Lundberg A & Glynn C (1986) Female pied flycatchers choose territory quality and not male char- acteristics. Nature 323: 152-153. Barba E, Gil-Delgudo J & Monrés JS (1995) The cost of being late: consequences of delaying great tit Parus major first clutches. J Anim Ecol 64: 642-651. Bertness MD & Callaway R (1994) Positive interactions in communities. Trends Ecol Evol 9: 191-193. Blake JG, Hanowski JM, Niemi GJ & Collins PT (1994) Annual variation in bird populations in mixed conifer-northern hardwood forests. Condor 96: 381-— Bo) Cody ML (ed) (1985) Habitat selection in birds. Acade- mic Press, London. Dall SRX & Cuthill IC (1997) The information costs of M. MONKKONEN and J. T. FORSMAN generalism. Oikos 80: 197-202. Dodds WK (1988) Community structure and selection for positive and negative species interactions. Oikos 53: 387-390. Doligez B, Danchin E, Clobert J & Gustafsson L (1999) The use of conspecific reproductive success for breeding habitat selection in a non-colonial, hole- nesting species, the collared flycather. J Anim Ecol 68: 1193-1206. Durango S (1947) Om vanan hos vissa faglar att bosatta sig intil insektsamhiillen eller andra fagelarter. Fauna Flora 42: 85-205. Elmberg J, Poysé H, Sjoberg K & Nummi P (1997) In- terspecific interactions and co-existence in dabbling ducks: observations and an experiment. Oecologia 111: 129-136. Enemar A, Nilsson L & Sjéstrand B (1984) The compo- sition and dynamics of the passerine bird community in a subalpine birch forest, Swedish Lapland. A 20- year study. Ann Zool Fenn 21: 321-338. Forsman J (2000) Heterospecific attraction in breeding bird communities. Implications to habitat selection and species interactions in a landscape perspective. Acta Universitatis Ouluensis A 345. Oulu University Press, Oulu. Forsman JT, Ménkkonen M, Helle P & Inkeréinen J (1998) Heterospecific attraction and food resources in migrants’ breeding patch selection in northern boreal forest. Oecologia 115: 278-286. Forsman JT & Monkkonen M (2001) Responses by breeding birds to heterospecific song and mobbing calls under varying predation risk. Anim Behav, in press. Fretwell SD (1972) Populations in a seasonal environ- ment. Princeton University Press, Princeton. Garcia EFJ (1983) An experimental test of competition for space between blackcaps Sylvia atricapilla and garden warblers Sy/via borin in the breeding season. J Anim Ecol 52: 795-805. Gunn JS, Desrochers A, Villard M-A, Borque J & Ibarz- abal J (2000) Playbacks of mobbing calls of black- capped chickadees as a method to estimate reproduc- tive activity of forest birds. J Field Ornithol 71: 472- 483. Gustafsson L (1987) Interspecific competition lowers fitness in collared flycatchers Ficedula albicollis: an experimental demonstration. Ecology 68: 291-296. Gustafsson L (1988) Inter- and intraspecific competition for nest holes in a population of the collared flycather Ficedula albicollis. Ibis 130: 11-16. Haartman L von (1967) Clutch size in the pied fly- catcher. Proceedings of the XIV International Or- nithological Congress, pp 155-164. 50 Haila Y & Jarvinen O (1990) Northern conifer forests and their bird species assemblages. In: Keast A (ed) Biogeography and ecology of forest bird communities. pp 61-85. SPB Academic Publishing, The Hague, Netherlands. Hanski I, Kouki J & Halkka AJ (1993) Three explana- tions of the positive relationship between distribution and abundance. In: Ricklefs RE & Schluter D (eds) Species diversity in ecological communities. pp 108-116. Chicago University Press, Chicago. Hanski IK, Fenske TJ & Niemi GJ (1996) Lack of edge effect in nesting success of breeding birds in managed forest landscapes. Auk 113: 578-585. Harvey PH, Stenning MJ & Campbell B (1985) Individ- ual variation in seasonal breeding success of pied fly- catcher (Ficedula hypoleuca). J Anim Ecol 54: 391— 398. Helle P & Fuller RJ (1988) Migrant passerine birds in European forest succession in relation to vegetation height and geographical position. J Anim Ecol 57: 565-579. Herrera CM (1978) On the breeding distribution pattern of European migrant birds: MacArthurs theme reex- amined. Auk 95: 496-509. Hildén O (1964) Ecology of duck populations in the is- land group of Valassaaret, Gulf of Bothnia. Ann Zool Fenn 1: 153-279. Hildén O (1965) Habitat selection in birds. A review. Ann Zool Fenn 2: 53-75. Hogstad O (1995) Do avian and mammalian nest preda- tors select for different nest dispersion patterns of fieldfares Turdus pilaris? A 15-year study. Ibis 137: 484-489. Hutto RL (1985) Habitat selection by nonbreeding mi- gratory landbirds. In: Cody ML (ed) Habitat selection in birds. pp 455-476. Academic Press, London. Jarvinen A (1990) Changes in the abundance of birds in relation to small rodent density and predation rate in Finnish Lapland. Bird Study 37: 36-39. Kalela O (1952) Elainpopulaatioiden optimitiheydesta ja ryhmien valisesta valinnasta. Arch Soc Zool Bot Fenn 6: 130-136. Koskimies P (1957) Terns and gulls as features of habi- tat recognition for birds nesting in their colonies. Ornis Fenn 34: 1-6. Krebs JR (1971) Territory and breeding density in the great tit, Parus major L. Ecology 52: 2-22. Lack D (1948) Notes on the ecology of robin. Ibis 90: 252-279. Lundberg A & Alatalo RV (1992) The pied flycatcher. T & A D Poyser, London. MacArthur RH (1959) On the breeding distribution pat- tern of North American migrant birds. Auk 76: 318- Heterospecific attraction 325. Martin PR & Martin TE (2001a) Ecological and fitness consequences of species coexistence: a removal ex- periment with wood warblers. Ecology 82: 189-206. Martin PR & Martin TE (2001b) Behavioral interactions between coexisting species: song playback experi- ments with wood warblers. Ecology 82: 207-218. Martin TE (1998) Are microhabitat preferences of coex- isting species under selection and adaptive? Ecology 79: 656-670. Martin TE & Roper JR (1988) Nest predation and nest- site selection of a western population of the hermit thrush. Condor 90: 51-57. Merila J & Wiggins DA (1995) Interspecific competi- tion for nest holes causes adult mortality in the col- lared flycatcher. Condor 97: 445-450. Mikkonen AV (1983) Breeding site tenacity of the chaffinch Fringilla coelebs and the brambling F: mon- tifringilla in northern Finland. Ornis Scand 14: 36— 47. Morse DH (1989) American warblers. Harvard Univer- sity Press, Cambridge. Monkk6nen M (1990) Removal of territory holders causes influx of small-sized intruders on passerine bird communities in northern Finland. Oikos 57: 281— 288. Monkk6nen M, Helle P & Soppela K (1990) Numerical and behavioral responses of migrant passerines to ex- perimental manipulation of resident tits (Parus spp.): heterospecific attraction in northern breeding bird communities? Oecologia 85: 218-225. Monkkonen M, Forsman J & Helle P (1996) Mixed- species foraging aggregations and heterospecific at- traction in boreal bird communities. Oikos 77: 127— 136. Monkkonen M, Helle P, Niemi GJ & Montgomery K (1997) Heterospecific attraction affects community structure and migrant abundances in northern breed- ing bird communities. Can J Zool 75: 2077-2083. Monkkonen M, Hardling R, Forsman JT & Tuomi J (1999) Evolution of heterospecific attraction: using other species as cues in habitat selection. Evol Ecol 13: 91-104. Norrdahl K & Korpimaki E (1998) Fear in farmlands: how much does predator avoidance effect bird com- munity structure? J Avian Biol 29: 79-85. 51 O’Connor RJ (1981) Comparisons between migrant and non-migrant birds in Britain. In: Aidley DJ (ed) Ani- mal Migration. pp 167-195. Cambridge University Press, Cambridge. Partridge L (1978) Habitat selection. In: Krebs JR & Davies NB (eds) Behavioural ecology. An evolution- ary approach. pp 351-376. Sinauer Associates, Sun- derland, Massachusetts. Pulliam HR (1988) Sources, sinks, and population regu- lation. Am Nat 132: 652-661. Reed TM (1982) Interspecific territoriality in the chaffinch and great tit on islands and the mainland of Scotland: playback and removal experiments. Anim Behav 30: 171-181. Sasvari L, Torok J & Toth L (1987) Density dependence effects between three competitive bird species. Oe- cologia 72: 127-130. Slagsvold T (1975) Competition between the great tit Parus major and the Pied Flycatcher Ficedula hy- poleuca in the breeding season. Ornis Scand 6: 170- 190. Slagsvold T (1980) Habitat selection in birds: on the presence of other birds species with special regard to Turdus pilaris. J Anim Ecol 49: 523-536. Solonen T (1997) Effects of sparrowhawk predation on forest bird communities in southern Finland. Ornis Fenn 74: 1-14. Stamps JA (1988) Conspecific attraction and aggrega- tion in territorial species. Am Nat 131: 329-347. Suhonen J, Norrdahl K & Korpiméki E (1994) Avian predation risk modifies breeding bird community on a farmland area. Ecology 75: 1626-1634. Svardson G (1949) Competition and habitat selection in birds. Oikos 1: 157-174. Timonen S, Monkkonen M & Orell M (1994) Does competition with residents affect the distribution of migrant territories? Ornis Fenn 71: 55—60. Wesolowski T, Tomialoj¢ L & Stawarczyk T (1987) Why low numbers of Parus major in Bialowieza For- est—removal experiments. Acta Ornithol 23: 303- 316. Wiens JA (1989) The Ecology of bird communities, Vol I. Cambridge University Press, Cambridge. Willson MF (1976) The breeding distribution of North American migrant birds: a critique of MacArthur (1959). Wilson Bull 88: 582-587. Lege tu at sk ) Ji (Fiat due (tr ean LOG. Onn zai “fle = Na hal sh onan ti a oreo \OamodiunoitsGtu {LBQ)) Lt wan’ a See), Yoly A cal rmoathy€himetord Grtgiobantion ere, auld! nalin 20 bGebde cop west! (Dade > dof Virlerwnerat bn dies {A \emie rh oabtidmaa ,casrt eA eomakee ot G4! A pg deoleumcphy beete- i) Tl theewheeesd Mu tines) asin (BACT) ATE ita, wee boul 4 {' Thi tk “iT? tesatt Sfihe zboular po til erie hn it oi Ne (lo AE. te fariootiersial! inh iotioethsd Liven ak Arabeniq’ UrviFiline: vt sHrebroqob wrens & Belly Acti (ee Ti sh ilebSd uv oabdotie bol? areal hi bab re Adiosibeq? bid svilitainos son feewied at sre on Tita ii, age et Thwin yign ith (Mio (Wraawied abihinanas (My Tobit rigett fe A wlek a oy Paver arth! bites! safes re Ch) ade aint ay. LAywt Salt vie i slay eas ; bi?) Hie leon antioegee ance AAR Do@ bigeaie Land use dren sy oi: xy alan Tilo PORT Tge ew C i/@annel 1d cvdat ily wired Ignite, rustwomege lo sioatta (YA) T wali Ai bil ne eens ifenunthite behead! | i Taebe took Dies igo nn eaitoeie Ot icc) LENO deh ego? uff i MA 20h isiotinai ai non tho) 6609 | cl Nolieniqingh 21 A ile bri ionenochs nO NT heats Ladd lyslebowtbestl hon lan acaaetidatg bk beg go tilaG ynolauc| pam | mialorwtn theot) nridusl bone noiuiisgmol (C671) O ovuetalive bt felt bebe) ttnet , rity A bl wpsilideiiMeod ionmotiuT now on! tanthe -elogiiass tiiw conker tect attr.) Paar Perea quai ta nit ow! Apaleed 2T ddan ¥ 4 : i "=k wimun wil wi j ee , AN aie ‘ nh’ s <-> he s 21 duh ta 4 aoa? bf reel i(W 1 acer Lou ee (ped nmr | Ty pe rol, nba other th Micha Codd Lag bo st 405) al N telvepactiordeprall (che) Gitacaivadhy = Qrehe! on Agel MALL H arillot - \.(tthe! Wi ease donabiova wAKbsIq cael i ah Des | i = al Pie _ << = — pans &laricea 0 oe Nort ey inmmiansaatiorane AT ere renrerer i nye yoo Woltiechiin ied ad \ Svinese Sateaoryohig hawt tan ) Tet i Canin: Vie nr alt pee Tht bk Aton, debi Ree OSL a Herre orbretuayi te lov ov coeulin oli - farost lek SA 2) ORR bS. oObne Siero amine GAL Ae T a , ” -i00\ wie Qn ia dt: eet Gohedeaiey wc her SF eR bt 1 oi +e tn \ticoactat a yal (eeE esther A n Te scrdeh perp: be ‘ cy matic, } HOES" onrnilt ‘wlthen daidebeiabh ie | CMT SFR Cae Ge Larva rrvb Eade ipvoniy inti oiled Sb Siena NaLerr chet Tae ree 10 iilian Te tot. fet are i gro up au r rhe oi Lewit-awenh) % sli Eero, ii AN EMteDIS! fax zn wert he FT ei dinners Roos o> ML (ed) Hela al imaiaginlts RL tate if Hiekh. Mom OMS Alas | poze | (hate t ak ‘tarred errainane il epunt etry) rung + OO LOT eu? od Ht ty F 24m LaicdouD BOTe reed to Seah Beate Media rite | heii aiken Ao ‘wortutiow (ipietne A WAeeMhes 124 me ini wo) ee 9 (oe fF 1 ai ee i _»ibern iremeaiteaet “tary fet! mf i canis a 43 ST ail a ma C800 0 pendent Jace ing L I Eechagy-Se3 ot fimacology ad K is MV AWD) The pad ( » 0 wy ‘ Pp Hit iV59) Qe ta reed ng hunt Tes eNan Hala ant bare -_ {ey = Ornithol. Sci. 1: 53-61 (2002) SPECIAL FEATURE Interspecific segregation and attraction in forest birds The effect of a typhoon on the flocking and foraging behavior of tits Shin-Ichi SEKI* and Tamotsu SATO Kyushu Research Center, Forestry and Forest Research Institute, 4-11-16 Kurokami, Kumamoto 860-0862, Japan Abstract A typhoon, that struck Kyushu, the southernmost of the four main islands of Japan, in September 1999, causing extensive wind damage to forests, was found to have affected the flocking and foraging behavior of Varied Parus varius and Great Tits P. major. After the typhoon had passed, the tits tended to participate in mixed- species flocks and preferred to forage in the lower parts, rather than in the upper parts, of the trees. Also the proportion of plant products in the diet of the Varied Tit was re- duced. The population and average flock size of the tits, however, remained stable even after the typhoon. The abundance of plant products as food resources remained unchanged despite severe damage to the trees, but the vegetation cover was reduced, which probably increased the predation risk. The increase of mixed-species flocking may have resulted from the increased risk of predation; mixed-species flocking is thought to increase vigilance and foraging efficiency while not increasing intraspe- cific competition. Changes in diet and preferred foraging sites were also consistent with the increased predation risk hypothesis. We conclude that the changes in forag- ing and flocking behavior after the typhoon were mainly due to the increased preda- ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2002 tion risk caused by the reduced vegetation cover. Key words Disturbance by storms is among the most important factors in determining the structure and species com- position of forest biotic communities (White 1979; Tanner et al. 1991). Bird populations are also affected by storms, both directly as a result of the stress of storms and indirectly as a result of habitat modifi- cation (Askins & Ewert 1991; Lynch 1991; Waide 1991; Wunderle et al. 1992). Although many studies have reported on the population dynamics of birds after storms, little attention has been paid to their be- havioral response. In Kyushu, the southernmost of the four main is- lands of Japan, typhoons are the most significant agent of forest disturbance (Yamamoto 1992). The passage of a powerful typhoon through Kyushu in 1999 provided us with a rare opportunity to docu- ment the effects of a storm on avian behavior. The aim of this paper is to describe the effects of this se- vere typhoon on the flocking and foraging behavior of the Varied Tit Parus varius and the Great Tit P. (Received 16 May 2001; Accepted 27 August 2001) * Corresponding author, E-mail: seki @ffpri.affrc.go.jp 53 Foraging site, Forest disturbance, Tit, Typhoon, Winter flocking major. Various factors affecting avian flocking and foraging behavior previously have been revealed, in- cluding: food abundance (Berner & Grubb 1985; Székely et al. 1989; Kubota & Nakamura 2000), weather conditions (Ekman 1984; Grubb 1987; Naka- mura & Shindo 2001), inter- and intraspecific compe- tition (Ekman 1987; Alatalo & Moreno 1987), abun- dance of congeneric species (Matthysen 1990), pred- ator abundance (Suhonen 1993; Kullberg 1998), and the distribution of protective cover (Krams 1996). Despite the number of these studies, few have exam- ined the effects of storms. To reveal the impact of the typhoon we compared the diets, foraging heights, and mixed-species flock attendance of the two species be- tween the winters preceding and following the ty- phoon. STUDY AREA AND METHODS The study was carried out at Tatsutayama Ex- perimental Forest in Kumamoto, Kyushu, Japan (32°49'N, 130°44'E, 28.4ha, 48-152 m asl). Tatsu- tayama is an isolated hill forest area of about 450 ha, S. SEKI and T. SATO and the Experimental Forest is located on the south- western slope of the hill. It is mostly covered with secondary evergreen broad-leaved forests 40-50 years of age, with some small coniferous plantations. The dominant tree species in this secondary forest is Castanopsis cuspidata, with an average height of about 17 m. On 24 September 1999, a severe typhoon (9918 Bart) struck Kyushu and its eye passed within 40 km of the Experimental Forest moving at 40kmh ', with a pressure at the eye of 945hPa. At the Kumamoto Local Meteorological Observatory, 2.5km west of our study site, a wind speed of over 10ms | was recorded for seven hours, with a maximum speed of 49.0ms | (Japan Meteorological Agency 1999). To examine the structural damage caused to trees by typhoon 9918 Bart, we established five 2m*50m transects within which we measured all the trees above breast height, and categorized the trees into three groups: 1) Sound; mostly healthy with slight or little defoliation; 2) Injured; trunk broken, trunk lean- ing, trunk down, limbs broken, or severely defoliated, and 3) Dead; died within two years of the typhoon. To examine the effect of the typhoon on the popu- lation and flocking behavior of tits, we conducted an- nual bird censuses from 15 November to 31 Decem- ber, from 1996 to 1999. Since the number of tits and their flocking and foraging behavior change continu- ously throughout the year, we confined the study to this short period when the social organization of tits is relatively stable (Saitou 1978; Gosler 1993; Kubota & Nakamura 2000). Moreover, since we could not find an appropriate control site, we compared the data sets from the four years defining the three years 1996-1998 as control years and 1999 as a disturbed year. To reduce the possibility of misidentifying the ‘among-year-fluctuation’ to be an effect of the ty- phoon, we regarded only those values in 1999 that were significantly different from those of the other three years, as having been disturbed by the typhoon. Although some typhoons affect our study site every year, no severe typhoons with maximum wind speeds of over 40ms | were recorded in the control years nor during the five years prior to the research. An elliptical, 3.8 km, census route was established in the study site, and this was surveyed six times each year between 0730 and 1030. The species and num- ber of all birds observed within 25 m on either side of the route were recorded. We carefully avoided re- peated counts of the same flock. Since the vegetation was not dense in our study site, we assumed that de- 54 tectability within 25m of the transect line was high and constant, and that the number of tits recorded within the area reflected their relative densities. Each tit encountered was categorized as either par- ticipating in a mixed-species flock, in a mono-spe- cific flock, or as being solitary. Three criteria were used to define a flock: (1) all members remain within about 20 m of each other; (2) flock members remain together for at least three minutes, and (3) members move at least 30m in the same direction. When more than two birds of two species were associating to- gether, they were defined as a mixed-species flock, following Bell’s definition (1986). In our study area, in addition to Paridae other species were also recorded as participating in mixed- species flocks. These included: Long-tailed Tit Aegit- halos caudatus, Japanese White-eye Zosterops japon- icus, Japanese Pygmy Woodpecker Dendrocopos kizuki, Goldcrest Regulus regulus, Ashy Minivet Pericrocotus divaricatus, Red-flanked Bushrobin Tarsiger cyanurus, and Japanese Bush Warbler Cettia diphone. It was not possible to determine the social organi- zation of the Varied and Great tits in detail, since in- dividual were not marked, therefore, we used the term ‘mono-specific units’ to describe the basic so- cial organization that included solitary individuals, mono-specific flocks, and the same species members belonging to mixed-species flocks. Mixed-species flock attendance rate was calculated for each species as the proportion of individuals attending the mixed- species flocks in relation to the total number of indi- viduals observed. To examine the effects of abiotic factors, we com- pared the weather parameters on census dates in the control and disturbed years, daily average tempera- ture and wind velocity, recorded at the Kumamoto Local Meteorological Observatory. These data were provided by the Japan Meteorological Agency. We recorded the diet and the foraging heights where successful foraging attempts by tits were ob- served, from November to January each year. Forag- ing heights were divided into four categories: (1) upper layer (>12m), (2) middle layer (6-12 m), (3) lower and shrub layers (0O—6m), and (4) on ground. Unfortunately, we could not divide these observations according to whether or not birds attended mixed- species flocks, because we recorded foraging behav- ior independently of the bird censuses. The abundance of acorns and other seeds was esti- mated using litter traps. Ten round traps, each with an Typhoon effects on a Tit flock opening of 0.58 m’, were placed in the C. cuspidata forest and we collected the contents over the three winter month of November, December and January. We sorted out seeds and mature acorns from the con- tents, oven-dried (70°C, 72h) and weighed them. Small immature acorns were excluded from the abun- dance measurement, since tits seldom fed on them (Higuchi 1975). We used the Kruskal-Wallis test to analyze the flock size data. To evaluate the differences among groups, we employed Dunn’s procedure following Zar’s manual (1999). We simply used chi-square test (contingency table) to analyze if the mixed-species flock attendance, foraging sites, and food items were independent of year, because our sample sizes were large enough to use this test without correction in most cases. When those variables were not independ- ent of year, we then evaluated the contribution of each cell of the contingency tables using adjusted residuals (d;,), which are approximate to z scores (Everitt 1977). RESULTS 1) Disturbance by the typhoon After the typhoon passed, 8.4% of the trees above | breast height died and 28.2% were severely injured | (their limbs and/or trunks were broken, or they were uprooted; see Table 1). The wind damage to the | canopy was greater than that to the sub-canopy, | 13.6% of the canopy trees (with a diameter at breast | height of over 20cm) died and 72.7% were severely . damaged (Table 1). Most of the trees in our study site were extensively | defoliated by the strong typhoon winds, and annual | litter fall in 1999 was 1.3- to 2.0-fold greater than _ during the three years prior to the typhoon (Sato un- published). As a result of this destruction, the average canopy cover was reduced from 95.4% to 87.2% | (Saitou, S. personal communication). We have only limited information on the winter predation risk relat- | ing to this decrease of cover. The mortality of young Table 2. | Fisheries Research. birds, however, was higher during the breeding sea- son following the typhoon than in the preceding sea- sons. For example, 52.3% of young Great Tits disap- peared within two weeks after fledging in 2000, whereas only 28.6% had disappeared in 1997 (chi- square test, y7=5.55, df=1, P=0.019). We also ob- served several attacks on family flocks of Great Tits in 2000, but not in other years. These attacks in- cluded three by Jungle Crows Corvus macrorhynchos (one was successful), and one successful attack by a Japanese Lesser Sparrowhawk Accipiter gularis (Seki unpublished). Despite the severity of the habitat modification re- sulting from the passage of the typhoon, the differ- ences among years were not significant for either temperature or wind velocity (Table 2). 2) Number of tits The number of individual Varied Tits recorded did not differ significantly among years (Table 3). The number of Great Tits observed in the three years 1997, 1998 and 1999, did not differ significantly, al- though the number of birds seen in 1996 differed sig- nificantly from the number in 1999 (Dunn’s proce- dure, 1996 vs 1999, Q=3.50, P<0.005; 1997 vs 1999, Q=1.00, P>0.50; 1998 vs 1999, Q=0.65, P>0.50). 3) Size and composition of flocks Neither the size of mono-specific units, nor the size of mixed-species flocks containing each species of tit, differed among the four years for either tit species Table 1. Structural damage to trees in the study area caused by typhoon 9918 Bart. Figures show the number of trees in each category. DBH class(cm) <10 10-20 20-30 30-40 >40 Sound 226 9 5 1 Injured 61 14 24 6 2 Dead 18 8 4 2 Weather parameters on census dates recorded at the Kumamoto Local Meteorological Observatory (mean+SD). Data 1996 1997 1998 1999 F P Temperature (°C) Voi E33) 9.1+4.9 8.4+2.4 5.8+3.9 0.83 0.50 Wind velocity (m/s) 1.6+0.3 1.9+0.6 1.6+0.2 2.140.4 2.24 0.12 S. SEKI and T. SATO Table 3. Comparison of observed number of individuals and the size of flocks in the control years and in the year disturbed by the typhoon (mean+SD). Number of Size of Size of Species Year individuals N'! mono-specific N* — mixed-species §_N? (per census) units” flocks* a) Varied Tit Control years 1996 23.8+8.6 6 2.6+1.8 55 16.7£9.8 10 1997 24.2+3.6 6 Dal eile 69 Si less 14 1998 18.2+3.2 6 PP \ ae UP) Sy 9.4+6.0 11 Disturbed year 1999 18.8+3.5 6 1.9+0.8 59 11.0+10.0 30 Kruskal-Wallis test (df=3) Hc=6.48 He=4.77 Hce=5.15 P=0.09 P=0.19 P=0.16 b) Great Tit Control years 1996 6.5+1.4* 6 1.5+0.8 26 11.9+9.9 14 1997 Moy) 2=3)67/ 6 1.7+0.8 54 OMe Se2) 19 1998 16.2+1.7 6 1.9+1.0 51 12.6+9.6 17 Disturbed year 1999 19.3+6.2 6 2.0+1.0 59 10.6+9.7 36 Kruskal-Wallis test (df=3) He= 14.04 Hc=5.65 He=1.32 P=0.003 P=0.13 P=0.72 ' Frequency of census * We used the term ‘mono-specific units’ to mean a basic social organization, including solitary individuals, mono-specific flocks, and the same species members belonging to mixed-species flocks. > Number of observed units or flocks each year * Size of mixed-species flocks containing each species of tit that also include the number of species other than tits: Long-tailed Tit, Japanese White-eye , Japanese Pygmy Woodpecker, Goldcrest, Ashy Minivet, Red-flanked Bushrobin, and Japanese Bush Warbler. *Significant difference from 1999 (P<0.05, Dunn Procedure). Table 4. Mixed-species flock attendance rates (%) in the control years and in the year disturbed by the typhoon. Control years Disturbed year 7-test 1996 N' 1997 N!' 1998 N! 1999 N! df x P Tits Varied Tit 30.8 143 234 145 17.4 109 SA 13) 3 37.26 <0.001 Great Tit 59.0 39 §©=— 3:1.9 92 36.1 97 68.1 116 3) 38.45 <0.001 Long-tailed Tit 50.8 252 483 201 56.6 234 91.1 169 3 89.00 <0.001 Other species Japanese White-eye Od 5545" 1258) 2405 Lee s58 35.3 292 3 91.63 <0.001 Japanese Pygmy Woodpecker 23.3 47 34.0 60 34.8 46 Silo ate 3 10.51 0.015 Goldcrest DDL 2 0.0 Sil 14 100.0 11 = = Ashy Minivet 0.0 19 10.5 20 0.0 12 13.3 15 = = Red-flanked Bushrobin 0.0 44 P.) 42 2.2 45 Sek) = = Japanese Bush Warbler 0.0 55 0.0 77 1.3 76 0.0 85 = = ' Sample size (N) shows the total number of individuals observed each year. (Table 3). was higher than in the three control years, not only The mixed-species flock attendance rate of the for both Parus species but also for the Long-tailed Great Tit was always higher than that of the Varied Tit, the Japanese White-eye and the Japanese Pygmy Tit within each year, even in the disturbed year (Table | Woodpecker (Table 4). The results in 1999 con- 4). The mixed-species flock attendance rate in 1999 _ tributed most to the dependence on year in all these 56 Typhoon effects on a Tit flock %1997 41998 @1999 70 60 50 40 30 20 10 0 Proprotion of use (%) Fig. 1. Upper Layer Middle Layer Lower & Shub Layers Ground 20 30 40 50 Proprotion of use (%) 60 70 The foraging heights of a) Varied Tit and b) Great Tit in the control years and in the year disturbed by ty- phoon 9918 Bart. The observed frequency of feeding attempts each year is shown in Table 5. species (adjusted residuals, Varied Tit, diiceg, jo99= 5.65, P<0.001; Great Tit, dig 1999=5-45, P<0.001; Long-tailed Tit, dyixeg 1999 = 9-06, P<0.001; Japanese White-eye, dire 1999 9-24, P<0.001; Japanese Pygmy Woodpecker, diixeq 1999 2-92, P=0.004). For the other four species, we could not perform statisti- cal analyses, because the number of individuals par- ticipating in mixed-species flocks was too small. Al- though we do not have quantitative data on the leader-follower relationships in the mixed-species flocks, if Long-tailed Tits were present they usually led the flock. In those flocks without Long-tailed Tits, we did not find any tendency for the Varied Tit or the Great Tit to be the leader (Seki, pers. obs.). 4) Foraging heights Varied Tits were only observed foraging in trees and never on the ground during the study period (see Fig. 1). The observed frequency of each forest layer used was not independent of year in the Varied Tit (chi-square test, y°=27.31, df=6, P=0.001). After the typhoon’s passage, the Varied Tit foraged signifi- cantly less in the upper layer and significantly more in the middle layer instead (adjusted residuals, upper layer, dipper, 1999 4.48, P<0.001; middle layer, d.niddle, 1999 2-29, P=0.001; lower and shrub layer, di wer, 1999 1-47, P=0.14). The Great Tit also foraged less in the upper layer after the disturbance and shifted to the lower part of the trees, although the ef- fect of year was not significant (chi-square test, 4 =15.12, df=9, P=0.09). 5) Diet components In our research area, plant materials accounted for 47.3 to 53.5% of the Varied Tit’s diet in the control years, whereas the proportion of plant materials in the diet decreased to 30.1% in the disturbed year (Table 57 5). Since unspecified items were assumed to be inver- tebrates (on the basis of the foraging site), we divided their diet into two parts, plant diet, and non-plant diet. The observed frequency of each group foraged was not independent of year (chi-square test, 12533; df=3, P<0.001), and the results in 1999 contributed most for the dependence on year (ad- justed residuals, dojant, 1999= — 3.43, P<0.001). The Great Tit also fed on plant products although the proportion was not high (Table 5). The Great Tit seldom ate acorns, probably because of physical con- straints. In fact, we observed only two incidences of Great Tits eating acorns, in both cases they were bro- ken due to drying. The diet composition of the Great Tit was independent of year during the study period (chi-square test, y’=1.09, df=3, P=0.78). 6) Plant abundance The abundance of plants, especially acorns, was highly variable among years (Table 6). Although acorns and seeds were not abundant in 1999, the abundance was still greater than that of the previous winter, which was one of the control years undis- turbed by severe typhoons. DISCUSSION We observed three changes in tit behavior after the severe disturbance caused by the typhoon. First, at- tendance rate in mixed-species flocks increased after the disturbance compared with the previous three years. Secondly, the proportion of plants in the Varied Tit’s diet was reduced, although there was no change in the Great Tit’s diet. Finally, the tits tended to prefer the lower part of the trees to the upper part as their foraging sites. The population and average flock size of the tits, however, were rather stable in spite of the S. SEKI and T. SATO Table 5. Annual variation in the winter diet (%) of Varied and Great Tits from 1996-1998. Varied Tit Great Tit Control years Disturbed year Control years Disturbed year 1996 1997 1998 1999 1996 1997 1998 1999 Plant materials Acorns Byosy 243) DILL) 19.4 45 00 0.0 0.0 Seeds 15.0 23.0 25.6 10.7 Gi) 7/9) IS 4 2 Non-plant materials Invertebrates Caterpillars 12.5 10.8 14.0 11.7 AMY-S) MS, IS) 20.0 Pupae PS SE) OO 49 eo) NMS Fell 4.6 Spiders WVs3}m lle4h 7s} 1.9 Deo MOS) Fai) 9.2 Lepidoptera(adults) 0.0 14 0.0 1.9 A) 246) WES 3.1 Hemiptera 13 14 0.0 1.9 0.0 53 3.8 3.1 Coleoptera(adults) OL0) pae4n23 5.8 0.0 0.0 0.0 3.1 Diptera(adults) 0.0 14 & 0.0 2.9 0.0 26 3.8 1.5 Hymenoptera(adults) 0.0 0.0 0.0 0.0 23 0.0 0.0 1.5 Other Invertebrates 1255 ASO 8.7 31.8 10.5 11.5 12.3 Unspecified WS 24-3) 2019 30.1 210 2613" 192 32.3 Total number of items 80 74 43 103 44 38 26 65 Table 6. Abundance (kgha ') of acorns and other seeds in the control years and in the year disturbed by the typhoon. Acorns Other seeds Control years 1996 456.77 1.66 1997 75.92 1.91 1998 1.64 1.34 Disturbed year 1999 4.38 1.43 extensive habitat modification. Foraging and flocking decisions are very sensitive to biotic and abiotic environmental changes (Matthy- sen 1990); since small passerines suffer from energy shortages during the cold and short winter days (Jans- son et al. 1981), they always face a trade-off between the benefit of efficient foraging and the cost of preda- tion risk when they make behavioral decisions (re- viewed in Lima & Dill 1990; Matthysen 1990; Suho- nen et al. 1993). Based on previous studies dealing with wind disturbance (Askins & Ewert 1991; En- gstorm & Evans 1990; Grant et al. 1997; Lynch 1991; Waide 1991; Wunderle et al. 1992), we had expected two major environmental changes: fluctuation of food abundance and the decrease of vegetation cover. Fluctuations in food abundance, however, are un- likely to have caused the behavioral changes in the present study. Although the strong typhoon winds in 58 1999 blew off large numbers of immature acorns and seeds (Sato unpubl.), the production of mature acorns and seeds during the following winter was still greater than that of poor crop years, such as 1998 (Table 6). The seed production of tree species in gen- eral is highly variable between years regardless of the effects of wind storms (Higuchi 1975; Kelly 1994). In 1998, Varied Tits selectively ate acorns despite their low abundance, but did not do so in 1999. Nei- ther Great Tit nor Varied Tit increased their mixed- species flock attendance rate in 1998, but in 1999 they did. The decrease of plant products after the ty- phoon seems unlikely to be the cause of behavioral changes of tits. Furthermore, potential invertebrate prey for tits did not markedly increase or decrease even in the disturbed winter. Increases in inverte- brates, which might have caused the Varied Tit to alter its diet, have often been reported after storms, but usually occur from spring following the damage (Thompson 1983; Furuta et al. 1984). Decreases in invertebrates, which might happen after canopy de- struction and thus lead to mixed-species flocking, could not have been so severe in the study area, be- cause Varied Tits risked shifting their diet from plant products to invertebrates in the disturbed winter. Reduced cover could have caused the observed be- havioral changes as a result of the increased risk of predation (Engstorm & Evans 1990; Thiollay 1997; Typhoon effects on a Tit flock Grant et al. 1997). The reduced degree of vegetation cover was observed also in our study area, and it con- tinued throughout the winter, since refoliation and sprouting could not proceed so rapidly in warm tem- perate forests (Bellingham et al. 1996). Although we have only limited information on the winter predation risk, the increased mortality of young during the breeding season after the typhoon would support our increased predation risk assumption. Higher predation risk increases the mixed-species flock attendance rate (Székely et al. 1989; Ekman 1989; Matthysen 1990; Suhonen et al. 1993), because heterospecific flocking is assumed to increase vigi- lance and foraging efficiency but does not increase intraspecific competition (reviewed in Lima & Dill 1990; Bednekoff & Lima 1998). Moreover, since in- dividuals in heterospecific flocks are probably not in- fluenced by long-term social bonds, interspecific as- sociations are more sensitive than intraspecific asso- ciations to environmental changes (Matthysen 1990), such as the disturbance caused by a typhoon. Our findings, that mixed-species flocking increased once vegetation cover had been reduced, were consistent with the increased predation risk hypothesis. The av- erage size of mixed-species flocks in 1999 did not in- crease in spite of the increased mixed-species flock attendance rates, presumably because of the smaller size of mono specific units of the Long-tailed Tit, nu- clear species for mixed-species flocks (averaging 7.4 in 1999 and 7.9-12.0 during 1996-98). Predation risk also affects the diet of prey species. Numerous studies have demonstrated that the pres- ence of predators increases the proportions of prey items found and consumed in a safer place with a safer method (Lima 1988a; Lima 1988b; Suhonen 1993; Suhonen et al. 1993; Krams 1996; Kullberg 1998, to name a few). The Varied Tit preferred the acorns of C. cuspidata before the disturbance, since other species did not take it in spite of its high ener- getic efficiency. However, acorns have some disad- vantages compared to invertebrates in their degree of safety from predators: (1) The Varied Tit is less vigi- lant while handling them since its ability to detect a predator is greater when it is raising its head (Lima et al. 1999); (2) predators might be able to locate the tit on the basis of its loud hammering noise (Székely et al. 1989); and (3) unspoiled acorns are distributed in the outer parts of trees where the vegetation cover was destroyed most severely. We suggest that the Varied Tit gave up foraging on acorns under the de- creased vegetation cover, as a trade-off between effi- 59 cient foraging and safety from predators while forag- ing on acorns. Furthermore, this shift in the diet could also be attributed to the increased mixed-species flock attendance rate of the Varied Tit, since the bene- fit of copying other species might be greater when they feed on cryptic invertebrates (Waite & Grubb 1988). The Great Tit, in contrast, did not show any changes in its diet, which might be because its diet consisted of invertebrates that are easier to handle than plant products, and because they are distributed in various parts of the trees. The changes in the preferred foraging heights after the typhoon might also be due to the predation risk. The reduction of cover was most notable in the upper layer, which the tits avoided after the disturbance. However, the new foraging heights, where the tits ap- peared after the typhoon, differed between the Varied and Great tits. There are several possible reasons for this difference, such as inter-specific relationships (Nakamura & Shindo 2001) and the distribution of food resources, although we cannot discuss these in detail as we know little about the quality of each for- aging substrate, the dominant-subordinate relation- ship in the mixed-species flocks, or the movement of leading species. We therefore conclude that the changes in foraging and flocking behavior after the typhoon were mainly due to the increased predation risk caused by reduced vegetation cover. ACKNOWLEDGMENTS We thank Akira Endo for helpful comments on an earlier draft of the manuscript, and Hiroyoshi Higuchi, Noritomo Kawaji, Teruaki Hino, Masahiko Nakamura, and Norio Sahashi for helpful discussion. We also thank Satoshi Saitou for valuable information. Unpublished meteorological data was offered by Japan Meteorologi- cal Agency, through database (common basic data), de- veloped by Computer Center for Agriculture, Forestry and Fisheries Research. REFERENCES Askins RA & Ewert DN (1991) Impact of Hurricane Hugo on bird populations on St. John, U.S. Virgin Is- lands. Biotropica 23: 481-487. Alatalo RV & Moreno J (1987) Body size, interspecific interactions, and use of foraging sites in tits (Paridae). Ecology 68: 1773-1777. Bednekoff PA & Lima SL (1998) Randomness, chaos S. SEKI and T. SATO and confusion in the study of antipredator vigilance. Tree 13: 284-287. Bell HL (1986) The participation by cuckoos in mixed- species flocks of insectivorous birds in South-Eastern Australia. Emu 86: 249-253. Bellingham PJ, Kohyama T & Aiba S-I (1996) The ef- fects of a typhoon on Japanese warm temperate rain- forest. Ecol Res 11: 229-247. Berner TO & Grubb TC Jr (1985) An experimental analysis of mixed-species flocking in birds of decidu- ous woodland. Ecology 66: 1229-1236. Ekman J (1984) Density-dependent seasonal mortality and population fluctuations of the temperate-zone willow tit (Parus montanus). J Anim Ecol 53: 119-134. Ekman J (1987) Exposure and time use in Willow Tit flocks: the cost of subordination. Anim Behav 35: 445-452. Ekman J (1989) Ecology of non-breeding social systems of Parus. Wilson Bull 101: 263-288. Engstorm RT & Evans GW (1990) Hurricane damage to red-cockaded woodpecker (Picoides borealis) cavity trees. Auk 107: 608-610. Everitt BS (1977) The analysis of contingency tables. Chapman and Hall, London. Furuta K, Ando S & Takahashi I (1984) A trial of mass trapping of Ips typographus japonicus Niijima after an extensive wind damage in Hokkaido. Appl Ent Zool 19: 518-519. Gosler A (1993) The great tit. Hamlyn, London. Grant GS, Craig P & Trail P (1997) Cyclone-induced shift in foraging behavior in flying foxes in American Samoa. Biotropica 29: 224-228. Grubb TC Jr (1987) Changes in the flocking behaviour of wintering English titmice with time, weather, and supplementary food. Anim Behav 35: 794-806. Higuchi H (1975) Comparative feeding ecology of two geographical forms of the varied tit, Parus varius varius in Southern Izu Peninsula and P. v. owstoni in Miyake I. of the Izu Is. Tori 24: 15-28. Jansson C, Ekman J & Bromssen A von (1981) Winter mortality and food supply in tits Parus spp. Oikos 37: 313-322. Japan Meteorological Agency (1999) Geophysical Re- view No 1202. Japan Meteorological Agency, Tokyo. Kelly D (1994) The evolutionary ecology of mast seed- ing. Trends Ecol Evol 9: 465-470. Krams IA (1996) Predation risk and shifts of foraging sites in mixed willow and crested tit flocks. J Avian Biol 27: 153-156. Kubota H & Nakamura M (2000) Effects of supplemen- tal food on intra-and inter-specific behaviour of the varied tit Parus varius. Ibis 142: 312-319. 60 Kullberg C (1998) Spatial niche dynamics under preda- tion risk in the willow tit Parus montanus. J Avian Biol 29: 235-240. Lima SL (1988a) Vigilance and diet selection: a simple example in the dark-eyed junco. Can J Zool 66: 593-596. Lima SL (1988b) Vigilance and diet selection: the clas- sical diet model reconsidered. J Theor Biol 132: 127- 143. Lima AL & Dill LM (1990) Behavioral decisions made under the risk of predation: a review and prospectus. Can J Zool 68: 619-640. Lima SL, Zollner PA & Bednekoff PA (1999) Predation, scramble competition, and the vigilance group size effect in dark-eyed juncos (Junco hyemalis). Behav Ecol Sociobiol 46: 110-116. Lynch JF (1991) Effects of Hurricane Gilbert on birds in a dry tropical forest in the Yucatan Peninsula. Biotropica 23: 488-496. Matthysen E (1990) Nonbreeding social organization in Parus. In: Power DM (ed) Current Ornithology, Vol 7. pp 209-249. Plenum Press, New York. Nakamura M & Shindo N (2001) Effects of snow cover on the social and foraging behavior of the great tit Parus major. Ecol Res 16: 301-308. Saitou T (1978) Ecological study of social organization in the great tit, Parus major L. I: Basic structure of the winter flocks. Jpn J Ecol 28: 199-214. Suhonen J (1993) Predation risk influences the use of foraging sites by tits. Ecology 74: 1197-1203. Suhonen J, Halonen M & Mappes T (1993) Predation risk and the organization of the Parus guild. Oikos 66: 94-100. Székely T, Szép T & Juhasz T (1989) Mixed species flocking of tits (Parus spp.): a field experiment. Oe- cologia 78: 490-495. Tanner EVJ, Kapos V & Healey JR (1991) Hurricane effects on forest ecosystems in the Caribbean. Biotropica 23: 513-521. Thiollay J-M (1997) Disturbance, selective logging and bird diversity: a Neotropical forest study. Biodivers Conserv 6: 1155-1173. Thompson D (1983) Effects of Hurricane Allen on some Jamaican forests. Commonw For Rev 62: 107-115. Waide RB (1991) The effect of Hurricane Hugo on bird populations in the Luquillo experimental forest, Puerto Rico. Biotropica 23: 475-480. Waite TA & Grubb TC Jr (1988) Copying of foraging locations in mixed-species flocks of temperate-decid- uous woodland birds: An experimental study. Condor 90: 132-140. White PS (1979) Pattern, process, and natural distur- bance in vegetation. Bot Rev 45: 229-299, Typhoon effects on a Tit flock Wunderle JM Jr, Lodge DJ & Waide RB (1992) Short- eration in primary evergreen broad-leaved forests of term effects of Hurricane Gilbert on terrestrial bird western Japan. Bot Mag Tokyo 105: 29-45. populations on Jamaica. Auk 109: 148-166. Zar JH (1999) Biostatistical analysis (fourth edition). Yamamoto S (1992) Gap characteristics and gap regen- Prentice Hall Press, New Jersey. 61 Te Weert! Dart arabe tiny © ot a ovdeT ga ied coq mleew wallets! Cee | Darth tines Auras HOON TEMA Bat Heetaeteten essere 3 of (nepealian Dabber? temdbhan | ‘Sie et | et eraee ie r oA ae | ‘LOOK The et —. ke 7 Previn etebys ‘teoa $l ites Vy ng a ye F Ply) ‘as éi fr iltonia} amt os es ~~ ve wn} 4 1) Degaitydepenaes oQnia ' Iv si Pil a aw te icioeeerale-~zik ; Ems 3 “inet PA s Hamper’ eval i liaes eur ~ F : dos MEN vk inet | a =) le Oia te Baty a ¥ 3 y ——- - 7 » © Gate-t bits ene Vv a . 2 ae OP ty, wT > - ie in ant baa \ Re (toe VY & Uraley JK ity i Pes Yienuiy 6 Me i Sf 1} ; ow 5 Onditonce, eheetiert op wIGy 2 re ae i mat Cis: rap? ‘ wig ta cries "=< ul vinnie ; rw het, Cm =O . Need Uys LOR, The thn of Hale = Flay a prénd=s “aan ity -Seperding ~, ieee, + tek 4h “S ey , wy Boichee-PYS => . tb bed ? on. Paeierh. poem, SPECIAL FEATURE Ornithol. Sci. 1: 63-69 (2002) Interspecific segregation and attraction in forest birds Foraging mode shifts of four insectivorous bird species under temporally varying resource distribution in a Japanese deciduous forest Masashi MURAKAMI** Biodiversity Group, Institute of Low Temperature Science, Hokkaido University, Sapporo, 060-0019, Japan ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2002 Abstract Temporal changes in the foraging habitat of four forest bird species and the distribution pattern of arthropod populations were investigated. The abundance and distribution of arthropods changed drastically with the season within the forest. Lepidoptera larvae were most abundant in the canopy in the first three weeks after budbreak; their numbers decreased rapidly during mid-June. In contrast, on the forest floor, the larvae were abundant from early to late June. The foraging height of the Narcissus Flycatcher Ficedula narcissina changed in parallel with the distribution pattern of Lepidoptera larvae. Three other species, the Great Tit Parus major, Marsh Tit P. palustris, and Eastern Crowned Warbler Phylloscopus coronatus, however, did not change their foraging heights; they continued to forage in the canopy. These dif- ferences are probably due to the greater preference of the flycatcher for Lepidoptera larvae compared with the other three species. The three other species switched from feeding on Lepidoptera larvae to spiders or other arthropods in mid June, when the number of Lepidoptera larvae decreased in the canopy. The results of this study sug- gest that the abundance and distribution of arthropods and differences in foraging tac- tics among bird species considerably affect avian foraging habitat. The foraging be- havior of three species of forest birds revealed species-specific responses to spatio- temporal fluctuations in the distribution of resources. Key words Forest bird, Foraging tactics, Lepidoptera larvae Since Lack’s (1971) findings of a clear-cut parti- tioning of foraging substrates in tree crowns among tit species, other studies have confirmed a close rela- tionship between bird foraging habit and resource distribution (Alatalo 1980; Holmes & Schultz 1988). Forest bird species have been shown to co-exist by partitioning their food resources by segregating their foraging habitats (MacArthur 1958; Lack 1971; Schoener 1974). However, studies of forest avian communities have usually been conducted under rather stable circumstances (Wagner 1981; MacNally 1994). The spatio-temporal variation in resource dis- tribution can considerably affect the foraging behav- iors of birds, and hence bird communities (Wiens 1989; Maurer 1990). In reality, the distribution of food resources for in- (Received 10 November 2000; Accepted 13 February 2001) * E-mail: masa @exfor.agr.hokudai.ac.jp * Present address: Tomakomai Research Station, Hokkaido Univer- sity Forests, Takaoka, Tomakomai, 053-0035, Japan. 63 sectivorous birds within a forest change drastically with the season (Raupp et al. 1988; Hunter 1991). In temperate deciduous forests, the abundance and dis- tribution of herbivorous insects, in particular Lepi- doptera larvae, which are the most preferred prey of insectivorous forest birds (Royama 1969), change dramatically in spring due to strengthening defence traits of tree leaves after budbreak (Feeny 1970; Murakami & Wada 1997). Moreover, most bird species require greater resources for feeding their nestlings and fledglings in this season (Holmes et al. 1979: Burke & Nol 1998). It has also been concluded that food often limits the reproduction and survival of forest birds during their breeding season (Martin 1987; Rodenhouse & Holmes 1992). It is, therefore, expected that temporal changes in foraging behavior among birds in response to the abundance and distri- bution of arthropods should affect the fitness of indi- vidual birds, and hence the structure of the avian community. Although van Noordwijk et al. (1995) M. MURAKAMI suggested a relationship between the timing of Great Tit breeding and the availability of Lepidoptera lar- vae as food for their fledglings, the responses of birds to temporal changes in resource abundance and distri- bution have been little studied. Further understanding of species-specific responses of birds to temporal changes in resource abundance and distribution in de- ciduous forests in spring will help to explain the com- position of local bird communities (cf. Holmes et al. 1979: Robinson & Holmes 1982; Hino 1994). Hejl and Verner (1990) suggested that some species of birds living in the same habitat undergo similar changes in foraging behavior and diet as tem- poral changes affect resource abundance and distribu- tion. In this study, I hypothesized that all forest bird species change their foraging habitat according to the distributional changes in Lepidoptera larvae. To eval- uate this hypothesis, I measured the abundance and distribution of arthropods within a forest during spring and early summer and quantified the differ- ences in the foraging behavior of four forest bird species, Great Tit Parus major, Marsh Tit P. palustris, Narcissus Flycatcher Ficedula narcissina, and Eastern Crowned Warbler Phylloscopus corona- tus, in response to the changes in resource distribu- tion. STUDY AREA AND METHODS Field studies were performed in a 9ha (300 300m) plot in Tomakomai Experimental Forest (TOEF) of Hokkaido University in northern Japan (42°43'N, 141°36’E; 50-95 m elevation). Oak Quer- cus crispula, maple Acer mono, and linden Tilia japonica dominate the vegetation of the study plot. A 20-m square grid was set up on the forest floor using color markers. The buds of the deciduous trees opened in mid May. Two resident insectivorous bird species, the Great Tit and the Marsh Tit, began to brood in mid May, and two migrant bird species, Nar- cissus Flycatcher and Eastern Crowned Warbler, began to brood in late May (cf. Ishigaki & Matsuoka 1972). Fledgling Great Tits and Marsh Tits were ob- served to leave their nests during late May and early June, and those of Narcissus Flycatcher and Eastern Crowned Warbler, during early and mid June. Sec- ondary nests were built by two pairs of Great Tit in the study area. Within the study plot, I measured the vertical fo- liage distribution on 8 July, 1995. The presence or ab- sence of foliage above each of the color markers 64 (n=256) for bird observation grids was recorded for each of the following height layers: 0-0.5m, 0.5- 1.5m, 1.5-3m, 3-5 m, 5—7m, 7-10 m, 10-15 m, 15- 20m, >20m). The percentage of foliage present in each layer, P;, was calculated as P;=(p,/256) x 100%, where p; is the number of observation points above which foliage is present at the ith height layer. The canopy top was 15 to 25m high, and saplings and current year seedlings of the dominant tree species grew in the shrub layer (0O-1.5m). The foliage was rather sparse at 1.5-5m under the dense foliage at 7-15 m. Sampling of arthropods. To reveal the seasonal changes in arthropod distributions in the canopy trees (canopy) and in the understory vegetation (forest floor), two different sampling methodologies were conducted simultaneously in the forest around the bird observation plot. Arthropods in the canopy were collected using the beating method, and those on the forest floor were collected by sweeping. One individual of oak was chosen randomly for each sampling period. The oak canopy was sampled weekly from 25 May to 6 July 1995. Climbing to the canopy layer (10-25m high) using Perry’s (1978) method, I beat branches repeatedly, and collected arthropods which dropped onto a tray (8080cm) beneath the branches. This was replicated randomly ten times for different parts of an individual tree. For each sample, I selected a single tree that had not pre- viously been sampled. Arthropods on the forest floor were collected by sweeping every week from 1 June to 13 July. For each sample, a 40cm diameter insect net was swept continuously for 30min. within a 400m? square on the forest floor. The same area was sampled only once during the study period. The arthropods col- lected were classified into two categories, Lepi- doptera larvae and other arthropods, and the number of individuals was counted separately for each cate- gory. Bird foraging ecology. Individuals of the four dominant bird species were identified by color rings, and their foraging behavior was observed for five consecutive days each week during the breeding sea- son from 24 May to 5 July 1995. An observer walked through the study plot on a systematic basis (cf. Kendeigh 1944) from 05:00 to 11:00hr. No more than 10 foraging maneuvers for each individual bird encountered were observed, these included both feed- ing for nestling or fledglings and foraging for them- selves. The feeding of nestlings or fledglings ac- Foraging mode shifts in forest birds counted for about 80% of all foraging maneuvers. On average, an individual bird was observed continu- ously for 3.5 min. during which it averaged 6.2 forag- ing maneuvers. Whenever the individual bird being observed made a foraging maneuver, the height at which it foraged, the prey type (Lepidoptera larvae or other arthropods), the kind of attack (sallying, glean- ing, or pecking, and the substrate were recorded; cf. Holmes et al. 1979). Foraging heights were estimated to the nearest two metres. An individual previously observed on the same day was left out so as to avoid a bias due to a particular individual. The kinds of at- tacks and substrates were combined into seven differ- ent foraging behaviors: 1) Air Sally, 2) Leaf Sally, 3) Leaf Glean, 4) Leaf Pecking, 5) Twig and Trunk Sally, 6) Twig and Trunk Glean, and 7) Twig and Trunk Pecking. The daily mean height of the foraging site was calculated for each individual bird. The per- centage of foraging on Lepidoptera larvae (Prey Type) and those of each foraging method were calcu- lated for each day. The daily data were summed for each week of the survey period. Statistical analyses. A two-way ANOVA (factor= period, arthropod category) was used to reveal sea- sonal changes in the abundance of the two arthropod categories in the canopy. The weekly changes and inter-specific differences in the foraging height and the proportion of Lepidoptera larvae in bird prey was analyzed by two-way ANOVA (factor=time, bird species). Furthermore, the seasonal and inter-specific difference in foraging behavior was analyzed by a two-way MANOVA (factor=time, species) based on the frequencies of each foraging method. Exact val- ues were log,, transformed and percentage data were arc-sin transformed to standardize variances and im- prove normality, if necessary to satisfy the assump- tions of the ANOVAs. All statistical tests were two- tailed. In all cases, statistical significance was evalu- ated at P<0.05. RESULTS Arthropod distribution. A two-way ANOVA re- vealed significant effects of both sampling period (F=12.84, df=6 and 126, P<0.001) and arthropod category (F=4.20, df=1 and 126, P=0.043; Fig. 1a). The interaction effect was also significant (F=4.25, df=6 and 126, P=0.002). The number of Lepi- doptera larvae showed a conspicuous peak during late May to mid June. Thereafter, it decreased rapidly and remained at a low level from late June to July. The ——#— Lepidoptera larvae -- a- -- other arthropods 15 b: Forest floor Number of individuals (m7) fo) Fig. 1. Seasonal changes in the numbers of arthropods in the canopy (a), and on the forest floor (b). Solid lines indicate Lepidoptera larvae and dotted lines other arthropods. Error bars are standard errors of the means. number of other arthropods showed a smaller peak than that of the Lepidoptera larvae from late May to early June. The number of Lepidoptera larvae was larger than that of other arthropods from late May to mid June. On the forest floor, the number of Lepidoptera lar- vae peaked in early June, then gradually decreased and reached a low level in late June (Fig. 1b). The number of other arthropods showed a rather small peak in mid June. Bird foraging habitat. In the study plot, 11 indi- vidual Great Tits, 16 Marsh Tits, 21 Narcissus Fly- catchers, and 19 Eastern Crowned Warbler s were recorded during the study period. Foraging height varied significantly seasonally (F=6.47, df=5 and 15, P<0.001; Fig. 2b) and among bird species (F= 137.0, df=3 and 15, P<0.001) with significant interaction (F=7.95, df=18 and 15, P<0.001). Great Tits, Marsh Tits, and warblers continued to forage in the canopy throughout the study period, whereas fly- catchers changed their foraging height. Flycatchers foraged in the canopy from late May to mid June, then on the forest floor from mid June to late June, and again in the canopy from late June to early July. Prey types were identified in 68% of the 3,857 for- aging maneuvers observed. The prey type varied sig- nificantly seasonally (F=64.96, df=5 and 96, P<0.001; Fig. 3) and among bird species (F=5.647, df=3 and 96, P=0.0013) with significant interaction M. MURAKAMI 30 (a) E 20 = A=) £ (b) o be = 2 10 t--4.-4--# 10% 7) <£ D c @ Great tit D N Marsh tit & y A Narcissus flycatcher ro) (@) P ee Crowned leaf warbler 0 u. ete ered 60 40 20 0O Jul Frequency (%) may dune sh Fig. 2. Vertical foliage distribution in the study plot (a), and seasonal changes in the foraging height of four-bird species (b). Error bars are standard errors of the means. Great tit 205 189 ee a 80 60 40 201 W 0 Marsh tit a 80 60] 4 a NS Pe May 159 W 188 168 Frequency (%) g NS June Fig. 4. July May 221 okt So o o o~ 80 — ® 6 © c 2 9 = 60 cs oO fom J o 2a as 40 a y @ Great tit 20 : Marsh tit A Narcissus flycatcher Crowned leaf warbler May July Fig. 3. Seasonal changes in the proportion of Lepidoptera larvae preyed on by four bird species. Error bars are standard errors of the means. Narcissus flycatcher 149 138 Crowned leaf warbler 185 181 94 116 Air Sally Leaf Sally Leaf Glean Leaf Peck Twig and Trunk Sally Twig and Trunk Glean BEHEBZOALD Twig and Trunk Peck June July Seasonal changes in relative frequencies of each of the foraging behaviors employed by four bird species. The letters on the top of each bar were the numbers of observation for each bird species in each period. (F=4.371, df=15 and 96, P<0.001). From late May to mid June, all four bird species foraged mostly on Lepidoptera larvae. In late June, the flycatchers con- tinued to forage on Lepidoptera larvae, whereas the three other species reduced their utilization ratio of Lepidoptera larvae. Then in July, all four species uti- lized the Lepidoptera larvae at a ratio of about 60%. The MANOVA analysis revealed that foraging be- havior differed both among bird species (Hotelling- 66 Lawley Trace: bird species, value=12.77, F=53.94, df=21, P<0.001) and across the seasonal period (value= 1.669, F=2.176, df=35, P<0.001; Fig. 4) with significant interaction (value=1.667, F=1.404, df=105, P=0.008). Flycatchers mostly foraged by sallying from leaves throughout the study period (more than 60% of all foraging maneuvers) and rarely utilized twigs and trunks (<15%). The three other species frequently foraged by gleaning from Foraging mode shifts in forest birds leaves. Great Tit also foraged by pecking on leaf rolling caterpillars, and Marsh Tit by pecking on twigs and trunks. From late June to July, these three species came to forage more frequently on twigs and trunks (about 40%) than in the previous period (about 20%). DISCUSSION Lepidoptera larvae were abundant in the oak canopy from late May to mid June, and then rapidly decreased. The spring foliage is known to open a “window” of high quality leaves for herbivorous in- sects but only for a short duration, which causes changes in the abundance and distribution of herbivo- rous insects (Feeny 1970; Kraft & Denno 1982; Mu- rakami & Wada 1997). In the present study, Lepi- doptera larvae disappeared rapidly from the canopy after early June, but became abundant on the forest floor during early to mid June. Many Lepidoptera lar- vae are known to migrate from the canopy to the floor for pupation or to seek alternative food re- sources when canopy leaves strengthen their defence traits during this season (Murakami & Wada 1997). In contrast, other arthropods, most of which are not herbivores (e.g. spiders and dipterans), did not make such a drastic distributional change. The present study showed that Narcissus Flycatch- ers shifted their foraging height nearly in parallel with the change in distribution of Lepidoptera larvae. They foraged in the canopy from late May to early June, then on the forest floor from mid to late June, confirming the results of my previous study con- ducted in 1994 (Murakami 1998). Such a shift was nearly synchronous with the decline in Lepidoptera larvae biomass in the canopy, but a little later than the peak of larval abundance on the forest floor, thus im- plying that the foraging habitat shift was due to the decline of Lepidoptera larvae in the canopy rather than the increase of the larvae on the floor. The three other bird species, however, continued foraging in the canopy even after this critical moment. During this study, Narcissus Flycatchers foraged intensively on Lepidoptera larvae, shifting their foraging site from the canopy to the forest floor in late June, when the abundance of larvae in the canopy decreased. In con- trast, the three other bird species continued to forage in the canopy, but shifted their foraging substrate there from leaves to twigs or trunks, and shifted their main prey from Lepidoptera larvae to other arthro- pods. Thus, the Narcissus Flycatcher and the three 67 other bird species coped differently with the abrupt change in the abundance of Lepidoptera larvae as the most important food resource in the canopy. The different responses among these four species should correspond to their species-specific foraging tactics. Rosenberg (1993) suggested that the foraging tactics utilized by birds considerably affect the acces- sibility of prey organisms on different substrates. Moreover, it has been suggested that differences in foraging methods affect a bird’s prey-type selection and that gleaners can obtain smaller prey than sally- ers (Holmes & Recher 1986). In this study, the fly- catcher mainly performed sallies, whereas the other species frequently fed by gleaning or pecking. The flycatcher probably changed its foraging habitat after having difficulty in finding alternative food resources in the canopy during mid and late June when Lepi- doptera larvae, which are far larger than other avail- able arthropods within the forest (Murakami unpubl. data), disappeared from the canopy. The three other species, however, did not change their foraging layer in this season. Instead, they began to frequently uti- lize other arthropods as alternative resources. In early July, the flycatcher resumed foraging in the canopy. When the density of Lepidoptera larvae was low both in the canopy and on the forest floor, the flycatcher, due to its innate foraging behavior (Wiens 1984), probably foraged more efficiently in the canopy than on the floor. Because the fledglings of all four bird species had already left their nests by this season, breeding phenology should have little or not effect on the differences observed in foraging site selection. The presence of species that prey on birds can also affect the foraging site selection of birds (Székely et al. 1989). A location that is covered by foliage may be safer than an exposed one (Ekman 1987). In the study plot, Japanese Lesser Sparrowhawk Accipiter gularis was observed five times during the study pe- riod (Murakami pers. obs.). The ground layer (0O—5 m) of this forest was not densely covered by foliage (Fig. 2), which indicates that birds on the ground may be at greater risk of predation by birds. Therefore, it is ex- pected that Narcissus Flycatchers selected a rather risky habitat in order to be able to forage on Lepi- doptera larvae. Interspecific competition among bird species may also affect foraging habitat selection (Alatalo et al. 1987). Although there was no direct evidence for interspecific competition in this study, the differences in foraging heights among the four bird species (Fig. 2) may indicate that interspecific competition was operating. M. MURAKAMI Given that previous studies have shown that sea- sonal shifts in foraging mode coincide among differ- ent bird species in the same habitat due to the change in prey availability (Alatalo 1980; Rotenberry & Wiens 1980; Hejl & Verner 1990), I hypothesized that all four bird species tracked the distributional change in Lepidoptera larvae. My results, however, indicated that each bird species responds differently to the changes in resource distribution according to their foraging tactics, which may be limited by their species-specific morphological structure as shown by Moreno and Carrascal (1993). Further investigation of the relations between such versatile foraging be- haviors of birds and fluctuation of resource abun- dance and distribution will provide a deeper insight into the mechanisms of species co-existence in bird communities (cf. Smith & Rotenberry 1990). ACKNOWLEDGMENTS I thank M. Toda and S. Nakano for valuable com- ments, and K. Ishigaki, T. Hiura, T. Aoi, and other staff and students of TOEF and of the Institute of Low Tem- perature Science for logistic assistance. Financial sup- port was provided by the Japanese Ministry of Educa- tion, Science, Sport and Culture (Nos. O9NP1501, 11440224). REFERENCES Alatalo RV (1980) Seasonal dynamics of resource parti- tioning among foliage-gleaning passerines in northern Finland. Oecologia 45: 190-196. Alatalo RV, Erikson D, Gustafsson L & Larsson K (1987) Exploitation competition influences the use of foraging sites by tits: experimental evidence. Ecology 68: 284-290. Burke DM & Nol E (1998) Influence of food abun- dance, nest-site habitat, and forest fragmentation on breeding ovenbirds. Auk 155: 96-104. Ekman J (1987) Exposure and time use in willow tit flocks: the cost of subordination. Anim Behav 35: 445-452. Feeny P (1970) Seasonal changes in oak leaf tannins and nutrients as a cause of spring feeding by winter moth caterpillars. Ecology 51: 565-581. Hejl SJ & Verner J (1990) Within-seasonal and yearly variations in avian foraging locations. In: Morrison ML, Ralph CJ, Verner J & Jehl JR Jr (eds) Avian For- aging: theory, morphology, and application. pp 202- 209. Allen Press, Lawrence. Hino T (1994) Interindividual differences in behavior 68 and organization of avian mixed-species flocks. In: Kawanabe H (ed) Mutualism and community organi- zation. pp 37-55. Oxford University Press, London. Holmes RT & Recher HF (1986) Search tactics of insec- tivorous birds foraging in an Australian eucalypt for- est. Auk 103: 515-530. Holmes RT & Schultz JC (1988) Food availability for forest birds: effects of prey distribution and abun- dance on bird foraging. Can J Zool 66: 720-728. Holmes, RT, Schultz JC & Nothnagle P (1979) Bird pre- dation on forest insects: an exclosure experiment. Sci- ence 206: 462-463. Hunter AF (1991) Traits that distinguish outbreaking and nonoutbreaking macrolepidoptera feeding on northern hardwood trees. Oikos 60: 275-282. Ishigaki K & Matsuoka S (1972) Birds in the Tomako- mai Experimental Forest of Hokkaido University. No 1. Res Bull Coll Exp For, Coll Agric, Hokkaido Univ 29: 43-54. Kendeigh SC (1944) Measurements of bird population. Ecol Monog 14: 67-106. Kraft SK & Denno RF (1982) Feeding responses of adapted and non-adapted insects to the defensive properties of Baccharis halimifolia L. (Compositae). Oecologia 52: 156-163. Lack D (1971) Ecological isolation in birds. Harvard University Press, Cambridge. MacArthur JW (1958) Population ecology of some war- blers in northern coniferous forests. Ecology 39: 599— 619. MacNally R (1994) Habitat-specific guild structure of forest birds in south-eastern Australia: a regional scale perspective. J Anim Ecol 63: 988-1001. Martin TE (1987) Food as a limit on breeding birds: a life-history perspective. Ann Rev Ecol Syst 18: 453- 487. Maurer BA (1990) Extensions of optimal foraging the- ory for insectivorous birds: implications for commu- nity structure. In: Morrison ML, Ralph CJ, Verner J & Jehl JR Jr (eds) Avian Foraging: theory, morphology, and application. pp 455-461. Allen Press, Lawrence. Moreno E & Carrascal LM (1993) Leg morphology and feeding postures in four Parus species: an experimen- tal ecomorphological approach. Ecology 74: 2037- 2044. Murakami M (1998) Foraging habitat shift in narcissus flycatcher, Ficedula narcissina, due to the response of herbivorous insects to strengthening defence of canopy trees. Ecol Res 13: 73-82. Murakami M & Wada N (1997) Difference in leaf qual- ity between canopy trees and seedlings affects migra- tion and survival of spring-feeding moth larvae. Can J For Res 27: 1351-1356. Foraging mode shifts in forest birds Noordwijk AJ van, McCleery RH & Perrins CM (1995) Selection for timing of great tit breeding in relation to caterpillar growth and temperature. J Anim Ecol 64: 451-458. Perry DR (1978) A method of access into the crowns of emergent and canopy trees. Biotropica 10: 155-157. Raupp MJ, Werren JH & Sadof CS (1988) Effects of short-term phenological changes in leaf suitability on the survivorship, growth, and development of gypsy moth (Lepidoptera: Lymantriidae) larvae. Env Ento- mol 17: 316-319. Robinson SK & Holmes RT (1982) Foraging behavior of forest birds: the relationships among search tactics, diet, and habitat structure. Ecology 63: 1918-1931. Rodenhause NL & Holmes RT (1992) Results of experi- mental and natural food reductions for breeding black-throated blue warblers. Ecology 73: 357-372. Rosenberg KV (1993) Diet selection in amazonian antwrens: Consequences of substrate specialization. Auk 110: 361-375. Rotenberry JT & Wiens JA (1980) Temporal variation in habitat structure and shrubsteppe bird dynamics. Oecologia 47: 1-9. Royama T (1969) Factors governing the hunting behav- iour and selection of food by the great tit (Parus major L.). J Anim Ecol 39: 619-668. Schoener TW (1974) Resource partitioning in ecologi- cal communities. Science 185: 27-39. Smith KG & Rotenberry JT (1990) Quantifying food re- sources in avian studies: present problems and future needs. In: Morrison ML, Ralph CJ, Verner J & Jehl JR Jr (eds) Avian Foraging: theory, morphology, and application. pp 3-5. Allen Press, Lawrence. Székely T, Szép T & Juhasz T (1989) Mixed species flocking of tits (Parus spp.): a field experiment. Oe- cologia 78: 490-495. Wagner JL (1981) Seasonal change in guild structure: oak woodland insectivorous birds. Ecology 62: 973— 981. Wiens JA (1984) Resource systems, populations, and communities. In: Price PW, Slobodchikoff CN & Gaud WS (eds) A new ecology. Novel approaches to interactive systems. pp 397-436. Academic Press, New York. Wiens JA (1989) The Ecology of Bird Communities. Cambridge University Press, Cambridge. | SPECIAL FEATURE Ornithol. Sci. 1: 71-80 (2002) Interspecific segregation and attraction in forest birds The effects of arthropod abundance and size on the nestling diet of two Parus species Mizuki MIZUTANI? and Naoki HIJII Laboratory of Forest Protection, Graduate School of Bioagricultural Sciences, Nagoya University, Nagoya 464-8601, Japan ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2002 Forest-dwelling insectivorous Abstract Feeding habits of Parus major and P. varius inhabiting coniferous planta- tions of Cryptomeria japonica and Larix kaempferi, each containing a small area of deciduous broad-leaved trees, were analyzed in relation to the abundance and size distribution of arthropods. In a C. japonica-dominated (CJ) area, C. japonica trees were mainly used by P. major only, while deciduous broad-leaved trees were used by both Parus species. In a L. kaempferi-dominated (LK) area, both Parus species used L. kaempferi trees and deciduous broad-leaved trees. The composition of nestling diets differed between Parus species. For prey size, the difference in the breadth was smaller and the overlap was larger between areas than between species. These results suggest that each Parus species preferred a specific size class of prey. That is, the sin- gle-prey loader P. major preferred large prey, whereas the multiple-prey loader P. var- ius preferred small prey. The abundance and size distribution of arthropods greatly differed among foraging microhabitats. Both Parus species selectively used foraging microhabitats according to their prey-size preference. Key words Diet selection, Foraging microhabitat, Nestling diet, Parus major, Parus varius, Size preference birds, including According to the optimal-foraging theory, preda- -Paridae, depend on arthropods for food, especially during the breeding season. The structure, biomass and dynamics of arboreal arthropod communities may vary depending on the tree species composition and structural features of forests (e.g. Stork et al. 1997), thus the characteristics and availability of arthropods as a food resource for insectivorous birds may differ among different forest types. In general, Parus species depend on caterpillars, mostly lepi- dopteran and hymenopteran larvae, for primary food resources (Perrins 1979), but their food composition differs among forest types (Gibb & Betts 1963; van Balen 1973). Although many studies have focused on the effects of differences in the availability of cater- pillars on the feeding habits or breeding ecology of Parus species (e.g. van Balen 1973; Yui 1988; Per- rins 1991; van Noordwijk et al. 1995; Seki & Takano 1998), little attention has been paid to other arthro- pods, or to the whole arthropod community, as poten- tial food resources. (Received 27 April 2001; Accepted 21 June 2001) * Corresponding author, E-mail: mmizuki@mcn.ne.jp 71 tors choose their diet to maximize their net rate of en- ergy intake (cf. Krebs & Kacelnik 1991). Thus, the characteristics of arthropods, such as abundance, bio- mass and other specific features associated with their susceptibility to capture, are relevant to the profitabil- ity of each prey item and thus to the foraging deci- sion by birds (Royama 1970; Hespenheide 1975). Also, it has been known for some time that birds dif- fer interspecifically in their prey-size preferences, presumably associated with their morphological traits (e.g. Betts 1955; Gibb & Betts 1963; Diamond 1973; Eguchi 1979; Quinney & Ankney 1985; Tordk 1986; Diaz 1994). In this paper, we describe the feeding habits of Parus major and P. varius in two coniferous planta- tions of the evergreen Cryptomeria japonica D. Don and the deciduous Larix kaempferi Carriér. Marked differences in the abundance, biomass, and composi- tion of the arthropod community has previously been revealed between these two plantations (Hijii et al. 2001; Mizutani & Hijii 2001). We demonstrate the effects of the abundance and size distribution of arthropods on microhabitat selec- M. MIZUTANI and N. HIJII tion by Parus species through their prey-size prefer- ence for their nestlings. MATERIALS AND METHODS 1) Study site The study was carried out in the Nagoya Univer- sity Forest at Inabu, Aichi Prefecture, central Japan (980-1230 m a.s.].; 35°11’N, 137°33’E). The annual air temperature averaged 8.3°C and the mean annual precipitation was 2250 mm (1981-1998). About 90% of the area of this experimental forest is dominated by plantations of C. japonica, L. kaempferi and Chamaecyparis obtusa Sieb. et Zucc. Small stands of deciduous broad-leaved trees such as Quercus crispula, Carpinus tschonoskii, Prunus grayana and Acer sieboldianum occur patchily, for the most part along ridges and streams. The height of dominant trees was 20-25 m and the age of the plantations was 21-40 years. The canopies of the plantations were closed; their understories were not dense and were heterogeneous, consisting mainly of Lindera praecox and L. triloba. For the purposes of this study, under- story was defined as the ground flora rising no more than two metres above the ground. Two adjoining study areas were established in the experimental forest: a C. japonica-dominated (C/) area and a L. kaempferi-dominated (LK) area (Fig. 1). More than 50% of each area was covered with a plan- tation of each dominant conifer, while about 10% of each area consisted of deciduous broad-leaved trees. The remainder of each area consisted mainly of Chamaecyparis obtusa plantations and bare areas, both of which were rarely used by either Parus species. Censuses were conducted during the nestling period of early broods of the two Parus species in 1999. 2) Foraging microhabitat From 23 May to 9 June 1999, the microhabitat use of foraging Parus species was surveyed using a line- census method. In each study area, two transects, each 50-m wide and 3.8-km long, covering 40% of the area, were surveyed six times. Whenever a bird was observed to forage on a prey item, its height above the ground, and the plant species from which it foraged, were recorded. Foraging microhabitat selec- tion for tree layers was analyzed using the selectivity index defined as the ratio of percentage microhabitat use to percentage cover area for each tree species (Manly et al. 1993). The selectivity index is 0 when the resource is not used at all, | when the resource is used as expected by chance, and larger than 1 when the resource is used selectively. 3) Nestling diet From 14 May to 16 June 1999, nestling diet was recorded with an 8-mm video camera at six clutches of P. major (3 at CJ area and 3 at LK area) and five clutches of P. varius (3 at CJ area and 2 at LK area) in nest boxes. For each observation, two trials, each of six hours, were made in the first half (S—9 days of age) and the latter half (11-15 days of age) of the nestling period. We converted the recorded video movie into a computer file in non-compressed video format, then extracted several still images for each nest-visit. From the still images, prey items (at the order level or as “caterpillars”), and prey-size (body length) were recorded. Totals of 1181 feeding records were obtained on video for P. major and 406 for P. varius. Among them, clear pictures suitable for prey identification and for prey-size determination amounted to 1124 for P. major and 330 for P. varius. Dry weights (W, mg) of prey items were estimated from body lengths (L, mm) and regression equations made based on field samples (caterpillars and or- thopteran insects) collected randomly in the study site [3 C. japonica L. kaempferi Ly deciduous Gf broad-leaved trees [_] others CJ area 500[m] Fig. 1. The vegetation of the two study areas. Nestling-diet of two Parus species during the nestling period of early broods of two Parus species or from the literature. For caterpillars the dry weight estimate was: W=0.0011L*” (n=103, r°=0.91, P<0.001) (1) for orthopteran insects it was ~ W=0.0120L7* (™m=88, r°=0.94, P<0.001) (2) and for other arthropods — W=0.0305L** (r°=0.94: Rogers et al. 1976). (3) __ Prey size was defined as the dry mass (estimated as -above) and classified according to size classes be- tween 10! and 107° mg dry weight (d.wt) at intervals of the power index 0.25. The breadth of the size-class distribution for prey items was evaluated using the Shannon-Wiener formula (Shannon & Weaver 1949), H’=— > p; log p; (4) j where p; is the proportion of prey individuals belong- ing to size class j7. The degree of overlap between size-class distributions was evaluated using the multi- plicative measures of niche overlap (Pianka 1973), > Phj Pij »_ =e » eed, oe j ij (5) where p,; and p,; are the proportions of prey individu- als of the jth size class used by the Ath and the ith species, respectively. The index @ varies from 0, when size-class distributions are completely distinct, to 1, when they fully overlap. 4) Arthropod sampling Arthropod sampling was conducted twice, on 18 May and 10 June 1999. Arthropods were collected from four foraging microhabitats: tree layers of C. japonica, L. kaempferi and deciduous broad-leaved stands using the branch clipping method, and from the understory using the beating method. Details of the protocol for the branch clipping method are de- scribed by Hijii et al. (2001). Arthropod sampling from C. japonica and deciduous broad-leaved stands was made in the CJ area, and that from a L. _kaempferi stand in the LK area. Since the two study areas were contiguous, and because the vegetation in the deciduous broad-leaved stands in both areas were similar, we assumed that the characteristics of the arthropods on the foliage of deciduous broad-leaved stands in the CJ area could be substituted for those in the LK area. On each of the two arthropod sampling dates, five branch-clipping samples were taken from C. japonica, five from L. kaempferi, and three from deciduous broad-leaved trees. Each sample consisted of two or three 60-80 cm long branches (leaves and associated twigs and branches) collected randomly from three trees from the coniferous stands, and two 30-40 cm long branches from 10 trees from the de- ciduous broad-leaved stands. Arthropod specimens from the understory were collected from each of the above three stands during one trial involving beating for 10 minutes in a 2X30m quadrat. The data from the understories of the three stands were combined for the analysis. All branch-clipping and beating sam- ples were treated immediately with an insecticide, scrutinized to collect all the arthropods (>1 mm) and then stored in 70% ethanol within 72 hours. These arthropods were identified, their body lengths were measured with a digital calliper or a micrometer under a binocular microscope, and then their individ- ual dry weights were estimated by using equations (1)-(3) above. The density of arthropods was evalu- ated from the number or biomass per unit foliar mass for branch-clipping samples, while that for beating samples was evaluated in terms of the number or bio- mass per trial. RESULTS 1) Foraging microhabitat Significant differences in the use of five microhabi- tats were found between Parus species in the CJ area (extended Fisher’s exact probability test, j=) Ie df=4, P<0.001), but not in the LK area (y’=1.92, df=4, P=0.75) (see Table 1). The proportion of un- derstory use did not differ significantly between Parus species nor between areas (Fisher’s exact prob- ability test, P>0.05 for each combination). With respect to canopy layers, both Parus species used the foliage of deciduous broad-leaved trees more frequently than expected by chance in both areas (Table 1). P. varius’s preference for deciduous broad-leaved trees was much higher than that of P. major in the CJ area, but slightly lower in the LK area. Among the coniferous microhabitats, P. major used C. japonica in the CJ area and L. kaempferi in the LK area almost randomly, whereas P. varius scarcely used C. japonica in the CJ area but used L. kaempferi in the LK area almost at random. M. MIZUTANI and N. HUI Table 1. Foraging frequency and microhabitat selection in canopy layers by P major and P. varius in each study area. The val- ues of coverage for each transect and of foraging frequency are shown as percentages in parentheses. The selection ratio is de- fined as the ratio of the proportion of foraging frequency (except in the understory), to the proportion of microhabitat coverage. The 7 value was calculated after Manly et al. (1993). Levels of statistical significance were obtained after applying the Bonfer- roni correction. ns, P>0.05; *, P<0.05; **, P<0.01; ***, P<0.001. Coverage of transect Foraging microhabitat [ha] Foraging frequency Selection ratio 7-value P major CJ area C. japonica 8.41 (54.6) 17 (53.1) 1.15 0.76" L. kaempferi 0 (0.0) 1 (3.1) _ — deciduous broad-leaved trees 2.15 (14.0) 6 (18.8) 1.59 eS" others 4.83 (31.4) 3 (9.4) 0.35 Dall understory — 5 (15.6) = — LK area C. japonica 0.08 (0.5) 0 (0.0) — —_— L. kaempferi 9.46 (63.3) 23 (52.3) 0.96 0.14" deciduous broad-leaved trees 2.11 (14.1) 11 (25.0) 2.05 8.00** others 3.30 (22.1) 4 (9.1) 0.48 3.42* understory = 6 (13.6) — —_— P varius CJ area C. japonica 8.41 (54.6) 11 (12.6) 0.25 58.68*** L. kaempferi 0 (0.0) 0 (0.0) — — deciduous broad-leaved trees 2.15 (14.0) 64 (73.6) SZ 315.43 *** others 4.83 (31.4) 5 (5.7) 0.20 2553 uae understory — 7 (8.0) — — LK area C. japonica 0.08 (0.5) 1 (0.9) — — L. kaempferi 9.46 (63.3) 62 (57.4) 0.98 0.08"* deciduous broad-leaved trees 2.11 (14.1) 26 (24.1) 1.84 12.62*** others 3.30 (22.1) 11 (10.2) 0.50 7.72** understory — 8 (7.4) — — 2) Composition and size distribution of prey items 35 In both study areas, P. major was always a single- = 30 CJ 3.41 £2.29 (n=226) prey loader, whereas P. varius carried multiple prey = MMB LK 2.84 + 1.88 (n=104) items at a time to its young. The number of prey = ZS 7 (mean + 1SD) items carried per visit by P. varius differed signifi- = 20 Y cantly between areas (Mann-Whitney U-test, U= ree ] 9.96 10°, P=0.02) (Fig. 2). $10 ae The main nestling diet of both Parus species con- io) y ] y sisted of caterpillars and orthopteran insects (Fig. 3). rd y y g He The nestling diet of P. varius was dominated by cater- eer | 6 7 8 9 10 11 12< pillars, which accounted for 76-94% (in number) of the whole nestling diet for each brood. In contrast, the main nestling diet of P. major comprised not only caterpillars (17-51%), but also orthopteran insects (42-75%; mainly Anoplophidae), and spiders (ca. 4%). The composition of P. major’s diet varied greatly between areas: the proportion of caterpillars in the diet was lower in number in the C/ area than in 74 5 Number of prey item per visit Fig. 2. Frequency distribution of the number of prey items per visit of P varius. the LK area. The composition of the nestling diet dif- fered significantly between Parus species (G-test with Bonferroni correction, G=634.7, P<0.001 for Nestling-diet of two Parus species [| caterpillars HMM orthopteran insects KY spiders other arthropods (a) ' (b) 49 . P. major HOOe ss ro 304 CY LK 75 @ 3 (n=571) (n=553) S 2 20 ci. = & < x) 10 2 8 , Ss 0 0 2 aa. E leon 40. =). varius waa 100 « D # = ® £ i 30 CJ LK 75 of oO (n=771) (n=295) ra 8 9 g SH 20 50 26 g 5 3 10 25 Z ® — of 2 0 0 2S 10-1 10°5 4 10° 491 10'S 4Q2 1025 49-1 10°5 4 1095 4Q1 1015 4Q2 1025 CJ lope : Prey body mass categories [mg d.wt] Fig. 3. Size distribution (a) and relative composition (b) of prey items used by P major (upper) and P varius (lower). Table 2. Mean body mass (mean+1 SD; mg dry wt) for all prey items and the major prey categories taken by Parus major and | P varius (sample sizes in parentheses). Levels of significance of differences in mean body mass, based on the Mann-Whitney U- test, are indicated (ns, P>0.05; *, P<0.05; ***, P<0.001). Caterpillar Orthoptera All prey items P major CJ area 43.0+£27.8 108.0+44.3 89.3+51.0 Xx (1) fo aa) fom er |e LK area 45.8+30.5 125.3+48.2 79.8+56.9 (257) (247) (553) P. varius CJ area 17.3+£16.3 104.1+47.9 27.6+35.9 Xx or | en) fo a7 Jo LK area 15.9+16.5 13.2285 24.7+28.0 (243) (45) (295) P. major XP. varius CJ area ss ns ee LK area KK OK KK 2K CJ; G=109.2, P<0.001 for LK) and between areas Parus species was 0.1 mgd.wt (Fig. 3). PR major only for P. major (G-test with Bonferroni correction, chose larger prey items than P. varius for all prey G=95.3, P<0.001 for P. major; G=4.25, P=0.24 for items (Table 2). The mean body mass of all prey P. varius) (Fig. 3). items used by P. major differed significantly between The minimum body mass of the prey items used by areas (Mann-Whitney U-test, U=1.37X10°, P< 75 M. MIZUTANI and N. HUI 0.001). Prey items from 10! to 107?°mg d.wt ac- counted for ca. 90% of the nestling diet of P. major. The mean body mass of all prey items used by P. var- ius was smaller than that used by P. major, and did not differ between areas (U-test, U=1.07X10°, P=0.12). About 90% of the nestling diet of P. varius consisted of prey items ranging in dry weight from 10°° to 10? mg. The caterpillars taken by P. major did not differ in size between areas (U-test, U=1.45 x 107, P=0.42), whereas those taken by P. varius did differ between areas (U-test, U=7.24 10*, P<0.01). P. varius took prey of a broader range of size classes, for all prey items, than P. major (Table 3). The difference in the breadth of prey size was smaller between areas than between species. The overlap in the prey-body-size distribution was larger between areas than between Parus species (Table 4). The same results were found for caterpillars, but not for Orthoptera. The smaller difference in the breadth, and the larger overlap for prey size, between areas than be- tween species, suggests that both P. major and P. var- ius had specific prey size class preferences. 3) Composition and size distribution of arthropods in foraging microhabitats The number of individuals of arthropods larger than 0.1 mg d.wt accounted for 94% of all individu- Table 3. Breadth of the size class distribution for all prey items and the major prey categories (H’). Caterpillar Orthoptera All prey items P major CJ area 1.58 1.19 1.62 LK area 1.67 1.08 1.83 P. varius CJ area 1.86 FSi 2.05 LK area 1.92 0.96 2.05 Table 4. Overlap of the size-class distribution for all prey items and the major prey categories (). Caterpillar | Orthoptera All prey items Overlap between areas (C/XLK) P. major 0.97 0.91 0.96 P. varius 0.95 0.71 0.93 Overlap between bird species (P majorXP. varius) CJ area 0.64 0.99 0.40 LK area 0.56 0.49 0.54 76 als. The number of individuals and biomass of all arthropods was highest in deciduous broad-leaved trees and lowest in C. japonica (Table 5). The mean body mass of major prey categories (caterpillars, orthopteran insects and spiders) also dif- fered among foraging microhabitats (Table 6). The caterpillars on L. kaempferi foliage were significantly smaller than those on deciduous broad-leaved tree fo- liage or in the understory. The spiders obtained from C. japonica foliage were significantly smaller than those in the three other foraging microhabitats were. The mean body mass for all major prey categories was largest in C. japonica foliage, where a large or- thopteran insect occurred, whereas mean body mass was smallest in L. kaempferi foliage, which was dom- inated by small caterpillars and spiders. These major prey types contributed relatively large proportions of the whole arthropod fauna: 28.1% (understory)— 42.7% (C. japonica) in terms of individuals, and 15.8% (C. japonica)—38.2% (deciduous broad- leaved trees) in terms of biomass (Fig. 4). The composition and size distribution of arthro- pods differed between foraging microhabitats (Fig. 4). The size class of caterpillars most frequently ob- served in L. kaempferi (10 °°-10°* mgd.wt) was smaller than that found in deciduous broad-leaved trees (10°°-10!”° mgd.wt). In the understory, cater- pillars of various sizes (10-°’°-10?*> mg d.wt) oc- curred. Orthopteran insects contributed only 1.4% of the whole arthropod fauna in terms of numbers of in- dividuals, but contributed 13% of the biomass in C. Japonica owing to the occurrence of one large indi- Table 5. Numbers of individuals and biomass of arthropods (>0.1 mg dry wt) taken by Parus major and P. varius in each foraging microhabitat (mean+1 SD). The different letters fol- lowing values in canopy samples indicate significant differ- ences between mean values according to the Steel-Dwass test (P<0.05). Number of Biowaasez individuals! Canopy C. japonica 34.6+£11.3* 511.4+481.8° L. kaempferi 429.1+142.2° 2207.0+1758.5° deciduous 1626.0+1277.0° 7933.4+7046.8° broad-leaved trees Understory 212.8+57.5 875.4+173.2 'Number of individuals and * biomass (mg dry wt) per unit fo- liar mass (kg ' dry wt) for the canopy or per trial for the un- derstory Nestling-diet of two Parus species Table 6. Mean body masses (mean+1 SD; mg dry wt) of prey in each foraging microhabitat (sample sizes in parentheses). The different letters following values indicate significant differences in the mean body mass between foraging microhabitats according to the Steel-Dwass test (P<0.05). Caterpillar Orthoptera’ Spider ples eee categories Canopy C. japonica 5.0+7.4 (3)! 139.9)" "@) 0.7+1.3 (56)* 5.5+35.3 (61) L. kaempferi MTB iSve(98)" 0.9+1.0 (56)° 2.043.2 (154)> deciduous broad-leaved trees 6.8+6.2 (84)° 2: 2-Aleien(G) L353. 50G3)2 Avie: 518 9u(h42)s Understory S45 1S 102). 16.7230: Gn) 1.4+2.1 (246)° 3.849.9 (359)? "No statistical analysis was performed because of the small sample size. {_] caterpillars HB orthopteran insects spiders _] other arthropods 2k P. V@iUS (b) _ iy _— 1 ae. japonica (n=143) IgG L. kaempferi (n=419) trees (n=352) Relative percentage [%] Relative percentage in number of individuals for each size class [%] deciduous broad-leaved | 40-71 10° 4 10° 4Q1 10'S 4Q2 107° a Biomass individuals Body mass categories [mg d.wt] Fig. 4. Size distribution (a) and relative composition in number of individuals and biomass (b) of arthropods collected from each foraging microhabitat. Only arthropods larger than the minimum size of prey items are shown. The horizontal bar above (a) shows the size range containing 90% of caterpillars preyed on by PR major and P varius. Ti M. MIZUTANI and N. HII vidual (276 mg d.wt). Spider was the most dominant group in number of individuals in all microhabitats, but most of them were less than 10°” mg d.wt. DISCUSSION 1) Prey size as a contributory factor of nestling-diet selection in two Parus species In this study, two Parus species selectively chose prey items of a particular size range specific to each species. The differences in the mean body mass of prey used by each Parus species in two study areas were markedly smaller than those between Parus species (Table 2). The inter-area differences in the breadth of prey size used by each Parus species was small (Table 3) and the inter-area overlap was large (Table 4) in spite of large differences in the body-size distribution (Fig. 4) and other characteristics (Tables 5 and 6) of arthropods between areas. It appears from these results that prey size is more important than prey species for diet selection by these Parus species. The difference in prey size in the diet between sympatric Parus species (Table 2) may be due to food-resource partitioning for a relaxation of inter- specific competition. Several authors have reported that preferred prey size differs among Parus species (Betts 1955; Gibb & Betts 1963; Eguchi 1985). Sym- patric Parus species tend to vertically segregate their foraging microhabitats during the non-breeding sea- son (Perrins 1979), and it is also reported that vertical partitioning between P. major and P. varius did not occur during the breeding season (Nakamura 1970). As reported in Nakamura (1970), the vertical parti- tioning between those Parus species was not found in our study. Because both species preferred deciduous broad-leaved trees, there may have been local compe- tition for foraging microhabitats between them. Nev- ertheless, they may have been able to avoid competi- tive interactions by taking different-sized prey. Food requirements by birds are strongly correlated with their body mass (Nagy 1987). Since the body mass of the two Parus species is almost the same (16.5 g fresh weight (f.wt) for P. major and 17.0 g f.wt for P. varius; Yui 1988), both Parus species should have similar feeding efficiencies. The inter- specific difference in prey-size preference could be explained by the fact that whereas P. major is always a single-prey loader, P. varius is a multiple-prey loader (Fig. 2). P. varius may compensate for a de- crease in its feeding efficiency by carrying many small prey items at a time. 78 2) Matching of foraging microhabitat use by Parus species with the characteristics of the arthropod com- munity The characteristics of arthropods as potential prey, such as abundance, biomass (Table 5) and the compo- sition and size distribution of individuals (Table 6, Fig. 4), greatly differed among foraging microhabi- tats. One of the strategies used by Parus species to adapt to mosaic environments with various vegeta- tion types and varying food availability, is selective use of foraging microhabitats with food resources sufficient for reproduction (cf. Dias & Blondel 1996). Considerable numbers of caterpillars occurred in each foraging microhabitat except in the foliage of C. Japonica, but their body-size distribution differed among microhabitats (Fig. 4). Orthopteran insects were found in each microhabitat except the foliage of L. kaempferi (Fig. 4). In number, their contribution to the whole arthropod fauna was very small, but their contribution to biomass was large because a few large individuals occurred. In this study, the sample size for the branch clipping method may have been insuffi- cient to evaluate the exact abundance and size distri- bution of orthopteran insects on trees, because of low efficiency of sampling. Our results were similar, nev- ertheless, to those based on chemical-knockdown samples, which showed that the proportion of or- thopteran insects was less than 1% in number, but 16% in biomass (average for the data of June and Au- gust; Hijii 1989). Because small orthopteran insects were found both in the foliage of deciduous broad- leaved trees and in the understory, larger individuals may also inhabit both microhabitats. Previously, chemical-knockdown sampling had failed to find or- thopteran insects in the foliage of L. kaempferi (Ter- akawa unpubl. data), a fact confirmed by our results. Although spiders have been reported as a secondary major food resource of Parus species in many habi- tats (e.g. Gibb & Betts 1963; Won et al. 1965; Minot 1981; Eguchi 1985; Arakida 1995), they were not a preferred nestling food in our study (Fig. 3). This is probably because small individuals constituted a large proportion of the spiders in number in this study site. There were significant differences in the use of for- aging microhabitats between Parus species in the CJ area, but not in the LK area (Table 1). Both Parus species selectively used the foliage of deciduous broad-leaved trees as a foraging microhabitat in both areas. In deciduous broad-leaved trees, various-sized caterpillars available to both P. varius (10°P=10!-2) Nestling-diet of two Parus species and P. major (10'-107) contributed large proportions to the arthropod fauna (Fig. 4). Thus, the foliage of deciduous broad-leaved trees would have been im- portant for both Parus species in providing stable food resources. Moreover, there may be more advan- tageous to P. varius capable of using smaller caterpil- lars, because smaller prey items were abundant, which may enhance the food availability to P. varius in this foraging microhabitat (Fig. 4). Thus, P. varius could use caterpillars as its main food category (Fig. 3) in both areas. The foliage of C. japonica was mainly used by P. major, but only scarcely used by P. varius (Table 1). Although the availability of caterpillars was very low, there were large orthopteran insects in the foliage of C. japonica (Fig. 4). For P. major, which tended to prefer large prey items (Table 2, Fig. 3), orthopteran insects on C. japonica trees can be a suitable food re- source, even if their abundance is low (Table 1). In the LK area, both Parus species randomly used the foliage of L. kaempferi, the preferences for which were lower than for deciduous broad-leaved trees (Table 1). The contribution of individual caterpillars to the overall arthropod fauna was almost the same, but the body-size distribution differed between these microhabitats. The profitability for P. major may be lower in L. kaempferi, because the proportion of small caterpillars was relatively high. Moreover, both species may not prefer coniferous L. kaempferi as much as deciduous broad-leaved trees owing to their morphological adaptations (Nakamura 1978). In both areas, both Parus species also foraged from the understory (about 10% of foraging behavior) (Table 1), where suitable caterpillars and orthopteran insects were available for both Parus species (Fig. 4). This foraging microhabitat would also be a useful food source for both Parus species. In conclusion, both Parus species selectively used foraging microhabitats with different arthropod com- munities according to their species-specific size pref- erences. The composition of their diets may reflect both the characteristics of the arthropod community in each area and the abundance and size of prey items in each foraging microhabitat within each area. ACKNOWLEDGMENT We are grateful to Drs K. Eguchi and T. Hino for re- viewing earlier versions of this paper and for their help- ful remarks and suggestions. Thanks are extended also to the members of the Laboratory of Forest Protection, 79 Nagoya University and the Nagoya University Forest for many helpful suggestions and assistance. REFERENCES Arakida Y (1995) Population increase and breeding ecology of the varied tit Parus varius in a secondary forest with nest boxes. Jpn J Ornithol 44: 37-65 (in Japanese with English summary). Balen JH van (1973) A comparative study of the breed- ing ecology of the great tit (Parus major) in different habitats. Ardea 61: 1-93. Betts MM (1955) The food of titmice in oak woodland. J Anim Ecol 24: 282-323. Diamond JM (1973) Distributional ecology of New Guinea birds. Science 179: 759-769. Dias PC & Blondel J (1996) Local specialization and maladaptation in the Mediterranean blue tit (Parus caeruleus). Oecologia 107: 79-86. Diaz M (1994) Variability in seed size selection by granivorous passerines: effects of bird size, bird size variability, and ecological plasticity. Oecologia 99: 1-6. Eguchi K (1979) Food and food consumption of nestling tits, Parus major minor and Parus varius varius, in the ever-green broad leaved forests in northern Kyushu. J Yamashina Inst Ornith 11: 1-18. Eguchi K (1985) Food size, energy intake and nutrient intake of nestling tits, Parus varius and Parus major. J Yamashina Inst Ornith 17: 74-83. Gibb JA & Betts MM (1963) Food and food supply of nestling tits (Paridae) in Breckland Pine. J Anim Ecol 32: 489-533. Hespenheide HA (1975) Selective predation by two swifts and a swallow in central America. Ibis 117: 82-99. Hijii N (1989) Arthropod communities in a Japanese cedar (Cryptomeria japonica D. Don) plantation: abundance, biomass and some properties. Ecol Res 4: 243-260. Hijiit N, Umeda Y & Mizutani M (2001) Estimating density and biomass of canopy arthropods in conifer- ous plantations: an approach based on a tree-dimen- sional parameter. For Ecol Manage 144: 147-157. Krebs JR & Kacelnik A (1991) Decision-making. In: Krebs JR & Davis NB (eds) Behavioural ecology: an evolutionary approach, 3rd edition. pp 105-136. Blackwell, Oxford. Manly BFJ, McDonald LL & Thomas DL (1993) Re- source selection by animals: statistical design and analysis for field studies. Chapman & Hall, London. Minot EO (1981) Effects of interspecific competition for food in breeding blue and great tits. J Anim Ecol 50: M. MIZUTANI and N. HUI 375-385. Mizutani M & Hijii N (2001) Mensuration of frass drop for evaluating arthropod biomass in canopies: a com- parison among Cryptomeria japonica, Larix kaempferi, and deciduous broad-leaved trees. For Ecol Manage 154: 327-335. Nagy KA (1987) Field metabolic rate and food require- ment scaling in mammals and birds. Ecol Monogr 57: 111-128. Nakamura T (1970) A study of Paridae community in Japan. Il. Ecological separation of feeding sites and foods. Misc Rept Yamashina Inst Ornith 6: 41-169 (in Japanese with English summary). Nakamura T (1978) A study of Paridae community in Japan. IV. Ecological segregation of species by the difference of use of bill in space and technique. Misc Rept Yamashina Inst Ornith 10: 94-118 (in Japanese with English summary). Noordwijk AJ van, McCleery RH & Perrins CM (1995) Selection for the timing of great tit breeding in rela- tion to caterpillar growth and temperature. J Anim Ecol 64: 451-458. Perrins CM (1979) British tits. Collins, London. Perrins CM (1991) Tits and their caterpillar food supply. Ibis 133 suppl: 49-54. Pianka ER (1973) The structure of lizard communities. 80 Ann Rey Ecol Syst 4: 53-74. Quinney TE & Ankney CD (1985) Prey size selection by tree swallows. Auk 102: 245-250. Rogers LE, Hinds WT & Buschbom RL (1976) A gen- eral weight vs. length relationship for insects. Ann Entomol Soc Am 69: 387-389. Royama T (1970) Factors governing the hunting behay- iour and selection of food by the great tit (Parus major L.). J Anim Ecol 39: 619-668. Seki S & Takano H (1998) Caterpillar abundance in the territory affects the breeding performance of great tit Parus major minor. Oecologia 114: 514-521. Shannon CE & Weaver W (1949) The mathematical theory of communication. Univ Illinois Press, Urbana. Stork NE, Adis J & Didham RK (1997) Canopy arthro- pods. Chapman & Hall, London. Torok J (1986) Food segregation in three hole-nesting bird species during the breeding season. Ardea 74: 129-136. Won P, Kim S & Kim C (1965) Breeding biology and chick food of varied tit Parus varius in Korea. Misc Rept Yamashina Inst Ornith 4: 198-206 (in Japanese with English summary). Yui M (1988) Mori ni sumu yachou no seitaigaku (Ecol- ogy of sylvan birds). Soubun, Tokyo (in Japanese). SPECIAL FEATURE Ornithol. Sci. 1: 81-87 (2002) Interspecific segregation and attraction in forest birds Prey distribution and foraging preference for tits Teruaki HINO**, Akira UNNO** and Shigeru NAKANO*** Faculty of Agriculture, Hokkaido ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2002 University, Sapporo 060-0809, Japan Abstract We examined the abundance and distribution of prey in four different height strata and eight tree species in a temperate forest, and analyzed the influence on foraging preference by three breeding tit (Parus) species. Densities of arthropod prey for tits in canopy foliage varied with tree species but not with height. Most of them were Lepidoptera larvae. Also, interspecific differences in choice of foraging substrate were found between tree species but not in height. These results demon- strate that tree species composition is a more important habitat factor than foliage height profile for coexistence of different tit species in forests. We examined four dif- ferent measures of prey abundance to find how tits chose tree species. The largest species, the Great Tit P. major, preferred the tree species with high total biomass, and the intermediate-sized Willow Tit P. montanus preferred those with high density per leaf area. Concentrated searching for prey on a few tree species with high total bio- mass may be a useful strategy for inflexible perch-gleaners such as P. major, and finer-scale searching on each leaf may be more practical for agile foragers such as P. montanus which often hang-glean to reach less accessible food. In spite of these dif- ferences, both species gained benefits from choosing the tree species on which they foraged most efficiently. In contrast, the smallest species, the Coal Tit P. ater, fre- quently foraged on food-poor tree species. Of the three tit species, P. ater was the most generalized forager, using diverse techniques on a variety of tree species and specializing at capturing small prey quickly. These foraging patterns may make it possible for the smallest species to coexist with the other tit species. Key words Foraging efficiency, Foraging technique, Parus, Prey distribution, Tree species preference As habitat factors determining bird species diver- sity in forests, some researchers have emphasized the importance of vertical foliage distribution (MacArthur & MacArthur 1961; Recher 1969) and others have emphasized tree species composition (Rice et al. 1984; Verner & Larson 1989). Either of these factors could help different species co-exist in different forests, and it remains unclear whether one or both factors provides a general mechanism for coexis- tence. Holmes et al. (1979) proposed a hypothesis that tree species composition determines bird species (Received 25 June 2001; Accepted 3 September 2001) * Corresponding author, E-mail: tkpk @affre.go.jp * Present address: Kansai Research Center, Forestry and Forest Products Research Institute, Kyoto 612-0855, Japan. ** Present address: Hokkaido Forestry Research Institute, Bibai 079-0198, Japan. *** Present address: Center for Ecological Research, Kyoto Uni- versity, Otsu 520-2113, Japan. 81 diversity within guilds, the presence and size of which are determined by foliage height distribution. Hino (1985) supported this idea in part by a correla- tion analysis between habitat variables and bird com- munities in shelterbelts of Hokkaido, but we need to know how bird species partition height strata and tree species based on food availability in forests. Many studies have shown the different use of height strata and/or tree species among insectivorous bird species within the same guild (Hartley 1953; MacArthur 1958; Morse 1970; Lack 1971; Alatalo 1982; Hino 1998). Few studies, however, have exam- ined prey abundance and distribution although the need for such studies has been recognized for many years (Wiens 1984; Morrison et al. 1990). Holmes and his colleagues demonstrated that foraging by birds on three different tree species was influenced by the prey abundance on and foliage structure of each tree species and the species-specific ability of the T. HINO et al. birds (Holmes & Robinson 1981; Holmes & Schultz 1988). Their conclusion was somewhat speculative, however, because they examined only three species of trees. Moreover, the prey distributions and their ef- fects on bird foraging at different height strata have never been studied in forest habitats. In this paper, we examine the abundance and distri- bution of prey in four different height strata and eight tree species in a temperate forest, and analyze the in- fluence on foraging preference by three breeding tit (Parus) species. We then reveal interspecific differ- ences in the degree to which prey abundance deter- mines foraging preference, and relate it to searching- pattern constraints on foraging techniques of each species. METHODS 1) Vegetation This study was conducted in the Nakagawa Ex- perimental Forest of Hokkaido University, northern Hokkaido, Japan (44°49'N, 142°16’E). A 6-ha study plot (200 mx300 m) was established along the moun- tain ridge (c. 350 m in altitude). This plot largely con- sists of deciduous broad-leaved trees dominated by Quercus mongolica and Betula ermanii with <10% of conifers Picea jezoensis and Abies sachalinensis (Appendix). Canopy height is 15—20m. The under- story is dominated by high density of two bamboo species Sasa kurilensis and Sasa senanensis 1.0-1.5 m in height. Vegetation was surveyed in the study plot in July 1990. Species, numbers of individuals and DBH (di- ameter at breast height) were recorded for all trees > 5 cm in DBH. The relative importance value (IV) of each tree species was calculated as the percentage of basal area (the sum of cross sectional areas of trunks determined from DBH), which is closely correlated with leaf surface area (Holmes & Robinson 1981). To determine the foliage distribution (in percentage cover) at five different height strata (O—3.6 m, 3.6—7.2 m, 7.2-10.8m, 10.8-14.4m, 14.4m+), four cate- gories of foliage volume (0: none, 1: 1-33%, 2: 34— 66%, 3: 67-100%) were recorded by eye at 459 points, averaged and multiplied by 30% for each stra- tum. 2) Insect distribution All arthropods, most of which were Lepidoptera larvae (97.6% in dry weight), were collected and counted from 300 leaves per unit, and their body 82 lengths were measured in the laboratory. The dry weight (W, mg: 60°C, 48h) of an arthropod was esti- mated from the body length (L, mm) with the follow- ing equation: W=0.033L?" (r=0.92, P<0.001, N=180). This equation was determined with a part of the samples (W, mg). For statistical analyses, the dry weights were transformed to In(W+0.5) to reduce skewness. The abundance of arthropods in the canopy foliage at different height strata was examined in June 1990. We collected samples directly from a nine-storied tower (14.4m high and 10m square) built of steel pipes near the study plot. This tower gave access to six trees of Quercus mongolica and seven of Betula ermanii, which were two dominant species. We chose two units randomly for arthropods sampling from each tree species at each story in June. The unit sam- ples from both species were combined and averaged for each of the same height strata as foliage distribu- tion except the 0—3.6m stratum for which no foliage was accessible. The expected percentage of prey bio- mass at each height stratum was calculated based on the average dry weight multiplied by the foliage vol- ume to compare with the foraging height use by birds. The abundance of arthropods in the canopy foliage of different tree species was examined in June 1991. We collected samples from the branches 1—3m in height with a clipping method. The tree species ex- amined were the eight species >2% in IV (Appen- dix): Quercus mongolica, Betula ermanii, Kalopanax pictus, Picea jezoensis, Abies sachalinensis, Tilia Japonica, Acer mono and Phellodendron amurense. We used four different measures of arthropod abun- dance as determinants of tree species selection by birds because the tree density and the leaf area varied among tree species. The two indices of density were the average dry weight (mg) per 100 leaves (D,) and per | m?—leaf area (D,), and the two indices of total biomass in the study site were D, XIV (T,) and D, Xx IV (T,), respectively. To calculate D,, we collected 25 leaves randomly from three trees of each species and measured their areas with a digitizer. We treated one shoot with a node for two conifer species and one leaflet for Phellodendron amurense with compound leaves as one piece of leaf. 3) Bird foraging Observations were made in the morning (0600- 1100 hours) and in the afternoon (1400-1700 hours) in June and July 1990 and in June 1991. The target Prey distribution and foraging preferences species were three breeding tit species, Great Tit Parus major, Willow Tit P. montanus and Coal Tit P. ater, the densities of which were 21.7, 34.8 and 43.5 pairs/100 ha, respectively (Hino & Nakano 1992). Body size was largest in P. major, medium in P. mon- tanus and smallest in P. ater (14.1 g, 10.7 g, 8.5g in Nakamura 1978). Each time a bird was observed for- aging for prey (including active searching), we recorded the height and tree species, and the prey length estimated in 5 mm intervals using beak size as a standard. We also differentiated between three types of feeding technique (perch-gleaning, hang-gleaning or sally-hovering; Remsen & Robinson 1990). To avoid bias from repeated observations of the same in- dividuals, we collected data from individuals in dif- ferent territories in the study area. In 1991, foraging time (searching and handling) was timed for each tree species. Foraging efficiency (mg/min) was calculated as a total dry weight of prey (estimated by body length) divided by the total foraging time for each tree species. Foraging rate was also calculated as the number of prey capured per minute. Tree species preference was calculated with In (r/p,), where r, was the proportion used by birds and p; was the IV/100 for each tree species i. For the di- versity of the tree species and feeding technique used by birds, we used the Simpson’s diversity index: 1/ Lp,’, where pi was the proportion of a category i. Correlation analyses were conducted to reveal which indices of food abundance on different tree species were related to the preferences and the foraging effi- ciencies by each tit species. We considered P<0.1 as significance level owing to small sample sizes (5-6). Bird classification followed the Ornithological Soci- ety of Japan (2000) RESULTS 1) Foraging height In Quercus mongolica and Betula ermanii in 1990, the average dry weight of arthropods per 100 leaves was 32.2mg (+28.6 SD, N=14) in the canopy fo- liage >3.6 m in height. The dry weights did not differ significantly among four height strata (F,,,=0.72, P=0.55), nor in comparisons between pairs of differ- ent strata (F-test with Bonferroni correction, P>0.20, Table 1). The foliage volume varied from 38% to 58% in percentage cover among height strata. The es- timated biomass of prey was least in the top stratum (14.4m+) and greatest in the stratum immediately below (10.8-14.4 m), with a threefold difference be- tween the two values (Table 1). y’-analyses revealed that the tit species all foraged made use of each stra- tum in proportion to the distribution of prey biomass in the study site, with no significant difference be- tween species (y?=15.1, df=8, P>0.05). 2) Foraging tree species In 1991, the average dry weights of arthropods per 100 leaves (D,) varied greatly among tree species from the lowest in Picea jezoensis through the high- est in Kalopanax pictus (Table 2). However, the or- ders of each species changed when we used different indices of prey abundance. In terms of dry weight per 1m (D,,), the highest values were obtained for Betula ermanii and then Tilia japonica, with large-leafed, Table 1. Prey abundance and foraging uses by tits at each height stratum : ; : ; é etek Caterpiller biomass Bird foraging use (%) category Foliage (%) Dry weight Expected : (m) per toUileaves me)’ RiomaseOaye P. major P. montanus Pater 14.4— 38.6 22.0+13.9 (4) 14.1 6.3 13.1 25.0 10.8-14.4 58.1 43.7+44.0 (8) 42.1 37.5 33.8 32.4 7.2-10.8 55.8 25.1£15.5 (8) 2B)? 29.2 26.2 19.4 3.6-7.2 38.1 32.8+18.3 (8) 20.7 Doll 22.8 17.6 0-3.6 Sl?) - - 0.0 4.1 5.6 No. of observations 48 145 108 7-value (df=3)° 6.63 51) 5.16 'Mean+SD (no. of sample size). * Percentages based on (Dry weight per 100 leaves) X (Foliage %) in the height categories 2-5. *Comparisons of bird foraging use to expected prey biomass in the height categories 2-5. Neither values were significant (P>0.05). T. HINO et al. Table 2. Four indices of prey abundance on each tree species tires D,: Dry weight D,: Dry weight T,: Total biomass T,: Total biomass species per 100 Leaves (mg) per | m? (mg) D, XIV! D,XIV! QM Mell 36.9 VIS 1303.4 BE 99.0 293.0 3089.5 9140.6 KP 182.4 137.4 1312.9 989.4 PJ 2.8 10.3 13.2 49.3 AS We3) 47.7 38.3 162.3 TJ Wd) 155.0 186.5 402.9 AM 29:3 SS) 61.5 113.1 PA 7.4 35.6 14.9 Te) ‘Importance Values (%) in Appendix. Table 3. Foraging uses and preferences of tits on each tree species P. major P. montanus Pater Tree species Use (%) Preference Use (%) Preference Use (%) Preference QM 28.0 0.26 15.1 —0.07 26.0 0.21 BE 50.0 0.74 50.0 0.74 28.0 0.33 KP 8.0 0.48 3) 0.58 0.0 —0.69 PJ 2.0 —0.69 0.0 —0.69 18.0 1.45 AS 0.0 —0.69 8.1 1.06 10.0 1.24 TJ+AM+PA 6.0 0.35 11.6 0.80 12.0 0.83 No. of observations 50 86 50 Simpson’s diversity 2.94 3.36 Splls) Kalopanax pictus (Appendix) falling to third place. Indices of total biomass (T, and T,), as expected, were high in the two dominant species, Betula er- manii and Quercus mongolica (>30% in IV, Appen- dix). Tree species preferences for foraging differed among tit species (Table 3). P. major foraged on the lowest diversity of tree species, preferring Betula er- manii and avoiding conifers. P. montanus showed op- posing preferences for each of the two coniferous species, preferring Abies sachalinensis and avoiding Picea jezoensis. P. montanus also made preferential use of deciduous species except Quercus mongolica. P. ater foraged on the greatest diversity of tree species, but showed special preferences for each of the two coniferous species and avoided Kalopanax pictus. Foraging efficiencies on each tree species also differed among the tit species (Table 4). P. major cap- tured prey on Betula ermanii and Quercus mongolica most effectively. P. montanus did so on Betula er- manit and Kalopanax pictus, and P. ater on Acer mono. 84 Correlation analyses were conducted to reveal which indices of food abundance on different tree species (Table 2) were related to the preferences (Table 3) and foraging efficiencies (Table 4) of each tit species (Table 5). P. major foraged preferentially on the tree species with high total biomass in the study site (T, and T,). P montanus foraged preferen- tially on tree species with high dry weight per leaf area (D,,). In both species, the indices of prey abun- dance for the tree species where they foraged most efficiently were consistent with those they preferred, although this relationship was not significant (Table 5). On the other hand, P. ater did not show positive relationships with any of the indices, but tended to avoid foraging on the tree species with high values of D, and T, (Table 5). 3) Foraging technique P. major was a specialized perch-gleaner, P. mon- tanus foraged most frequently by hang-gleaning, and P. ater used both perch- and hang-gleaning with the same frequency. The diversity of foraging techniques Prey distribution and foraging preferences Table 4. Foraging efficiencies of tits on each tree species P major P. montanus P. ater ies Species Efficiency Time Efficiency Time Efficiency Time (mg/min) (sec) (mg/min) (sec) (mg/min) (sec) QM 28.6 457 35) 384 14.4 672 BE 34.9 1185 46.4 1241 12.0 748 KP 21.9 83 44.3 265 - — PJ 0.0 36 = - 49 253 AS - — 21.9 219 5.9 97 TJ+AM+PA 5.0 201 36.1 409 49.2 162 All species 28.9 2018 34.7 2631 14.8 1932 Table 5. Results of correlation analyses between foraging preferences (Table 3) or efficiencies of tits (Table 4), and four indices of caterpiller abundance on each tree species (Table 2) N .D, DD, TT, 1, Preference Preference P major 6 NS NS + (4) Pmontanus 6 NS (+) NS NS Pater 6 —- NS —- NS Efficiency P major 5 Gi) NS + 3 (+) Pmontanus 5 NS (+) NS NS NS Pater 5 NS NS NS NS NS Positive correlation: +P<0.05, (+) 0.050.1. N: sample size. Table 6. Foraging techniques of tits Foraging technique P major Pmontanus P ater Perch-gleaning 85.4 23.3 41.9 Hang-gleaning 2.4 58.3 40.0 Sally-hovering 12.2 18.3 18.1 Number of observations 41 120 105 Simpson’s diversity 1.34 2.33 2.71 was highest in P. ater and lowest in P. major (Table 6). In total, foraging efficiency was high in P. mon- tanus and P. major, and low in P. ater (Table 4). This difference was related to prey size captured by each tit species: P. ater caught significantly smaller prey (9.0mg+20.4 SD, N=53) than P montanus (23.8 mg+35.4 SD, N=64, U-test with Bonferroni’s cor- rection: z=—3.21, P<0.05) or P. major (27.8mg= 40.8 SD, N=35, z=—2.62, P<0.005). On the con- 85 trary, foraging rate was highest in P. ater (1.65), the second in P. montanus (1.46) and lowest in P. major (1.04). The species with the most diverse technique took prey at the fastest rate (r=0.99, P<0.05, N=3). DISCUSSION During the breeding season, densities of arthropod prey for tits varied with tree species but not with height. The three tit species showed interspecific dif- ferences in their use of tree species but not in use of height. These quantitative results demonstrate that tree species composition is a more important habitat factor than foliage height profile for coexistence of different tit species in forests. Similar results were found by Holmes et al. (1979) in North American forests and by Hino (1985) in northern Japanese shel- terbelts. These studies show that tree species compo- sition determines bird species diversity within guilds, though abundance may be affected further by foliage height distribution. Our results suggest that each tit species responded to a different measure of prey abundance in selecting tree species for foraging. The largest species, P. major preferred the tree species with high total bio- mass in a forest, and the intermediate-sized, P. mon- tanus preferred those species with high average den- sity per leaf area. This interspecific difference may be related to their searching patterns associated with for- aging technique. P. major was a specialized perch- gleaner taking prey mainly from the upper surface of leaves, as reported in other studies (Morse 1978; Hino 1993). P. montanus hang-gleaned more fre- quently but also foraged by perch-gleaning and sally- hovering. Since most of caterpillars are found on un- dersides of leaves (Greenberg & Gradwohl 1980; Holmes & Schultz 1988), hang-gleaning would be a T. HINO et al. more effective (but energy-expensive) technique for birds to access them. Thus, P. montanus are expected to forage on different tree species more flexibly than P. major while searching for prey from tree to tree, as indicated by using more diverse species of trees and fine-scale searching on each leaf. On the other hand, concentrated searching for prey on a small number of tree species with high total biomass may be useful for less flexible perch-gleaners like P major because prey accessibility is constrained (Holmes & Robinson 1981; Holmes & Schultz 1988). In spite of these dif- ferences, both species gained benefits from choosing the tree species where they foraged most efficiently (as reported by Partridge 1976a, b). In contrast, the smallest species, P. ater often used food-poor tree species. We can suggest two possible reasons to explain this observation. One possibility is that interference competition from the larger species may deny P. ater access to the most profitable forag- ing sites. Many studies have shown that P. ater are forced to food-poor substrates in the presence of the other tit species (Alatalo et al. 1985; Alatalo & Moreno 1987; Suhonen et al. 1993; Fyhn & Sorensen 1997). These studies all were conducted in winter when interspecific competition was severe owing to scarcity of food resources. Our study indicates that interspecific competition may also play an important role for foraging-site selection by tits during the breeding season. The other possibility is that P. ater with a morphology adapted to conifers (Partridge 1976b), may forage on Picea jezoensis and Abies sachalinensis frequently regardless of prey abun- dance. This could be partly true but P. ater was nei- ther a specialized nor an efficient forager on conifers. Goldcrests Regulus regulus, which also bred in low density in the study site (4.3 pairs/km’, Hino & Nakano 1992), are known as specialized foragers in conifers (Nakamura 1980). This species foraged on coniferous trees much more frequently (83% in 1990, 55% in 1991) and more efficiently (20.4 mg/min) than P. ater, with more frequent sally-hovering (56%) (Hino et al. unpubl. data). Of the three tit species, P. ater was the most generalized forager. This species took prey from a variety of tree species with diverse foraging techniques: by perch-gleaning like P. major, hang-gleaning like P. montanus and sally-hovering like Regulus regulus. Instead of low foraging effi- ciency, P. ater captured small prey with high speed. This flexible foraging pattern may make it possible for this smallest tit species to coexist with the other tit species. 86 ACKNOWLEDGMENTS We greatly thank Yukio Akibayashi, Shunji Natsume, Satoru Okuyama and other staff of the Nakagawa Ex- perimental Forest, Hokkaido University for their co- operation, and also to Woo-Shin Lee, Mitsuru Saito, Shoichiro Yamamoto, Shigeo Kuramoto, Sawako Tokuda, Kyosuke Ohkawara for their assistance during the research. We are grateful to Richard Loyn and Chan- Ryul Park for their useful comments on the manuscript. This study was partly supported by a JSPS Fellowship for Japanese Junior Scientists REFERENCES Alatalo RV (1982) Multidimensional foraging niche or- ganization of foliage-gleaning birds in northern Fin- land. Ornis Scand 13: 56-71. Alatalo RV & Moreno J (1987) Body size, interspecific interactions, and use of foraging sites in tits (Paridae). Ecology 68: 1773-1777. Alatalo RV, Gustafsson L, Linden M & Lundberg A (1985) Interspecific competition and niche shifts in tits and the goldcrest: an experiment. J Anim Ecol 54: 977-984. Fyhn M & Sorensen L (1997) Niche shifts of coal tits Parus ater in Denmark. J Avian Biol 28: 68-72. Greenberg R & Gradwohl J (1980) Leaf surface special- ization of birds and arthropods in Panamanian forest. Oecologia 46: 115-124. Hartley PHT (1953) An ecological study of the feeding habits of the English titmice. J Anim Ecol 22: 261-— 288. Hino T (1985) Relationships between bird community and habitat structure in shelterbelts of Hokkaido, Japan. Oecologia 65: 442-448. Hino T (1993) Interindividual differences in behaviour and organization of avian mixed-species flocks. In: Kawanabe H, Cohen JE & Iwasaki K (eds), Mutual- ism and community organization. pp 87-94. Oxford University Press, Oxford. Hino T (1998) Mutualistic and commensal organization of avian mixed-species flocks in a forest of western Madagascar. J Avian Biol 29: 17-24. Hino T & Nakano S (1992) Breeding bird community of a deciduous broad-leaved forests in northern Hokkaido, Japan. Res Bull Hokkaido Univ For 49: 195-200 (in Japanese with English summary). Holmes RT, Bonney RE Jr & Pacala SW (1979) Guild structure of the Hubbard Brook bird community: a multivariate approach. Ecology 60: 512-520. Holmes RT & Robinson SK (1981) Tree species prefer- ences of foraging insectivorous birds in a northern Prey distribution and foraging preferences hardwoods forest. Oecologia 48: 31-35. Holmes RT & Schultz JC (1988) Food availability for forest birds: effects of prey distribution and abun- dance on bird foraging. Can J Zool 66: 720-728. Lack D (1971) Ecological isolation in birds. Blackwell Scientific Publications, Oxford. MacArthur RH (1958) Population ecology of some war- blers of northeastern coniferous forest. Ecology 39: 599-619. MacArthur RH & MacArthur J (1961) On bird species diversity. Ecology 42: 594-598. Morrison ML, Ralph CJ, Verner J & Jehl JR Jr (1990) Avian Foraging: theory, morphology, and applica- tions. Allen Press, Lawrence. Morse DH (1970) Ecological aspects of some mixed- species foraging flocks of birds. Ecol Monogr 40: 119-168. Morse DH (1978) Structure and foraging patterns of flocks of tits and associated species in an English woodland during the winter. Ibis 120: 298-312. Nakamura T (1978) A study of Paridae community in Japan IV: Ecological segregation of species by the difference of use of bill in space and technique. Misc Rept Yamashina Inst Ornithol 10: 94-118 (in Japan- ese with English summary). | Nakamura T (1980) Ecological separation and adaptive space of warbler guild inhabiting the coniferous for- est in Shiga Heights. Bull Inst Nature Ede Shiga | Heights 19: 45-59. Partridge L (1976a) Individual differences in feeding ef- ficiencies and feeding preferences of captive Great Tits. Anim Behav 24: 230-240. Partridge L (1976b) Field and laboratory observations on the foraging and feeding techniques of blue tits (Parus caeruleus) and coal tits (P. ater) in relation to their habitats. Anim Behav 24: 534-544. Recher HF (1969) Bird species diversity and habitat di- versity in Australia and North America. Am Nat 103: 75-80. Remsen JV Jr & Robinson SK (1990) A classification scheme for foraging behavior of birds in terrestrial habitats. In: Morrison ML, Ralph CJ, Verner J & Jehl JR Jr (eds) Avian Foraging: theory, morphology, and applications. pp 144-160. Allen Press, Lawrence. Rice J, Anderson BW & Ohmart RD (1984) Compari- son of the importance of different habitat attributes to avian community organization. J Wildl Manag 48: 895-911. Suhonen J, Halonen M & Mappes T (1993) Predation risk and the organization of the Parus guild. Oikos 66: 94-100. The Ornithological Society of Japan (2000) Check-list of Japanese birds: 6 ed. OSJ, Obihiro. Verner J & Larson TA (1989) Richness of breeding bird species in mixed-conifer forests of Sierra Nevada, California. Auk 106: 447-463. Wiens JA (1984) Resource systems, populations, and communities. In: Price PW, Slobodchikoff CN & Gaud WS (eds) A new ecology. pp 397-436. John Wiley Sons, New York. | Appendix Average leaf area (+SD, N=25) and importance value (IV) of each tree species Code Tree species QM Quercus mongolica BE Betula ermanii KP Kalopanax pictus PJ Picea jezoensis AS Abies sachalinensis TJ Tilia japonica AM Acer mono PA Phellodendron amurens Leaf area (cm’) Importance value (“%) 75.0+49.5 35.3 33.8412.1 31.2 132.7+106.1 Ye 26.84 12.3 4.8 23.6+15.9 3.4 49.3+19.7 2.6 54.4+26.4 Del 20.9+14.1 2.0 87 SPECIAL FEATURE Ornithol. Sci. 1: 89-93 (2002) Interspecific segregation and attraction in forest birds The effects of food-supply on Southeast Asian forest birds Navjot S. SODHI* Department of Biological Sciences, National University of Singapore, Blk S2, 14 Science Drive 4, Singapore 117543, Republic of Singapore ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2002 Abstract Southeast Asian forests are being lost at an alarming rate. This unprece- dented deforestation is resulting in avifauna losses. Despite this, Southeast Asian avi- fauna remains poorly studied. A few studies measured the food-supply and correlated it with the Southeast Asian forest bird ecology. These correlative studies (qualitative as well as quantitative) show that food-supply can affect the bird diversity, abun- dance/density, breeding ecology, body condition, ranging behaviour and/or flocking behaviour. However, there has been no experimental study conducted to determine the effects of food-supply on the forest bird ecology. In this geographic area, exciting research avenues remain available to study the avian feeding ecology and to explore a relationship between food-supply and forest bird ecology. Descriptive, correlative as well as experimental data on these aspects are required to enhance the knowledge of avian ecology as well as for avian conservation purposes. Key words Southeast Asia (primarily containing Myanmar, Thailand, Cambodia, Laos, Vietnam, Malaysia, Sin- gapore, Brunei and Indonesia) is a region of high bird diversity with over 1200 bird species present (Inskipp et al. 1996, Robson 2000). Within this region, at least 70% of resident bird species may be partly or exclu- sively dependent upon the primary forest (Wells 1985). The percentage of threatened species restricted to forested areas within Southeast Asia varies from 17% (Singapore) to 72% (Indonesia) (Collar et al. 1994). Despite this, forest loss in Southeast Asia has been extensive; with rate of deforestation three times higher here than other tropical areas (Food and Agri- culture Organization 1999). This forest loss has prob- ably resulted in concomitantly heavy avian extinc- tions (Diamond et al. 1987, Castelletta et al. 2000). The avian extinctions do not appear to occur ran- domly following forest loss (e.g. Karr 1980, Brash 1987). Frugivores and insectivores, for example, are particularly vulnerable to extinctions after deforesta- tion (Castelletta et al. 2000). One of main contribut- ing factors in such cases could be a decline in food- supply. Therefore, it may be critical to understand the relationships between the ecology of forest birds (de- (Received 22 March 2001; Accepted 11 June 2001) * E-mail: dbsns@nus.edu.sg 89 Bird ecology, Food-supply, Rainforest birds, Southeast Asia, Tropics pending on primary or old secondary forest to sur- vive) and food abundance/density. My objectives here are to: 1) summarize results of studies exploring food-supply and bird ecology relationships, 2) iden- tify if there are any general patterns based on previ- ous studies and 3) provide some future directions for studies to determine the effects of food-supply on bird ecology. Only studies where food abundance/ density was quantitatively measured and related with Southeast Asian bird ecology are discussed here. Published studies were searched by using various databases such as the Web of Science and BIOSIS. These searches were supplemented by gleaning through the literature cited by relevant studies and searching through some of the regional journals such as Tropical Biodiversity. Bird classification used in this manuscript follows Sibley & Monroe (1990). CASE STUDIES 1) Malaysia Fogden (1972) estimated fruit and insect abun- dance in the Semengo Forest Reserve (Sarawak, Malaysia). He qualitatively related the food abun- dance with various aspects of bird ecology. Birds’ breeding in December to May coincided with a pe- riod when insects were more abundant. Frugivorous N. S. SODHI bird species had similar breeding phenology as insec- tivorous bird species despite the fact that fruits were not consistently abundant during any particular time of the year. This may be due to the fact that many fru- givorous species supplement their diets with protein rich insects during the breeding season. In Semengo, insects were least abundant in No- vember and most bird species finished moulting by the end of October (Fogden 1972). Exception to this included the species for which insect abundance was suspected not to vary widely such as bark-foraging woodpecker species (e.g. the Rufous Woodpecker Celeus brachyurus) and leaf-litter foraging babbler species (e.g. the Short-tailed Babbler Malacocincla malaccensis). Two bulbul species (Pycnonotus spp.) in Semengo fed on both insects as well as fruits (Fogden 1972). Fruits must have constituted an important part of the diet for these bulbuls because they consumed them even when insects were plentiful. Probably fruits rep- resent an easily harvestable source of energy. These two bulbul species had suppressed breeding activity in low fruit abundance year than in high fruit abun- dance year. The proportion of juveniles in the popula- tion was seven fold higher in high fruit abundance year (14%) than in low fruit abundance year (2%). There also seemed to be a recovery in the body masses of these bulbul species when fruits abundance increased (Fogden 1972). Thus, Fogden’s qualitative study implies broad consequences of food-supply on the bird ecology. Wong (1986) studied the effect of food resources (flowers, fruits and arthropods) on understory bird communities in primary versus disturbed (regener- ated) areas of the Pasoh Forest Reserve (Peninsular Malaysia). Wong found that the number of plant indi- viduals that produced flowers or fruits used by birds was at least three times higher in primary than in the disturbed forest. Foliage arthropod abundance did not differ between the primary and disturbed areas but the period of low arthropod abundance was four months longer in disturbed than in the primary forest. Both, bird species richness (73 versus 83 species) and abundance (575 versus 703 individuals), were lower in disturbed than in the primary forest. Wong argued that this was because lower resource abundance could support fewer birds in disturbed than in the primary forest. In Pasoh, foliage arthropods did not show marked temporal and spatial variation as shown by fruiting and flowering plants. Therefore, frugivores would 90, have to range over wider areas than insectivores. Consistent with this prediction, Wong (1986) found that the abundance of frugivores varied twice more than that of insectivores both in primary and the dis- turbed forest. Wong’s study highlights the importance of food-supply on the avian community diversity and abundance. 2) Indonesia Leighton and Leighton (1983) monitored fruit den- sity in the Kutai Nature Reserve (Kalimantan, In- donesia). In a semi-quantitative manner, they corre- lated fruit density with frugivore behaviour and movements. Leighton and Leighton found that all hornbill species (the Rhinoceros Hornbill Buceros rhinoceros, Helmented Hornbill Rhinoplax vigil, Bushy-crested Hornbill Anorrhinus galeritus, Black Hornbill Anthracoceros malayanus, White-crowned Hornbill Aceros comatus, Wrinkled Hornbill A. cor- rugatus and Wreathed Hornbill A. undulates) only bred when the fruit density peaked (January to May). Based on bird sightings, it appeared that the Green Imperial Pigeon Ducula aenea, Jambu Fruit Dove Ptilinopus jambu, Hill Myna Gracula_ religiosa, Green Broadbill Calyptomena virdis, A. undulates and A. corrugatus emigrated from the study area dur- ing low fruit density periods (September-October). These results suggest that birds may be adjusting their breeding activities and/or foraging areas by tracking food resources. Kinnaird et al. (1996) studied the effect of tempo- ral and spatial differences in fruit abundance on the Sulawesi Red-knobbed Hornbill (Aceros cassidix) in the Tangkok DuaSundra Nature Reserve (Sulawesi, Indonesia). The Sulawesi Red-knobbed Hornbill feeds on fruits in the forest canopy. Kinnaird et al. (1996) found that fig (Ficus spp.)-fruit biomass corre- lated significantly with the temporal population fluc- tuations of the hornbill. The flock size also increased during months when fig-fruit biomass was high. Ad- ditionally, hornbill densities were higher in areas with higher than in lower fig-tree densities (Kinnaird et al. 1996). In line with some previous studies (e.g. Wong 1986), food-supply seemed to effect population dy- namics in this case. 3) Singapore Sodhi et al. (in press) studied the effects of food abundance on bird abundance in two large (>480 ha) forest fragments, MacRitchie and Nee Soon, in Sin- gapore. Line transects of approximately four and half Food and Asian forest birds km were surveyed eight times at each site between 14 July and 24 September 1997. All surveys were made during fair weather conditions and between 0700 and 0930h. Foliage along the transects was sweep sam- pled for the presence of understorey arthropods. Each transects was sampled for arthropods on five different occasions. Sweep sampling was conducted at inter- vals of 150m and 16 sweeps using butterfly net were made at each sampling station. Sampling stations were changed for subsequent samplings. All arthro- pods found were used in calculating their mean abun- dance. The number of fruiting trees along the transect was also counted twice. Sodhi et al. found that the mean number of under- story insectivore bird individuals (belonging to species: the White-rumped Shama Copsychus mal- abaricus, Dark-necked Tailorbird Orthotomus atrogu- laris, Abbott’s Babbler M. abbotti, Short-tailed Babbler and Chestnut-winged Babbler Stachyris arythroptera) did not differ significantly between MacRitchie (73.63+5.76 [standard error] individu- als) and Nee Soon (87.88+7.57 individuals). And as expected, the mean number of foliage arthropods also did not differ between the two forests (11.8+0.39 [standard error] and 9.66+0.67 individuals, respec- tively). However, more frugivore bird individuals (belonging to species: the Red-crowned Barbet Mag- | alaima rafflesii, Long-tailed Parakeet Psittacula longicauda and Pink-necked Green-Pigeon Treron vernans) were found in MacRitchie (maximum num- ber=45) than in Nee Soon (10). More fruiting trees were also available in MacRitchie (maximum number=352) than in Nee Soon (285). Castelletta (unpubl. data) measured the arthropod _ abundance and insectivore bird density in 13 forest patches (7—935ha) in Singapore. Eleven insectivo- : rous bird species that feed in the understory were considered (the Laced Woodpecker Picus vittatus, Oriental Magpie Robin Copsychus saularis, White- rumped Shama, Yellow-bellied Prinia Prinia flaviven- tris, Common Tailorbird Orthotomus sutorius, Dark- | necked Tailorbird, Rufous-tailed Tailorbird O. sericeus, Ashy Tailorbird O. ruficeps, Abbott’s Bab- bler, Chesnut-winged Babbler and White-chested Babbler Trichastoma rostratum). Castelletta analysed _ correlations between the mean insectivore density (all | species combined) and all arthropod abundance. Sep- | arate analyses were conducted for arthropods less than | cm and those more than | cm in length. No sig- nificant correlation was found between the mean den- sity of insectivores and arthropod abundance for all l 91 the comparisons. One of the reasons for this lack of correlation may be that the diet of the study species is poorly understood thus making the analyses crude. However, Castelletta found that arthropod abundance was highest during the peak of the breeding season (April to September) of insectivorous forest birds. The studies from Singapore show that some food types can effect bird abundances and breeding season while others may either exert minimal effect or their effect may be difficult to demonstrate. WHAT ARE THE GENERAL PATTERNS? Only a few studies have been conducted on the ef- fects of food-supply on Southeast Asian forest birds. However, these limited number of studies imply the importance of food-supply on forest bird ecology (Table 1). Timing of breeding and moulting, bird di- versity and abundance/density, ranging behaviour and flock size seem to be effected by food-supply. The only study that found no significant correlative effect of food-supply on forest bird abundance is from Sin- gapore (Catelletta unpublished data). As mentioned, this study suffers from the fact that the diet of the studied bird species remained poorly documented. Some of the studies also imply or show that different prey types exert different influences on the bird ecol- ogy. Arthropod resources usually do not show re- markable temporal and spatial variation in abundance as shown by other resource types such as fruits and flowers. The temporal and spatial variation in the abundance of fruits particularly effects the distribu- tion, dispersal and movements of the frugivores (e.g. Fogden 1972, Wong 1986, Kinnaird et al. 1996). FUTURE DIRECTIONS One of the problems with Southeast Asian forest birds is that basic descriptive data are lacking even to conduct critical correlative studies. For example, the diet composition of many birds is poorly understood and data on breeding ecology (e.g. clutch sizes) are not readily available. All the studies determining the effects of food-sup- ply on Southeast Asian forest birds are either qualita- tive or at the best correlative. Demonstrating a pre- cise effect of food-supply on bird ecology may be difficult trom correlative studies and there may be a need to conduct well-designed experiments (Newton 1998). Correlative studies do however, present hy- potheses and ideas for experimental testing. As men- N. S. SODHI Table 1. The number of studies correlating the food-supply with Southeast Asian forest bird ecology. No. of studies No. of studies : Total no. of . ha! : : Bird ecology : with quantitative with experimental studies } : comparisons evidence Population characteristics Diversity 1 1 0 Abundance/density 4 4 0 Breeding ecology Phenology 3 0 0 Reproductive success 0 0 Other ecological aspects Body condition 1 0 0 Moulting 1 0 0 Flocking 1 1 0 Ranging behaviour 2 0 0 tioned, even correlative studies linking the food-sup- ply with Southeast Asian forest bird ecology are few. There is an urgent need to conduct the studies that at- tempt to find a correlation between food supply and bird movements, reproductive success and recruit- ment. With relatively well-developed radio-tracking techniques, it is now possible to track selected bird species and determine how they use home ranges or territories in relation to food distribution. Mass flowering or fruiting is a striking feature of dipertocarp forests of Southeast Asia (Ashton et al. 1988). This phenomenon occurs during irregular in- tervals of 2-10 years and it causes many different families to flower simultaneously. While the ecologi- cal causes of this phenomenon are poorly understood (Corlett 1990), such an event can have profound im- pacts on bird ecology and adaptations. To my knowl- edge, the effects of mass flowering or fruiting on Southeast Asian forest bird ecology remains poorly documented. To my knowledge no experimental study has been conducted to determine the effects of food-supply on Southeast forest birds (Table 1). For some of the rela- tively well-studied species, food supplementation (e.g. addition of fruits to a frugivore’s range) or food reduction (e.g. removing or killing of exotic vegeta- tion) experiments should be carried out. Therefore, in summary, exciting research avenues remain available on the feeding ecology and on the effects of food- supply on Southeast Asian forest bird ecology. In light of heavy deforestation, descriptive, correlative as well as experimental data are urgently needed on these research aspects. Researchers can rely for theo- retical and methodological considerations on the re- 92 search conducted on the effects of food-supply on forest birds in other geographic areas (e.g. Holmes & Schultz 1988, Rodenhouse & Holmes 1992, Burke & Nol 1998). CONCLUSIONS As mentioned, forests within Southeast Asia are lost at an unprecedented rate. Previous studies reveal that forest loss and fragmentation can cause local ex- tinctions in birds in this region (e.g. Ford & Davison 1996, Castelletta et al. 2000). Despite this, the ecol- ogy of Southeast Asian forest birds remain poorly studied. Because reduction in food-supply may be one of the variables causing the decline or local ex- tinction of Southeast Asian forest birds, it is critical to understand the relationships between food-supply and bird ecology. Avian extinctions may occur over 100 years following habitat loss (Brooks et al. 1999). Therefore, following habitat loss and degradation, there may be time to save some of the avifauna through proper conservation actions. For example, some of the supplemental food experiments may be critical to determine if the conservation of some of the threatened birds is feasible (e.g. maintaining or increasing their abundance with supplemental food). ACKNOWLEDGMENTS I thank Richard Corlett, Marjorie Castelletta, Jean- Marc Thiollay, Malcolm Soh and an anonymous re- viewer for making constructive comments on an earlier draft. Food and Asian forest birds REFERENCES Ashton PS, Givnish TJ & Appanah S (1988) Staggered flowering in the Dipterocarpacae: new insights into floral induction and the evolution of mast fruiting in the aseasonal tropics. Am Nat 132: 44-66. Brash AR (1987) The history of avian extinction and forest conversion on Puerto Rico. Biol Conserv 39: 97-111. Brooks TM, Pimm SL & Oyugi JO (1999) Time lag be- tween deforestation and bird extinction in tropical forest fragmentation. Conserv Biol 13: 1140-1150. Burke DM & Nol E (1998) Influence of food abun- dance, nest-site habitat, and forest fragmentation on breeding ovenbirds. Auk 115: 96-104. Castelletta M, Sodhi NS & Subaraj R (2000) Heavy ex- tinctions of forest avifauna in Singapore: lessons for biodiversity conservation in Southeast Asia. Conserv Biol 14: 1870-1880. Collar NJ, Crosby MJ & Stattersfield AJ (1994) Birds to watch 2. BirdLife International, Cambridge. Corlett RT (1990) Flora and reproductive phenology of the rain forest at Bukit Timah, Singapore. J Trop Ecol 6: 55-63. Diamond JM, Bishop KD & Balen S van (1987) Bird survival in an isolated Javan woodland: island or mir- ror? Conserv Biol 2: 132-142. Fogden MPL (1972) The seasonality and population dy- namics of equatorial forest birds in Sarawak. Ibis 114: 307-343. Food and Agriculture Organization (1999) State of world’s forests. FAO, Rome. Ford HA & Davison GWH (1996) Forest avifauna of Universiti Kebangsaan Malaysia and some other for- est remnants in Selangor, Peninsular Malaysia. Malayan Nat J 49: 117-138. Holmes RT & Schultz JC (1988) Food availability for forest birds: effects of prey distribution and abun- 93 dance on bird foraging. Can J Zool 66: 720-728. Inskipp T, Lindsey N & Duckworth W (1996) An anno- tated checklist of the birds of the Oriental region. Ori- ental Bird Club, Bedfordshire. Karr JR (1980) Geographic variation in the avifaunas of tropical forest undergrowth. Auk 97: 283-298. Kinnaird MF, O’Brien TG & Suryadi S (1996) Popula- tion fluctuation in Sulawesi red-knobbed hornbills: tracking figs in space and time. Auk 113: 431-440. Leighton M & Leighton DR (1983) Vertebrate re- sponses to fruiting seasonality within a Bornean rain forest. In: Sutton SL, Whitmore TC & Chadwick AC (eds) Tropical rain forest: ecology and management. pp 181-196. Blackwell Scientific Publications, Cam- bridge. Newton I (1998) Population limitation in birds. Acade- mic Press, London. Robson C (2000) A field guide to the birds of South-east Asia. New Holland, London. Rodenhouse NL & Holmes RT (1992) Results of experi- mental and natural food reductions for breeding black-throated blue warblers. Ecology 73: 357-372. Sibley CG & Monroe BL Jr 1990. Distribution and tax- onomy of birds of the world. Yale University Press, New Haven. Sodhi NS, Castelletta M, Loh WZ, Lee BPY & Subaraj R (in press) Tropical lowland rainforest birds on a highly urbanized islands: monitoring, losses and les- sons. In: Shaw WW, Harris L & VanDruff L (eds) Proceedings of 4th International symposium on urban wildlife conservation. University of Arizona, Tucson. Wells DR (1985) The forest avifauna of Western Malaysia and its conservation. In: Diamond AW & Lovejoy TE. (eds.) Conservation of tropical forest birds. pp 213-232. ICBP, Cambridge. Wong M (1986) Trophic organization of understory birds in a Malaysian dipterocarp forest. Auk 103: 100-116. Ornithol. Sci. 1: 95—99 (2002) ORIGINAL ARTICLE Spotted-throat individuals of the Rufous Vanga Schetba rufa are yearling males and presumably sterile Satoshi YAMAGISHI'*, Shigeki ASAI', Kazuhiro EGUCHI?’ and Masaru WADA? ' Department of Zoology, Graduate School of Science, Kyoto University, Kyoto 606-8502, Japan ? Department of Biology, Faculty of Science, Kyushu University, Fukuoka 812-8581, Japan 3 College of Liberal Arts and Sciences, Tokyo Medical and Dental University, Ichikawa, Chiba 272-0827, Japan Abstract The Rufous Vanga Schetba rufa is endemic to Madagascar and lives in one-female groups. During the 1994-1999 breeding seasons, a total of 294 nestlings were banded. Among these nestlings, 51 stayed within the study area as spotted- throat individuals. In the next breeding seasons, 35 of 45 spotted-throat individuals were subsequently observed as black-throated males, and once they became black- throated males, these individuals never reverted to the previous spotted-throat pattern. In contrast, 30 banded nestlings were recovered as yearling females with white throats, and the female’s color pattern never changed thereafter. All the spotted-throat males were helpers or floaters. All the males of one group consisting of an adult male with a black throat and two males with spotted throats were captured and sacrificed humanely. The testes were dissected from each specimen and were histologically ex- amined. The testes of the spotted-throat males contained only spermatogonia, and no spermatids or spermatozoa were present. In contrast, the testes of the black-throated male were well-developed and contained enlarged seminiferous tubules with lumen, where numerous spermatozoa were evident. Considering these facts, spotted-throat males of this species are assumed to be sterile. We suggest that, due to their underde- veloped testes, the spotted-throat males (one-year-old males) of the Rufous Vanga are physically incapable of breeding. ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2002 Key words Schetba rufa Cooperative breeding, Delayed maturation, Madagascar, Plumage, The family Vangidae is monophyletic and under- went extensive in situ radiation in Madagascar (Yama- gishi et al. 2001). The Rufous Vanga Schetba rufa is a member of this family. In a previous paper (Yama- gishi et al. 1995) we reported that this endemic species lives in groups of two to four individuals and that the groups contain one adult female, one or two adult males and sometimes also an individual with a spotted throat. We described how those birds with spotted throats helped the pairs with brood care dur- | ing part of the breeding cycle, and suggested that these individuals were immature males. In the present study we confirmed that individuals with spotted throats were yearling males. Furthermore, we report that the spotted-throat males never attempted to copulate nor became breed- (Received 27 August 2001; Accepted 20 October 2001) * Corresponding author, E-mail: yama@ci.zool.kyoto-u.ac.jp 95 ers. These observations implied that they had not reached sexual maturity. Was this absence of breed- ing in one-year-old males due to an inability of repro- duction or to ecological constraints? The alternatives would lead us to different interpretations of delayed dispersal of yearling males. To answer the question of sexual maturity of yearling males, we here provide evidence concerning the change in plumage pattern of males and also provide anatomical evidence of spermatozoa maturation. From this viewpoint we then discuss the reproductive ability of helper males of this species. STUDY AREA AND METHODS A study of Rufous Vanga ecology and behavior was conducted at Botanical Garden A in the Ampi- joroa Forest Reserve (16°15’S, 46°48’E, c. 200m asl) about 110km southeast of Mahajanga in Mada- S. YAMAGISHI et al. gascar. We marked nestlings with individually dis- tinct combinations of color bands during 1994-1999. Then, every breeding season from 1995 to 2000 we recorded which groups each individual joined, its sta- tus in the groups, and the color patterns of its throat. On 15 October 1998, we captured all the males of a group consisting of an adult male with a black throat and two males with spotted throats (spotted- throat males | and 2). At that time of year, the cap- tured group was at the late stage of nest-building, at which time we considered that the testes of adult males had achieved full functional maturity. After capture, birds were sacrificed humanely, body mass was measured with an electric balance to the nearest 0.1 g and wing length was measured with a ruler to the nearest 0.1 mm. At the field station the testes were dissected from each specimen, and fixed in 10% for- malin solution until histological examination. Right and left testes from each bird were weighed to the nearest 0.1 mg with an electric balance. Each testis was subjected to standard histological procedures for light microscopy; the testis was embedded in paraffin, and sectioned at 4 4m, the sections then stained with hematoxylin and eosin. RESULTS Among the 259 groups recorded during the seven years of study, 161 (62%) were pairs without male helpers, and 98 groups (38%) had one to four male helpers. Among the 98 groups containing helpers, 54 (55%) contained at least one male with a spotted throat. During the 1994-1999 breeding seasons, we banded a total of 294 nestlings, of which 51 stayed within the study area as spotted-throat individuals (Fig. 1-a and Table 1). In the next breeding seasons, 35 of 45 spotted-throat individuals were subsequently observed as black-throated males (Fig. 1-b), and once they became black-throated males, they never re- verted to the previous spotted-throat pattern (Table 1). In contrast, 30 individuals banded as nestlings were recovered as yearling females with white throats (Fig. l-c) in the pre-breeding season or in the breed- ing season (some of them disappeared after the re- covery). The female’s color pattern never changed thereafter. All the spotted-throat males were helpers or floaters. In general, after a male developed a black throat, he established a new territory and became a breeder. Of those 24 males whose age at first breed- ing was confirmed, 15 bred at two years old. During seven years of observations, we also noted seven in- stances where individually identifiable helper males with black throats attempted sneak copulations with breeding females. Of these incidents, in three cases the sneaking males were three years old, in two cases the same two-years-old male attempted mating, and in the remaining two cases the ages of the males were unknown. Furthermore, in the case of the two-years- old male, at least one nestling which the helping male cared for did not have an allele derived from the breeding male on a microsatellite locus developed as markers of paternity tests (Asai et al. 1999), but shared another allele with the helping male which at- tempted mating (Asai unpubl. data). We never ob- served males with spotted throats attempting to copu- late with females, nor did we find them forming a pair. These observations suggest that the helper status Table 1. Observed male nestlings within the study area and change of their throat color. Observed individuals hess Ps ral te 1995 1996 1997 1998 1999 2000 1994 14 2, 2, 2 2 2 2 1995 38 8 7 i 6 3 1996 68 ils) 12 10 6 1997 53 gos 7 1998 16 eae 7 1999 45 Total 294 Schadowed numbers indicate spotted-throat males. 96 | | | Spotted-throat individuals of the Rufous Vanga of a black-throated male per se does not inhibit his reproductive activity. In comparison, among 21 year- ling females observed in our study site during the Fig. 1. The plumage of the head and throat in Rufous Van- gas. a: a male with a spotted throat, b: a male with black throat, and c: a female. Testis showing seminiferous tubules of three Rufous Vangas from the same breeding group, captured at the study site at the late nest-building stage. a: Black-throated male, b: Spotted-throat male 1, and c: Spotted-throat male 2. Note that spermatozoa were only found in the black-throated male. SG: spermatogonium, SC: spermatocyte, SZ: spermatozoa. Table 2. Combined testes mass and right testicular volume of each Rufous Vanga captured at the late nest-building stage. Body mass Wing length Tarsus length Combined tested Right testicular (g) (mm) (mm) mass (mg) volume (mm?) Black-throated male 37.6 108.7 24.3 171.5 75.4 Spotted-throat male 1 39.4 103.9 24.4 60.8 38.5 Spotted-throat male 2 39.8 105.6 23.8 14.5 13.1 97 S. YAMAGISHI et al. breeding season, 16 (76%) were breeders and 5 (24%) were auxiliary females. Table 2 shows the combined testicular mass and measurements of body mass, wing length and tarsus length for each of the three dissected males. Histolog- ical pictures of testis sections are given in Fig. 2. When these pictures are considered together with the data of testicular mass, it is indisputable that the spot- ted-throat male 2 was an immature individual with completely regressed testes. The spotted-throat male | had heavier testes than spotted-throat male 2, but histological examination indicated that spermatogen- esis in this individual had been arrested at the early stages. The testes contained only spermatogonia, and no spermatids or spermatozoa were present (Fig. 2- b). In contrast, the testes of the black-throated male were well-developed and contained enlarged seminif- erous tubules with lumen, where numerous spermato- zoa were evident (Fig. 2-a). This histological evi- dence indicates that the black-throated male was re- productively active, whereas the two spotted-throat males were not. DISCUSSION The observations that spotted-throat males never attempted to copulate nor became breeders, and that helper males of age two years or more could poten- tially copulate with females suggest that sexual matu- rity of males is not dependent on social constraint with respect to the helper status, but is instead age re- lated. However, observational field data cannot strictly identify whether it is physiological restraint of maturity or ecological constraint on chance of mating which resulted in the absence of mating by the year- ling males. In this context, although the direct evi- dence is lacking in black throat helpers, our anatomi- cal evidence is more informative. Morphological characters of the three sacrificed males were identical (Table 2), indicating that all had achieved mature adult size. However, testicular masses were different; only the black-throated male had mature testes, whereas the two yearling males with spotted throats had small regressed testes. The testicular masses were different between two yearling males (Table 2). Unfortunately, we have in- sufficient data to explain the individual difference of testicular masses between these two yearlings. Such a difference, considering that these were immature in- dividuals, might fall within the range of individual variations related to physiological conditions, or 98 might result from ecological constraints that acted to either advance or inhibit the developmental process of testes. In the former case, the difference of testicu- lar masses might not be functional. In the latter case, the development of testes might correlate with the ex- tent of helping activity of yearling males (cf. Eguchi et al. in press). However, even if the difference of tes- ticular masses affected helping behavior, the anatomi- cal data indicated clearly that the yearling males had not reached sexual maturity (Fig. 2). Related helper males of the cooperatively breeding African White-browed Sparrow Weaver also have smaller testes (around 100mg) than breeding males (>300mg) in January/February (Wingfield et al. 1991). In this species, plasma levels of luteinizing hormone and testosterone in the helpers are also lower than in breeding males, indicating that helpers are not fully mature. Similarly, in the cooperatively breeding Florida Scrub-jay, helpers are subordinate and nonbreeders, and have smaller testes than the breeding males, with whom they share a territory (Schoech et al. 1996). Suppression of hormonal levels of helpers is known in several cooperative breeding species (Schoech et al. 1996; Wingfield et al. 1991; Poiani & Fletcher 1994; Mays et al. 1991). However, whereas those studies emphasized the role of spontaneous re- straint by subordinates, we suggest that, due to their underdeveloped testes, the spotted-throat males (one- year-old males) of the Rufous Vanga are physically incapable of breeding. In the cooperatively breeding Mexican Jay, year- ling males also never breed, and have low reproduc- tive steroid levels during the breeding season (Vleck & Brown 1999). Two-year-old males do regularly breed whether or not they are nest owners. Vleck and Brown (1999) consider that yearling males delay sex- ual maturity in order not to pay any physiological costs of high testosterone levels. Furthermore, year- ling males of the Mexican Jay may be tolerated by other males to a greater extent than those males with high testosterone levels, because they are readily identifiable through their juvenile bill coloration (Vleck & Brown 1999). The spotted-throat coloration of the male Rufous Vanga might have a similar effect as the bill coloration of the Mexican Jay, and is as- sumed to be an example of delayed plumage matura- tion (Rohwer et al. 1980; Greene et al. 2000). De- layed maturation is expected to delay the dispersal of males, and therefore may be closely related to the maintenance of the cooperative breeding system of Spotted-throat individuals of the Rufous Vanga the Rufous Vanga. Additionally, it is evident that fe- males achieve sexual maturity at age one, and thus the delayed maturation of males affects the opera- tional sex ratio. Although the population sex ratio containing yearling males was biased toward males (0.60 on average), the operational sex ratio was 0.55 on average (Asai et al. unpubl.). ACKNOWLEDGMENT We thank the director, Dr. A. Randrianjafy, and other staff at the Botanical and Zoological Park Tsimbazaza, for their cooperation throughout our project in Mada- gascar. Thanks also to Drs. T. Hino, M. Nakamura, H. Nagata, H. Hotta, H. Kofuji, T. Mizuta and Messrs. T. Masuda, F. Iwasaki, S. Fukushima, Y. Takeda, F. Rako- tondraparany, J. R. Ramanampamonjy, B. Raveloson and Mss. H. Amano, M. Tanimura and J. Razanatsoa for | their help in collecting data and to Conservation Interna- | tional for providing facilities in Ampijoroa. This study | was partly supported by a grant under the Monbusho In- | ternational Scientific Research Program (Field Re- | search, No. 06041093 and No.11691183 to Yamagishi). REFERENCES | Asai S, Shimoda C, Nishiumi I, Eguchi K & Yamagishi S (1999) Isolation of microsatellite loci for paternity testing in the rufous vanga Schetba rufa. Mol Ecol 8: 513-514. | Eguchi K, Yamagishi S, Asai S, Nagata H & Hino T (in press) Helping does not enhance reproductive success of cooperatively breeding rufous vanga in Madagas- car. J Anim Ecol. 99 Greene E, Lyon BE, Muehter VR, Ratcliffe L, Oliver SJ & Boag PT (2000) Disruptive sexual selection for plumage coloration in a passerine bird. Nature 407: 1000-1003. Mays NA, Vleck CM & Dawson J (1991) Plasma luteinizing hormone, steroid hormones, behavioral role, and nest stage in cooperatively breeding Harris’ hawks (Parabuteo unicinctus). Auk 108: 619-637. Poiani A & Fletcher T (1994) Plasma levels of andro- gens and gonadal development of breeders and helpers in the bell miner (Manorina melanophrys). Behav Ecol Sociobiol 34: 31-41. Rohwer S, Fretwell SD & Niles DM (1980) Delayed maturation in passerine plumages and the deceptive acquisition of resources. Am Nat 115: 400-437. Schoech SJ, Mumme RL & Wingfield JC (1996) De- layed breeding in the cooperatively breeding Florida scrub-jay (Aphelocoma coerulescens): inhibition or the absence of stimulation? Behav Ecol Sociobiol 39: 77-90. Vleck CM & Brown JL (1999) Testosterone and social and reproductive behaviour in Aphelocoma jays. Anim Behav 58: 943-951. Wingfield JC, Hegner RE & Lewis DM (1991) Circulat- ing levels of luteinizing hormone and steroid hor- mones in relation to social status in the cooperatively breeding white-browed sparrow weaver, Plocepasser mahali. J Zool Lond 225: 43-58. Yamagishi S, Honda M, Eguchi K & Thorstrom R (2001) Extreme endemic radiation of the Malagasy vangas (Aves: Passeriformes). J Mol Evol 53: 39-46. Yamagishi S, Urano E & Eguchi K (1995) Group com- position and contributions to breeding by rufous vangas Schetba rufa in Madagascar. Ibis 137: 157— 161. ORIGINAL ARTICLE Ornithol. Sci. 1: 101—110 (2002) Nest-site selection of the Red-billed Leiothrix and Japanese _Bush Warbler in Japan ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2002 Hitoha E. AMANO!* and Kazuhiro EGUCHI’ ' Graduate School of Social and Cultural Studies, Kyushu University, Ropponmatsu, Fukuoka 810-8560, Japan ? Department of Biology, Faculty of Science, Kyushu University, Fukuoka 812-8581, Japan Abstract The Red-billed Leiothrix Leiothrix lutea has been introduced from China and is rapidly increasing in deciduous broad-leaved forests of Japan. We studied nest- site characteristics and nest-site selection of this species and the Japanese Bush War- bler Cettia diphone, a sympatric native species, in southwestern Japan. Both species placed nests exclusively in bamboo thickets and on bamboo stalks. The Red-billed Leiothrix built pendulous nests in the canopy of high concealment. The Japanese Bush Warbler placed nests on the crossing of bamboo stems and selected places of high stem density. The Japanese Bush Warblers placed nests in denser vegetation than the Red-billed Leiothrix. The segregation of nesting microhabitat was also evident in both species to coexist in bamboo thickets. Existence of few inhabitants in bamboo Key words Many birds have been transported out of their na- tive ranges for trading, which has facilitated estab- lishment of naturalised populations in various regions of the world (Long 1981). Biological invasion raises several ecological questions. First, what factors affect / invasion success (Case 1991; Veltman et al. 1996). Second, what are the direct or indirect influences of introduced species on native species (Diamond & | Case 1985; Lodge 1993). The second question is es- pecially important for people who deal with the con- servation of native avifaunas. In spite of its impor- “tance, there have been few studies on this subject, | particularly during the establishment of introduced birds. Introduced birds may diminish the number of na- tive species through interspecific competition (Moun- -tainspring & Scott 1985; Jones 1996). However, exis- tence of interspecific competition is hard to detect (Lodge 1993), and evaluations of the influences of in- : troduced birds on native ones have caused disputes | (e.g. Moulton & Pimm 1983; 1986; Simberloff & Boecklen 1991; Simberloff 1992; Moulton 1993). It | is important to clarify the ecological characteristics of (Received 17 September 2001; Accepted 12 November 2001) * Corresponding author, E-mail: hitorcb@mbox.nc.kyushu-u.ac.jp : : thickets may contribute to the invasion success of the Red-billed Leiothrix. Cettia diphone, Introduced birds, Leiothrix lutea, Nest-site selection both introduced and native species in order to discern the reasons for successful introduction and competi- tion between them. A study of habitat selection is one approach to achieve this clarification (Sol et al. 1997: Of forest-living birds which are believed to be less successful in establishment (Long 1981), two timali- idid species (Melodious Laughing-thrush Garrulax canorus and Red-billed Leiothrix Leiothrix lutea) and the Japanese White-eye Zosterops japonica have suc- cessfully established in native forests of Hawaii (Mountainspring & Scott 1985). The Red-billed Leio- thrix has established also in deciduous broad-leaved forests of Japan in recent years (Yamashina Institute for Ornithology 1993; Eguchi & Masuda 1994). No- tably, the same species have established in native forests of different regions which are believed to be resistant habitats to exotic birds. The Red-billed Leiothrix ranged originally from southern China to the Himalayan region (Ali & Rip- ley 1972; Long 1987) and is a popular caged bird in Western countries. In Hawaii, North America and the European continent, naturalised populations have been reported (Long 1981; Lever 1987). In Japan, the number of naturalised individuals has been increasing rapidly in deciduous broad-leaved forests above Hitoha E. AMANO and Kazuhiro EGUCHI 1,000m in elevation during the past two decades (Eguchi & Masuda 1994; Tojo 1994). However, in- formation on the ecology of the Red-billed Leiothrix in natural habitat is generally lacking in native and introduced regions (but see Fisher & Baldwin 1947; Ralph et al. 1998). In deciduous broad-leaved forests in Japan, the Red-billed Leiothrix forages in the lower layer of the forest and nests in thickets of dwarf bamboo (Eguchi & Masuda 1994). Only one sympatric species, the Japanese Bush Warbler Cettia diphone, nests in the same habitat where competition between them is likely. In Hawaiian islands, the Red-billed Leiothrix and Melodious Laughing-thrush diminished the num- ber of native species through interspecific competi- tion (Mountainspring & Scott 1985). Whether such an influence upon Japanese Bush Warblers has been exerted by the Red-billed Leiothrix is an important concern. Two major resources, nest-sites and foods, may be limiting, if competition exists. Our objective is to clarify characteristics and selection of nest-site in the two species and to determine the effect of nest- site microhabitat on nesting success. We discuss the possible impact of the Red-billed Leiothrix on native species with reference to the utilisation of nesting habitat. METHODS The study was conducted from April to August, 1997-1999 in the Ebino Plateau, Miyazaki and Kagoshima Prefectures, southwestern Japan (1,200 m elevation; 31°56’N, 130°51’E). The main study area was a mixed forest (16ha) composed of Abies firma, Tsuga sieboldii, Pinus densiflora, Quercus crispula, Hydrangea paniculata, Symplocos coreana and S. myrtacea. In the shrub layer, dwarf bamboo Sasamor- pha borealis of ca. 2m in height, was spread predom- inantly throughout the forest, with small patches of bare ground. A road of ca. 10m in width ran through the forest (Fig. 1). In 1997, additional data were col- lected at a 6ha site, 800m northwest of the main study area. The vegetational structure in the second site was the same as in the main study area. The an- nual mean precipitation exceeds 5,000mm on the Ebino Plateau, of which more than one-third occurs during June and July (data from Miyazaki Branch, Weather Service of Japan). It was about twenty years since the Red-billed Leiothrix was first recorded in this area (Kamitanigawa pers. comm.). Nests were located by systematic searches in the 102 Yy Vj Dan OD , OS Uy eo YJ ™ % 4a | Xs Epi Uff» Fig. 1. Distribution of nests at a study site in 1998. Closed circles are the nests of Red-billed Leiothrix and open circles are those of Japanese Bush Warbler. The thick line is a road and thin lines are paths. All of the study area is covered with a continuous canopy of tall trees with dwarf bamboo (hatched areas) and bare ground (open areas) comprising the under- story. bamboo thicket, using song and behaviour of birds as a cue. Nests were marked with plastic flags. Nests were checked at 3-7 days intervals until failure or fledging. Chicks of both species fledge between 10 to 15 days after hatching (pers. obs.). If chicks disap- peared during this stage, we searched for them around nest-sites to confirm whether they fledged or not. If at least one chick fledged, the nest was as- signed as “successful”. Because we could not distin- guish between first and second nesting attempts for most nests, we combined them in the analyses. After the termination of breeding, each nest was visited for measurement of nest-site characteristics. Variables measured were; size of nest, nest materials, nest height from the ground, height of bamboo canopy, length of the longest stalk supporting a nest, distance from the root of supporting bamboo to the nest, the number of twigs supporting a nest, the num- Nest-site selection of Leiothrix ber of stems supporting a nest, distances from nearest stream, road, and edge of thicket, and density of veg- etation. The density of vegetation was evaluated as follows. A red plastic board (2540cm) was placed 50cm above the ground adjacent to the stalks sup- porting the nest, and a photograph of the board was taken from 1 m away. We scanned these photographs and measured the area of the red portion of the board using the public domain NIH Image program (devel- oped at the U.S. National Institutes of Health and available on the Internet at http://rsb.info.nih.gov/nih- image/) on a Macintosh computer. The density of vegetation is indicated by the area of the board ob- scured by the vegetation (=1,000—the area of red portion of the board). In 1999, we did not conduct nest measurements. The density of the bamboo thicket varied from rel- atively open to quite dense. In order to examine the relationship between the density of bamboo and dis- tance from the road, we sampled bamboo density in ten 50X50cm quadrats (2 quadrats sets 2m apart at 10m intervals) along a 50m transect. Five and four transects were set east and west of the road, respec- tively. We counted the number of live bamboo stems in each quadrat. Data from two quadrats were summed for each point. Similarly, four transects were set parallel to the road between neighbouring streams. Because the foliage of bamboo is concentrated around the terminal end of the stem, nests in the fo- liage are of low visibility while those in the middle of stem are of high visibility. In 1999, in order to exam- ine the nest-site selection of both species, the stem density and the visibility index were measured at the nests and at control points. To quantify the stem den- sity, we counted the number of stems in a quadrat of 5050 cm. Three visibility indices were measured as follows. A white board (30*45cm) with 5x5cm grids was placed at heights of 1 m (approximate bush warbler nest level) and 1.8m (approximate leiothrix nest level). At a distance of 1.5m from the board, we counted the number of intersections visible with the naked eye or through a video camera. The index ranged from 0 (no visibility) to 54 (perfect visibility). In order to quantify the degree of visibility at height of 1.8m (the level of bamboo canopy), we used two different indices, the one from a point | m above the ground and the other from a point 1.8m above the ground. Control points were chosen at a distance of 10m north of each nest. If the point chosen fell on bare ground, another point was chosen to the south of 103 the nest. Logistic regression analysis of data collected in 1999 was used to determine characteristics related to nest-site selection. The dependent variable was nomi- nal: places preferred for nesting and control points. Independent variables were stem density and the three visibility indices. Only completed nests were used for the analyses. Except for the interspecific comparison of the density of vegetation around nests and the analysis of nesting success, only data from the main study area were used. However, because all information was not al- ways obtained for all nests, sample sizes were differ- ent among analyses. Means are shown with SD. A significant level is at ~=0.05. RESULTS 1) Nest shape and nest attachment Many more Red-billed Leiothrixs’ nests were found than those of the Japanese Bush Warbler: 67 nests (49 at the main site and 18 at the second site) in 1997, 83 in 1998, and 84 in 1999 for the Red-billed Leiothrix, compared to 25 nests (17 at the main site and 8 at the second site) in 1997, 41 in 1998, 22 in 1999 for the Japanese Bush Warbler. New nests of the Red-billed Leiothrix were cup-shaped, 9.7+1.0.cm in diameter and 9.341.2cm in height for 135 nests measured (data combined for 1997 and 1998) and were made of blades of bamboo, moss, roots of plants and, sometimes, plastic cords. Nests of the Japanese Bush Warbler were elliptical ball-shaped, 10.1+1.4 cm in diameter and 14.9+2.4cm in height for 56 nests measured, and were made entirely of blades of bamboo. Both species placed nests on bamboo stalks in dense bamboo thickets. Only one nest of the Red- billed Leiothrix was found in the understory lacking bamboo and dominated by Symplocos myrtacea. The Red-billed Leiothrix placed their nests in fo- liage at the top of bamboo stalks. Although the length of stalks ranged from 2 to 3m, stalks bent due to the weight of the nest so the nest height fell in the range of 1-2.5m above the ground (Table 1). Nests were hung at the forks of stems with roots of plants and plastic cords. Each nest was supported by two or three stems. In a few cases, only one stem supporting the nest. The Japanese Bush Warbler placed nests at the middle height of stems, 0.5—2 m from the ground. Nests were attached to four or five stem crossings and were held to the stems rather than tied. Therefore, more stems were necessary to support a nest of the Hitoha E. AMANO and Kazuhiro EGUCHI Japanese Bush Warbler than that of the Red-billed Leiothrix (Table 1). The nest height of Japanese Bush Warblers was lower than that of the Red-billed Leio- thrix, though not significant, in 1997. However, this does not mean that the Japanese Bush Warblers pre- ferred short bamboo thickets. The height of vegeta- tion around the nests was not significantly different between these two species (Table 1). The Japanese Bush Warbler preferred nesting lower on the bamboo stalks than the Red-billed Leiothrix. A high density of stems were necessary for the Japanese Bush Warbler to fasten the nest, while den- sity of stems was not so important for the Red-billed Leiothrix. Many nests of Japanese Bush Warblers were located where the density of vegetation was high (Fig. 2). Mean values of indices of vegetation density were 743.94107.0 (N=68) in 1997 and 529.9+157.7 (N=70) in 1998 for the Red-billed Leiothrix, and 808.5+116.3 (N=26) in 1997 and 604.8+ 148.6 (N=27) in 1998 for the Japanese Bush Warbler. In both years, Japanese Bush Warblers placed nests in denser vegetation than Red-billed Leiothrix (U4, ¢g=572, P=0.01 in 1997, U,,7,=698, P=0.05 in 1998; Mann-Whitney U-test). Overlap in the spatial distribution of nests between both species was large (Fig. 1). However, nests of the Red-billed Leiothrix were distributed throughout the study area while those of Japanese Bush Warblers were concentrated near the road. The proportion of nests within 10m of the road was greater for Japan- ese Bush Warblers than for Red-billed Leiothrix in both 1997 and 1998 (Table 2). Streams and bare ground also interrupted the bamboo thicket. The pro- portions of nests within 10m of streams or bare ground were not different between the two species (Table 2). The bamboo thicket was densest near the road (Fig. 3). However, there was not a consistent ten- dency between the density of bamboo stalks and the distance from a stream; 27.5+15.0 stalks per 5,000 cm’ (N=13) within 10m of a stream and 22.6+17.0 stalks (N=13) more than 10m from a stream (U,3 ,3=65.5, P>0.30, Mann-Whitney U-test). Be- cause tall trees were lacking near the road, the bam- boo thicket grew well. However, because there was a continuous canopy of tall trees over the streams, the density of the bamboo thicket was not as high there, as near the road (27.5+15.0 stalks (N=13), versus 95.6+33.7 stalks (N=9), respectively, U,,,=0, P< 0.0001). 2) Nest-site selection In the bamboo thickets, both species selected nest- sites non-randomly. Nests were placed at points of higher concealment and higher density of bamboo stems than control points by both species (Table 3). Table 1. Comparison of nest placement between Red-billed Leiothrix and Japanese Bush Warbler. 1997 1998 Red-billed | Japanese Bush Red-billed Japanese Bush Leiothrix Warbler Leiothrix Warbler N 40 16 72 32 mean SD mean SD U [pe mean SD mean SD U pe Height (cm) Nest 165.8 34.7 146.9 27.3 232 >0.10 159.7 33.3 137.8 32.2 731.5 <0.01 Bamboo 236.3 29.9 238.8 25.8 313 >0.80 218.8 41.55 2209 34.8 1087 >0.60 Length (cm) Nest* 214.4 385 142.2 37.6 219.5 <0.001 Bamboo? 257.1 37.1 220.9 37.5 524.5 <0.001 Relative nest height® 3.8 0.5 3.0 0.6 112 <0.001 3.7 0.5 2.9 0.5 334 <0.001 Nest Twig 6.3 2.3 6.1 16 314.5 >0.90 6.5 2.1 7.8 2.3. 750.5 <0.01 attachment® Stem 2.6 1.2 4.0 1.4 138 0.001 2.9 1.0 4.8 14 309.5 <0.001 * distance from the root of supporting bamboo to the nest > length of the longest stalk supporting a nest “ divided into four portions (1st, 2nd, 3rd and 4th from ground to top) * number of twigs (stems) supporting a nest “ Mann-Whitney’s U test 104 Nest-site selection of Leiothrix Japanese Bush Warber 1997 rey c 30 o z 8 20 = ire 10 0 0 200 400 600 800 1000 Red-billed Leiothrix 1997 50 40 > e 30 o > Oo 20 = LL tt) 200 400 600 800 1000 Index of vegetation density Fig. 2. Comparisons of the density of vegetation around ese Bush Warbler. The index of vegetation density is shown Japanese Bush Warber 1998 o i] i=} o Py So o a So o o So o = o Red-billed Leiothrix 1998 0 200 400 600 800 1000 Index of vegetation density nests between the Red-billed Leiothrix and the Japan- as an area of a board covered with vegetation. Table 2. Locations of nests of Red-billed Leiothrix and Japanese Bush Warbler. 1997 1998 Red-billed Japanese Bush Red-billed Japanese Bush Leiothrix Warbler Leiothrix Warbler N 50 18 Vi 36 Distance from road Less than 10m 6 10 21 26 More than 10m 44 8 56 10 Pp* 0.001 <0.001 Distance from bare ground Less than 10m 24 7 30 11 More than 10m 26 11 47 MS Pe >0.50 >0.40 Distance from streams Less than 10m 22 5 46 24 More than 10m 28 13 31 12 P? >0.20 >0.50 * Fisher’s exact probability-test For Red-billed Leiothrix, the differences were signifi- cant for the stem density and visibility indices in the canopy but not for the visibility index at the lower height. However, logistic regression indicated that the overall contribution of these variables was low (R’=0.14) and no particular characteristic was re- 105 lated to the separation between nests and control points (Table 4). On the other hand, the Japanese Bush Warbler selected high stem density and well concealed places. The density of bamboo stems and visibility index in the canopy were related to the sep- aration between nests and control points (Table 4). Hitoha E. AMANO and Kazuhiro EGUCHI 3) Nesting success and vegetation density at nest- site In the Red-billed Leiothrix, fate of 51 nests was traced and five nests were successful (9.8%) in 1997, five from 63 nests (7.9%) in 1998, and only two from 200 160 120 80 Number of bamboo stems per 5000 cm? 40 0 10 20 30 40 50 Distance from the road (m) Fig. 3. The relationship between the density of bamboo stems and distance from the road. The line connects means. 68 nests (2.9%) in 1999, respectively. In the Japanese Bush Warbler, only one nest was successful from 18 nests (5.3%) in 1997, and no nest was successful in both 1998 (from 33 nests) and 1999 (from 14 nests). More than a half of nests studied were left no egg laid. In the Red-billed Leiothrix, the proportions of nests in which eggs were laid were 43.1% (=22/51), 34.9% (=22/63), 39.7% (=27/68), in 1997, 1998 and 1999, respectively. In the Japanese Bush Warbler, the proportions of nests in which eggs were laid were 50.0% (=9/18), 36.4% (=12/33), 21.4% (=3/14), in 1997, 1998 and 1999, respectively. We could not de- termine whether these failed nests were deserted be- fore clutch initiation or were depredated soon after that. Most failures after clutch initiation were due to total loss of clutches or broods, which suggests nest predation. Combined data for three years: 58 of 59 nests in which causes of failure were confirmed in the Red-billed Leiothrix and 20 of 23 nests in the Japan- ese Bush Warbler. We observed the Jay Garrulus glandarius and snakes (species unknown) predating nestlings. In the Red-billed Leiothrix, neither the density of vegetation nor nest height related to nesting success. The indices of vegetation density of the successful nests and failed nests were 769.9+166.8 vs. 740.84 101.9 in 1997 (U; 45=83, P>0.40; Mann-Whitney U- test), 586.34224.2 vs. 528.8+147.3 in 1998 (Us y= 90, P>0.30) and the nest height were 174.0+48.3 cm Table 3. Comparison of the stem density and the visibility index between nests and control points for Red-billed Leiothrix and Japanese Bush Warbler. Red-billed Leiothrix Japanese Bush Warbler N 31 22 mean SD mean SD No. of stems Nest 27.6 139 38.3 11.4 control 19.7 25) Daye \\ iiss; Pe <0.05 0.001 Visibility index | m above ground Nest 16.4 10.3 13.3 7.6 control PINES) 11.3 18.4 10.0 Be >0.10 <0.05 canopy from below Nest 5.4 4.1 5.6 6.9 control 9.6 6.7 8.0 6.6 Be <0.05 <0.05 canopy level Nest 4.4 4.2 3.1 3.3 contro] 8.8 7.6 8.2 8.0 lee <0.01 0.001 * Wilcoxon’s signed-ranks test 106 Nest-site selection of Leiothrix Table 4. Logistic regression analysis of the stem density and the visibility index between nests and control points for Red- billed Leiothrix and Japanese Bush Warbler. Red-billed Leiothrix Japanese Bush Warbler N 62 44 Characteristics Parameter estimate P Parameter estimate P Intersect —0.77 0.53 2.64 0.15 Density of stems —0.02 0.48 —0.11 0.01 Visibility index at 1 m above ground 0.01 0.72 0.01 0.93 Visibility index in canopy from below 0.08 0.29 —0.13 0.14 Visibility index in canopy 0.07 0.29 0.27 0.04 R°=0.14 R?=0.36 mse lO 292532. 5.cm. (U5 ,,=90, P>0.:60) im 1997), 152.0+32.7 cm vs. 156.1£31.3 cm in 1998 (U; 4.=98, P>0.40; Mann-Whitney U-test). In the Japanese Bush Warbler, sample sizes of nesting success were too small to test. In 1999, because few nests were successful, we could not show the relationships between nesting suc- cess and nest characteristics (the stem density and in- dices of visibility). DISCUSSION 1) Utilisation of bamboo thickets The Red-billed Leiothrix builds nests on bamboo not only in the Ebino Plateau but also in other areas of Japan (this study; Nakamura & Tojo unpubl.; Eguchi & Amano unpubl.). Only a few species are known to nest in bamboo thickets in Japan, including the Japanese Bush Warbler, Grey Bunting Emberiza variabilis, Siberian Meadow Bunting E. cioides, Bull-headed Shrike Lanius bucephalus and the Japan- ese White-eye (Kiyosu 1951). In Mt. Tsukuba, _ Ibaraki prefecture, the Japanese Bush Warbler, Japan- ese White-eye, Siberian Meadow Bunting and Long- | tailed Tit Aegithalos caudatus nest in thickets of bamboo Sasamorpha borealis (Nakamura & Tojo un- publ.). On the Kayanodaira Plateau, Nagano Prefec- ture, the Japanese Bush Warbler and Grey Bunting are the dominate nesting species in thickets of dwarf bamboo Sasa senanensis and S. kurilensis (Ezaki pers. comm.). However, only two native species, the Japanese Bush Warbler and Grey Bunting, nest pre- _ dominantly in bamboo thickets in Japan. Although the avifauna is rich in deciduous broad-leaved forests, it is likely that the microhabitat which the Red-billed Leiothrix uses for nesting has been vacant. 107 The Red-billed Leiothrix is not a specialist nesting on bamboo stalks in its original habitats (Ali & Rip- ley 1972; Long 1987). In Hawaii, Red-billed Leio- thrix nests in dense undergrowth, but does not spe- cialise in a particular substrate tree species (Fisher & Baldwin 1947). Also in Japan, there have been a few cases where Red-billed Leiothrix nested on shrub trees, such as Symplocos myrtacea, Abies firma, Eurya japonica, Litsea glauca, and Camellia japon- ica (Eguchi & Amano unpubl.). The Monk Parakeet Myiopsitta monachus, a habitat generalist in its origi- nal range, became a specialist nesting solely on palms of Phoenix sp. in Spain (Sol et al. 1997). Sol et al. (1997) suggested that the preference of a habitat oc- cupied by no other species allows the parakeet to in- crease rapidly in the initial stage of invasion and that gradually it expands the width of habitat preference after successful establishment. The Red-billed Leio- thrix also could increase in number by nesting in bamboo thickets where few competitors live and by specializing in nesting substrate and position of the nest. 2) Difference in nest-site selection Both the Red-billed Leiothrix and Japanese Bush Warbler selected dense vegetation for nesting. How- ever, the species selected different nesting microhabi- tats. Red-billed Leiothrix built nests in the canopy of bamboos, while the Japanese Bush Warbler placed nests at the middle height of stems. Such a difference may be due primarily to a difference in manner of nest attachment between two species. The Red-billed Leiothrix built pendulous nests at the top of bamboo stalks, while Japanese Bush War- blers built nests at the crossing of bamboo stems. Japanese Bush Warblers require a bamboo thicket of Hitoha E. AMANO and Kazuhiro EGUCHI high density for fastening nests firmly. Therefore, nests of Japanese Bush Warblers are distributed in the vicinity of the forest edge, such as near roads where the bamboo thicket is densest. On the other hand, be- cause the density of bamboo stalks is not important for attaching nests, the Red-billed Leiothrix selects a wider range of understory density for nest-sites than the Japanese Bush Warbler. The Japanese Bush Warbler also places nests in the middle of the bamboo stem in other regions of Japan. On the Kayanodaira Plateau where the Red-billed Leiothrix has not invaded yet, the Grey Bunting places nests at the top of bamboo stalks in areas of relatively sparse thickets, while the Japanese Bush Warbler nests in denser thickets (Ezaki pers. comm.). Thus, two species with different nest-site preferences could coexist in bamboo thickets. 3) Nest-site selection and nesting success Dense vegetation may provide shrub-nesting species excellent protection against predators and non-random nest-site selection has been found in other such species (e.g. Black-throated Blue Warbler, Holway 1991; Hooded Warbler, Kilgo et al. 1996; Wood Thrush, Hoover & Brittingham 1998). The Japanese Bush Warbler selected places of high stem density. Dense bamboo stalks may impede the move- ment of predators. The Red-billed Leiothrix placed nests in the canopy of bamboos of high concealment. However, logistic regression suggested that this species may not always select a place of highest con- cealment in a territory for nesting. The Red-billed Leiothrix sometimes built an exposed nest just above a stream. Apparently, nest concealment is not a pri- mary factor determining nest-site selection in this species. Although the Red-billed Leiothrix and Japanese Bush Warbler selected somewhat of nest-site charac- teristics, the nesting success was very low in these species due to predation, particularly evident in al- most all nests of the Japanese Bush Warbler. The den- sity of vegetation at the nest site and nest height did not influence nesting success. We observed nest pre- dation by the Jay and snakes (species unknown). An- other potential predator may be the Jungle Crow Corvus macrorhynchos. Avian predators such as corvids depend on visual cues, and activities of par- ent birds may be important cues for such predators (Holway 1991; Hoover & Brittingham 1998). We often observed jays flushing from canopies of bam- boos. Some jays may have searched exclusively for 108 nests of birds nesting in the bamboo thicket when eggs and nestlings of these species were available. On the other hand, snakes search for nests using ol- factory cues (Burhans & Thompson 1998). The effect of nest-site characteristics to predation may vary de- pending on the predator species. If a guild of preda- tors is composed of species searching in different ways, selection of a specific type of nest-site may not be advantageous (Filliater et al. 1994; Hoover & Brit- tingham 1998). High nest predation and failure to de- tect a significant relationship between nesting success and nest characters in this study may be due to a var- ied predator community. Moreover, because the Red-billed Leiothrix is a newcomer in this area, there may have not been enough time for a counter adaptation against a new guild of predators to appear. However, this species renested rapidly after predation and breeding season was long, with egg-laying occurring from April to September (pers. obs.). Rapid renesting and a long breeding season may be adaptations to high nest pre- dation as reported in other shrub-nesting species (Martin 1995). Further studies evaluating a role of each taxonomic group of predators on mortality of birds nesting in the bamboo thicket are needed. a) bler So far, no distinct influence of the Red-billed Leio- thrix on native species, such as a decline in number or habitat shrinkage, has yet been reported. The loca- tions of nests in the bamboo thicket differ between the Red-billed Leiothrix and Japanese Bush Warbler. Even in the case where nests of these species were close to each other, direct interaction suggesting in- terspecific territoriality was not observed. It is un- likely that competition for nest-sites occurs between these species. However, indirect interference competition is prob- able. In the community of Acrocephalus warblers in the reed beds of Europe, such an increase of preda- tion due to the coexistence of species of similar nest- site preference is also considered probable (Hoi et al. 1991). When we searched for nests of the Red-billed Leiothrix, nests of the Japanese Bush Warbler were also found. A high density of nests may attract vari- ous kinds of predators into the breeding areas. In this study, the density of Red-billed Leiothrix nests, which was two to three times higher than that of Japanese Bush Warblers, may have caused an in- Influence on breeding of the Japanese Bush War- crease in the number of predators and predation on nests of shrub-living species (the functional response; Holling 1959). The breeding success of Japanese Bush Warbler may be low due to indirect interference competition caused by the increase of Red-billed Leiothrix. Further studies are needed including a monitoring in an area of initial invasion or eliminat- ing experiments for the Red-billed Leiothrix. The Red-billed Leiothrix has invaded various re- gions of the world (Long 1981). Comparisons of habitat selection among naturalised populations and among naturalised populations and populations in their original range may indicate flexibility of habitat preference in this species, as well as changes in habi- tat selection during the process of invasion. Further studies are needed both in the original and new habi- tats of the Red-billed Leiothrix . ACKNOWLEDGMENTS This study was financially supported by the Ministry of Education and Culture of Japan (Grant No. 08454252) to K. Eguchi and by the Sasakawa Scientific Research Grant from the Japan Science Society to H. E. Amano. We are very grateful to Mr. N. Kamitanigawa for lodging in the study area and providing information on the Red-billed Leiothrix, Y. Ezaki, S. Hamao, S. Nakamura, A. Oshiro and H. Tojo for providing infor- _ mation on the Red-billed Leiothrix or other shrub-nest- ing species, and K. Kawano and H. Kofuji for field as- _ sistance. Dr. B. J. Sinclair reviewed an earlier draft and improved the English. REFERENCES Ali S & Ripley SD (1972) Handbook of the Birds of India and Pakistan, Vol 7. Oxford University Press, Bombay. Burhans DE & Thompson FR III (1998) Effects of time and nest-site characteristics on concealment of song- bird nests. Condor 100: 663-672. Case TJ (1991) Invasion resistance, species build-up and community collapse in metapopulation models with interspecies competition. Biol J Linn Soc 42: 239-266. Diamond JM & Case TJ (1985) Community ecology. _ Harper and Row, New York. Eguchi K & Masuda T (1994) A report on the habitats | of Peking robin Leiothrix lutea in Kyushu. Jpn J Or- nithol 43: 91-100 (in Japanese with English sum- mary). | Filhiater TS, Breitwisch R & Nealen PM (1994) Preda- 109 Nest-site selection of Leiothrix tion on northern cardinal nests: Does choice of nest- site matter? Condor 96: 761-768. Fisher HI & Baldwin PH (1947) Notes on the red-billed leiothrix in Hawaii. Pac Sci 1: 45-51. Hoi H, Eichler T & Dittami J (1991) Territorial spacing and interspecific competition in three species of reed warblers. Oecologia 87: 443-448. Holling CS (1959) Some characteristics of simple types of predation and parasitism. Can Ent 91: 385-398. Holway DA (1991) Nest-site selection and the impor- tance of nest concealment in the black-throated blue warbler. Condor 93: 575-581. Hoover JP & Brittingham MC (1998) Nest-site selection and nesting success of wood thrushes. Wilson Bull 110: 375-383. Jones C (1996) Bird introductions to Mauritius: status and relationship with native birds. In: Holmes JS & Simons JR (eds) The Introduction and naturalisation of birds. pp 113-123. Stationery Office Publishing Centre, London. Kilgo JC, Sargent RA, Chapman BR & Miller KV (1996) Nest-site selection by hooded warblers in bot- tomland hardwoods of South Carolina. Wilson Bull 108: 53-60. Kiyosu Y (1951) The Birds of Japan Vol 1. Kodansha, Tokyo (in Japanese). Lever C (1987) Naturalised birds of the world. Long- man, London. Lodge DM (1993) Biological invasions: lessons for ecology. Trends Ecol Evol 8: 133-137. Long JL (1981) Introduced birds of the world. Reed, Wellington. Long Z (1987) Leiothrix. In: Chen T, Long Z & Zheng B (eds) Fauna Sinica. pp 154-162. Science Press, Beijing (in Chinese with English summary). Martin TE (1995) Avian life history evolution in rela- tion to nest-sites, nest predation and food. Ecol Monogr 65: 101-127. Moulton MP (1993) The all-or-none pattern in intro- duced Hawaiian passeriforms: the role of competition sustained. Am Nat 141: 105-119. Moulton MP & Pimm SL (1983) The introduced Hawai- ian avifauna: biogeographic evidence for competition. Am Nat 121: 669-690. Moulton MP & Pimm SL (1986) The extent of competi- tion in shaping an introduced avifauna. In: Diamond JM & Case TJ (eds) Community Ecology. pp 80-97. Harper and Row, New York. Mountainspring S & Scott JM (1985) Interspecific com- petition among Hawaiian forest birds. Ecol Monogr 55: 219-239. Ralph CJ, Fancy SG & Male TD (1998) Demography of an introduced red-billed leiothrix population in Hitoha E. AMANO and Kazuhiro EGUCHI Hawaii. Condor 100: 468-473. Simberloff D (1992) Extinction, survival, and effects of birds introduced to the Mascarenes. Acta Oecol 13: 663-678. Simberloff D & Boecklen W (1991) Patterns of extinc- tion in the introduced Hawaiian avifauna: a reexami- nation of the role of competition. Am Nat 138: 300- 327: Sol D, Santos DM, Feria E & Clavell J (1997) Habitat selection by the monk parakeet during colonization of a new area in Spain. Condor 99: 39-46. 110 Tojo H (1994) Population increase of the red-billed leio- thrix Leiothrix lutea in the Massif Tsukuba. Jpn J Or- nithol 43: 39-42 (in Japanese with English summary). Veltman CJ, Nee S & Crawley MJ (1996) Correlates of introduction success in exotic New Zealand birds. Am Nat 147: 542-557. Yamashina Institute for Ornithology (1993) Report of the bird migration research center 1993. Kokusai Bunken Insatsu, Tokyo (in Japanese with English summary). ‘ bs - “bee Belin; him aia ir »h is ol _— i re) brn ‘an ia j ee CR ieing behias ti ithe; i f fixn sa a siti ¥ A 2 ae! stecny hy ; Fobeig \s . se Ga teeribernt! 4 oi x Mf hed larritactr pernse RE i874 Wr ex bine ferrscnw hcites | meres . vi wi , writ 7 i) tT. awl Wem - AreG) BS OF b OV ere Me vas i'8" *, 7 i 2 2 ” a — ee aru oc {i worttuaA ahi! is “i oripiel warn of (Tt hots, Paarsetittiastte vol srl) Btorerl Warmcuuliary | mopennt “ - hartecntetnicior) tot af si PUP TGR i geri wit aces @ & 5 VINO N Go 7A. a | epebes what rfid) Bs tats rcpepy!4 os estan is ‘agra AE ureveto >. nT Thetaky MAIN NP Mgt OE drifter see Reb reuhe niin DA Mii «iu. } Pettiiry Qudivints: edt (blocrht-ae ‘boryess ny Fatt att ems eoomitalitte geynuni Gs erellib diiw edjuA oo Instructions to Authors Ornithological Science is wholly owned and published by the Ornithological Society of Japan (OSJ), and published in English twice a year (January and July). Membership to OS] is not required for contributors to Ornithological Science. Manuscripts may be submitted internationally for free publication up to 20 pages, although color pages will be charged separately to contributors. The contribution can be classified into four categories: Reviews, Original Articles, Short Communications and Comments. Com- ments are only for the papers published on Ornithological Science. Accepted papers automatically become the copyrighted property of OSJ and may not be reproduced, in whole or part, without a written permission from OS]. Submission Manuscripts have to be original and so far unpublished. The contributors are required to prepare three complete photocopies (the original not required at the time of reception) of manuscript, tables and figures. Floppy disks of 3.5-inch containing a text file of the manuscript is also required after acceptance. The Editorial Board of Ornithological Science will then forward the submitted materials to unidentified referees, and taking ac- count of their opinions, the Editor-in-Chief will decide the validity of the submitted material for publication. In case the contribu- tors were requested to reconsider part of their materials, they are supposed to amend the material within a month before return- ing the materials with a note describing how the amendments had been carried out. Should the contributors’ and the referees’ opinions show discrepancy, the contributors are allowed to express their view separately. Any material overdue the above time period is automatically interpreted as having been withdrawn by the contributors. International submission of papers prepared by Microsoft WORD and PDF files may be accepted via e-mails. Here follows the address for submission with contacts Dr. Teruaki Hino, Associate Editor Kansai Research Center, Forestry and Forest Products Research Institute, Kyoto 612-0855 JAPAN. TEL: +81-75-611-1201, FAX: +81-75-611-1207, E-mail: tkpk@affre.go.jp Organization of Manuscript The Editorial Board suggest the following preferred organization of an original article: (1) Title, Author name(s) and Affilia- tion(s) (2) Abstract (3) Introduction (4) Materials and Methods (5) Results (6) Discussion (7) Acknowledgments (8) References (9) Tables (10) Figures and Legends. It can be modified and changed for clarification of the whole structure and for effectiveness of arguments. In a Short Communication, no abstract is required. Physical Features of Manuscript The manuscripts should be printed on A4 sheets, with one-inch margins for refereeing, 20 lines per sheet. No rigid limit for the length of manuscripts are set, but the Editorial Board deems anything exceeding 20 pages in a final published form to be long enough. Short Communications have to be less than four pages in time of publication. All manuscript sheets must be numbered consecutively. The use of footnotes to the text is not recommended. Extra Charges A charge of 2,500 yen will be added to every page that exceeds 20 free pages. Each color page will be charged 50,000 yen. The Editorial Board suggests adherence to the following guidelines of manuscript submission. Contributors please feel free to ask the Editorial Board for more details. General Instructions 1. Cover page The first page of the manuscript should contain the title, author names, affiliations, addresses, the contact name and address (phone and/or facsimile number together with an e-mail address) for correspondence. Authors with different affiliations should be identified by the use of a superscript linking the name and affiliation. Also on the same page, a running title (less than 40 letters) and the number of tables and figures accompanying the manuscript. To discriminate the family name(s) from their first and/or the middle ones, contributors are encouraged to spell their family names with all capitals. 2. Abstract and key words Every submission should accompany an abstract less than 300 words followed by alphabetically arranged key words less than six. 3. Names of birds Contributors are requested to give scientific name at the first mention of a species both in the abstract and in the article. Scien- tific and English name of birds should follow an authoritative source for the appropriate regions. Capitals should be used for the initial letters for all single words or hyphenated English names (e.g., Black-billed Magpie) but not a group name (e.g., crows). 4. References References in the text should be cited as follows: One author: Tokugawa (1995) or (Tokugawa 1995) Two authors: Oda and Smith (1996) or (Tokugawa 1995; Oda & Smith 1996) More than two authors: Mori et al. (1997) or (Mori et al. 1997; Richard et al. 1999a, b) Manuscripts accepted for publication but not yet published should be listed in the reference section as ‘in press.’ All contributors are advised to prepare a list of references cited in the text arranged alphabetical by the first author names. Book: Cambell RC (1974) Statistics for biologists. 2nd ed. Cambridge University Press, London. Chapter in an edited book: Dawson WH (1996) Energetic features of avian thermoregulatory responses. In: Carey C (ed) Avian energetics and nutri- tional ecology. pp 85-124. Capman & Hall, New York. Oring LW (1982) Avian mating system. In: Farner DS, King JR & Parkes KC (eds) Avian biology. Vol 1. pp 1-92. Academic Press, New York. Journal article: Yamaguchi N & Kawano KK (2001) Effect of body size on the resouce holding potential of male varied tits Parus varius. Jpn J Ornithol. 50: 65-70. Papers written in a language other than English has to be accompanied by a language designators such as “(in Japanese)” to- gether with a translated title. Yamagishi S (1981) Mozu no yomeiri-toshikouen no mozu no seitai wo saguru (The bridal of shrikes -ecological research of bull-headed shrikes in an urban park). Dai-Nippon-Tosho, Tokyo (in Japanese). The sources of the reference should be given following the commonly accepted abbreviations for journal titles (refer to the ‘In- ternational List of Periodical Title Abbreviations’). The use of “in preparation”, “submitted for publication”, or “personal commu- nication” is not allowed in the reference list. Also, “unpublished data” and “personal communication” should appear parenthet- cally following the relevant name(s) in the text. 5. Units and Abbreviations All measurements should be given in SI units. Also, the Editorial Board suggests 24-hour clock (e.g., 0930 and 1500) and the geographical location 35°20'25"N, 136°10'20’E. Statistical symbols used in the contribution are to be capitalized except df (e.g., SD, SE, P, N, df). The Board further prefers unitalicized expressions for Latin phrases frequently used. 6. Charts and Illustrations The contributors are supposed to take the size of the printed page into consideration when they prepare their tables and fig- ures: the actual print is on 165X223 mm in double columns with 80mm width. ORNITHOLOGICAL SCIENCE Volume 1 PREFACE Higuchi H From Japan through Asia to the world: building bridges in ornithological science. EDITORIAL Ueda K “Ornithological Science”, the new English publication from Japan. SPECIAL FEATURE Interspecific segregation and attraction in forest birds Hino T Introduction. Loyn RH Patterns of ecological segregation among forest and woodland birds in south-eastern Australia. Recher HF, Davis WE Jr & Calver MC Comparative foraging ecology of five species of ground-pouncing birds in western Australian woodlands with comments on species decline. Monkkonen M & Forsman JT Heterospecific attraction among forest birds: a review. Number 1 29 41 January 2002 Contents Seki S & Sato T The effect of a typhoon on the flocking and foraging behavior of tits. Murakami M Foraging mode shifts of four insectivorous bird resource temporally varying Japanese species under distribution in a deciduous forest. Mizutani M & Hijii N The effects of arthropod abundance and size on the nestling diet of two Parus species. Hino T, Unno A & Nakano § Prey distribution and foraging preference for tits. Sodhi NS The effects of food-supply on Southeast Asian forest birds. ORIGINAL ARTICLES Yamagishi S, Asai S, Eguchi K & Wada M Spotted-throat individuals of the Rufous Vanga Schetba rufa are yearling males and presumably sterile. Amano HE & Eguchi K Nest-site selection of the Red-billed Leiothrix and Japanese Bush Warbler in Japan. Published by the Ornithological Society of Japan Printed by Kokusai Bunken Insatsusha Co., Ltd. a3 63 7h 8] 89 95 101 Ana TAM annz U 1 3 ) pu o } / *” : i RS ORNITHOLOGICAL SCIENCE Vol. 1 No. 2 . The Ornithological Society of Japan ORNITHOLOGICAL SCIENCE Official journal of the Ornithological Society of Japan Editor-in-Chief Keisuke Ueda, Rikkyo University, Tokyo Associate Editors Teruaki Hino, Forestry and Forest Products Research Institute, Kyoto Hidetsugu Sakai, Nihon University, Tokyo Editorial Board Masahiko Nakamura, Joetsu University of Education, Joetsu Isao Nishiumi, National Science Museum, Tokyo Kazuo Okanoya, Chiba University, Chiba Eiichiro Urano, Yamashina Institute for Ornithology, Abiko Advisory Board Alexander V. Andreev, Institute of Biological Problems, Magadan Walter J. Bock, Columbia University, New York Jiro Kikkawa, The University of Queensland, Brisbane Woo-Shin Lee, Seoul National University, Suwon Bernd Leisler, Max-Planck-Gesellschaft, Radolfzell Anders P. Mller, Universite Pierre et Marie Curie, Paris Richard Noske, Northern Territory University, Casuarina Pilai Poonswad, Mahidol University, Bangkok Lucia Liu Severinghaus, Academia Sinica, Taipei Navjot S. Sodhi, National University of Singapore, Singapore Jeffrey R. Walters, Virginia Polytechnic Institute and State University, Blacksburg John C. Wingfield, University of Washington, Seattle Jeong-Chil Yoo, Kyung-Hee University, Seoul Editorial Policy Ornithological Science publishes reviews, original articles, short communications and comments covering all aspects of ornithology. Manuscripts are judged on the basis of their contribution of original data and ideas or interpretation. All articles are peer-reviewed by at least two researchers expert in the field of the submitted paper. Manuscripts are edited where necessary for clarity and economy. Ornithological Science aims to publish as rapidly as is consistent with the requirements of peer-review and normal publishing constraints. Submission Manuscripts and editorial correspondence should be ad- dressed to: Dr. Teruaki Hino, Associate Editor, Forestry and Forest Products Research Institute, Kyoto 612-0855, Japan. Tel: +81-75-611-1201, Fax: +81-75-611-1207, e-mail: tkpk @affre.go.jp For detailed instructions concerning the submission of man- uscripts, please refer to the Instructions to Authors at the in- side of the back cover of each issue of the journal or visit the OSJ web page: http://wwwsoc.nii.ac.jp/osj/ Subscriptions Membership of the Ornithological Society of Japan (OSJ) is open to anyone interested in ornithology and the aims of the OSJ. The annual fee for an ordinary member is Japanese Yen 5,000. Members are entitled to receive two journals. Ornithological Science (in English; two issues per year) and Japanese Journal of Ornithology (with English summary; two issues per year). To join the OSJ, please apply to: The OSJ Office, Laboratory of Biodiversity Science, School of Agricultural and Life Sciences, The University of Tokyo, Yayoi 1-1-1, Tokyo 113-8657, Japan. Tel: +81-3-5841-7541, Fax: +81-3-5841-8192, e-mail: osj@lagopus.com Copyright Submission of a manuscript implies: that the work de- scribed has not been published before; that is not under con- sideration for publication elsewhere; that its publication has been approved by all co- authors, if any, as well as by the re- sponsible authorities at the institute where the work has been carried out; that, if and when the manuscript is accepted for publication, the authors agree to automatic transfer of the copyright to the OSJ; and that the manuscript will not be pub- lished elsewhere in any language without the consent of the copyright holders. ISSN: 1347-0558. All articles published in the Ornithological Science are pro- tected by copyright, which covers the exclusive rights to re- produce and distribute the article (e.g. as offprints) as well as all translation rights. No material published in this journal may be reproduced photo-graphically or stored on microfilm, in electronic data bases, video disks, etc., without first obtaining written permission from the copyright holders. Special re- quests should be addressed to: Dr. Keisuke Ueda, Editor-in- Chief, Laboratory of Animal Ecology Faculty of Science, Rikkyo University, Ikebukuro, Tokyo 171-0021, Japan, Tel: +81-3-3985-2596, Fax: +81-3-3985-2596, e-mail: keisuke @rikkyo.ac.jp Printer: Kokusai Bunken Insatsusha Co., Ltd., Takada-no-baba 3-8-8, Shinjuku-ku, Tokyo 169-0075, Japan. Tel: +81-3-3362- 9741, Fax: +81-3-3368-2822. Cover design: Eiichiro Urano + (N/A: 381] Sdequieur ur epnjouy) (furs + Jo afin 2 ) [feuw-Y] (N/A: 4ST] Stequiew ul epnypouyT) XV (Jus + fu %bQpiney ) (N/A : 4ST] Stequiew UT epnyouT) (Jay + Jupp apae A ) TAL (su0Y EA - YOM ARM) * PUME}OD [SsoJppy] _ at dF} [owen] 2M WEay (9 SEES ze YA ‘sjoyoeig UI UONdO pajydaJas VIDI1D asea[g “ARMA ODVALY SY Ju OF BOVE *F) LUCMAKSRHE LEVER YHD Lic r Breese oO ex ¢ 22 tH A eh A Tel FW 15463 A 31H EC ef Woy te} ER iced aI 8 ik ws & naeoaen H bi AE ORR ORE YN ae Zh me OBR AO “FEAR % i a fe Aa? fF St a Be re eA a Fu ai ia Special Message to Members of the Ornithological Society of Japan SAA SEROLOORA’YT—h HA BSS Hla T 113-8657 FR ARC ae SHAE 1-1-1 RRACK ht aE a SAE AED 2S REP ET eB RUT ah: 03-5841-7541 FAX: 03-5841-8192 e-mail: osj@lagopus.com HABYS Cs, COU, FBS 2002 FBMeRTTSACLELERNELE, TO ZBISSADAIMHESNAZLELDCH), SBOAAMAAMICMABICART SC LIE CEnTHEVD ET. AiBloik, AAW BS BAL CHAM ISHM STORE REVERB AZATHOEST. K TY7—-bCE, CNYELOWROGMS AA SNSAMEIMEDRICLAEBAETS CFC, SROAGASAOWRSE BRACE KEER< Bo CTHVET. CHALACE, VPORMEBIKSBAREEX, TOMMeE FAX CRKUTRIEFE Cm, MCUADILME CRKROLEGHLEOTCS, iis, fl CUR MEI4SO A -— 4X— Y http://wwwsocuniiac.jp/osj/ (= & BRK XC >. CHOSeAAALT, BFA -—N ORRLTOIKE EW TEDMEVDEGA, SAAEIIs, CESFBOZMREMEMEADERMHSEM, SSBMSTRRVYE Jeeta, LAU BMOWEUET. Fi TCBRAROBSMS, SB CHB CO4RAeTOXEMIATELICK OEGOC, CTKXK FAW, igh Ie vets: WPOARE CHASES. BMRB S eC, TRO (Bee - AAT) DES SWILOe THA HSV. 1. KG 2. & OW 3. fer (Fea eC RESWADOATHAS KEW) aT. fae SEMIS, CO MO5-O BE (LE S6SERO DF ery 7eECHASESYW) 4. Baas (Zieteakcny - Ao] ) 5. FAX #3 — — (4 egal - An) 6. X-NPTFFVA @ (4 ieteakay - Aa) Special Message to Members of the Ornithological Society of Japan Ornithological Society of Japan Members List Questionnaire Office of the Ornithological Society of Japan Laboratory of Biodiversity Science, School of Agriculture and Life Sciences University of Tokyo, Yayoi 1-1-1, Bunkyo-ku, Tokyo 113-8657, JAPAN Phone: +81—3-5841-7541 Fax: +81-3-5841-8192 Email: osj@lagopus.com The Ornithological Society of Japan (OSJ) plans to publish a comprehensive Members List. This list will be distributed to all OSJ members, but will not be available for distribution outside the Society. At this time, we would like to ask for members’ names and addresses to include on this list. As well as creating the membership list, the information received will be used to check that our records are up to date, and thus to ensure delivery of material from the Society. The questionnaire below also appears on the Society homepage, at http://wwwsoc.nii.ac.jp/os}. If you use the online form, please feel free to forward your response to us by email, at osj@ lagopus.com. Otherwise, we would be pleased to receive your reply by fax or regular mail. If re- plying by fax or mail, please use the postcard included in this issue, or fill in and send the section below. We hope to include all members of the Society on the Members List. If we do not receive a re- sponse from you, we will include your records as they currently appear in the Society’s files. Thank you very much. Please answer the questions and check /circle the appropriate responses below: 1. Name: 2. Address including postal code (Fill in this section only if the address this questionnaire has been mailed to differs from the address you wish to use in the Member’s List): C) Work address (1) Home address Include in Members List 3. Phone number: - ~ rg N 4. Fax Number: ~ - nf N 5 bd N cS) Email address: 2 fd bes ~peigg Pina, : ORIGINAL ARTICLE Teruaki HINO* ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2002 Kansai Research Center, Forestry and Forest Products Research Institute, Kyoto 612-0855, Japan Ornithol. Sci. 1: 111-116 (2002) Breeding bird community and mixed-species flocking ina deciduous broad-leaved forest in western Madagascar Abstract The breeding bird population of a deciduous broadleaved forest in west- ern Madagascar was censused by means of territory mapping. Despite the foliage structure being simpler, neither species richness nor density was less than those in mature temperate forests. Species diversity was higher in the western Madagascan forest owing to the higher species evenness. Tree-cavity nesters and bark foragers were few because woodpeckers, nuthatches, and tits have not colonized Madagascar. The scarcity of birds nesting on or near the forest floor may be attributable to abun- dance of nest-predators such as large lizards and snakes in these areas. The bird com- munity was dominated in abundance by the members of mixed-species flocks, almost all of which forage in the canopy. Mixed-flocking can be beneficial for these birds to avoid predation by raptors, which were frequently observed in the canopy. Since most of the flock members had relatively similar territory sizes resulting in similar densi- ties, the high species evenness in this community may have resulted from mixed- flocking by canopy-foraging species. Key words Species evenness Bird species diversity is generally higher in tropi- cal forests than in temperate forests. This latitudinal gradient of species diversity has been mainly ex- plained by external factors such as the structural complexity of habitat (MacArthur et al. 1966), cli- matic stability (Stiles 1978) and predictability or di- versity of food resources (Karr 1971; Schoener 1971). In contrast, Powell (1989) explained the high species richness in the neotropical avifauna as arising from the internal structure of the community itself. Multispecies territoriality (i.e., the year-round com- munal defence of territory) by the core species of mixed-species flocks reduces the densities of small species because they have larger territories than ex- pected from their body size. As a result, a greater number of small species can coexist owing to the under-utilization of food resources in such a commu- nity. Madagascar, which lies within the tropical region, supports various kinds of forests including: rain for- est, deciduous broad-leaved forest, and subarid scrub. (Received 5 February 2002; Accepted 4 July 2002) * E-mail: tkpk @affre.go.jp 111 Forest bird community, Madagascar, Mixed-species flocks, Predators, Although the avifauna is highly unidue with >50% of the breeding species endemic (Langrand 1990), there have been no quantitative studies of Madagas- can forest bird communities. Multispecies bird flocks are observed all year round in Madagascan forests (Eguchi et al. 1993; Hino 1998, 2000) as well as in other tropical forests (Bell 1983; Powell 1989; Jullien & Thiollay 1998). Hino (1998, 2000) has shown that the core species of mixed flocks in deciduous broadleaved forest in Madagascar gain mutual bene- fits relating to foraging and/or anti-predation. The de- ciduous forest is an appropriate habitat to examine the factors, other than foliage structure, that may ex- plain the differences between bird communities in tropical and temperate forests. In this paper, I de- scribe the characteristics of breeding bird communi- ties in a deciduous broad-leaved forest in western Madagascar by comparing them with those of tem- perate forests in Japan. Then I consider the effect of predators and multispecies flocking to explain the characteristics of the Madagascan bird community. T. Hino METHODS This study was conducted in a 450550-m quadrat (called Jardin A) in a deciduous broadleaved forest in the Ampijoroa Forest Reserve (16°35’S, 46°82’E; ca. 200 m asl), about 110km southeast of Mahajanga. In this forest, 113 tree species are listed by Razafy (1987). The mean annual precipitation of the area is about 1500 mm, 97% of which is recorded during the wet season (November to April), and the mean an- nual temperature is 26.8°C (Razafy 1987). Ten censuses of breeding bird populations were conducted using the territory mapping method, walk- ing steadily around the whole study site from 10-20 October 1994. All birds seen or heard within 50 m of either side of the census trail were recorded on a scale map. The territories overlapping the border of the study area were recorded as half territories (0.5). Raptorial species (e.g. Otus rutilus, Accipiter mada- gascariensis, A. francessi, Polyboroides radiatus) and forest-edge species (e.g. Falculea _palliata, Merops superciliosus, Caprimulgus madagascarien- sis, Coracopsis vasa), which were observed during the census, were excluded from the analysis of this study. At each encounter, whether each individual partici- pated in mixed-species flocks (foraging with two or more different species for more than 10 min; Jullien & Thiollay 1998) was recorded from 10 October to 12 November. The mixed-flocking propensity of each species was calculated as the percentage of the num- ber of times the species was found foraging in a mixed flock relative to the total number of times this species was encountered (Jullien & Thiollay 1998). Categorization of foraging and nesting habits fol- lowed Langrand (1990), Yamagishi et al. (1997), Morris and Hawkins (1998) and my field observa- tions, and body lengths followed Langrand (1990). Species diversity (D) and evenness (E) were calcu- lated using Simpson’s indices: D=1/2P? and E= D/S, where P, is the proportion of abundance for species i and S is the number of species. Vegetation structure was surveyed in 120 sample plots (Sm 5m) at 50m intervals on 15-17 Novem- ber 1994. The number of trees or woody plants stems in each of four categories of diameter at breast height (DBH: 30cm) that typi- cally forage on the ground and those that eat plant material. Almost all of the most abundant species were regular or occasional members of mixed-species flocks (Fig. 2). A significant positive correlation was found between abundance and flocking propensity (Kendall’s t=0.506, P<0.0001). DISCUSSION Although the deciduous broad-leaved forest stud- ied in western Madagascar was mature, the character- istics of its vegetation structure (i.e. abundant small trees and low foliage height) corresponded to those of a young forest in a temperate region (Aber 1979). In No. of pairs/km? Japan° No. of species* No. of pairs/km? 436.4 7-11 61-83 50.5 10-12 195-279 al (yy) 111-209 2.0 l 2-9 294.9 10-12 185-229 20.2 5— 6 53-61 133.3 QI Il 109-180 S2Al 2a 39-100 427.3 20-21 317-404 WS y= 9 74-152 0.0 l= 2 2-6 499.0 29-31 422-491 16.9 13.3-15.0 0.58 0.46—-0.52 A=animals (arthropods, small reptiles), P=plant materials (seeds, fruits, nectar), O=ominifarious food 0.5 was given to each group for the birds categolized to two groups (e.g., F/B in Table 2) 12 Flocking propemsity 10 BM 40%< B 20-40 % 8 O <20% Relative abundance (%) fo>) 10 15 Species rank 20 25 Fig. 2. Mixed-flocking propensity and the relative abun- dance of species arranged in decreasing order of abundance. general, bird species diversities are lower in young forests than in mature forests owing to their simpler structure (Bongiorno 1982; Helle 1985). 114 Forest Bird Community in Western Madagascar Neither the species richness nor the density of breed- ing birds (except raptorial or forest-edge species) in the study area differed from those of mature temper- ate forests in Europe and Japan (about 30 species and 500-700 pairs/km/’; Hino 1990, 1993). Moreover, species diversity was high owing to the high species evenness. That is, the Madagascan study area had a more diverse bird community than would be pre- dicted from an examination of its vegetation struc- ture. Razafy (1987) listed 113 species of trees in the forest reserve including the study area. This number of tree species is much higher than that seen in tem- perate forests (<40 species at most, Hino 1990). One of the factors explaining the high bird diversity of the study area may be the high tree species diversity of the forest. Many studies have shown that more di- verse bird communities are found in forests with more diverse tree composition (e.g. Rice et al. 1984; Verner & Larson 1989). The composition of nesting-site groups differed considerably between Madagascan and Japanese bird communities. Tree-hole nesters were few in Mada- gascar because ancestral woodpecker, nuthatch, and tit species failed to colonize the island. In particular, the absence of woodpeckers (which excavate their own nesting cavities annually thereby creating a valu- able cavity resource in trees), must have had a con- siderable negative influence on the nesting habits of other species. The scarcity of species nesting on or near the forest floor may be attributable to the abun- dance of terrestrial nest-predators such as large lizards Oplurus cuvieri and snakes Leioheterodon madagascariensis. These reptiles are considered to be major predators on the eggs and/or fledgling of Schetba rufa (Eguchi et al. 2001), Terpsiphone mu- tata (T. Mizuta pers. com.) and Coua coquereli (T. _ Masuda pers. com.) although they build their nests in _ the low (1-5 m high) canopy. The composition of foraging—site groups also dif- fered in abundance between Madagascan and Japan- ese bird communities. The reason why the bark-for- aging birds were few will be the same one why the cavity nesters are few, that is, the failure in coloniza- tion of ancestral woodpecker, nuthatch, and tit species. Although some of the Vangidae species ob- served during this study, such as Xenopirostris damii and Leptopterus viridis, have evolved as bark for- agers thorough adaptive radiation (Yamagishi & Eguchi 1996), this niche seems not to have been full occupied by birds in Madagascar. The present bird community was dominated, in 115 terms of abundance, by the members of mixed- species flocks. Flocking propensities were more than 48% among regularly flocking species and more than 25% among occasionally flocking species despite ob- servations being made during the breeding season. In the non-breeding season, flocking propensities were almost double (Hino 1998). Almost all flocking species forage in the canopy, where raptors (e.g. O. rutilus, A. madagascariensis, A. francessi, P. radiatus) were frequently observed. Mixed-species flocking may be an effective strategy for avoiding predation as well as of achieving increased foraging efficiency (reviews in Morse 1977, Barnard and Thompson 1985). In fact, mobbing of raptors by flock members was often observed in the study area. The anti-predatory value of mixed-flocking may be enhanced if the flock members move consistently to- gether within a communal territory, that is, multi- species territoriality in neotropical rain forests (Munn & Terborgh 1979; Powell 1989; Jullien & Thiollay 1998). Powell (1989) demonstrated that multispecies territoriality should increase species richness. Species evenness should also increase through this system since the community is composed of species with the same density. Although the mixed-flock members in the study area did not hold communal territories rig- orously, similar sized territories appear to be main- tained by regular flock-members (except the most dominant species Newtonia brunneicauda) resulting in similar densities (20-38 pairs/km?). N. brunne- icauda may have adjusted its home range size to that of the other flock members by forming conspecific flocks consisting of two pairs with neighboring terri- tories (Hino 2000). In the present avian community, therefore, the high species evenness may have re- sulted from mixed-flocking among canopy-foraging species. ACKNOWLEDGMENTS I am grateful to A. Randrianjafy, J.R. Ramanampa- monjy and F. Rakotondraparany of the Tsimbazaza Botanical and Zoological Park for their kind coopera- tion in the course of our project in Madagascar and to UNESCO/PNUD and Conservation International for providing convenient facilities in Ampijoroa. I thank S. Yamagishi and K. Eguchi for their helpful advice in re- lation to this research. This study was supported by a Grant-in-Aid for Overseas Scientific Survey from the Ministry of Education, Science and Culture of Japan (06041093). T. Hino REFERENCES Aber JD (1979) Foliage-height profiles and succession in northern hardwood forests. Ecology 60: 18-23. Barnard CJ & Thompson DBA (1985) Gulls and Plovers. Croom Helm, London. Bell HL (1983) A bird community of lowland rainforest in New Guinea: mixed-species feeding flocks. Emu 82: 256-275. Bongiorno SF (1982) Land use and summer bird popu- lations in northwestern Galicia, Spain. Ibis 124: 1—20. Eguchi K, Yamagishi S & Randrianasolo V (1993) The composition and foraging behavior of mixed-species flocks of forest-living birds in Madagascar. Ibis 135: 91-96. Eguchi K, Nagata H, Asai S & Yamagishi S (2001) Nesting habits of the Rufous Vanga in Madagascar. Ostrich 72: 201-218. Fujimaki Y (1986) Breeding bird community in a decid- uous broad-leaved forest in southern Hokkaido, Japan. Jpn J Ornithol 35: 15-23. Fujimaki Y (1988) Breeding bird community of a Quer- cus mongolica forest in eastern Hokkaido, Japan. Jpn J Ornithol 37: 69-75. Helle P (1985) Habitat selection of breeding birds in re- lation to forest succession in northeastern Finland. Ornis Fenn 62: 113-123. Hino T (1990) Palaearetic deciduous forests and their bird communities: comparisons between east Asia and west-central Europe. Keast A (ed) Biogeography and ecology of forest bird communities. pp 87-94. SPB Academic Pub, Hague. Hino T (1993) Bird fauna and its distribution in Hokkaido. In: Higashi S, Osawa A & Kanagawa K (eds) Biodiversity and ecology in the northernmost Japan. pp 89-102. Hokkaido University Press, Sap- poro. Hino T (1998) Mutualistic and commensal organization of avian mixed-species foraging flocks in a forest of western Madagascar. J Avian Biol 29: 17-24. Hino T (2000) Intraspecific differences in benefits from feeding in mixed-species flocks. J Avian Biol 31: 441-446. Hino T & Nakano S (1992) Breeding bird community of a deciduous broad-leaved forests in northern 116 Hokkaido, Japan. Res Bull Hokkaido Univ For 49: 195—200 (in Japanese with English summary). Jullien M & Thiollay JM (1998) Multi-species territori- ality and dynamics of neotropical forest understory bird flocks. J Anim Ecol 67: 227-252. Karr JR (1971) Structure of avian communities in se- lected Panama and Illinois habitats. Ecol Monogr 41: 207-229. Langrand O (1990) Guide to the birds of Madagascar. Yale University Press, New Haven. MacArthur RH, Recher H & Cody M (1966) On the re- lation between habitat selection and species diversity. Am Nat 96: 319-332. Morris P & Hawkins F (1998) Birds of Madagascar. Pica Press, East Sussex. Morse DH (1977) Feeding behavior and predator avoid- ance in heterospecific groups. Bioscience 27: 332-339. Munn CA & Terborgh JW (1979) Multi-species territo- riality in neotropical foraging flocks. Condor 81: 338-347. Powell GVN (1989) On the possible contribution of mixed species flocks to species richness in neotropi- cal avifaunas. Behav Ecol Sociobiol 24: 387-393. Razafy FL (1987) La REserve ForestiEre d’Ampijoroa: Son modEle et son bilan. MEmoire de fin d’Etudes, University Madagascar, Antananarivo (in French). Rice J, Anderson BW & Ohmart RD (1984) Compari- son of the importance of different habitat attributes to avian community organization. Wildl Manag 48: 895-911. Schoener TW (1971) Large-billed insectivorous birds: a precipitous diversity gradient. Condor 73: 154-161. Stiles EW (1978) Avian communities in temperate and tropical alder forests. Condor 80: 276-284. Verner J & Larson TA (1989) Richness of breeding bird species in mixed-conifer forests of Sierra Nevada, California. Auk 106: 447-463. Yamagishi S & Eguchi K (1996) Comparative foraging ecology of Madagascar vangids (Vangidae). Ibis 138: 283-290. Yamagishi S, Matsuda T & Rakotomanana (1997) A field guide to the birds of Madagascar. Kaiyusha, Tokyo (in Japanese and Malagasy). Ornithol. Sci. 1: 117-122 (2002) ORIGINAL ARTICLE Secondary users of Great Spotted Woodpecker (Dendro- Copos major) nest cavities in urban and suburban forests in Sapporo City, northern Japan Nobuhiko KOTAKA'** and Shigeru MATSUOKA? ' Laboratory of Animal Sociology, Faculty of Science, Osaka City University, Sugimoto, Sumiyoshi-ku, Osaka 558-8585, Japan ? Woodland Bioecology Group, Hokkaido Research Center, Forestry and Forest Products Research Institute, Sapporo 062-8516, Japan Abstract Old nest cavities excavated by Great Spotted Woodpeckers (GSW) Den- drocopos major were examined in two study areas (urban and suburban forests) in Sapporo, the capital city of Hokkaido, northern Japan. Five avian and one mammalian secondary cavity user (SCU) species occupied 47 of 101 GSW cavities inspected. The species composition differed between urban and suburban forests. Avian SCU species occupied GSW cavities more frequently in the urban than in the suburban forests. Tree Sparrows Passer montanus and Chestnut-cheeked Starlings Sturnus philippensis were the only dominant cavity breeding species in the severely fragmented urban forests. Flying Squirrels Pteromys volans were the most dominant users of GSW cav- ities in the suburban forests. The density of GSW cavities depends not only on natural processes but also on human activities. The suitability of the GSW cavities for certain SCU species decreases with time. To maintain the diversity of cavity-nesting wildlife in urban and suburban areas of Sapporo, preservation of existing trees with GSW cay- ities as well as providing suitable habitat conditions to support continued production ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2002 of new cavities is essential. Key words Urban area Cavity-nesting species comprise a major compo- nent of the forest wildlife community (Scott et al. 1980). Martin and Eadie (1999) proposed as a direct analogy to “food webs’ that cavity-nesting bird com- munities are organized in ‘nest webs’ with the cavi- ties as the central resource, around which inter-spe- cific and intra-specific interactions occur. In such a web structure, a certain species may have dispropor- tionate importance if it constitutes a key component of the cavity resources. Primary cavity nesters (PCNs, e.g. woodpeckers) are such key components that excavate cavities used as nests or roosts by sec- ondary cavity users (SCUs) including not only bird but also many other wildlife species. The SCUs rely on the cavities created by PCNs or on natural holes formed through other processes. The number of (Received 15 April 2002; Accepted 9 July 2002) * Corresponding author: E-mail: nkotaka@ msb.biglobe.ne.jp * Present address: Yambaru Wildlife Center, Ministry of the Envi- ronment, Hiji 263—1, Kunigami, Okinawa 905—1413, Japan. 117 Dendrocopos major, Nest webs, Sapporo City, Secondary cavity user, available nesting cavities has been considered to be a major factor limiting the population size of SCU bird species (Haartman 1957; Perrins 1979; Newton 1994). Thus, the density and diversity of woodpeck- ers may have a strong influence on the richness and abundance of other SCU species (Martin & Eadie, 1999). In recent years in Hokkaido, northern Japan, GSW is the only PCN species regularly breeding in urban areas (Yamauchi et al. 1997). In fact, GSW is now the most abundant and often the only PCN species in the urban area of Sapporo, the capital city of Hokkaido (Kotaka & Kameyama in press). In urban areas, GSW may affect entire communities of cavity-nesters through the excavation of nest cavities. However, in- formation on the importance of GSW cavities for other species so far has been just anecdotal, and few studies have quantified utilization of GSW cavities by SCU species in urban areas of modern large cities such as Sapporo. The composition of the nest webs N. KOTAKA and S. MATSUOKA may also vary with habitat features such as forest type or landscape pattern. In this paper, we compare the availability of GSW cavities and their actual occupation frequency by SCU species in two different types of forest (urban and suburban area) in Sapporo to develop a better un- derstanding of the ecological role of GSWs within the nest webs. STUDY AREAS AND METHODS The GSW nest cavities were surveyed in two dif- ferent types of landscapes in Sapporo; a highly frag- mented forest in an urban area (HKD) and a less frag- mented forest in an agricultural suburban area (HJO). Both areas are characterized by a relatively flat to- pography. The former (HKD) includes the University of Hokkaido Campus and its Botanical Garden lo- cated in the center of Sapporo (43°04'N, 141°20’'E) and totalling 271 ha. This area is characterized by highly fragmented woodland (21.3% of tree cover- age) comprising woodland with dense undergrowth, open woodland with little undergrowth, hedgerows, farmland, sports grounds, lawns and buildings. Domi- nant tree-species are Acer mono and Ulmus japonica. The latter (HJO) is situated in the grounds of the Hokkaido National Agricultural Experiment Station and the Forestry and Forest Products Research Insti- tute, totalling about 380 ha, located in the southeast- ern part of Sapporo about 8 km from HKD. This area is characterized by farmland used mainly for crop production and grazing. The tree coverage is about 39.8%, and dominant tree species are Quercus mon- golica and Acer mono. Both study sites were searched for all new GSW nests (HKD in 1994-1997 and HJO in 1995-1998). Locations of the nest cavity-trees were mapped, and each tree was classified as living or dead. The GSW cavities were revisited and examined at least twice in late May and June during the period 1995-1998 in HKD and in 1999 in HJO. Use of the cavities was ex- amined by climbing the trees to inspect cavities with a dentist’s mirror. If adults, eggs or young of certain species were found, the cavity was classified as ‘oc- cupied’. The users of those cavities were identified through observation with binoculars. The nest cavi- ties inspected were further classified into two cate- gories according to their age, “1 year-old” if they were examined one year after excavation and “2-4 year-old” if they were examined two to four years after excavation. When we observed cavity usurpa- 118 tions between cavity users during our survey, we recorded which species were involved. All visits were made during daylight so we have no informa- tion on cavity utilization at night. Some GSW cavi- ties were lost due to natural events or logging, and we classified the former as “broken” and the latter as “logged.” Line transects for bird censuses were estab- lished in both study sites: 50m wide and 1.4km long in HKD and 50m wide and 3.4km long in HJO. At each study site, line transect surveys were conducted from 05:00 to 08:00 in the morning during late May (1995-1998 for HKD and 1999 for HJO). RESULTS 1) Cavity loss, reuse by GSWs and availability to SCUs Through the survey, we found 24 (5, 7, 6, and 6 for 1994, 1995, 1996 and 1997) and 42 (7, 8, 15 and 12 for 1995, 1996, 1997 and 1998) new GSW nests in HKD and HJO, respectively. The proportion of dead trees among the nest trees was significantly higher in HJO (45.2%) than in HKD (4.2%) (Fisher’s exact test, P<0.01). We inspected 59 nest cavities (1 year-old=24; 2-4 year-old=35) in total for HKD (1995-1998) and 42 nest cavities (1 year-old=12; 2-4 year-old=30) for HJO (1999). Some nest cavities or whole trees with nest cavities were lost due to human activity or natu- ral events. The losses amounted to 12.5% of thel year-old and 17.1% of the 2-4 year-old cavities in HKD. The primary cause of cavity losses in this area was logging in the course of road or building con- struction (6 of 8); the other two were lost when trunks broke at a cavity (Fig. 1). In HJO, all cavity losses (16.7% for 1 year-old and 20.0% for 24 year- old cavities) were caused by natural events (mainly strong winds) (Fig. 1). We found four cases of cavity reuse for breeding by GSW (1 in HKD and 3 in HJO, Fig. 1). Cavities were not considered available for SCU when GSWs were using them for nesting. Finally, we found that the proportions of GSW cavities available for SCU species were 87.5% (1 year-old) and 80.0% (24 year-old) in HKD, and 66.7% (1 year-old) and 76.7% (2-4 year old) in HJO. 2) Cavity use by SCUs and nest webs Cavity-nesting bird species are a major component of the bird communities in both of our study sites. From the line transect surveys, the proportion of cay- Secondary Users of Great Spotted Woodpecker Cavities 2-4 year-old N=35 HKD 1 year-old N = 24 Fig. 1. forests. ity-nesting bird species among all recorded individu- als was 75.5% (total bird density: 8.65/ha) for HKD and 62.0% (5.08/ha) for HJO (Table 1). Two species of PCN, GSW and Japanese Pygmy Woodpecker Dendrocopos kizuki were observed at our study sites. For both sites, the GSW was the dominant PCN species (Table 1). Overall, five avian and one mam- malian SCU species used 47 (34 for HKD and 13 for HJO) GSW nest cavities (Fig. 2). Within the available nest cavities, 1 year-old cavi- ties were occupied more frequently than 24 year-old cavities (Fisher’s exact test, HKD: P=0.06; HJO: P<0.05, Table 2). The number of nest cavities in the two different “cavity age groups” used by each species at both study sites, elucidating the shift from | GSW to SCU, can be seen in Figure 2. The composition of the “nest webs” differed fun- damentally between HKD and HJO. In HKD, avian species that used GSW nest cavities were Tree Spar- row Passer montanus, Chestnut-cheeked Starling _ Sturnus philippensis and Nuthatch Sitta europaea. The Tree Sparrows and Chestnut-cheeked Starlings were the dominant GSW cavity users and occupied 94.1% (16 for Tree Sparrow and 16 for Chestnut- cheeked Starling) of 34 cavities used by SCUs. Al- though the density of the Tree Sparrows (4.01/ha) was about five times as high as that of the Chestnut- cheeked Starlings (0.80/ha), we found the occupation rate of the GSWs nest cavities by Chestnut-cheeked Starlings was disproportionately higher than that by Tree Sparrows (G-test, df=1, G=37.4, P<0.001, Fig. _2). One mammalian species (the Flying Squirrel Pteromys volans) was the most dominant user of GSW cavities in HJO, followed by avian species (Nuthatch, Great Tit Parus major and Russet Spar- 1 year-old N=12 HJO 119 2-4 year-old N= 30 Status and utilization of 1 year-old and 2-4 year-old GSW cavities in urban (HKD) and suburban (HJO) Table 1. Bird density (no/ha) of HKD (averaged 1995- 1998) and HJO (1999) Bird Species HKD HJO (no/ha) (no/ha) PCNBs_ Dendrocopos major 0.16 0.20 D. kizuki 0.01 0.04 Total PCNBs 0.17 0.24 SCUs Passer montanus 4.01 — Sturnus philippensis 0.80 = Parus major 0.65 0.71 Sturnus cineraceus 0.62 — Parus palustris 0.10 0.75 Ficedula narcissina 0.09 0.24 Sitta europaea 0.02 0.31 Parus varius 0.01 0.08 Passer rutilans — 0.51 Parus ater — 0.08 Total SCUs 6.29 2.68 Others 2.18 DMN H Total 8.65 5.08 row Passer rutilans). The Flying Squirrel occupied 61.5% (8 of 13) of the GSW cavities used by SCU. It is to be noted that 85.7% (1 year-old) and 57.1% (24 year-old) of all the available GSW cavities ex- amined were occupied in HKD, while in HJO only 25.0% (1 year-old) and 13.0% (2-4 year-old) were occupied by avian SCU species (Fisher’s exact test, 1 year-old: P<0.005; 2-4 year-old: P<0.005). The Nuthatch was the only SCU species that used the GSW cavities at both study sites. Although Great Tits were found at both of our study sites, we could not find any nest belonging to this species among the available cavities in HKD. One of the five Great Tit N. KOTAKA and S. MATSUOKA Nest webs HK wae Passer montanus A Sturnus philippensis te Sitta europea D ——_ a O D. majorO O © D. major Parus major O==F Passer rutilans © Mammalian Pteromys volans @ =H haanaividual mu = 10 % occupation for 1 year-old cavity Seaman — 10 % occupation for 2-4 year-old cavity Fig. 2. Turnover of species using GSW nest cavities in urban (HKD) and suburban (HJO) forests. Arrows con- nect species that use resources provided by GSW. The size of the circles indicates the bird density (no/ha) of each species and the width of the arrows shows the percentage of occupancy of available GSW cavities for each species. The sample sizes for the SCU species are 21 (1 year-old) and 28 (2-4 year-old) for HKD and 8 (1 year- old) and 23 (2-4 year-old) for HJO. Table 2. Utilization of GSW cavities in relation to nest age in urban (HKD) and suburban (HJO) forests. HKD HJO Nest age lyear 2-4 year lyear 2-4 year Occupied 18 16 6 7 Empty 3 12 2 16 nests found in HKD was in a cavity of unknown ori- gin and the others were in natural or artificial holes. 3) Interactions and cavity usurpations In HKD, cavity usurpations between SCU species were observed. Chestnut-cheeked Starlings drove five pairs of Tree Sparrows out of their nest sites, and Chestnut-cheeked Starlings and Tree Sparrows re- placed one and two pairs of Great Tits, respectively. GSWs aggressively harassed Chestnut-cheeked Starlings, Tree Sparrows, and Great Tits when these species approached their nests. In spite of these at- tempts to defend their nest sites, 28.6% (6 of 21) of newly excavated GSW nest cavities were usurped by Chestnut-cheeked Starlings. Three of the six pairs whose cavities were taken over attempted to excavate new nest cavities, but the others did not. Only one of the pairs that made new cavities succeeded in raising 120 their offspring while the others abandoned their nests. In HJO, we did not observe direct competition be- tween SCU species and GSW for nest cavities. DISCUSSION In our study sites, we found that five avian and one mammalian species used GSW cavities. Thus, the GSW may function as a keystone species, in the urban as well as suburban areas of Sapporo, by pro- viding a critical resource—nesting cavities-for SCUs. Woodpecker cavities can enhance breeding success and reduce predation risk (Rendell & Robertson 1989; Li & Martin 1991), and they may provide ther- moregulatory advantages for some SCU species, in- cluding Flying Squirrels (Carey et al. 1997). With the exception of the Nuthatch, none of the species using GSW nest cavities in HKD was found to use them in HJO, and vice versa. It is interesting to note that in HKD where the woodland was highly fragmented, the GSW cavities were frequently occu- pied by avian SCU species, but almost none by mam- malian SCU species. The competition for GSW cavities may be severer in HKD, reflecting its poor habitat quality. In this area Chestnut-cheeked Starling is the largest user of GSW cavities. In HKD, Chestnut-cheeked Starlings Secondary Users of Great Spotted Woodpecker Cavities occupied the same number of GSW cavities as Tree Sparrows did, because they are probably the domi- nant species in competition for the cavities. In HKD, Chestnut-cheeked Starlings even usurped GSW newly excavated cavities. Ingold (1994) sug- gested that woodpeckers (Red-bellied Woodpecker Melanerpes carolinus, Northern Flicker Colaptes au- ratus, Red-headed Woodpecker M. erythrocephalus) can avoid competition with European Starling Stur- nus vulgaris and may not have suffered reductions in fecundity, because at least some of these pairs were successful in building new nests later in the season. However, the success rate of the re-nesters was only 17% (1 of 6) in HKD; pairs of GSW were unable to avoid competition with Chestnut-cheeked Starlings and suffered apparent reductions in fecundity. In HJO, the Flying Squirrel was the main user of GSW cavities. In North America, Flying Squirrels are seen aS major predators or competitors of nesting Red-cockaded Woodpeckers Picoides borealis (Loeb 1993; Loeb & Hooper 1997; but see Mitchell 1999). Dominant competitors (e.g. Chestnut-cheeked Star- lings or Flying Squirrels) may affect utilization and nest-site selection of other SCU species as well as PCNs through their ability to win the competition for cavities. Johnsson et al. (1993) found that the most competitive species, Jackdaws Corvus monedula usu- ally used the best old Black Woodpecker Dryocopus martius holes, while other subordinate species used inferior holes in a Swedish forest. Further research on species-specific cavity selection and ability of com- peting for nest cavities would help to determine func- tional relationships between GSW and SCU species. The significant difference in the utilization fre- quencies between 1 year-old and 2-4 year-old cavi- ties suggests that the suitability of GSW nest cavities deteriorates with time elapsed after excavation. Some of the old GSW nest cavities could obviously not be used by SCU species any more because they were clogged with mushrooms, or because the size and shape of the entrance had changed as the living tis- sues of the tree around it kept growing. Old GSW cavities in living trees seemed to have narrower openings than newly excavated ones, probably due to the lateral growth of the stems. Another factor that has been reported to prevent re- peated use of nest holes is infection with pathogenic organisms or parasites. If old nest material contains larger parasite loads (Perrins 1979; Mller 1989), the lower occupation rate for older (24 year-old) cavi- ties might be an indirect effect of parasite avoidance. 121 In addition to losses of GSW cavities due to natu- ral events, human activities can be of major impor- tance, especially in or near urban areas. Human im- pact is responsible mostly for decreases in the number of available nest cavities, whereas natural processes influence also the suitability of a given nest hole for repeated use. Most woodpecker species use trees for nesting, foraging and communication, and they are extremely sensitive to extensive forest harvesting (Winkler et al. 1995). If local extinction of a keystone species, such as GSW, occurred, we predict that the wildlife com- munity will face a catastrophic ecological cascade. To maintain the diversity of wildlife in urban areas, city planners should protect existing GSW cavity trees and maintain a sustainable breeding habitat for con- tinued production of new cavities. ACKNOWLEDGMENTS We thank Dr. K. Ishida for useful comments on an earlier manuscript. We also thank Y. Takada for helping our field surveys and her valuable input, and two anony- mous referees for contributing to the manuscript. REFERENCES Carey AB, Wilson TM, Maguire CC & Biswell BL (1997) Density of northern flying squirrels in the Pa- cific Northwest. J Wildl Manag 61: 684-699. Haartman L von (1957) Adaptation in hole-nesting birds. Evolution 11: 339-347. Ingold DJ (1994) Influence of nest-site competition be- tween European starlings and woodpeckers. Wilson Bull 106: 227-241. Johnsson K, Nilsson SG & Tjernberg M (1993) Charac- teristics and utilization of old black woodpecker Dry- ocopus martius holes by hole-nesting species. Ibis 135: 410-416. Kotaka N & Kameyama S (in press) Use of geographic information systems for the evaluation of the great spotted woodpecker Dendrocopos major breeding habitat in the urban area of Sapporo City. In: Pechacek P & d’Oleire-Oltmanns W (eds) International Wood- pecker Symposium Proceedings. Forschungsbericht Nationalparkverwaltung Berchtesgaden, Berchtesgaden. Li P & Martin TE (1991) Nest-site selection and nesting success of cavity-nesting birds in high elevation for- est drainages. Auk 108: 405-418. Loeb SC (1993) Use and selection of red-cockaded woodpecker cavities by southern flying squirrels. J Wild] Manag 57: 329-335. N. KOTAKA and S. MATSUOKA Loeb SC & Hooper RG (1997) An experimental test of interspecific competition for red-cockaded wood- pecker cavities. J Wildl Manag 61: 1268-1280. Martin K & Eadie JM (1999) Nest webs: a community- wide approach to the management and conservation of cavity-nesting forest birds. For Ecol Manag 115: 243-257. Mitchell RL (1999) Effects of southern flying squirrels on nest success of red-cockaded woodpeckers. J Wildl Manag 63: 538-545. Moller AP (1989) Parasites, predators and nest boxes: facts and artifacts in nest box studies of birds. Oikos 56: 421-423. Newton I (1994) The role of nest sites in limiting the numbers of hole-nesting birds: a review. Biol Conserv 70: 265-276. Perrins CM (1979) British tits. Collins, London. 122 Rendell WB & Robertson RJ (1989) Nest-site character- istics, reproductive success and cavity availability for tree swallows breeding in natural cavities. Condor 91: 875-855. Scott VE, Whelan JA & Svoboda L (1980) Cavity-nest- ing birds and forest management. In; DeGraaf RM (ed) Management of western forests and grasslands for nongame birds. pp 311-324. US For Serv Gen Tech Rep INT-86. Winkler H, Christie DA & Nurney D (1995) Woodpeck- ers: An Identification Guide to the Woodpeckers of the World. Houghton Mifflin, New York. Yamauchi K, Yamazaki S & Fujimaki Y (1997) Breed- ing habitats of Dendrocopos major and D. minor in urban and rural areas. Jpn J Ornithol 46: 121-131 (Gn Japanese with English summary). Ornithol. Sci. 1: 123-131 (2002) ORIGINAL ARTICLE Foraging niches of introduced Red-billed Leiothrix and native species in Japan Hitoha E. AMANO!” and Kazuhiro EGUCHI? ' Graduate School of Social and Cultural Studies, Kyushu University, Ropponmatsu, Fukuoka 810-8560, Japan ? Department of Biology, Faculty of Science, Kyushu University, Fukuoka 812-8581, Japan Abstract In Kyushu, southwestern Japan, the introduced population of the Red- billed Leiothrix Leiothrix lutea has increased rapidly and its range expanded consid- erably since early 1980s. In order to clarify the influences of Red-billed Leiothrix on native bird species, we examined the similarities and differences in foraging patterns among species occurring in a deciduous broadleaved forest on the Ebino Plateau, dur- ing the breeding seasons from 1997 to 2000. Leiothrix foraged in a lower vegetational layer with bamboo, intermediate in height between the foraging levels of the Japanese Bush Warbler Cettia diphone and various Parus species. Foraging height, extent of foraging on deciduous trees and foraging technique were major factors best distinguishing Leiothrix from native species. Seg- regation of foraging niche was distinct and no apparent niche shift, due to invasion of the new species, was detected. Aerial insects tended to be more abundant just above bamboo, mainly about one meter above the canopy, than above bare ground. Thus, jumping, a specific technique used by Leiothrix, is effective for capturing aerial in- sects or agile invertebrates resting on leaves and twigs. Aerial insects were found to be abundant in the foraging space preferred by Leiothrix. Gleaning and hanging, tech- niques mainly used by native species, are suitable for capturing prey of low mobility such as Lepidoptera larvae. Probably due to morphological constraints, Parus spp. and Japanese Bush Warblers seldom foraged by jumping, indicating that they exploit quite different food resources from those utilized by Leiothrix despite their foraging spaces overlapping to some extent. In the deciduous broadleaved forests of Kyushu, an avian guild of foraging aerial in- sects in intermediate and lower layers of the forests is poor. Such a community may be subject to the successful invasion of the Red-billed Leiothrix into native forests. ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2002 Key words Ecological isolation, Foraging niche, Interspecific competition, Intro- duced birds, Leiothrix lutea The Red-billed Leiothrix Leiothrix lutea, originally occurring from southern China to the western Hi- malayas (Ali & Ripley 1972; Long 1981; Lever 1987; MacKinnon & Phillipps 2000), has been intro- duced to Japan where many naturalized populations have been found in deciduous broadleaved forests in central and southwestern Japan since the early 1980s (Tojo 1994; Eguchi & Amano 2000). The expansion of its range and the increase in its numbers has been particularly remarkable in Kyushu, southwestern Japan (Eguchi & Amano 2000). Where it is indige- /nous, the Red-billed Leiothrix occurs in various habi- (Received 27 April 2002; Accepted 9 July 2002) * Corresponding author, E-mail: hitorcb@mbox.nc.kyushu-u.ac.jp 123 tats including deciduous broadleaved forests, bamboo thickets, tea plantations, and secondary forests near human habitations from 900m to 3,000m asl (De Schauensee 1984; Long 1987). In Japan, Leiothrix in- habits deciduous broadleaved forests with a dense un- derstory of dwarf bamboo, and utilizes low layers of forest vegetation (Eguchi & Masuda 1994). In Kyushu, the habitat of Leiothrix overlaps that of a number of native species including the Japanese Bush Warbler Cettia diphone, Long-tailed Tit Aegithalos caudatus and various Parus species such as the Great Tit Parus major, Coal Tit P. ater, Varied Tit P. varius, and Willow Tit P. montanus (Eguchi & Masuda 1994). These native species also occur widely in de- ciduous broadleaved forests in Japan (Nakamura H. E. AMANO and K. EGUCHI 1970, 1978). Introduced birds may diminish the number of na- tive species through interspecific competition (Moun- tainspring & Scott 1985; Jones 1996). Usually intro- duced bird species succeed in establishing self-main- taining populations only in habitats disturbed by peo- ple (Case 1996) where they compete only with intro- duced species (Moulton & Pimm 1983; Moulton 1993) The Red-billed Leiothrix has, however, in- vaded native forests in Japan and has increased rap- idly in number. As a result of this, interspecific com- petition between Leiothrix and native species is likely in Japan. Instances of interspecific competition are difficult to detect (Lodge 1993). One approach to clarifying such competition is the study of habitat selection (Sol et al. 1997). If competition exists, two major re- sources, nest-sites and food, may be limiting. Amano and Eguchi (2002) revealed differences in nest-site characteristics between the Red-billed Leiothrix and Japanese Bush Warbler in forest, and suggested that competition between them for nest-sites did not exist, however, no quantitative evaluation has so far been made of either their food resources or their foraging spaces. Foraging niches are highly segregated among sym- patric Paridae (Gibb 1954; Hogstad 1978; Morse 1978; Nakamura 1970, 1978; Ogasawara 1970, 1975). Such segregation is partly derived from inter- specific competition (Lack 1971). The Willow Tit, for example, often shifts its foraging height when it oc- curs in mixed-species flocks in the presence of domi- nant Great Tit (Alatalo 1981). Niche shift, in the pres- ence of competitive species, has also been observed in another Paridae community with a different species composition (Herrera 1978). Thus a shift of foraging location is important evidence of interspe- cific interference competition. In this paper, we aim to clarify the patterns of for- aging space and foraging techniques of the intro- duced Red-billed Leiothrix and of several sympatric native forest bird species. In addition, we will show Leiothrix utilizes a unique foraging space that native species rarely use. STUDY AREA AND METHODS 1) Study area The study was conducted from April to September during each breeding season from 1997 to 2000, on the southwestern Kyushu, Japan Ebino Plateau, 124 (1,200m in elevation; 31°56’N, 130°51’E). The study area was situated in a mixed forest (16 ha) com- posed of Abies firma, Tsuga sieboldii, Pinus densi- flora, Quercus crispula, Hydrangea paniculata, Sym- plocos coreana, and S. myrtacea. The shrub layer was dominated throughout the forest by the dwarf bam- boo Sasamorpha borealis (ca. 2m in height). Only small patches of ground were bare. A road, approxi- mately 10m wide ran through the forest. The annual mean precipitation exceeds 5,000mm on the Ebino Plateau, of which more than one-third occurs during June and July (data from Miyazaki Branch, Weather Service of Japan). The Red-billed Leiothrix was first recorded in this area about twenty years earlier (N. Kamitanigawa pers. comm.). This species breeds from April to Sep- tember and emigrates to lower areas for the winter (pers. obs.). During the breeding season, six resident native species regularly occur in the same habitat as Leiothrix; four Paridae (Great, Varied, Willow, and Coal tits), Long-tailed Tit, and Japanese Bush War- bler (see Appendix | for the morphological character- istics of these species). Other sympatric species, in- cluding two trunk-specializing foragers (Japanese Pygmy Woodpecker Dendrocopos kizuki and Nuthatch Sitta europaea), and one sallying forager (Blue-and-white Flycatcher Cyanoptila cyanomelana), were excluded from the analysis, because of their low abundance, scarcity of observations, or great differ- ence in foraging techniques from Leiothrix. 2) Observation of foraging birds We searched for birds as we walked along forest paths. When we encountered birds foraging, we recorded the following information: species, time of day, foraging height, height of trees on which birds foraged, foraging location, foraging technique and whether or not there was a dwarf bamboo understory beneath/around the foraging location. We compared the percentage occurrence of foraging above or inside dwarf bamboo, because small patches of dwarf bam- boo and of bare ground are abundant in the study area. We divided foraging location into six cate- gories: foliage (including leaf, flower, fruit, bud and twig), branch, trunk, undergrowth (defined as shrubs if they were 2 m tall or shorter), ground, and air. For- aging techniques were defined as follows: (1) glean, a technique in which a prey item was picked up from a substrate by a standing or walking bird; (2) jump, a technique in which a bird jumped upon a prey item and snatched it from a substrate; (3) hang, a tech- Foraging niche of the Red-billed Leiothrix nique in which a bird hanging on a substrate picked a prey item; (4) hover, a technique in which a hovering bird picked a prey item from a substrate; (5) hawk, capturing an aerial prey item in the air; (6) peck, a technipue in which a bird pecked a substrate and picked a subsurface prey item; and (7) others, other miscellaneous techniques were includeed. Heights were estimated to the nearest meter. We followed individual birds as long as they re- mained in sight, because the foliage was dense and visibility in the forest was poor. We changed individ- uals after one foraging record had been collected. If only single birds or pairs were present, we collected two further foraging records from each bird once they had changed foraging trees or foraging locations. No more than one record was collected for an individual while in the same tree. Observations were made from 0830 to 1700. For the analyses, data were pooled for all individuals of each species, for all months and years. 3) Measurement of abundance of invertebrate prey We compared the abundance of aerial insects be- tween areas above dwarf bamboo and areas away from dwarf bamboo. In May 1997, we set 500-ml aluminum cans daubed with sticky glue 3-4 m above the ground as traps. Pairs of traps were set, one of each pair was set above dwarf bamboo and the other was set 10m away, and away from dwarf bamboo (Fig. la). In total, 50 pairs of traps were distributed throughout the study area. Two days after setting a) 3~4m Fig. 1. them, we recovered the traps and collected, and counted, all the invertebrates stuck to their surfaces. In 2000, in order to assess the vertical distribution of aerial insects above dwarf bamboo, we set sticky traps (commercial sticky fly traps; 70X3.5cm) 3m and 6m above the ground, and | m and 4m above the canopy of dwarf bamboo (Fig. 1b). These traps were set at 15 points for two days. Then, after recovering them, we counted the numbers of invertebrates on each trap. These invertebrates were identified either to the family or order level and divided into two size classes; larger than 5mm but smaller than 10mm, and 10 mm or larger. RESULTS 1) Foraging height and tree height There were significant interspecific differences in foraging heights at which birds foraged (F5 y= 12.67, P<0.0001, ANOVA, Fig. 2). The Red-billed Leiothrix foraged mainly 4+2m (median and 25-75% percentiles) above the ground, and the Japanese Bush Warbler foraged mainly 342m above the ground, although the difference between these two species was not significant (P=0.23, Scheffe’s test). The Varied Tit and Coal Tit frequently foraged in the upper layer of the forest above 8 m, signifi- cantly higher than Leiothrix (P<0.001 for each com- parison, Scheffé’s test). There were no significant dif- ferences in foraging height, however, between Leio- thrix and Long-tailed Tit (P=0.97, Scheffe’s test), b) 6m — hii igi a HATTA shrubs mn Traps for collecting aerial insects. a) 500 ml can traps daubed with sticky glue, b) sticky ribbon traps (70 x 3.5 cm) set at 1 m and 4m above a canopy of bamboo shrubs. H. E. AMANO and K. EGUCHI 104 136 160 20 15 10 Foraging height (m) L! Cd Ac Pmon Pm Pa Pv Fig. 2. Foraging heights of each species. Ll, Leiothrix lutea; Cd, Cettia diphone; Ac, Aegithalos caudatus; Pmon, Parus montanus; Pm, PR major; Pa, P ater; Pv, P varius. Crossbars represent 50% (median), boxes 25-75% percentiles and bars 10-90% percentiles. Circles are outliers. Pairs of a same char- acter indicate a significant difference; a P<0.001, b,c,d P<0.0001, e,f P<0.01, g P<0.05, using Scheffé’s F test. Height of trees (m) a LI! Cd Ac Pmon Pm Pa Pv Fig. 3. Heights of trees in which each species foraged. See Fig. 2 for abbreviations. Pairs of a same character indicate a significant difference; a,g P<0.01, bye P<0.0001, c,f P<0.05, d P<0.001, using Scheffe’s F test. Willow Tit (P=0.98) or Great Tit (P=0.98). There were significant differences among bird species in the height of trees in which they foraged (F, 7gg= 10.90, P<0.0001, ANOVA, Fig. 3). Japanese Bush Warblers mainly used trees or shrubs lower than 5 m. Varied and Coal tits used trees that were signifi- cantly higher than those used by Leiothrix (P<0.0001 and P<0.01, respectively, Scheffé’s test). There were no significant differences, however, be- tween the heights of trees selected by Leiothrix and the remaining species (P=0.20 Japanese Bush War- bler; P=1.00 Long-tailed Tit; P=0.89 Willow Tit; 126 and P=0.98 Great Tit; all Scheffé’s test). 2) Foraging location The Red-billed Leiothrix, as well as the other species, mainly foraged in deciduous broadleaved trees, while Long-tailed Tits and Coal Tits also fre- quently foraged in coniferous trees (Fig. 4a). Japan- ese Bush Warblers often foraged in the undergrowth, mainly in dwarf bamboo. Variation in foraging location was rather small (Fig. 4b). Each species foraged in foliage, mainly from leaves and twigs. In substrates other than fo- liage, Japanese Bush Warblers, Willow, and Great tits often foraged in the undergrowth or on the ground, while Coal and Varied tits foraged on branches and trunks. Great Tits also frequently caught prey in the air (aerial catching). Both Red-billed Leiothrix and Japanese Bush War- bler exclusively used that part of the forest where there was dwarf bamboo in the understory, while the levels of occurrence in such areas was relatively low for the Long-tailed and Great tits, and intermediate for Willow, Coal, and Varied tits (Table 1). The over- all difference was significant among species (df=6, x’ =150.5, P<0.0001). 3) Foraging technique Species varied considerably in their foraging tech- niques (Fig. 4c). The major technique used by each species was gleaning. Red-billed Leiothrix, however, often foraged by jumping on prey items. Hanging was used frequently by Long-tailed, Willow, Coal, and Varied tits. Willow Tits and Coal Tits often pecked at buds and flowers, and also picked prey items. Great Tits also caught aerial insects gathering around flowers before leafing out in spring, whereas Japanese Bush Warblers foraged exclusively by gleaning. 4) Vertical distribution of invertebrates From the results referred to above, it is clear that the Red-billed Leiothrix mainly used the lower layer of the forest with dwarf bamboo in the understory and it foraged more frequently by jumping than any of the other species. Foraging by jumping may be an effective technique for catching aerial insects resting on leaves or twigs. Leiothrix often foraged just above dwarf bamboo. Although there was no significant dif- ference in the abundance of small invertebrates (S—10 mm), large Mecoptera, a dominant taxon among large (=10mm) invertebrates (and eaten by Leiothrix, Foraging niche of the Red-billed Leiothrix © conifers w@ evergreen broadleaved © deciduous broadleaved a) 299 44 49 99 126 100% 80 60 undergrowth @ others undergrowth i ground @ others peck — jump hover BB hawk hang O glean Cd Ll Ac Pmon Fig. 4. Pm Pa Pv a) Use of trees by each species, b) use of substrates by each species, c) foraging techniques used by each species. Figures above columns are the numbers of observations. See Fig. 2 for abbreviations. Table 1. Percent occurrence of foraging above or inside of bamboo shrubs. Leiothrix Cettia Aegithalos Parus Parus Parus Parus lutea diphone caudatus montanus major ater varius Yes 89.2 88.2 36.6 66 40.8 60.3 63.3 No 10.8 11.8 63.4 34 59.2 39.7 36.7 No. of observations 295 34 41 50 103 136 158 pers. obs.), were more abundant just above dwarf bamboo shrubs than above bare ground (0.44+1.07 vs. 0.12+0.39; mean number+SD; N=50, P=0.048, Wilcoxon signed rank test). In a comparison between the abundance of inverte- brates in an upper layer (4m above dwarf bamboo canopy) and a lower layer (1m above dwarf bam- 127 boo), the total abundances of both small and large in- vertebrates were greater in the lower level than in the upper in each month, but not significant (see Fig. 5). The only exception was for the abundance of small invertebrates in May (Fig. 5a). At taxonomic levels, the abundances of invertebrate in the lower layer also tended to be greater than in the upper layer, but the H. E. AMANO and K. EGUCHI a b Small size Diptera Apr. 5-10 mm Apr. Hymenoptera May Others Aug. 9 Diptera May Hymenoptera Large size Apr. 10. ie | gaha naered a ’ '* ¢ eo ¢ ¢ ’ WET a" freeride i ah oe v rrrtnes ies Kati rs aah a ur — Ww hor Sos Wrilte) & Am thu a ue ! ovine sade or iam pay : trove 4 se PRN ono TRE JAM pe @ |: Cabhmpicbodacie sehbaae ate u2igo (aeiMocnprinntaan doney soxbqa pirll (dW lg penpals hl 4herttatrhorbhai® de ’ sTapap. Inco 2Aig2 ch Iwirt Ame Wir bree it ails (Sie be ree ' 2-AR 4 a aa « Spe rr. Thi 5 130 pres " “ ~ grin to Boo Leki itr yon emia: Tatmoé y Gieweraid apdodal Me pene (Giada: grilyralause brs , TACK MH AR MET servile ict TA Pai atta te > ee 6 ere =e | i > apices! ammaaguted 50) Bie Perr etry ee ee on and aluiee at |) v1) (Cbrdbee or Me VLivvtvegm TO fect): Fie Parad Sncihs “aad Gh, Sucskdue Gieiae | ee O00T8 2 peace RP, Hapa, Ape Dy ie: Sasakiwe Aesentlic- Tewcrsaioais sigolodinO: bs dy ant oye inant =e vain Lance £> OPE Reale: falufwin ” fe hah A ah ed Bh Sy Teh ate densi naam anaes snl rie ia need ica alan i nes ree th cs eit eo ie ie eg Bie = oe Jolieton AE) Cheah ciland - n rr rete = cam reais bellid. ae: sett, NewdSurau: nonsleg ee | YOU Te ‘Leck rfgifeeT vem: siityeatt Gira) attri eS el . = mua ititgaibiutiw soangnl ni) £ | ; r= “4 (1@0% + iy , PERERENCES ~~ [3 leads Me aaa ee > i > ein be ae od qiiviyottinad Guid Yorekstats Sh Plbed MoS otnfibcgiion in its 2 neritic! 4) ybhednun’/ ew ayn uo "} > Pini y A Wallington u i _ gb r sta, > ite, Z (1987) Lelothna. be Cheat Lam Bia Ie Mrarad beck 2 iis) Fauna Stites iy : Seek kiana ion |) Gidong Boaverih ing (ia hue wee ea it oh a 4 a ee) eae ‘Kinacn J & Poigen | é } ae 7s. v™ Aids (tu a 5 a r i i T 5 rei ia ; 7 a $6 AC TR Wn cheb be Ki ; [i as ms MeDsa =e : z ' “(dw =a IO". wid 1 HAE ‘ve . ay is aha we ahi his 1m 0 —. 7 a = YS Tw dll-or-noge Mv wer i: iil > j dua cd reves Pecwrifimurcs he " Yin ta ics Sofi ore ion Wanisibed Aru Mat 41: 10S ee ] i: » Peyemrf. Moaion WP & Pir SL (1983) The} odo Di. . An reife: he “eg tophic Videnod hit : 4 Yuba Sr i Arn Mal lietes 47) e-2- i & naw Vhrwse eaietcr | rg : m Mar Kop Yan vshacea dye Orni bt ime Bo : ) Japanese with English saenrbaty)j $4 m Nekanmira, 7 (2978) 4 otidy ob Rae ORIGINAL ARTICLE Ornithol. Sci. 1: 133-142 (2002) Tree species preferences of insectivorous birds in a Japanese deciduous forest: the effect of different foraging techniques and seasonal change of food resources Akira UNNO** Faculty of Agriculture, Hokkaido University, Sapporo 060-8589, Japan ORNITHOLOGICAL © The Ornithological Society Abstract I examined the effects of arthropod abundance and of bird foraging tech- niques on the tree species preferences of seven insectivorous bird species in a temper- ate deciduous forest. It is hypothesized that bird species with a wide range of foraging techniques respond more flexibly to the spatial distribution and seasonal change of prey than those with specialized foraging techniques. This hypothesis was supported by the fact that tits, bird species with a wide range of foraging techniques, changed their techniques when foraging in tree species with different foliage structures. They also used various tree species in late summer when food requirements increased owing to the addition of nestlings and fledglings. Bird species with a narrow range of foraging techniques, such as flycatchers and white-eyes, did not change their tech- niques among tree species and had strong tree species preferences in all research peri- ods. Key words foraging technique Tree species diversity is one of the most important habitat factors determining bird species diversity in temperate forests, because diverse composition of tree species should facilitate the coexistence of differ- ent species of birds (Holmes et al. 1979; Rice et al. 1984; Hino 1985). However, the mechanism of coex- istence has not been sufficiently understood. Al- though many studies have shown different use of tree species among insectivorous birds (e.g., Hartley 1953; Nakamura 1970; Morse 1978; Recher et al. 1991), most such studies have not surveyed the avail- ability of food for birds on different tree species (but see Holmes & Robinson 1981; Diaz et al. 1998; Hino et al. 2002). The abundance and distribution of prey, and the foliage structure, which vary among tree species, influence prey detectability and accessibility by birds (Holmes & Schultz 1988). Thus, the prey availability for each bird species must be determined separately for each tree species. Since different tree species provide different foraging opportunities for (Received 14 July 2001; Accepted 12 July 2002) * E-mail: unno@hfri.bibai.hokkaido.jp * Pressent address: Hokkaido Forestry Research Institute, Naka- gawa, Hokkaido 098-2805, Japan 133 Foliage structure, Prey abundance, Tree species preference, Variety of birds, tree species composition within a forest should influence bird species composition and diversity (Holmes & Schultz 1988). Prey abundance for birds varies among tree species and changes temporally during the breeding season (Feeny 1970; Nager & van Noordwijk 1995; Dias & Blondel 1996; Murakami 1998). Foraging techniques, determined by morphological characteristics of each bird species (Moreno & Carrascal 1993; Carrascal et al. 1995), affect bird preferences for foraging habitat (Nakamura 1978; Holmes and Schultz 1988; Hino et al. 2002). Different foliage structure among tree species often requires foraging birds to use different foraging techniques (Whelan 1989). Under these cir- cumstances, we hypothesize that birds with a variety of foraging techniques can respond flexibly to tempo- ral change and spatial distribution of food abundance. So far, few studies have examined this hypothesis (but see Hino et al. 2002; Murakami 2002). This study examined the use of tree species and the foraging techniques of seven forest bird species in temperate deciduous forest. Temperate deciduous forests are most appropriate for the study of tree species preferences of birds because the number of A. UNNO tree species is not as high as in tropical forest and not as low as in coniferous boreal forest. I analyzed how seasonal changes in tree species preferences of birds were affected by their foraging techniques, by prey abundances and by the foliage structure of trees. MATERIALS AND METHODS 1) Study site This research was conducted in the experimental forest of the Hokkaido Research Center, the Forestry and Forest Products Research Institute, located in Sapporo, Japan (42°59’'N, 141°23’E). A 9 ha-study area was established in a secondary deciduous broadleaved forest, which had been burned about 70 years ago. The vegetation was surveyed in fifteen 25m-square plots selected randomly in the study area. Species, number of individuals, and diameter at breast height (DBH) were recorded for all trees. Tree species composition was calculated as the percentage of basal area given by DBH. The dominant tree species are Japanese White Birch (Betula platyphylla, 46.2%), Oak (Quercus mongolica v. grosseserratus, 13.7%), Casterarealia (Kalopanax pictus, 10.0%), Printed Maple (Acer mono, 5.0%), Japanese Linden (Tilia japonica, 5.0%) and Alder (Alnus hirsuta, 4.3%). The understory is dominated by a high density of dwarf bamboo (Sasa senanensis and S. kurilensis) 0.5—2 m in height. Budding of most trees starts in early May. Re- search was conducted from 21 May to 5 July in 1992 and from 22 May to 19 June in 1993. Five research periods of two weeks each were established for bird observations and prey sampling: 92-1 (21 May-—5 June), 92-2 (6 June—20 June), 92-3 (21 June—5 July) in 1992; 93-1 (22 May-—5 June), 93-2, (6 June—19 June) in 1993. 2) Bird foraging The target species were seven insectivorous bird species. These were three resident species: Great Tit (Parus major), Marsh Tit (P. palustris) and Long- tailed Tit (Aegithalos caudatus), and four migrant species appearing in May: Japanese White-eye (Zos- terops japonica), Narcissus Flycatcher (Ficedula nar- cissina), Eastern Crowned Leaf Warbler (Phyllosco- pus coronatus) and Brown Flycatcher (Muscicapa dauurica). Observations of Marsh Tits may have in- cluded Willow Tits (Parus montanus), which were very low in abundance, because those two species are difficult to distinguish in the field. The data on fledg- lings were excluded from the analysis. I observed bird foraging behavior through binocu- lars (8—16 times zoom) from 0500 to 1000 except on rainy days or on days with strong wind. When a for- aging bird was encountered, I followed it as long as possible, and recorded the tree species it visited and its foraging techniques. Foraging techniques were classified into three types: perch-gleaning (gleaning prey from leaves or branches while perching on branches); hang-gleaning (taking prey from leaves or twigs by hanging upside down from twigs or leaves); and hovering. In this study, one tree individual was regarded as one patch for foraging. Even if birds searched but did not capture prey on a tree during the observation, thus, the datum was dealt with as one sample of use of the tree species. Even if a bird for- aged many times successively on the same tree, the datum was dealt with as one sample. When a bird for- aging on a particular tree species flew to the same tree species less than 3m away, or with and overlap- ping crown, or when a bird being followed was lost to sight for a few seconds in foliage (when no other conspecific birds were found nearby), the datum was also dealt with as one sample. In contrast, all data re- lating to foraging techniques were dealt with as inde- pendent samples even when foraging occurred suc- cessively on the same tree. No more than 10 samples of foraging technique observations were taken from each individual bird. The data included both feeding nestlings or fledglings and foraging for themselves. To avoid bias from repeated observations of the same individuals, I collected the data while walking steadily within the study area. The preferences of birds for each tree species were calculated using Ivlev’s electivity index (Ivelv 1955): E=(p;—r,/(p;+r,), where p; represents the propor- tions of the ith tree species used by a particular bird species and where r; is the proportion of tree species composition occupied by the ith tree species in the study area. Following Holmes and Robinson (1981), the ‘tree preference index’ was obtained as a sum of the percentage deviations of bird use from the tree species composition for six dominant species (B. platyphylla, Q. mongolica, K. pictus, A. mono, T. japonica, A. hirsuta). This index shows that the higher the value is, the more specialized the bird is in tree species use. The similarity in tree species use be- tween two bird species was calculated with Pianka’s index (Pianka 1973): 134 . _ — Tree Species Preferences of Birds je =D Pi af Stow" [Lew } where p;, and p;, are the proportions of the th tree species used by the jth and the kth bird species, re- spectively. The similarity between tree species use by birds and tree species composition in the study plot was also calculated using Pianka’s index. The cluster analyses were performed using Mountford’s method (Mountford 1962). The variation in foraging tech- niques was shown using Shannon’s H’ (Shannon & Weaver 1949): H'=—2 q,log, q;, where q; is the pro- portion of the jth foraging technique. 3) Arthropod sampling Spiders and insects were searched for, counted and measured, at 1-2 m in height in four species of trees (A. mono, A. hirsuta, Q. mongolica and T. japonica) in 1992 and six species (addition of B. platyphylla, and K. pictus) in 1993. Each sample consisted of 400 | leaves with branches and twigs for five of these tree | species, with the exception of K. pictus of which 50 leaves were sampled owing to their very large size. / Seven units were sampled for each tree species in all | research periods in 1992, ten in 93-1, and nine in 93- , 2. Dry mass (W, mg, 60°C, 48h) of arthropods was estimated from the body length (L, mm) with the fol- | lowing equation: W=0.12L'™ (r=0.83, P<0.001, | N=78), which was determined based on part of the | samples. Leaf areas were measured (using a digitizer) | for 25 leaves for each tree species in July to calculate | the arthropod dry mass per | m? leaf area. | 4) Statistical analyses Chi-square tests were conducted to reveal differ- ences in tree species use among bird species during each research period. Some tree species were com- bined to make expected frequencies large enough for chi-square tests. Chi-square tests and Fisher’s exact probability tests with multiple comparison methods Were conducted to compare the frequencies of the tree species used by a bird species with the frequen- cies expected from tree species composition during each research period. Chi-square tests with multiple comparison methods were also conducted to reveal the differences in foraging techniques used by each bird species between tree species. Mantel-Haenszel | tests were conducted to detect similarities of relative frequencies of foraging techniques on different tree species among bird species. The arthropod dry mass 135 per 1 m*-leaf area data were log-transformed to re- duce skewness for ANOVA. Two-way ANOVAs were conducted to reveal the seasonal changes in arthropod abundances on different tree species (factor=period, tree species). One-way ANOVAs were conducted to reveal the differences in arthropod abundances among tree species during each period (factor=tree species). Spearman’s rank correlation tests were conducted to reveal the relationship be- tween arthropod abundances and selectivity of tree species for each bird species during each research pe- riod. Sequential Bonferroni methods (Rice 1989) were used for multiple comparisons in nonparametric tests and Fisher’s PLSDs were used in ANOVA. Sta- tistical significances were evaluated at P<0.05 except for correlation analyses. P<0.1 was considered as the significance level for correlation analyses owing to the small sample sizes involved (4 or 6). RESULTS 1) Tree species preferences of foraging birds The uses of tree species differed significantly among bird species during all research periods (Table 1). In the 92-1 period, A. mono was preferred by three species of tits and Q. mongolica was preferred by Great Tit, but B. platyphylla was avoided by most species of birds. In the 92-2 period, neither prefer- ence nor avoidance for tree species was shown by any bird species except for Eastern Crowned Leaf Warbler preferring Q. mongolica. In the 92-3 period, K. pictus was preferred by Marsh Tit and Eastern Crowned Leaf Warbler, but B. platyphylla was avoided by Marsh Tit and Japanese White-eye. In the 93-1 period, A. mono was preferred by three species of tits and Q. mongolica was preferred by Eastern Crowned Leaf Warbler, Japanese White-eye and Nar- cissus Flycatcher, but B. platyphylla was avoided by all species of birds. In the 93-2 period, B. platyphylla was avoided by Great Tit and Narcissus Flycatcher. On the whole, the preferred or avoided tree species appeared to be consistent for each bird species during each research period, although the use of tree species differed significantly among bird species (Table 1). The tree preference index for each bird species was compared between the first (92-1, 93-1) and the later research periods (92-2, 92-3, 93-2). This index de- creased for Long-tailed, Great, and Marsh tits, but in- creased or remained unchanged for the Narcissus Fly- catcher in both years (Table 2). This index increased for Eastern Crowned Leaf Warbler in 1992 but de- A. UNNO Table 1. Preferential use of tree species by foraging bird species and tree species composition in the study area. + or — repre- sent differences between the percent of each bird species using a tree species and the percent of the tree species composition of the same tree species . Some bird species were not shown in each period owing to sample size being too small for analysis. AM=4cer mono, AH=Alnus hirsuta, QM=Quercus mongolica, TJ=Tilia japonica, BP=Betula platyphylla, KP=Kalopanax pictus. Tree species Number of +? test AM AH QM TJ BP KP observations 92-1 period Long-tailed Tit tp Mee Stel +8.3 selfs) = Sy CS) 59 7 =59.0 Great Tit aS DO) SP iloO SOs SB 5, 142 df» =20 Marsh Tit aH Sse tle(ss +9.7 Sy = 25.09 "= Ss 186 P<0.001 Eastern Crowned Leaf Warbler +6.5 (Sie sh”) —3.1 = Osi —6.2 52 Narcissus Flycatcher +6.7 Fe ar OA = 3m SHADES este lleT WH 92-2 period Long-tailed Ti +4.5 +5.2 +19.6 O10) 2214 S010 21 7 =33.8 Great Tit +9.6 poe —0.3 = 3)! =O)3) S42 103 df =12 Marsh Tit —3.4 1110) i mio —6.3 —8.4 61 P<0.001 Eastern Crowned Leaf Warbler +2.5 SB PAS PS Pil? aS) 40 Narcissus Flycatcher +7.9 TPP aPilfsh{o) seis = 297) =35 31 92-3 period Long-tailed Tit —5.0 —0.9 +7.0 +1.9 —8.9 —10.0 29 oP =55.1 Great Tit POR AAO =05 OSs ares is. ea 114 df=15 Marsh Tit —4.1 +2.1 +2.8 8) IGEN Sa 7A0) obs 109 P<0.001 Eastern Crowned Leaf Warbler —1.6 +6.0 —3.4 =) Bh a A 58 Japanese White-eye —5.0 ae OMN ipa OAO) +2.1 —42.6* +18.5 28 Narcissus Flycatcher +11.0 apo arllat3) —-1.0 -—302 +10.0 25 93-1 period ’ Long-tailed Tit +28.37 —43 £x+1916 ap ill) s=s4bil = 1lOW 33 =55.8 Great Tit ap less" ap il) +9.6 SAG ie DS et ee iV 146 df =24 Marsh Tit APTS ie eA ar lod) “Ose = DAGe eee Oullin 162 P<0.001 Eastern Crowned Leaf Warbler +10.1 SFOS EMT 2OtO* TEC 2 ete OP ee OU 53 Japanese White-eye +11.7 SOlomalGES4A* else 42 SF e100 27 Narcissus Flycatcher —0.2 sO imei D SFO WectS Siero Sal eee taOn/ 84 Brown Flycatcher +10.4 Ll) ae AAD) cy | Sen) 74) stale 26 93-2 period Great Tit +93 +1.0 +5.1 7.0") 161 10 133 7 =26.5 Marsh Tit 45) —1.8 +8.1 —1.6 —8.4 =3.3 119 df =12 Eastern Crowned Leaf Warbler +2.3 +0.6 +15.5 —0.1 —9.6 =i] 41 P<0.001 Narcissus Flycatcher SN ae ett) Stele eA are OL 39 Tree species composition (%) 5.0 4.3 13.7 5.0 46.2 10.0 * P<0.05, comparisons of the tree species composition for each tree species and the use of same tree species for each bird species by each period and bird species (7° test or Fisher’s exact test with sequential Bonferroni method). ** Comparisons with use of tree species at each period. ') Tree species categories were combined to make expected frequencies large enough for chi-square tests. ') Six tree categories; AM, QM, TJ, BP, KP, AH+other tree species. ) Four tree categories; AM, QM, BP, AH+KP+TJ-+other tree species. ) Four tree categories; QM, BP, KP, AM+AH+TJ+ other tree species. ») Five tree categories; AM, QM, TJ, BP, AH+KP-+ other tree species. ) Five tree categories; AM, QM, TJ, BP, AH+KP+ other tree species. 136 Tree Species Preferences of Birds Table 2. Tree Preference Index during each period. Some bird species were not shown for each period owing to sample sizes being too small for analysis. Period 92-1 92-2 92-3 93-1 93-2 Long-tailed Tit 73.8 66.7 3)3)57/ 97.4 Great Tit Tiel 30.0 31.5 69.4 39.5 Marsh Tit 51.0 24.7 46.6 60.7 28.3 Eastern Crowned Leaf Warbler 46.9 59.3 79.8 76.3 30.8 Japanese White-eye 93.0 100.5 Narcissus Flycatcher 53.3 63.4 70.2 72.0 79.9 Brown Flycatcher S37) creased in 1993. In the first research periods of both years, the uses of tree species were similar among bird species and differed from the tree species composition in the study plot (Fig. 1). In the later research periods, bird species were divided into two groups: the first con- sisted of bird species that used trees in relation to the tree species composition of the area, while the second consisted of bird species that foraged in trees unre- lated to their species composition (Fig. 1). Two Parus species usually belonged to the first group, while Narcissus Flycatcher and Japanese White-eye usually belonged to the second group. Long-tailed Tit and | Eastern Crowned Leaf Warbler were classified into different groups in different research periods. 2) Foraging techniques on each tree species Long-tailed Tit used all three foraging techniques _ with almost the same frequency (Table 3). Great Tit and Marsh Tit foraged by perch-gleaning most fre- quently but also hang-gleaned (30%). Eastern Crown Leaf Warbler foraged by hovering most frequently, but also perch-gleaned (30%). Japanese White-eye al- most always foraged by perch-gleaning, and Narcis- sus Flycatcher and Brown Flycatcher almost always foraged by hovering. The variety of foraging tech- niques was maximal for the Long-tailed Tit, interme- diate for the two Parus species, Eastern Crowned _ Leaf Warbler, and Japanese White-eye, and lowest for the two flycatcher species. Foraging techniques used by each bird species were compared among tree species (Table 4). Signifi- _ cantly different frequencies of gleaning (perch-glean- ing+hang-gleaning) and hovering among tree species were found in Long-tailed Tit (y’=6.20, df=2, P=0.045), Great Tit (y?=15.79, df=4, P=0.003) and 137 Eastern Crowned Leaf Warbler (y°=16.92, df=4, P=0.002). Long-tailed Tit hovered more frequently on B. platyphylla than on A. mono (y’=6.24, df=1, P<0.05). Great Tit hovered more frequently on B. platyphylla than on A. mono and Q. mongolica (BP vs. AM: 7°=11.13, df=1, P<0.001; BP vs. QM: y=9.48, df=1, P<0.05). Eastern Crowned Leaf Warbler hovered more frequently on B. platyphylla than on Q. mongolica (7°=15.66, df=1, P<0.001). In all seven bird species, on the whole, hovering on B. platyphylla was more frequent than on A. mono or Q. mongolica (Mantel-Haenszel test with sequential Bonferroni method, BP vs. AM: 7°=13.91, df=1, P<0.01; BP vs. QM y?=36.57, df=1, P<0.001). Moreover, significantly different frequencies of perch- and hang-gleaning among tree species were found in Great Tit (y’=15.83, df=4, P=0.003) and Marsh Tit (7°=36.91, df=4, P<0.001). Both tits for- aged by hanging more frequently on A. mono and T. japonica than on Q. mongolica (Great Tit: AM vs. QM: 7°=8.61, df=1, P<0.05; QM vs. TJ: 7°=8.98, df=1, P<0.05; Marsh Tit: AM vs. QM: 7’=17.96, df=1" P=0)001;, OM vs. Tl 77—7.99, df=1, P= 0.05). Marsh Tit also hang-gleaned more frequently on T. japonica than on B. platyphylla (7°=7.99, df=1, P<0.05). In the four gleaning species (Long- tailed Tit, Parus species, and Japanese White-eye), hang-gleaning on A. mono was more frequent than on Q. mongolica or B. platyphylla (Mantel-Haenszel test with sequential Bonferroni method, AM vs. QM: 7 =16.02, df=1, P<0.001; AM vs. BP: 7’=10.86, df=1, P<0.01), and hang-gleaning on T. japonica was more frequent than on Q. mongolica, B. platy- phylla or K. pictus (TJ vs. QM: 7°=31.75, df=1, P<0.001; TJ vs. BP: y’?=14.00, df=1, P<0.01; TJ vs. KP: 7°=8.07, df=1, P<0.05). 92a 92-2 92—3 93-1 Similarity Fig. 1. tree species composition Eastern Crowned Leaf Warbler Narcissus Flycatcher Marsh Tit Long-tailed Tit Great Tit Eastern Crowned Leaf Warbler Narcissus Flycatcher Long-tailed Tit Great Tit Marsh Tit tree species composition Eastern Crowned Leaf Warbler Narcissus Flycatcher Japanese White-eye Marsh Tit Great Tit Long-tailed Tit tree species composition tree species composition Narcissus Flycatcher Eastern Crowned Leaf Warbler Japanese White-eye Brown Flycatcher Long-tailed Tit Marsh Tit Great Tit Narcissus Flycatcher Great Tit Eastern Crowned Leaf Warbler Marsh Tit tree species composition Similarity between use of tree species by bird species and the tree species composition shown by means of @ index (Pianka 1973). Dendrograms were constructed according to Mountford (1962). These four bird species, Long-tailed tit, Parus species, and Eastern Crowned Leaf Warbler, which changed their foraging techniques between different tree species, used the widest range of foraging tech- niques (Table 3). 3) Prey abundances on each tree species Arthropod abundances (dry mass/1 m? leaf surface area) changed seasonally, with the fluctuation pat- terns differing among tree species in each year (Fig. 138 2, two-way ANOVA, 1992: periods, F,, ,=4.78, P=0.011; tree species, F,, ,=1.64, P=0.19; periods x tree species, F,, ,=2.45, P=0.033; 1993: periods, F,,3 ;=8.86, P=0.004; tree species, F,,;,; ;=8.02, P< 0.0001; periods Xtree species, F,,,; ;=1.13, P=0.35). Arthropod abundances differed among tree species in 92-1, 93-1, and 93-2 but not in 92-2, or 92-3 (one- way ANOVA; 92-1: F,7 ,=5.26, P=0.006; 92-2: F,, 3= 1.88, P=0:16; 92-3°F5, ,=0.62; P=O:61; 93am F,, ;=4.63, P=0.002; 93-2: F,, ,=4.51, P=0.002). Tree Species Preferences of Birds Table 3. Bird foraging techniques. i 0 Sieauing 2) Hovering Diversity index Number of 0 », ! . Perch-gleaning Hang-gleaning a) Soanno mea SO SEU Long-tailed Tit De, 45.5 28.8 1.54 132 Great Tit 60.0 32.3 Tcl 1.25 637 Marsh Tit 59.7 333 7.0 1.24 1079 Eastern Crowned Leaf Warbler 28.5 4.7 66.8 1.11 240 Japanese White-eye 80.2 17.0 2.8 0.83 274 Narcissus Flycatcher Tes Nog 90.8 0.51 106 Brown Flycatcher 10.7 0.0 89.3 0.49 56 Table 4. Foraging techniques on each tree species. Parentheses show the number of observations. Because of small sample sizes comparison of Long-tailed Tit, Brown Flycatcher, and Japanese White-eye’s foraging techniques on Caster arealia and | Japanese linden were omitted (see Table 1 for AM, AH, QM, TJ, BP, and KP). Gleaning Gleaning Hovering (%) (%) Perch-gleaning Hang-gleaning (%) (%) Long-tailed Tit AM 83.7 (36) 16.3 (7) * 37.2 (9) 62.8 (27) QM 70.6 (24) 28.4 (10) 58.8 (10) 41.2 (14) BP 56.0 (14) 44.0 (11) 68.0 (6) 32.0 (8) Great Tit AM 94.6 (194) 5.4(11) ** 63.9 (120) 36.1 (74) i QM 94.7 (144) 5.3 (8) * 78.3 (111) 21.7 (33) TJ 86.5 (32) 13.5 (5) 54.0 (15) 46.0 (17) BP 82.8 (82) 17.2 (17) 75.8 (58) 24.2 (24) KP 92.3 (36) 7.7 (3) 64.1 (22) 35.9 (14) Marsh Tit AM 90.5 (134) 9.7 (14) 56.7 (70) 43.3 (64) b a QM 95.8 (256) 4.2 (8) 78.4 (199) 21.6 (57) la TJ 78.4 (34) 21.6 (3) 45.9 (14) 54.1 (20) | an BP 92.1 (232) 7.9 (20) 69.4 (155) 30.6 (77) KP 95.2 (80) 4.8 (4) 70.0 (55) 30.0 (25) Eastern Crowned Leaf Warbler AM 35.0 (7) 65.0 (13) 95.0 (6) 5.0 (1) QM 44.1 (49) 55.9 (62) *k 97.3 (46) 2.7 (3) TJ 25.0 (4) 75.0 (12) 100.0 (4) 0.0 (0) BP 5.7 (11) 84.3 (59) 95.7 (8) 4.3 (3) KP 23.1 (3) 76.9 (10) 92.3 (2) 7.7 (1) Japanese White-eye AM 96.0 (24) 4.0 (1) 76.0 (18) 24.0 (6) QM 100.0 (51) 0.0 (0) 90.2 (46) 9.8 (5) BP 66.7 (2) 33.3 (1) 100.0 (2) 0.0 (0) Narcissus Flycatcher AM 10.3 (3) 89.7 (26) 100.0 (3) 0.0 (0) QM 11.1 (9) 88.9 (72) 98.8 (8) 1.2 (0) TJ 6.7 (1) 93.3 (14) 100.0 (1) 0.0 (0) BP 4.6 (2) 95.4 (42) 100.0 (2) 0.0 (0) KP 5.7 (2) 94.3 (33) 97.3 (1) 2.7 (1) Brown Flycatche AM 15.4 (2) 84.6 (11) 100.0 (2) 0.0 (0) QM 19.1 (3) 80.9 (17) 100.0 (3) 0.0 (0) BP 0.0 (0) 100.0 (18) 100.0 (0) 0.0 (0) *: P<0.05, **: P<0.01. 139 A. UNNO 92-2 92-3 Dry weight (mg/mz) 93-1 93-2 BP KP AM AH QM TJ Fig. 2. Arthropod dry mass per | m? leaf area on various tree species (mean+SE). Refer to Table for AM, AH, QM, TJ, BP, and KP. BP and KP were not surveyed in 1992. Letters above vertical lines indicate the results of multiple comparison tests (Fisher’s PLSD, P<0.05); the same letters indicate non- significant differences. In the first research periods of each year, arthropod abundances were highest in A. mono (Fig. 2). Of the 27 cases of correlation analyses for each bird species during each research period, there were seven positive relationships (26%) between selectiv- ity by birds and arthropod abundance on different tree species (Table 5). Of the seven significant relation- ships, six (86%) were found for the four bird species, Long-tailed tit, Parus species, and Eastern Crowned Leaf Warbler, using wide range of foraging tech- niques (Table 3) and five (71%) were found during the first research periods. DISCUSSION Tree species supporting high prey abundance can be expected to be preferred by foraging birds (Hino et al. 2002). Prey abundance may explain why A. mono was most preferred by tit species during the first re- search periods each year. In most cases, however, tree species preferences were not related to prey abun- dances on trees. Some researchers have indicated that tree species preferences of birds are influenced not only by food abundance but also by food accessibil- ity, that is foliage structure of trees and manoever- ablity of birds both play a role (Holmes & Robinson 1981; Whelan 1989; Dias et al. 1998). For example, Q. mongolica was preferred but B. platyphylla was avoided by most bird species despite them supporting similar abundances of prey. This result may be attrib- utable to the different accessibility of food in relation the foliage structure of these tree species. Almost all birds captured prey by hovering on B. platyphylla more frequently than on Q. mongolica. B. platyphylla has fine twigs, long petioles, and horizontally distrib- uted leaves, which make searching and capturing prey while perching difficult, and make sallying or hanging from branches necessary. In contrast, Q. mongolica, which has thick twigs, very short petioles and a more three-dimensional distribution of leaves, provides more opportunities for foraging by perch- gleaning. Because perch-gleaning is a less energy-ex- pensive foraging technique than hang-gleaning and hovering, capturing prey on Q. mongolica is less en- ergy-expensive for birds than taking the same prey from B. platyphylla. Thus, birds should prefer Q. mongolica rather than B. platyphylla. Likewise, the food accessibility for gleaner species (Long tailed Tit, Parus species, and Japanese White-eye) may be in- fluenced by the foliage structure of A. mono and T: japonica. These birds foraged frequently by hanging 140 from twigs on these trees. Because leaves are distant from twigs, owing to the upward pointing long peti- oles of these trees, hang-gleaning would be an effi- cient technique to capture prey from the undersides of leaves where most caterpillars are found (Green- berg & Gradwohl 1980; Holmes & Schulz 1988). The present results supported my hypothesis that there is a correlation between the variety of foraging Tree Species Preferences of Birds Table 5. Results of correlation analyses between Ivlev’s electivity index of foraging bird species and arthropod abundances on each tree species. Spearman’s rank correlation tests were one-tailed. Because of a few observations, results of Long-tailed Tit in 93-2, Japanese White-eye in 92-1, 92-2 and 93-2, and Brown Flycatcher in 92-1, 92-2, 92-3, and 93-2 were omitted. Periods (Number of tree species) 92-1 (4) 92-2 (4) 92-3 (4) 93-1 (6) 93-2 (6) Long-tailed Tit NS NS NS Gi) Great Tit ar NS NS aE ar NS Marsh Tit NS NS NS (+) ++ Eastern Crowned Leaf Warbler NS + NS NS NS Japanese White-eye NS NS Narcissus Flycatcher NS NS NS (+) NS Brown Flycatcher NS Positive correlation: ++; P<0.01, +; P<0.05, (+); 0.10.05: Mann- Whitney’s U-test). Therefore, data for males and fe- males were pooled for the analyses. RESULTS The width of the nape patch and the coloration of the chest band varied from II to V and from I to III, respectively. Out of the 26 specimens, five exhibited characteristics reported in the original description of P. s. orti (a narrow NP (I-III) in combination with a dark CB (1)), while eleven showed characteristics given in the original description of P. s. kobayashii (a wide NP (IV—V) in combination with a faint CB (II- III)). The remaining ten specimens had a narrow NP combined with a faint CB. I have named these three color morphs as orii-type, kobayashii-type and inter- mediate-type. No specimens were found to have a wide NP in combination with a dark CB. All three color morphs occurred in the western part (Yonaguni- jima) and in the eastern part (Iriomotejima, Hateru- majima, Kuroshima, and Ishigakijima) of the Yaeyama Group. The proportions of the three color morphs did not vary significantly between these two 156 areas of the Yaeyama Group (P>0.05: Fisher’s exact test). There was, however, seasonal variation in the num- bers of the three color morphs (see Table 1). The Orii-type was found only from September through December, while the kobayashii-type was found only from April through August. The intermediate-type bridged the appearance of the other types, being found only from December through June. The nape patch of the specimens collected from September through March was significantly narrower than that of specimens collected from April through August (P<0.01: Mann-Whiteny’s U-test). The chest band became significantly fainter from September through August (P<0.001: Kendall’s rank correlation test). DISCUSSIONS The present results strongly indicate that P. s. orii undergoes a seasonal plumage change, just as the continental subspecies P. s. sinensis and P. s. hoyi do. It seems that two color morphs, the orii-type and the kobayashii-type, which were originally described as different subspecies by Kuroda (1923, 1930), actually correspond to winter and summer plumages. This seasonal change in plumage color may result from molting in spring and autumn and from feather Plumage Color of Pycnonotus sinensis orii Fig. 2. Dorsal (A,C,E) and ventral (B, D, F) views of Pycnonotus sinensis orii. A, B: orii-type from Yonaguni- jima (KMNH 16). C,D: kobayashii-type from Ishigakijima (YIO 12740, holotype of Ps. kobayashii). E,F: kobayashii-type from Yonagunijima (YIO 24194). Sy T. YAMASAKI Table 1. NP Month < September December March April June August Nw ABN DW W Cororo- orNK AN HUNnNOSCS —=NOSCCSO SSeoqgwt CB Seasonal changes in NP width, CB intensity, and the number of color morphs in Pycnonotus sinensis orii. Color morph Ill orii-type intermediate-type kobayashii-type 0 3 0 0 0 2 4 0 0 0 2 0 0 0 2 2 8 0 2) 7 2 0 0 2 wear. The color change patterns identified in the pres- ent study suggest that P. s. orii molts its nape feathers in spring and autumn, and its chest feathers only in autumn. The chest feathers may gradually become worn over the period from autumn to the next sum- mer. Future direct observations of molt and feather wear will, it is hoped, confirm the present conclusion. ACKNOWLEDGMENTS I thank K. Maehara for help with collecting the KUZ specimens used in this study, and M. Kajita and H. Morioka for their valuable suggestions. I am also grate- ful to T. Hikida, who patiently helped correct an earlier draft of the manuscript. Special thanks are due to H. Hirotani, Y. Kato, and K. Nakamura (KMNH), M. Motokawa (KUZ), I. Nishiumi (NSMT), K. Takehara (OPM), and T. Hiraoka and K. Momose (YIO) for al- lowing me to examine specimens under their care. REFERENCES Kuroda N (1923) Descriptions of new subspecies from Japan. Bull Br Ornithol Club 43: 105-109. Kuroda N (1930) Ishigakijima san shirogashira no ichi sinn ashu ni oite (An apparently new form of Pyc- 158 nonotus from the South Riu Kiu Islands). Tori 6: 271-274. (in Japanese with English summary) Mauersberger G & Fischer S (1992) Intraspecific vari- ability of the Light-Vented Bulbul, Pycnonotus sinen- sis (Gmelin, 1789). Mitt Zool Mus Berl 68 (Suppl): Ann Ornithol 16: 167-177. Nakamura K & Hanawa S (1987) Ryukyu retto san shi- rogashira no bunrui to bunpu hensen (Classification and change of distribution of Light-vented Bulbul within the Ryukyu Archipelago). In: Japan Environ- ment Agency (ed) Showa 61 nendo tokushu chorui chosa. pp 39-58. Japan Environment Agency. Tokyo (in Japanese). Rand AL & Deignan HG (1960) Family Pycnonotidae. In: Mayr E & Greenway JC (eds) Check-list of bird of the world. A continuation of the work of James L. Pe- ters. Vol 9. pp 221-300. Museum of Comparative Zo- ology, Cambridge. Traylor MA (1967) A _ collection of birds from Szechuwan. Fieldiana Zool 53: 1-67. Ornithological Society of Japan (1942) A hand-list of the Japanese birds. 3rd ed. Ornithological Society of Japan, Tokyo. Ornithological Society of Japan (2000) Check-list of Japanese birds. 6th ed. Ornithological Society of Japan, Obihiro. SHORT COMMUNICATION Ornithol. Sci. 1: 159-162 (2002) Foraging areas of Short-tailed Shearwaters during their northward migration along the Pacific coast of northern Japan ‘ # ORNITHOLOGICAL =n ITO SCIENCE © The Ornithological Society of Japan 2002 I studied the distribution of Short-tailed Shearwa- ters Puffinus tenuirostris off the Sanriku and Joban districts of northeastern Japan in April 1999 during their northward migration. Short-tailed Shearwaters breed in southeastern Australia and Tasmania and mi- grate to the North Pacific in the non-breeding season. During migration, part of the population passes along the coast of northern Japan between April and June (Serventy 1953; Shuntov 1974; Degawa & Watabe 1983; Watabe et al. 1987). In this area, Euphausia pacifica is an abundant zooplankton species and is considered to be a key species in food web among marine organisms (Taki et al. 1996; Taki 1998). The distribution pattern and foraging ecology of the Short-tailed Shearwaters during their nomadic phase have been studied in the Bering Sea, where millions of these birds feed predominantly on euphausiids (Ogi et al. 1980; Schneider et al. 1986; Hunt et al. 1996), however, little is known about their ecology during the northward migration period. In the present paper, I discuss the foraging distribution of Short- tailed Shearwaters during this northward migration __ by analyzing the density and behavior of shearwaters _ in relation to prey availability. METHODS Observations were made from the wing deck _ (height: 7 m) of the 692 ton R/V Wakataka-Maru (To- | hoku National Fisheries Research Institute:), from 13 to 23 April 1999 during daylight hours while the ship was underway at a speed of 12 knots (about 22 km/h). I identified, counted, and recorded the behavior of all _ seabirds seen within an area with a radius of 1,000m from directly ahead of the ship to 90° off the side of (Received 19 April 2002; Accepted 9 July 2002) * E-mail: shin@fish.hokudai.ac.jp 159 Faculty of Fisheries, Hokkaido University, Hakodate, Hokkaido, 041-8611 Japan the ship with the best visibility using 856mm and 20X60 mm image-stabilizer binoculars. Bird behav- ior was classified as either flying or sitting/foraging. The flight directions of flying birds were recorded as well as the time of each bird sighting. Ship speed and sea-surface temperature (SST) data were recorded at one minute intervals. The ship’s position was recorded every 10 minutes. Abundance data were standardized to number of birds/km* every 10 min- utes. The method used for observing the birds was the same as that described by Tasker et al. (1984). Based on oceanographic data collected by the R/V Wakataka-Maru and the Japan Fishery Information Service Centre (JFISC) SST isothermal map (JFISC 1999; 15-18 April), the sampling area was divided into the following water masses: the Oyashio Area (OA), characterized by SST <10°C; Transition Area (TA), characterized by SST >11°C; a thermal front (TF) near the 10-11°C SST isothermal line, and a Warm-Core Ring (WCR), which formed from the Kuroshio Extension was characterized by a SST of 11°C, and the WCR remained at 39-40°N, 142- 145°E (Fig. 1a). RESULTS AND DISCUSSION The total standardized number of all birds recorded during the 3,860min of observation was 6,130.2 birds belonging to 32 species. The total standardized number of Short-tailed Shearwaters was 2,164.4, which was 35.4% of the standardized total of all birds observed. Thus, this shearwater was the predominant species in this survey area and period (Table 1). The densities of Short-tailed Shearwaters varied widely throughout the survey area, and were high within 50-100km of the coast (Average+SD: 5.6+13.8 birds/km?; Maximum: 140.4 birds/km?, Fig. 1b). Sitting/foraging shearwaters were found on the TF Table 1. Shin ITO Species composition and total number of seabirds recorded during the survey off Sanriku and Joban, northern Japan, in April 1999. Species name Total Order Gaviiformes Family Gaviidae Pacific Loon Gavia pacifica WS Yellow-billed Loon Gavia adamsii 2.0 Order Procellariiformes Family Diomedeidae Laysan Albatross Diomedea immutabilis 40.1 Black-footed Albatross Diomedea nigripes 7? Family Procellariidae Northern Fulmar Fulmarus glacialis 26.6 Streaked Shearwater Calonectris leucomelas 786.2 Flesh-footed Shearwater Puffinus carneipes S5leZ Sooty Shearwater Puffinus griseus 347.9 Short-tailed Shearwater Puffinus tenuirostris 2,164.4 Family Hydrobatidae Fork-tailed Storm-petrel Oceanodroma furcata 0.5 Leach’s Storm-petrel | Oceanodroma leucorhoa 14.4 Sooty Storm-petrel Oceanodroma tristrami 667.1 Madeiran Storm-petrel Oceanodroma castro 13.9 Order Pelecaniformes Family Phalacrocoracidae Temminck’s Cormorant Phalacrocorax 0.6 filamentosus Pelagic Cormorant Phalacrocorax urile 4.7 Order Charadriiformes Family Phalaropodidae Red Phalarope Phalaropus fulicarius 345.0 Family Stercorariidae South Polar Skua Catharacta maccormicki 0.3 Pomarine Jaeger Stercorarius pomarinus 11.6 Long-tailed Jaeger Stercorarius longicaudus 33.8 Parasitic Jaeger Stercorarius parasiticus 0.8 Family Laridae Herring Gull Larus argentatus 20.0 Slaty-backed Gull Larus schistisagus 65.7 Glaucous Gull Larus hyperboreus 0.5 Black-tailed Gull Larus crassirostris 310.2 Black-legged Kittiwake Larus crassirostris 599.9 Common Tern Sterna hirundo 0.3 Family Alcidae Thick-billed Murre Uria lomvia 10.4 Marbled Murrelet Brachyramphus 0.7 marmoratus Ancient Murrelet Synthliboramphus 22258 antiquus Crested Auklet Aethia cristatella 0.3 Parakeet Auklet Aethia psittacula 0.6 Rhinoceros Auklet Cerorhinca monocerata 53.6 Total 6,120.2 160 and OA waters. Only a few flying birds and no sitting birds were observed in the WCR (Figs. 1b & Ic). Higher proportions (82%) of birds flew northward were recorded during survey period suggesting that they were on migration to northern sea area as de- scribed by Ito and Ogi (1999) (Table 2). However at 141 142 143 144 145°E Fig. 1. Oceanographic features during 15—18 April and dis- tribution of Short-tailed Shearwaters off Sanriku and Joban in April. (a) OA: Oyashio Area; TA: Transition Area; TF: ther- mal front; WCR: Warm-Core Ring; (b) flying and sitting/for- aging shearwaters; (c) sitting/foraging shearwaters only. +: position where a 10 minute count was made; the size of the circle represents the abundance of birds at each position (birds/km’). Foraging Areas of Short-tailed Shearwaters Table 2. Behavior of Short-tailed Shearwaters recorded for each sighting. N (%) Flying Northward 1,771.4 81.8 Flying Southward 14.0 0.6 Sitting/Foraging 378.9 eS Total 2,164.4 100.0 least five flocks consisting of 100—1,000 Short-tailed Shearwaters (including the birds foraging outside of the sampling radius) were observed on the water at 39°15'N, 142°00’E on 22 April 1999. In these flocks more than 80% of birds were actively diving into the water and they continued foraging until the ship ap- proached closely. Some of them (at least 20 birds) vomited euphausiids before flying away from the ship. Taki et al. (1996), and Taki and Ogishima (1997) reported that E. pacifica adults are the most abundant zooplankton in April and they tend to occur in the Oyashio where the water is at S—10°C (Kotani et al. 1996; Taki et al. 1996; Taki & Ogishima 1997). E. pacifica avoids areas of high SST and rarely oc- curs in the WCR (Taki 1998). In mid April, E. paci- fica form daytime surface swarms in the coastal area off Sanriku (Endo 1984; Odate 1991) Thus, the distri- bution of some shearwaters during their northward migration is presumably related to the presence of adult euphausiids in cold water masses off the San- riku and Joban district. ACKNOWLEDGMENTS I thank Professor Haruo Ogi, Faculty of Fisheries, Hokkaido University, Hokkaido, Japan, for his encour- agement throughout this study. I also thank Dr. Kazushi Miyashita, Faculty of Fisheries, Hokkaido University, Dr. Kenji Taki, Tohoku National Fisheries Research In- _ stitute, Miyagi, Japan, and the captain and crew of R/V Wakataka-Maru for their assistance during the survey. I also thank Dr. John R. Bower, Faculty of Fisheries Hokkaido University, for kindly helping with an earlier draft of this manuscript. REFERENCES Degawa M & Watabe Y (1983) Distribution of the Short-tailed Shearwater Puffinus tenuirostris in Japanese waters. Bull Appl Ornith 3: 19-27. Endo Y (1984) Daytime surface swarming of Euphausia 161 pacifica (Crustacea: Euphausiacea) in the Sanriku coastal waters off noretheastern Japan. Mar Biol 79: 269-276. Hunt GL, Coyle KO, Hoffman S, Decker MB & Fulint EN (1996) Foraging ecology of Short-tailed Shearwa- ters near the Pribilof Islamd, Berig Sea. Mar Ecol Prog Ser 141: 1-11. Ito S & Ogi H (1999) Flight speed measurement of Short-tailed Shearwaters Puffinus — tenuirostris recorded by ship-loading radar. J Yamashina Inst Or- nith 31: 88-93 (In Japanese with English summary). Japan Fishery Information Service Centre (1999) To- hoku fisheries and oceanographic condition bulletin chart of SST isothermal map during the 15—18 April 1999 period No. 659 (In Japanese). Kotani Y, Kuroda K & Taki K (1996) Ecological studies on Euphausia pacifica Hasen and seasonal change of it environment off Onagawa, Miyagi Prefecture II. Zooplankton biomass and copepod community struc- ture. Bull Tohoku Nat Fish Res Inst 58: 77-87 (In Japanese with English summary). Odate K (1991) Fishery Biology of the Krill, Euphausia pacifica, in the Northern Coasts of Japan. Suisan Kenkyu Sosho (Libr Fish Stud) 40: 1-100 (In Japan- ese with English summary). Ogi H, Kubodera T & Nakamura K (1980) The pelagic feeding ecology of the Short-tailed Shearwater Puffi- nus tenuirostris in the subarctic Pacific region. J Yamashina Inst Ornith 12: 157-182. Schneider DC, Hunt GLJr & Harrison NM (1986) Mass and energy transfer to seabirds in the southern Bering Sea. Cont Shelf Res 5: 241-257. Serventy DL (1953) Movements of pelagic sea-birds in the Indo-Pacific region. Proc 7th Pacific Sci Congr 4: 394-407. Shuntov VP (1974) Sea birds and biological structure of the ocean. National Technical Information Service U.S. Department of Commerce, Spring Field (Trans- lated from Russian). Taki K, Kotani Y & Endo Y (1996) Ecological studies on Euphausia pacifica Hasen and seasonal change of it environment off Onagawa, Miyagi Prefecture III. Distribution and diel vertical migration. Bull Tohoku Nat Fish Res Inst 58: 89-104 (In Japanese with Eng- lish summary). Taki K & Ogishima T (1997) Distribution of some de- velopmental stages and growth of Euphausia pacifica Hasen in the Northwestern Pacific on the Basis of Norpac Net samples. Bull Tohoku Nat Fish Res Inst 58: 77-87 (In Japanese with English summary). Taki K (1998) Horizontal Distribution and Diel Vertical Migration of Euphausia pacifica Hansen in Summer in and around a Warm-Core Ring off Sanriku, North- Shin ITO western Pacific. Bull Tohoku Nat Fish Res Inst 60: Watabe Y, Oka N & Maruyama N (1987) Seaonal ap- 49-61. pearance of Short-tailed (Puffinus tenuirostris) and Tasker ML, Jones PH, Dixon T & Bleake BF (1984) Sooty (Puffinus griseus) Shearwaters on the Tokyo- Counting seabirds at sea from ships: A review of Kushiro line, Japan. J Yamashina Inst Ornth 19: methods employed and a suggestion for a standard- 117-124. ized approach. Auk 101: 567-577. 162 SHORT COMMUNICATION Ornithol. Sci. 1: 163-166 (2002) Sexual differences in the external measurements of Black- tailed Gulls breeding on Rishiri Island, Japan ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2002 Osaka, 558-8585, Japan Sex ratio is an important parameter to consider in ecological and conservation § seabirds _ studies (Weimerskirch & Jouventin 1987). Visually assessing with certainty the sex of live seabirds, however, may be an impossible task when no obvious differences in plumage or body size exist between the two sexes. Several previous studies of Laridae species have shown that males are significantly larger than females (external measurements) and discriminant functions using measurements sex Laridae with great reliability (Shugart 1977; Ryder 1978; Fox et al. 1981; Schnell et al. 1985; Coulson et al. 1983; Evans et al. 1993; Bosch 1996; Rodriguez et al. 1996; Palomares et al. 1997). _ The Black-tailed Gull Larus crassirostris is a medium-sized gull, endemic to the northwest Pacific, breeding extensively in coastal regions around the Japanese archipelago. Although Black-tailed Gulls are one of the most common seabird species in Japan, little is known about their biology or morphology. _ Similarly to other Laridae species, Black-tailed Gulls have no sexual differences in their plumage or col- oration. This study is aimed at describing the external _ measurements of Black-tailed Gulls and determining a discriminant function using the measurements to fa- cilitate the sexing of the gulls in the field. METHODS A management program controlling the number of 'Black-tailed Gulls was carried out by the Rishiri (Received 14 May 2002; Accepted 12 July 2002) * Corresponding author, E-mail: niizuma @ affrc.go.jp * Present address: Yokohama Zoological Gardens, Kamishirane- cho 1175-1, Asahi-ku, Yokohama 241-0001, Japan. ** Present address: Japan Society for the Promotion of Science, Hokkaido National Fisheries Research Institute, Katsurakoi 116, Kushiro 085-0802, Japan. Michiyo CHOCHI'*, Yasuaki NIIZUMA?**** and Masaoki TAKAGP ' Department of Ecology and Systematics, Graduate School of Agriculture, Hokkaido Univer- sity, Sapporo 060-8589, Japan ? Institute of Low Temperature Science, Hokkaido University, Sapporo 060-0819, Japan 3 Laboratory of Animal Sociology, Faculty of Science, Osaka City University, Sumiyoshi-ku, town office on Rishiri Island (45°O5N, 141°07E), off the northwest coast of Hokkaido, Japan. In this re- gion, gulls are a nuisance to townspeople, inflicting damage on the commercial fishery and affecting flights to and from the airport nearby. Carcasses of gulls culled during the course of this program were used to investigate the relationship between the exter- nal measurements and sex. On Rishiri Island, about 7,900 pairs bred in 1998 and 8,600 pairs in 1999 (Hokkaido Government 1999, 2000), on a 9.12ha area covered with dwarf bamboo Sasa nipponica. Gulls were shot around the colonies from 14—18 June 1998 during the incubation period, and from 17—24 June 1999 during the incuba- tion and hatching periods (Hokkaido 1999, 2000). A total of 237 carcasses was collected and frozen less than one hour after being shot. External measure- ments were carried out after the carcasses had been thawed. Ten external measurements were performed fol- lowing the procedures described in Bosch (1996). Head length (HL), long bill length (LBL), short bill length (SBL), nalospi (Nal: the distance from the tip of bill to the nostril), bill depth (BD), tarsus length (Tar), foot length (FL), and middle toe length (MTL) were measured to the nearest 0.01 mm using vernier calipers. Tail length (Tail L: the length between the uropygial gland and the tip of one rectrix) and wing length (WL: natural chord) were measured to the nearest 0.1 mm and | mm, respectively, using rulers. Body mass was weighed to the nearest 0.1 g using an electrical balance. Several studies have found sexual dimorphism in body mass for Laridae species (Ingolf- sson 1969; Ryder 1978; Threlfall & Jewer 1978; Hunt et al. 1980; Fox et al. 1981; Coulson et al. 1983; Monaghan et al. 1983; Bosch 1996; Rodriguez et al. 1996; Palomares et al. 1997). However, due to its variability throughout the breeding season (Jones M. CHOCHL et al. Table 1. Average values for external measurements and MD index of Black-Tailed Gull. Standard deviations are expressed as=SD, and ranges are given in Parentheses. Significance levels were improved using the sequential Bonferroni technique. Adult Male (N=139) Female (N=98) t-value P-value MD index Head Length (mm) 112.78+2.66 103.99+2.53 25.55 <0.01 8.11 (105.60—118.99) (96.82-111.35) Long Bill Length (mm) 74.68+2.84 69.03+2.77 15.24 <0.01 7.86 (62.70-85.20) (63.10—77.40) Short Bill Length (mm) SP 2NEEDN2 47.67+2.16 16.11 <0.01 9.09 (46.75—57.10) (42.40-55.35) Nalospi (mm) 24.53+1.59 DESO ERS 10.40 <0.01 9.12 (21.44-28.80) (18.70-28.90) Bill Depth (mm) 17.1141.24 15.09+1.01 13.32 <0.01 12.55 (14.50-20.67) (13.30-18.90) Tarsus Length (mm) 57.62+2.15 SSE IG) 15.76 <0.01 8.00 (49.85-63.35) (45.60-57.45) Foot Length (mm) 104.47+3.52 97.36+3.46 15.42 <0.01 7.05 (89.45—114.12) (89.15—111.98) Middle Toe Length (mm) Sy) WaE D5) 48.28+2.34 12.72 <0.01 7.65 (43.45-57.04) (39.95—-52.01) Tail Length (mm) 143.85+5.83 136.36+4.84 10.43 <0.01 5.35 (126.5—157.0) (124.3-149.0) Wing Length (cm) 39.01+0.80 37.39+0.75 15.65 <0.01 4.24 (36.5—157.0) (35.6-39.2) 1994; Croxall 1995), sexual difference in body mass is not described here (Lorentsen & R@v 1994). Once these measurements were collected, birds were dis- sected and sexed based on their reproductive organs. Each external measurement was compared be- tween the sexes using Student’s f-test. Sequential Bonferroni’s f-test was performed to increase signifi- cance power. To compare the degree of sexual differ- ence among measurements, the mean difference (MD) index was used (Agnew & Kerry 1995). The MD index was expressed as the difference of the av- erage measurements between male and female: MD=200X(X.m—X f)/(X.m+X.f) where X.m and X.f are the mean of the measurements for males and females, respectively; a larger index in- dicating greater sexual dimorphism. Finally, a discriminant function analysis was per- formed using HL and BD for sex determination of the gull, because these two measurements are readily taken during field studies (Fox et al. 1981; Palomares et al. 1997). RESULTS Male gulls were significantly larger than females for all external measurements (Table 1). MD indices for parameters of the head region, i.e. HL, LBL, SBL, Nal, and BD were larger than those of the extremi- ties, i.e. WL and Tail L. The average body mass of males was 641.2 g+44.0 (SD) and 537.5 g+42.3 for females. Using HL and BD measurements, the following discriminant function was obtained; D=150.63—1.22 HL—1.14BD (F, 534=372.88, P<0.0001) Both two variables contributed significantly to the function (HL, F, 534=324.21, P<0.0001; BD, F, 534= 25.41, P<0.0001). Following this function, birds were classified as males when D0. This discriminant function proved reliable for 95.7% of the males (133 out of 139), for 98.0% of the females (96 out of 98), and for a total of 96.6% of the birds (Fig. 1). DISCUSSION This study demonstrated that Black-tailed Gulls, breeding on Rishiri Island have marked sexual differ- ences in their external measurements and can be sexed by a discriminant function using HL and BD with a total reliability of 96.6%. This discriminant function will prove useful for sexing Black-tailed Sexual differences in Black-tailed Gull 120.00 115.00 110.00 105.00 Head length (mm) 100.00 95.00 12.00 18.00 20.00 14.00 16.00 22.00 Bill depth (mm) Fig. 1. Relationship between bill depth (BD) and head length (HL) to the function 1.22 HL=150.63—1.14 BD. Open circles show females and closed circles show males. Gulls in the field. Rapid sexing of a bird would con- tribute to shortening handling time thereby allowing researchers to release birds quickly after a minimum of disturbance. Before applying this method to other populations or age classes, however, sexing accuracy requires further testing, especially since Laridae species show inter-colonial variation in external measurements (Threlfall & Jewer 1978; Monaghan et al. 1983; Jehl 1987; Evans et al. 1993), as well as age related differences (Coulson et al. 1981; Allaine & Lebreton 1990; Palomares et al. 1997). The greater difference in size between the sexes found in the head region indicates that this region, especially the bill, may play an important role, probably in sexual dis- play and territorial defence by males (Ingolfsson 1969). The smaller sexual dimorphism in the wing re- gion may be related to the species’ flight performance (Schnell et al. 1985; Croxall 1995). ACKNOWLEDGMENTS We would like to thank the Rishiri town office for au- thorizing us to use the gulls’ carcasses and for support- ing our work on Rishiri Island. We are grateful to M. Sato and M. Senda for assistance in the field and Y. Ropert-Coudert for helpful comments on the manu- script. 165 REFERENCES Agnew DJ & Kerry KR (1995) Sexual dimorphism in penguins. In: Dann P, Norman I & Reilly P (eds) The Penguins: ecology and management. pp 299-318. Surrey Beatty and Sons, Australia. Allaine D & Lebreton J-D (1990) The influence of age and sex on wing-tip pattern in adult Black-headed Gulls Larus ridibundus. Ibis 132: 560-567. Bosch M (1996) Sexual size dimorphism and determi- nation of sex in Yellow-legged Gulls. J Field Ornithol 67: 534-541. Coulson JC, Duncan N, Thomas CS & Monaghan P (1981) An age-related difference in the bill depth of Herring Gulls Larus argentatus. Ibis 123: 499-502. Coulson JC, Thomas CS, Butterfield JEL, Duncan N, Monaghan P & Shedden C (1983) The use of head and bill length to sex live gulls Laridae. Ibis 125: 549-557. Croxall JP (1995) Sexual size dimorphism in seabirds. Oikos 73: 399-403. Evans DR, Hoopes ED & Griffin CR (1993) Discrimi- nating the sex of Laughing Gulls by linear measure- ments. J Field Ornithol 64: 472-476. Fox GA, Cooper CR & Ryder JP (1981) Predicting the sex of Herring Gulls by using external measurements. J Field Ornithol 52: 1-9. Hokkaido Government (1999) Government report for revival plan of community with seabirds (in Japan- ese). Sapporo. Hokkaido Government (2000) Government report for revival plan of community with seabirds Hokkaido Government, Sapporo (in Japanese). Hunt GL Jr., Wingfield JC, Newman A & Farner DS (1980) Sex ratio of Western Gulls on Santa Barbara Island, California. Auk 97: 473-479. Ingolfsson A (1969) Sexual dimorphism of large Gulls (Larus spp.). Auk 86: 732-737. Jehl JR Jr. (1987) Geographic variation and evolution in the California Gull (Larus californicus). Auk 104: 421-428. Jones IL (1994) Mass changes of Least auklets Aethia pusilla during the breeding season: evidence for pro- grammed loss of mass. J Anim Ecol. 63: 71-78. Lorentsen S-H & R@v N (1994) Sex determination of Antarctic Petrels Thalassocia antarctica by discrimi- nant analysis of morphometric characters. Polar Biol. 14: 143-145. Monaghan P, Coulson JC, Duncan N, Furness RW, Shedden CB & Thomas C (1983) The geographical variation of the Herring Gull Larus argentatus within Britain and in northern Europe; a biometrical ap- proach. Ibis 125: 412-417. M. CHOCHI et al. Palomares LE, Arroyo B, Marchamalo J, Sainz JJ & Voslamber B (1997) Sex- and age-related biometric variation of black-headed Gulls Larus ridibundus in Western European populations. Bird Study 44: 310- STR Rodriguez EF, Pugesek BH & Diem KL (1996) A sex- ing technique for California Gulls breeding at Bam- forth Lake, Wyoming. J Field Ornithol 67: 519-524. Ryder JP (1978) Sexing Ring-billed Gulls externally. Bird-Banding 49: 218-222. Schnell GD, Worthen GL & Douglas ME (1985) Mor- 166 phometric assessment of sexual dimorphism in skele- tal elements of California Gulls. Condor 87: 484-493. Shugart GW (1977) A method for externally sexing gulls. Bird-Banding 48: 118-121. Threlfall W & Jewer DD (1978) Notes on the standard body measurements of two populations of Herring Gulls (Larus argentatus). Auk 95: 749-753. Weimerskirch H & Jouventin P (1987) Population dy- namics of the Wandering Albatross, Diomedea exu- lans, of the Crozet Islands: Cause and conservation of the population decline. Oikos 49: 315-322. SHORT COMMUNICATION Ornithol. Sci. 1: 167—170 (2002) External features and molecular sexing of the anomalous Pintail, Anas acuta, found at Hyo-ko Waterfowl Park, Niigata Prefecture, Japan ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2002 Fumio SUGIMORIP 951-8580, Japan Akira CHIBA', Koichi MURATA’, Shigeki MIZUNO?, Ryuhei HONMA* and ’ Department of Biology, Nippon Dental University School of Dentistry at Niigata, Niigata ? Department of Wildlife Science, College of Bioresource Sciences, Nihon University, Fujisawa 1866, Kanagawa 252-8510, Japan > Department of Agricultural and Biological Chemistry, College of Bioresource Sciences, Nihon University, Fujisawa 1866, Kanagawa 252-8510, Japan * Niigata Prefecture Wild Bird Conservation Group, Shiunji Town, Niigata 957-0231, Japan ° Research Center, Yamashina Institute of Ornithology, Abiko, Chiba 270-1145, Japan Phenotypic variation of the plumage including al- binism and melanism is widely known among birds. On the other hand, morpho-chromatic variation of the plumage caused by hybridization has also been found in nature. Such an example is occasionally found in the Anatidae (Kuroda 1939), and possible hybrids of the Pintail Anas acuta with other species, e.g., Mal- lard (A. platyrhynchos), Teal (A. crecca), Baikal Teal (A. formosa), Gadwall (A. strepera), Wigeon (A. penelope), Shoveler (A. clypeata) and Pochard (Aythya ferina), were documented previously (Kuroda 1939). In fact, a presumptive hybrid between the Pin- tail and Bikal Teal was observed also in Hyo-ko Wa- terfowl Park, located in a suburb of Niigata City, Ni- igata Pref., Japan (Chiba, unpubl. data). During the course of bird-banding study in this Park, we inciden- tally found 2 unusual male Pintails, which were mor- phologically different from any known hybrids. The present study, therefore, was conducted to clarify the characters of these anomalous Pintails on the basis of macroscopic examination and molecular sexing. MATERIALS AND METHODS 1) Birds Thirteen individuals of Pintail were used in this study (Table 1). They were captured (by permission of the Ministries of the Environment and Culture, Japan) at Hyo-ko Waterfowl Park (37°50'N, 139°14’E) in the suburb of Niigata City, Niigata Pre- (Received 10 June 2002; Accepted 26 August 2002) * Corresponding author: E-mail: chiba@ngt.ndu.ac.jp 167 fecture, Japan, during 2 winter seasons, from January 2000 to March 2001. The birds were caught hu- manely by hand-made net or by hand, marked with a metal ring, and kept in plastic cages (80 cm 50 cm 25cm) for a while. Then, their external features were macroscopically examined and recorded in photo- graphs, and measurement of the body was made. Sex and age were checked mainly based on their plumage and cloacal structure. The contour feathers as a source of DNA, 5 to 7 feathers per bird, were plucked from the mid breast region by using sterilized for- ceps, put into clean plastic bottles, and stored in a re- frigerator at —10°C before extracting the DNA. After examination, the birds were released back to the wild. 2) Isolation of gnomic DNA Genomic DNA was isolated from 3 contour feath- ers plucked from the mid breast of each bird. The method of DNA extraction followed Walsh et al., (1991), Murata & Masuda (1996) and Murata et al. (1998), with a slight modification. Briefly, a length of calamus about 3mm was removed from the proximal end with clean scissors and cut into smaller pieces, which were incubated in 200 uL of 5%(w/v) Chelex® (Bio-Rad) at 56°C overnight, and then boiled in a water-bath for 8min. After centrifugation at 12,000 rpm for 5 min, the supernatant was used as template DNA for PCR. 3) PCR for gender determination Primer sets used in this study for amplifying the Z/W chromosome-specific DNA sequences of Pintail were those used for sexing domestic duck (A. A. CHIBA et al. platyrhynchos var. domestica) as reported by Itoh et al. (2001). The sequences of sexing primer for ampli- fying the W chromosome-specific DNA (~190 bps) were 5’-ACAGTTTGTCTGTCTCCGGGGAA-3’' (AWS03) and 5’-AGCTGGAYTTCAGWSCATCTT- CT-3' (USP3), and those of internal control primer for amplifying the Z/W chromosome-common DNA (~250 bps) were 5'-CTCTGTCTGGAAGGACTT-3’ (INT-R) and 5’-ATAGAAACAATGTGGGAC-3’ (INT-F). Detailed information about these primers and related sequences was given elsewhere (Itoh et al. 1997, 2001; Ogawa et al. 1997). PCR was carried out in a 25-uWL mixture contain- ing a 0.2 mM concentration of each dNTP, 50 pmoles of primers AWS03, USP3, INT-R and INT-F, 3 uL of DNA extract, 0.25 units of Jag polymerase (Amer- sham), and 2.5 uL of 10 PCR buffer. The PCR con- ditions used for the thermal cycler (PE Applied Biosystems, 9700) were as follow: initial DNA de- naturation at 95°C for 3 min followed by 35 cycles of 95°C, 80sec for denaturation; 59°C, 90sec for an- nealing; 72°C, 60 sec for extension, and lastly 72°C, 9 min for final elongation. PCR products (8 WL) were electrophoresed on a 2% agarose gel (A-6013, SIGMA) in 0.5 TBE (44.5mM Tris-borate, 44.5 mM Boric acid, 0.5 mM EDTA) buffer at 100V for 35 min, stained with ethidium bromide, and visualized under a UV transilluminator. RESULTS Externally, 2 birds in question, Sp-1 (ring number, 10A-80409) and Sp-2 (10A-75898), were character- ized by nuptial plumage of the male type (Fig. 1A). Measurement values of the wing, tail and body weight of these birds approximately corresponded to Fig. 1. External features (A) and external aspect of the disclosed vent (B) of an anomalous Pintail (Sp-2). Cor- responding part of an adult male with normal plumage (C) and that of an adult female with normal plumage (D) are also shown. dl, dorsal lip of cloaca; p, phallus (artificially everted and erected); vl, ventral lip of cloaca. Scale bar, | cm t fe Anomalous Pintail Table 1. for comparison Measurements of anomalous males (SP-1 and SP-2) of the Pintail, Anas acuta, and of control (normal plumage) birds Wing* Tail* Bill* Tarsus* Body Weight** Sp-1 (10A-80409) 259.0 148.0 46.5 40.8 810.0 Sp-2 (10A-75898) 259.0 173.0 47.3 42.5 790.0 Control Normal males (N=5) 260.8+5.4 145.4+38.5 50.6+2.4 41.8+2.0 836.0+50.0 Normal females (N=5) 224.6+10.6 109.8+11.2 47.7+2.0 38.9+2.3 722.0+65.5 “Buff” mutant female 260.0 110.0 45.3 39.0 670.0 *, mm; **, g; Numerals are presented as the average +SD those of control males rather than to those of control 1 Pomona 5 66 females (Table 1). However, their plumage was un- clear in comparison with normal male plumage; i.e, ™~ 250bp > their heads were light brown in color, the breast and ™ 190bp > upper belly not brilliantly white, but finely striped, the back and sides roughly striped, the black spot on the scapulars indistinct, and the dark undertail coverts and creamy caudal belly unclearly demarcated (Fig. 1A). As so far studied, no morphological evidence of hybridization with other species has been detected in the present birds. Interestingly, neither phallus (Fig. 1C) nor its equivalent in the cloacal region (Fig. 1B), which macroscopically appeared just the same as that seen in the female duck (Fig. 1D). The W chromosome-specific DNA fragments (~ 190 bp) were PCR-amplified from feather extracts of 2 unusual males, Sp-1 and Sp-2, one “Buff” mutant fe- male, and normal-plumage females. In contrast, only Z/W chromosome-common DNA fragment (~250 bp) was amplified from normal-plumage males (Fig. 2). No DNA fragment was amplified from negative control (distilled water). Thus, the results indicated that the 2 atypical males in question have DNA se- quences common to the sequences of Z and W chro- mosomes. DISCUSSION The present study provided new data available for characterization of unusual individuals of Pintail. The molecular sexing data suggested two possibilities, i.e., (1) that the anomalous birds studied are geneti- cally female (Z/W) irrespective of their male-type plumage and (2) that they represent the individuals of sex chromosomal aberrations, e.g., ZZW. Apart from these possibilities, one may presume that the birds in question are immature (under-yearling) males. How- ever, the third possibility can be excluded, for they 169 Fig. 2. Electrophoretic pattern of DNA for gender determi- nation of Pintail. The W chromosome-specific DNA fragments (~190 bp) were amplified from plucked contour feathers of anomalous Pintails (lanes 1 and 2), “Buff” mutant female (lane 3) and normal-plumage female (lane 5), while only Z/W chromosome-common DNA fragment (~250 bp) was ampli- fied from normal-plumage male feathers (lane 4). No sex-spe- cific DNA band was observed in the negative control using distilled water (DW, lane 6). had no phallus in the cloaca, although no anatomical evidence on the gonads was obtained. In favor of this view, a study made long ago showed that immature Mallard males killed between late July and early No- vember had a macroscopically distinct phallus (Hohn 1960). Furthermore, recovery data of the marking showed that at least one of them, Sp-2, was a 2+- year-old adult bird. If the anomalous birds studied are genetic females, we have to explain why did the genetic females have a male character for their plumage? Previous studies cited in a review paper (Witschi 1961) may be helpful for discussing this point. In the duck, Wolff and Wolff (1949) and Wolff (1950) showed that the de- velopment of accessory sex organs, syrinx and phal- lus, depends on the gonads in the prehatching stages: gonadectomy caused various degrees of masculiniza- tion in the female embryos, i.e., the enlargement of syrinx and the development of the phallic tubercle. Unfortunately, however, these studies provided no in- formation about the effect of gonadectomy on the plumage. In birds, it is generally known that go- nadectomy or deprivation of sex steroids causes the sex-related characters to turn to the phenotype of the A. CHIBA et al. homozygote (genetic male), not heterozygote (ge- netic female), of the sex chromosomes. The present birds may be a case of this phenomenon. If so, we may speculate that the present birds may have dys- functional ovary or they may have been physiologi- cally ovariectomized, presumably in an earlier life stage. Currently, we have no data to exclude the second possibility, i.e., the sex chromosomal aberration. It is also known that the domestic fowls of ZZW-genotype show masculinization in the post-hatching early life stages (Naito 1998). In any case, we need further in- formation about the anomalous Pintail from studies on the gonads, chromosomes, genes, plasma concen- tration of sex steroids, and so on. Recent progress made by studies on the Z/W sex chromosomes of birds has been reviewed with respect to the mecha- nisms of sex determination and sex differentiation (Naito 1998; Ellegren 2001; Mizuno 2001). So far as we surveyed during the recent 3 years, the incidence of the present anomaly in the study area is estimated to be 0.02—0.03%. Exact causal factor(s) of the anomaly found in the Pintail remain unknown, but it seems to be important to examine possible rela- tionship between the anomaly and the global pollu- tion. Future comprehensive studies on sexually anomalous birds, both in the field and laboratory, may contribute to various aspects of avian biology and en- vironmental chemistry. ACKNOWLEDGMENTS We are grateful to Mr. I. Satoh, Director of Hyo-ko Waterfowl Park, and a local bird-banding group for their help with the field study. REFERENCES Ellegren H (2001) Hens, cocks and avian sex determina- tion. A quest for genes on Z or W? EMBO Rep 2: 192-196. Hohn EO (1960) Seasonal changes in the mallard’s penis and their hormonal control. Proc Zool Soc Lond 134: 547-555. Itoh Y, Ogawa A, Murata K, Hosoda T & Mizuno S (1997) Identification of the sex of Oriental white stork, Ciconia boyciana, by the polymerase chain re- 170 action based on its sex chromosome-specific DNA se- quences. Genes Genet Syst 72: 51-56. Itoh Y, Suzuki M, Ogawa A, Munechika I, Murata K & Mizuno S (2001) Identification of the sex of a wide range of Carinatae birds by PCR using primer sets se- lected from chicken EE0.6 and its related sequences. J Hered 92: 315-321. Kuroda N (1939) Gan to Kamo (Geese and ducks of the worlds). Shyukyosha-Shoin, Tokyo (in Japanese). Mizuno S (2001) Current studies on mechanisms of sex determination and sex differentiation with the ZW sex chromosomes of birds: comparison with studies on mammalian systems. Seikagaku 73: 1411-1427 (in Japanese). Murata K & Masuda R (1996) Gender determination of the Linne’s two-toed sloth (Choloepus didactylus) using SRY amplified from hair. J Vet Med Sci 58: 1157-1159. Murata K, Itoh Y, Ogawa A & Mizuno S (1998) Sexing the Oriental white stork Ciconia boyciana by PCR using a single plucked feather as a source of DNA. Jpn J Ornithol 46: 157-162. Naito M (1998) Sexual differentiation of gonads and ga- metogenesis in the avian species. Tanpakushitsu Kakusan Koso 43: 470-477 (in Japanese). Ogawa A, Solovei I, Hutchinson N, Saitoh Y, Ikeda J, Macgregor H & Mizuno S (1997) Molecular charac- terization and cytological mapping of a non-repetitive DNA sequence region from the W chromosome of chicken and its use as a universal probe for sexing Carinatae birds. Chromosome Res 5: 93-101. Walsh PS, Metzer DA & Higuchi R (1991) Chelex 100* as a medium for simple extraction of DNA for PCR- based typing from forensic material. Biotechniques 10: 506-513. Witschi E (1961) Sex and secondary sexual characters. In: Marshall AJ (ed) Biology and comparative physi- ology of birds. Vol 2. pp 115-168. Academic Press, New York. Wolff Em (1950) La différenciation sexuelle normale et le conditionnement hormonal des caractéres sexuels somatiques préces, tubercule génital et syrinx, chez l’embryon de canard. Bull Biol Fr Belg 84: 119-193. Wolff Et & Wolff Em (1949) Application de la methode de castration 4 l’embryon de canard: sur deux tests de activité précoce des gondes embryonnaires, la sy- rinx et le tubercule génital. Compt Rend Soc Biol 143: 529-531. ISSN 1347-0558 -ORNITHOL OGICAL SCIENCE Vol.1 2002 The Ornithological Society of Japan ORNITHOLOGICAL SCIENCE Official journal of the Ornithological Society of Japan Editor-in-Chief Keisuke Ueda, Rikkyo University, Tokyo Associate Editors Teruaki Hino, Forestry and Forest Products Research Institute, Kyoto Hidetsugu Sakai, Nihon University, Tokyo Editorial Board Masahiko Nakamura, Joetsu University of Education, Joetsu Isao Nishiumi, National Science Museum, Tokyo Kazuo Okanoya, Chiba University, Chiba Eiichiro Urano, Yamashina Institute for Ornithology, Abiko Advisory Board Alexander V. Andreev, Institute of Biological Problems, Magadan Walter J. Bock, Columbia University, New York Jiro Kikkawa, The University of Queensland, Brisbane Woo-Shin Lee, Seoul National University, Suwon Bernd Leisler, Max-Planck-Gesellschaft, Radolfzell Anders P. Mller, Universite Pierre et Marie Curie, Paris Richard Noske, Northern Territory University, Casuarina Pilai Poonswad, Mahidol University, Bangkok Lucia Liu Severinghaus, Academia Sinica, Taipei Navjot S. Sodhi, National University of Singapore, Singapore Jeffrey R. Walters, Virginia Polytechnic Institute and State University, Blacksburg John C. Wingfield, University of Washington, Seattle Jeong-Chil Yoo, Kyung-Hee University, Seoul CONTENTS OF VOLUME 1 Number 1 PREFACE Higuchi H From Japan through Asia to the world: building bridges in ornithological science. 1 EDITORIAL Ueda K “Ornithological Science”, the new English publication from Japan. SPECIAL FEATURE Interspecific segregation and attraction in forest birds Hino T Introduction. 5 January 2002 Loyn RH Patterns of ecological segregation among forest and woodland birds in south-eastern Australia. 7 Recher HF, Davis WE Jr & Calver MC Comparative foraging ecology of five species of ground-pouncing birds in western Australian woodlands with comments on species decline. 29 Monkkonen M & Forsman J Heterospecific attraction among forest birds: a review. 4] Seki S & Sato T The effect of typhoon on the flocking and foraging behavior of tits. 53 Printer: Kokusai Bunken Insatsusha Co., Ltd., Takada-no-baba 3-8-8, Shinjuku-ku, Tokyo 169-0075, Japan. Tel: +81-3-3362- 9741, Fax: +81-3-3368-2822. Cover design: Eiichiro Urano Murakami M | Foraging mode shifts of four insectivorous birds species under temporally varying resource | distribution in a Japanese deciduous forest. 63 Mizutani M & Hijii N | The effects of arthropod abundance and size | | \ Sodhi NS The effects of food-supply on Southeast Asian forest birds. 89 ORIGINAL ARTICLES Yamagishi S, Asai S, Eguchi K & Wada M on the nestling diet of two Parus species. 7] Spotted-throat individuals of the Rufous Vanga Schetba rufa are yearling males and Hino T , Unno A & Nakano S presumably sterile. 95 Prey distribution and foraging preference for tite. s3 Amano HE & Eguchi k Nest-site selection of the Red-billed Leiothrix and Japanese Bush Warbler in Japan. 101 | Number 2 September 2002 ! ORIGINAL ARTICLES Brazil M ' Hino T Common Raven Corvus corax at play; records . Breeding bird community and mixed-species from Japan. 150 flocking in a deciduous broad-leaved forest in Kawakami K & Higuchi H | western Madagascar. fil The first record of cavity nesting in the | Kotaka N & Matsuoka S Ogasawara Islands Honeyeater Apalopteron Secondary users of Great Spotted Woodpecker familiare on Hahajima, Bonin Islands, Japan. 153 (Dendrocops major) nest cavities in urban and Vamnasalal suburban forests in Sapporo City, northern Seasonal variation of plumage color in Japan. 117 Japanese Light-vented Bulbul Pycnonotus _ Amano HE & Eguchi K sinensis orii in the Yaeyama Group, Southern Foraging niches of introduced Red-billed Ryukyus. 155 Leiothrix and native species in Japan. 123 te S Unno A Foraging areas of Short-tailed Shearwaters __ Tree species preferences of insectivorous birds during their northward migration along the in a Japanese deciduous forest: the effect of Pacific coast of northern Japan. 152 different foraging techniques and seasonal Chochi M, Niizuma Y & Takagi M change of food resources. 133 Sexual differences in the external | measurements of Black-tailed Gulls breeding SHORT COMMUNICATIONS on Rishiri Island, Japan. 163 | exam ie eee eeenl H bs Chiba A, Murata K, Mizuno S, Honma R & _ Bird predation by domestic cats on Hahajima : f : Island, Bonin Islands, Japan. 143 SEM F | External features and molecular sexing of the Tojo H, Nakamura S & Higuchi H anomalous Pintail, Anas acuta, found at Hyo- | Gape patches in Oriental Cuckoo Cuculus ko Waterfowl Park, Niigata Prefecture, saturatus nestlings. 145 Japan. 167 Instructions to Authors Ornithological Science is wholly owned and published by the Ornithological Society of Japan (OSJ), and published in English twice a year (January and July). Membership to OSJ is not required for contributors to Ornithological Science. Manu- scripts may be submitted internationally for free publication up to 20 pages, although color pages will be charged sepa- rately to contributors. The contribution can be classified into four categories: Reviews, Original Articles, Short Communi- cations and Comments. Comments are only for the papers published on Ornithological Science. Accepted papers automatically become the copyrighted property of OSJ and may not be reproduced, in whole or part, without a written permission from OSJ. Submission Manuscripts have to be original and so far unpublished. The contributors are required to prepare three complete photo- copies (the original not required at the time of reception) of manuscript, tables and figures. A floppy disk of 3.5-inch con- taining a text file of the manuscript is also required after ac- ceptance. The Editorial Board of Ornithological Science will then forward the submitted materials to two unidentified referees, and taking account of their opinions, the Editor-in-Chief will decide the validity of the submitted material for publication. In case the contributors are requested to reconsider part of their materials, they are supposed to amend the material within a month before returning the materials with a note describing Should the contributors’ and the referees’ opinions show discrepancy, how the amendments have been carried out. the contributors are allowed to express their view separately. Any material overdue the above time period is automatically interpreted as having been withdrawn by the contributors. In- ternational submission of papers prepared by Microsoft WORD and PDF files may be accepted via e-mails. Organization of Manuscript The Editorial organization of an original article: (1) Title, Author name(s) Board suggests the following preferred and Affiliation(s) (2) Abstract (3) Introduction (4) Materials and Methods (5) Results (6) Discussion (7) Acknowledgments (8) References (9) Tables (10) Figures and Legends. It can be modified and changed for clarification of the whole structure and for effectiveness of arguments. In a Short Communication, no abstract is required. lhe manuscripts should be printed on A4 sheets, with one- inch margins for refereeing, 20 lines per sheet. No rigid limit for the length of manuscripts are set, but the Editorial Board deems anything exceeding 20 pages in a final published form to be long enough. Short Communications have to be less than four pages in time of publication. All manuscript sheets must be numbered consecutively. The use of footnotes to the text is not recommended. Extra Charges A charge of 2,500 yen will be added to every page that ex- ceeds 20 free pages. Each color page will be charged 50,000 yen. General Instructions 1. Cover page The first page of the manuscript should contain the title, au- thor names, affiliations, addresses, the contact name and ad- dress (phone and/or facsimile number together with an e-mail address) for correspondence. Authors with different affilia- tions should be identified by the use of a superscript linking the name and affiliation. Also on the same page, a running title (less than 40 letters) and the number of tables and figures accompanying the manuscript. To discriminate the family name(s) from their first and/or the middle ones, contributors are encouraged to spell their family names with all capitals. 2. Abstract and key words Every submission should accompany an abstract less than 300 words followed by alphabetically arranged key words less than six. 3. Names of birds Contributors are requested to give scientific name at the first mention of a species both in the abstract and in the article. Sci- entific and English name of birds should follow an authorita- tive source for the appropriate regions. Capitals should be used for the initial letters for all single words or hyphenated English names (e.g., Black-billed Magpie) but not a group name (e.g., Crows). 4. References References in the text should be cited as follows: One author: Tokugawa (1995) or (Tokugawa 1995) Two authors: Oda and Smith (1996) or (Tokugawa 1995; Oda & Smith 1996) More than two authors: Mori et al. (1997) or (Mori et al. 1997; Richard et al. 1999a, b) Manuscripts accepted for publication but not yet pub- lished should be listed in the reference section as ‘in press.’ All contributors are advised to prepare a list of references cited in the text arranged alphabetical by the first author names. (Book) Cambell RC (1974) Statistics for biologists. 2nd ed. Cam- bridge University Press, London. (Chapter in an edited book) Dawson WH (1996) Energetic features of avian thermoregula- tory responses. In: Carey C (ed) Avian energetics and nu- tritional ecology. pp 85-124. Capman & Hall, New York. Oring LW (1982) Avian mating system. In: Farner DS, King JR & Parkes KC (eds) Avian biology. Vol 1. pp 1-92. Academic Press, New York. (Journal article) Yamaguchi N & Kawano KK (2001) Effect of body size on the resouce holding potential of male varied tits Parus varius. Jpn J Ornithol. 50: 65-70. Papers written in a language other than English has to be ac- companied by a language designators such as “(in Japanese)” together with a translated title. Yamagishi S (1981) Mozu no yomeiri-toshikouen no mozu no seitai wo saguru (The bridal of shrikes - ecological re- search of bull-headed shrikes in an urban park). Dai-Nip- pon-Tosho, Tokyo (in Japanese). The sources of the reference should be given following the commonly accepted abbreviations for journal titles (refer to the ‘International List of Periodical Title Abbreviations’). 2 66 The use of “in preparation”, “submitted for publication”, or “personal communication” is not allowed in the reference list. Also, “unpublished data” and “personal communication” should appear parenthetically following the relevant name(s) in the text. 5. Units and Abbreviations All measurements should be given in SI units. Also, the Editorial Board suggests 24-hour clock (e.g., 0930 and 1500) and the geographical location 35°20'25"N, 136°10'20”E. Sta- tistical symbols used in the contribution are to be capitalized except df (e.g., SD, SE, P, N, df). The Board further prefers unitalicized expressions for Latin phrases frequently used. 6. Charts and Illustrations The contributors are supposed to take the size of the printed page into consideration when they prepare their tables and figures: the actual print is on 165223 mm in double columns with 80 mm width. ORNITHOLOGICAL SCIENCE Volume 1 Number 2 September 2002 Contents ORIGINAL ARTICLES Hino T Breeding bird community and mixed-species flocking in a deciduous broad-leaved forest in western Madagascar. 111 Kotaka N & Matsuoka S Secondary users of Great Spotted Woodpecker (Dendrocops major) nest cavities in urban and suburban forests in Sapporo City, northern Japan. LIF Amano HE & Eguchi K Foraging niches of introduced Red-billed Leiothrix and native species in Japan. 123 Unno A Tree species preferences of insectivorous birds in a Japanese deciduous forest: the effect of different foraging techniques and seasonal change of food resources. 133 SHORT COMMUNICATIONS Kawakami K & Higuchi H Bird predation by domestic cats on Hahajima Island, Bonin Islands, Japan. 143 Tojo H, Nakamura S & Higuchi H Gape patches in Oriental Cuckoo Cuculus saturatus nestlings. 145 Brazil M Common Raven Corvus corax at play; records from Japan. 150 Kawakami K & Higuchi H The first record of cavity nesting in the Ogasawara Islands Honeyeater Apalopteron familiare on Hahajima, Bonin Islands, Japan. 153 Yamasaki T Seasonal variation of plumage color in Japanese Light-vented Bulbul Pycnonotus sinensis orii in the Yaeyama Group, Southern Ryukyus. 155 Ito S Foraging areas of Short-tailed Shearwaters during their northward migration along the Pacific coast of northern Japan. 159 Chochi M, Niizuma Y & Takagi M Sexual differences in the external measurements of Black-tailed Gulls breeding on Rishiri Island, Japan. 163 Chiba A, Murata K, Mizuno S, Honma R & Sugimori F External features and molecular sexing of the anomalous Pintail, Anas acuta, found at Hyo- ko Waterfowl Park, Niigata Prefecture, Japan. 167 Published by the Ornithological Society of Japan Printed by Kokusai Bunken Insatsusha Co., Ltd. ISSN 1347-0558 is oer? > 7 Ike ¥ 'Aiiss= Fe 23 JUL 2003 - ORNITHOLOGICAL SCIENCE Vol.2 No.1 February 2003 ~ The Ornithological Society of Japan ORNITHOLOGICAL SCIENCE Official journal of the Ornithological Society of Japan Editor-in-Chief Keisuke Ueda, Rikkyo University, Tokyo Associate Editors Teruaki Hino, Forestry and Forest Products Research Institute, Kyoto Hidetsugu Sakai, Nihon University, Tokyo Editorial Board Masahiko Nakamura, Joetsu University of Education, Joetsu Isao Nishiumi, National Science Museum, Tokyo Kazuo Okanoya, Chiba University, Chiba Eiichiro Urano, Yamashina Institute for Ornithology, Abiko Advisory Board Alexander V. Andreey, Institute of Biological Problems, Magadan Walter J. Bock, Columbia University, New York Jiro Kikkawa, The University of Queensland, Brisbane Woo-Shin Lee, Seoul National University, Suwon Bernd Leisler, Max-Planck-Gesellschaft, Radolfzell Anders P. Mller, Universite Pierre et Marie Curie, Paris Richard Noske, Northern Territory University, Casuarina Pilai Poonswad, Mahidol University, Bangkok Lucia Liu Severinghaus, Academia Sinica, Taipei Nayjot S. Sodhi, National University of Singapore, Singapore Jeffrey R. Walters, Virginia Polytechnic Institute and State University, Blacksburg John C. Wingfield, University of Washington, Seattle Jeong-Chil Yoo, Kyung-Hee University, Seoul Editorial Policy Ornithological Science publishes reviews, original articles, short communications and comments covering all aspects of ornithology. Manuscripts are judged on the basis of their contribution of original data and ideas or interpretation. All articles are peer-reviewed by at least two researchers expert in the field of the submitted paper. Manuscripts are edited where necessary for clarity and economy. Ornithological Science aims to publish as rapidly as is consistent with the requirements of peer-review and normal publishing constraints. Submission Manuscripts and editorial correspondence should be ad- dressed to: Dr. Teruaki Hino, Associate Editor, Forestry and Forest Products Research Institute, Kyoto 612-0855, Japan. Tel: +81-75-611-1201, Fax: +81-75-611-1207, e-mail: tkpk @affre.go.jp For detailed instructions concerning the submission of man- uscripts, please refer to the Instructions to Authors at the in- side of the back cover of each issue of the journal or visit the OSJ web page: http://wwwsoc.nii.ac.jp/osj/ Subscriptions Membership of the Ornithological Society of Japan (OSJ) is open to anyone interested in ornithology and the aims of the OSJ. The annual fee for an ordinary member is Japanese Yen 5,000. Members are entitled to receive two journals. Ornithological Science (in English; two issues per year) and Japanese Journal of Ornithology (with English summary; two issues per year). To join the OSJ, please apply to: The OSJ Office, Laboratory of Biodiversity Science, School of Agricultural and Life Sciences, The University of Tokyo, Yayoi 1-1-1, Tokyo 113-8657, Japan. Tel: +81-3-5841-7541, Fax: +81-3-5841-8192, e-mail: osj@lagopus.com Copyright Submission of a manuscript implies: that the work de- scribed has not been published before; that is not under con- sideration for publication elsewhere; that its publication has been approved by all co- authors, if any, as well as by the re- sponsible authorities at the institute where the work has been carried out; that, if and when the manuscript is accepted for publication, the authors agree to automatic transfer of the copyright to the OSJ; and that the manuscript will not be pub- lished elsewhere in any language without the consent of the copyright holders. ISSN: 1347-0558. All articles published in the Ornithological Science are pro- tected by copyright, which covers the exclusive rights to repro- duce and distribute the article (e.g. as offprints) as well as all translation rights. No material published in this journal may be reproduced photo-graphically or stored on microfilm, in elec- tronic data bases, video disks, etc., without first obtaining writ- ten permission from the copyright holders. Special requests should be addressed to: Dr. Keisuke Ueda, Editor-in-Chief, Laboratory of Animal Ecology Faculty of Science, Rikkyo University, Ikebukuro, Tokyo 171-0021, Japan, Tel: +81-3-3985-2596, Fax: +81-3-3985-2596, e-mail: keisuke @rikkyo.ac.jp Printer: Kokusai Bunken Insatsusha Co., Ltd., Takada-no-baba 3-8-8, Shinjuku-ku, Tokyo 169-0075, Japan. Tel: +81-3-3362- + $1-3-3368-2822. Cover design: Elichiro Urano 1741, Fax Ornithological Society of Japan i Membership Application a Office of the Ornithological Society of Japan, Laboratory of Biodiversity Science, School of Agriculture and Life Sciences University of Tokyo, Yayoi 1-1-1, Bunkyo-ku Tokyo 113-8657, JAPAN Phone: +81-3-5841-7541 Fax: +81-3-5841-8192 ts E-mail: osj@lagopus.com on (LC) Contact details have changed LU) No change) Name: Address (include postal code): - Seater eee ee ee Fax: E-mail: Type of membership (Please check one of the following boxes): * Ordinary member Annual dues ¥5,000 (Regular member) =: Student member Annual dues ¥3,000 (Members with confirmation of student status) ¢ Sustaining member Annual dues ¥10,000 (Member supporting the Society economically) © Institutional member Annual dues ¥10,000 (Universities, companies, non-governmental organizations, research groups) Certification of student status (to be completed by student’s supervisor): ete See Oh. . Certify that is a student registered at the institution below. Name of university: Supervisor's signature: Donations to the Society ¥ Purchase of “Check-list of Japanese Birds” (¥4,380/copy) ¥ Subtotal ¥ / : Total of membership dues and other payments: ¥ ‘ at (Please check selected method of payment) Siiemationsl Postal Money Order (Address to the Ornithological Society of Japan) * Direct deposit into the Society’s bank account Account information: Bank: Mizuho Bank, Nezuekimae Branch Account type and number: Regular account 1743740 Account administrator: President of the Ornithological Society of Japan, Dr. Hiroyoshi Higuchi * Credit card (Please check either VISA or Mastercard. We are unable to accept any other cards.) - VISA - Mastercard Credit card number: Expiry (mo/yr): / Name on card (First-last): _ Signature: SPECIAL FEATURE Ecology of seed dispersal INTRODUCTION Seed dispersal is a process in which immobile plants expand the species range in a wide scale of the distri- bution. Plants bear fruits with high nutrient value as a reward to animals which transport them. Animals dis- perse the seeds far from their parent plants. It is an important process to avoid inbreeding around parent plant and includes escaping high mortality of seedling around the parent plant via intraspecific competition with their kin. As a result, most of the fruits have been adapted to dispersal by birds and other animals. The relationship between the fruiting phenology of fruited plants and frugivorous birds has been investi- gated in northern temperate region. However, those were mainly restricted to Europe and North America; therefore, we have not so much information on the bird-plant relationship in eastern paleoarctic region. Re- cently, some plant ecologists and ornithologists are bearing their interests to this subject in Japan. So, we would like to introduce recent advances in studies of seed dispersal in this feature. We have five articles on the seed dispersal in this issue. These are contributed by 13 Japanese plant ecologists and ornithologists and one Madagascar ornithologist. These five articles include essential and important information on seed dispersal in eastern pleoarctic and tropic Madagascar. Kominami et al. showed the relationships between plant types and frugivorous birds in a primary lucido- phyllous forest in southern Japan. They classified 111 tree species into 17 categories according to their pheno- logical traits. Out of fifteen species of birds recorded in the area, eight are main agents for seed dispersal. A group of summer fruit species provided continuous and familiar food for many bird species and maintains a stability of food resources for birds in the evergreen forest. In their paper Takanose and Kamitani showed a seasonal correspondence between the fruiting phenology of freshy-fruited plants and the abundance of frugivorous birds in deciduous forests in central Japan. In this area, the majority of fruits ripened in the fall when migrating frugivorous birds were most abundant. The Brown- eared Bulbul Hypsipetes amaurotis was an important seed disperser, because of its high frequency of occur- rence throughout the fruiting season, and its large gape size which allowed it to swallow all sizes of fruits found at the area. As suggested by these two studies made in central and southern Japan, habitat differences is important factor to explain different syndrome in bird-plant relationship. Food hoarding is another important process of seed dispersal. In his paper, Hayashida is discussing the seed dispersal of a pine species by the nutcracker. Seed caching is essential process of this syndrome. Nutcrackers store seeds of the Japanese stone pine. All mature cones disappeared by mid-October. Almost all the cones were carried away by nutcrackers. The nutcracker is the most important agent for the seed dispersal of the Japanese stone pine. Hayashida found that pine seedlings were growing in clusters. It means that most seedlings originated from nutcracker caches. It is most likely to give rise to some strong bird-plant reciprocal dependence. It is a good example of co-evolution syndrome between birds and plants. The interaction between fruits and seed dispersers affect diversity of fruit characteristics mediated by the length of retention time in bird’s gut. In her paper, Fukui investigated a seed retention time of 16 fruit species in guts of the Brown-eared Bulbul. She measured some fruit characteristics: fruit size, seed size and water con- tent. The bulbuls defecated large seeds more rapidly than small ones. It suggests that the bulbuls have a mecha- nism to eliminate bulky seeds from their guts rapidly in order to overcome the gut limitation. Large seeds have an advantage of dispersal quantity and small ones have that of dispersal distance contrary to our expectation. Rakotomanana and other Japanese researchers studied the role of the Velvet Asity Philepitta castanea in Madagascar rain forest. Ecology of seed dispersal in this region and the role of this endemic bird species have been left unknown. They found that the seeds transported by the Velvet Asity were less adapted to bird disper- sal. It suggests the presence of dispersal agents other than birds in this area. Community structure of the en- demic fauna and flora of Madagascar must be intensively studied. We could invite many plant ecologists outside our society as authors when we planned to publish this issue. It was the first step for collaboration between ornithologists and a plant ecologists in Japan. I hope that these papers facilitate ornithologists and plant ecologists who have interest in seed dispersal syndrome in eastern Palearctic region. Keisuke UEDA — Corresponding editor Laboratory of Animal Ecology Rikkyo University, Ikebukuro, Tokyo 171-8501, Japan Ornithol. Sci. 2: 3—23 (2003) SPECIAL FEATURE Ecology of seed dispersal Classification of bird-dispersed plants by fruiting phenology, fruit size, and growth form in a primary lucidophyllous forest: an analysis, with implications for the conservation of fruit- bird interactions Yohsuke KOMINAMI'*, Tamotsu SATO!, Keiko TAKESHITA!, Tohru MANABE’, Akira ENDO? and Naohiko NOMA? ' Kyushu Research Center, Forestry and Forest Product Research Institute, Kumamoto 860-0862, Japan ? Kitakyushu Museum and Institute of Natural History, Kitakyushu 805-0061, Japan ? Japan Science and Technology Corporation, Tokyo 102-0081, Japan * School of Environmental Science, The University of Shiga Prefecture, Hikone 522-8533, Japan Abstract To understand the patterns of fruit-bird interactions and to identify species ORNITHOLOGICAL with significant roles that are irreplaceable in these interactions (key species), we SCIENCE classified plant types according to traits relating to frugivory by birds, and analyzed © The Omithologiéal Society the relationships between plant types and frugivorous birds in a primary lucidophy]- of Japan 2003 lous forest in Japan. At the 4-ha study site, 111 plant species were bird-dispersed and 15 common bird species were frugivorous. The growth form of plant species was di- vided into overstory, understory, and liana. The phenological pattern of fruiting was divided into “summer”, “fall”, and “persistent” from the temporal pattern of the seed rain. Fruits were classified in terms of size, as small, a size widely eaten by birds, and large, a size that is difficult for small birds to eat. Seventeen types of plant were iden- tified in the study site, which were classified according to growth form, phenological pattern, and fruit size. Of these fruits, 14 species were considered to be major species, that is species that are both abundant and important for certain birds, and a further 20 species were identified as complementary species, that is species that compensate for a low diversity or for a temporal lack of the major species. Of the birds, eight species were considered major dispersal agents. The patterns of relationship between fruits and birds overlapped in various ways. No strong relationship in which species of fruits and birds are dependent almost entirely on each other were found. An important species set composed of three key species (Eurya japonica, Cleyera japonica, and Cornus controversa) and a group of summer fruits provided continuous and familiar food for many bird species. The patterns of relationship suggest that conservation of the overall composition of fruit types improves the stability of food resources for birds and facilitates dispersal success for the plants themselves. Key words Birds, Forest conservation, Frugivory, Fruit-bird interaction, Lucido- phyllous forest, Seed dispersal Interactions between fruiting plants and frugivo- rous birds (fruit-bird interactions) play an important role in both plant regeneration and food availability for animals. Frugivorous birds contribute to the dis- persal success of plants and improve their chances of regeneration (e.g., Howe 1986; Willson 1992; Her- rera 2002; Terborgh et al. 2002). Fruits are essential foods for frugivorous and omnivorous animals. Fruits (Received 30 July 2002; Accepted 18 October 2002) * Corresponding author, E-mail: kominami @affrc.go.jp also support a winter population of bird species that is mainly insectivorous in summer (Sorensen 1981; Herrera 1984). Therefore, a sudden fluctuation in the interaction between fruiting plants and frugivores has a critical effect both on plant and animal populations (Howe 1977; Hawthorne & Parren 2000; Loiselle & Blake 2002; McConkey & Drake 2002). Fruit-bird interactions are crucial for the conserva- tion of warm-temperate evergreen broad-leaved for- est (lucidophyllous forest), which is the most com- mon forest vegetation in southwest Japan. Over 60% Y. KOMINAML et al. of woody plant species in lucidophyllous forests pro- duce fruits that are eaten and dispersed by birds (Nakanishi 1994; Kominami et al. 1995), indicating that the regeneration of many plant species.depends on frugivory of birds. In recent years, much of the primary lucidophyllous forest in Japan has been lost and most of the remnant forests are small isolated stands (Sato 1983; Hattori & Asami 1998). Some studies in tropical forests have reported that human disturbance causes the impoverishment of frugivores and the disruption of adequate seed dispersal (Kan- nan & James 1999; McConkey & Drake 2002). Simi- lar concerns must be raised in relation to Japan’s forests. To conserve lucidophyllous forests, it is nec- essary to determine the effects of disturbance on the dispersal success of plants and on food availability for birds. Understanding the relationship patterns between fruits and birds, and identifying irreplaceable species that play vital roles in fruit-bird interactions (key species) may help to assess the sensitivity of this in- teraction to habitat disturbance in lucidophyllous for- est. Many obvious relationships between fruiting plants and frugivores have been studied, mainly in tropical forests. For example, only large frugivores can successfully disperse large seeds (Wheelwright 1985; McConkey & Drake 2002; Peres & van Roos- malen 2002). Therefore, interactions between fruit and frugivores in tropical forests are often sensitive to the loss of certain species (Howe 1977; Hawthorne & Parren 2000; Loiselle & Blake 2002; McConkey & Drake 2002; Peres & van Roosmalen 2002). Fruit- bird interactions in temperate forests, however, are more diffuse than in tropical forests (Sorensen 1981; Herrera 1984; Skeate 1987; Malmborg & Willson 1988; Debussche & Isenmann 1989; Jordano 1993; but see Wheelwright 1988). Consequently, it has been suggested that component species of the interaction are interchangeable (Noma & Yumoto 1997). We therefore attempted to classify fruit types for an overview of fruit-bird interactions, and to identify the key species of fruiting plants in lucidophyllous forest. Here, we classify fruit types, using three important traits, and distinguish key species in a primary luci- dophyllous forest. First, it is well known that fruit size restricts seed dispersers, because many frugivo- rous birds prefer fruits that are smaller than their gape width (Herrera 1985; Wheelwright 1985; Jordano 1987; Debussche & Isenmann 1989; Hegde et al. 1991; Fukui 1995). Second, the phenological traits of fruiting are important in defining fruit-bird interac- tions in temperate forests (Thompson & Willson 1979; Stiles 1980; Sorensen 1981; Skeate 1987; Noma & Yumoto 1997). The fruit maturation pattern and the fruiting period of bird-dispersed plants are often adapted to seasonal fluctuations in fruit con- sumption by birds (Herrera 1985; Noma & Yumoto 1997). Third, the growth forms of plants are also es- sential in the relationship. Growth forms determine the fruiting layer. Birds often prefer a specific layer within the forest when foraging for fruit (Hoppes 1987; Malmborg & Willson 1988). Thus, these three traits of fruits can be related to consumption by par- ticular groups of birds. Although plant species can be subdivided into a few large groups based on these traits, a classification using all three traits may illus- trate more detailed structure of fruit types. The classification of fruit types makes an overview of fruit-bird interactions possible. We attempted to detect relationship among species of fruits and birds that are almost entirely dependent on each other. The role of component species in such strong relation- ships is not compensated for by other species (Howe 1977; McConkey & Drake 2002). Then we attempted to identify fruit species associating with a wide vari- ety of birds. Such fruit species play a driving role in the fruit-bird interactions of a community (Skeate 1987; Debussche & Isenmann 1989). We considered both component species in strong relationships and abundant species in overall relationships as key species. The list of key species thus obtained may contribute to an assessment of the impact of species loss on fruit-bird interactions. STUDY SITE AND METHODS 1) Study site The study was conducted at the Aya Research Site, a long-term ecological research site in a well-pre- served lucidophyllous forest, covering over 300 ha, in Miyazaki Prefecture, southwestern Japan (32°03'N, 131°12’E). The study site has been described in detail by Tanouchi and Yamamoto (1995), Sato et al. (1999), and Nagamatsu et al. (2002). The annual mean temperature is 14.2°C and the warmth index is 111. The annual precipitation is 3,070 mm. The most abundant soil type is a moderately moist brown forest soil. The dominant canopy species are Distylium racemosum, Persea japonica, P. thunbergii, Quercus acuta, and Q. salicina. The canopy height ranges from 25 to 32m. The subcanopy layer is dominated by Actinodaphne longifolia, Camellia japonica, Cin- Classification of bird-dispersed plants namomum japonicum, Cleyera japonica, Eurya Japonica, and Neolitsea aciculata. A 4-ha permanent plot for long-term ecological research has been estab- lished on a north- to northwest-facing slope at eleva- tions from 380 to 520 m within the study site. 2) Endozoochorous plants Of the 203 species of higher plants found at the study site (Sato et al. 1999), 140 are woody plants, 43 are herbs, and 20 are ferns. We classified the dis- seminule forms of the plant species according to van der Pijl (1982). Seed dispersal by frugivores (endo- zoochory) was considered the disseminule form for fruits with a fleshy portion, such as the pericarp, aril- lus, perianth, or receptacle. Birds also disperse the dry fruits of several species of Euphorbiaceae and Rutaceae, which have a lipid-rich seed coat (Ueda & Fukui 1992). We considered endozoochory to be the disseminule form of 111 plant species in the study site (Appendix 1). The scientific names of plants fol- lowed Ohwi and Kitagawa (1983). 3) Growth form Initially, we classified the growth forms of plant species according to three categories: the overstory (canopy trees, subcanopy trees, and epiphytes), the understory (small-trees, shrubs, and herbs), and the liana (woody climbers and large herbaceous climbers). At the study site, trees over 10m high and epiphytes bear fruit in the canopy or subcanopy lay- ers, whereas smaller trees, shrubs, and herbs bear fruit in the understory. Liana species were classified as a separate category, because the fruiting layer fluc- tuated with the host plant. 4) Seed rain To measure the seed rain at the study site, we used 263 seed traps. A 1.2-ha core plot (100120 m) that included all types of microtopography was estab- lished within the 4-ha plot. We divided the core plot into 120 grids (each 10X10 m), and positioned seed traps at the intersections and centers of every grid. Each trap consisted of a polyester cloth cone sup- ported by a circular fiberglass frame, with an area of 0.58 m’, placed 1 m above the ground. Seeds that fell into the traps were collected monthly from May 1991 to October 1995 (54 months). All seeds larger than 1 mm in diameter were identified and counted. The col- lected seeds were classified as either pulpless seeds or seeds with pulp. We considered seeds with pulp to have fallen because of their weight only, and the pulpless seeds to have been dispersed by birds. Rain- wash and pulp-consuming insects might also have re- moved pulp, but as the polyester cloth of the traps al- lowed good ventilation, the trapped material dried quickly, and neither pulp-consuming insects nor rot- ten fruit were observed in our traps. 5) Fruit phenology The phenological patterns of fruiting and dispersal in the study site were divided into the following three types, based on phenological studies of seed dispersal by birds in temperate forests (Thompson & Willson 1979; Stiles 1980; Skeate 1987; Nakanishi 1991; Noma & Yumoto 1997): “summer” fruit begin to ripen from late spring to summer, and the fruit are often removed quickly; “fall” fruit ripen from late summer to fall, and most fruit are removed in the fall; and “persistent” fruit are borne for a long period, from fall to winter. We used the seed rain data for the largest seed crop year to determine the phenological type of 50 species. Seed rain data for a combined pe- riod of three or four years were used to determine the type for five species for which the seed rain in any one year was less than 15 seeds. We were unable to distinguish the seeds of Persea japonica from P. thunbergii in the seed rain. Seed rain data of Persea in 1992 were used to determine the phenological type of P. japonica, because most Persea seeds with pulp were found to have fallen beneath P. japonica trees, and pulpless seeds were distributed round P. japonica trees in 1992. No suitable data were available for de- termining the type of P thunbergii. Of 55 species with insufficient seed rain data (<15 seeds in five years in total) at the study site, the types of six species were determined using the phenological pat- terns of fruiting and dispersal reported in other luci- dophyllous forest in southwestern Japan (Noma & Yumoto 1997). In all, the phenological types of 62 species (56% of the endozoochorous species in the flora) were classified. 6) Fruit size To measure the fruit size of endozoochorous species in the flora of the study site, we used fruit and seed specimens collected in Japan. The shorter diam- eters of fruits and seeds were measured, because the shorter diameter of an ovoid fruit is more likely to be significant in preventing birds from swallowing the fruit than its longer diameter (Karasawa 1978; Fukui 1995). For Celastraceae and Magnoliaceae fruits, measurements of seeds with an arillus were treated as Y. KOMINAMT et al. “fruit size” while those without the arillus were treated as “seed size”. The size of the receptacle was treated as “fruit size” for Podocarpus macrophyllus (Podocarpaceae). The seed coat of Evodia meliaefolia (Rutaceae) was thin; therefore, we used the same value for “seed size” (seed size without seed coat) and “fruit size” (seed size with seed coat). We meas- ured the fruit size of 57 species (51% of the endozoo- chorous species) and the seed size of 99 species (89%). For the other species, we used the diameters reported in Karasawa (1978), Kitamura and Murata (1981), Satake et al. (1982), Takushi (1983), Satake et al. (1989), Nakanishi (1996), Noma and Yumoto (1997), and Nakayama et al. (2000). We used the fruit and seed sizes of 107 species for our analyses. 7) Fruits eaten by birds We recorded the bird species observed in the core plot during the monthly seed-trap collections from 1991 to 1995. The fruit species eaten by these birds were identified from descriptions of their food habits in the literature (Mizobuchi 1958; Fennell 1965; Oga- sawara 1968; Chiba 1969; Nakamura 1970; Chiba et al. 1972; Kiyosu 1978; Nakagoshi 1982; Moriyama et al. 1985; Nimura 1988; Narita et al. 1989; Naka- nishi 1991; Fukui 1995; Noma & Yumoto 1997). The sizes of fruits and seeds eaten by birds were meas- ured or determined from the literature. The growth forms of all plants were defined as either overstory, understory, or liana. Plant species from cool temper- ate forests were included for analysis of the fruit size and growth form preferences for each bird species. The scientific and English names of birds followed The Ornithological Society of Japan (2000). 8) Plant abundance Tree censuses, systematic sampling by quadrat method, and data from seed traps were used to quan- tify the abundance of each plant species in the study site. All trees and lianas greater than Scm in dbh were identified within the 4-ha plot (4,856 individu- als) in 1993. The number of individuals in the 4-ha study plot was used to quantify the abundance of each tree and liana species. We established 441 quadrats (4m? each) at the every intersections of 10*10-m grid in the 4-ha plot, and recorded shrub and herb species in each quadrat in 1991. The num- ber of quadrats including a focal species was used as an indication of the abundance of the species. Fruit abundance of each species was evaluated from the total number of fruits that fell into 263 traps in 54 6 months. To convert the numbers of seeds in trap data to numbers of fruits, we counted the numbers of seeds per fruit using specimens of 89 species. For 20 species, we used the seed numbers per fruit reported in Nakayama et al. (2000). The number of Persea fruits was combined for P. japonica and P. thunbergii. RESULTS 1) Growth forms of fruit species in the Aya Re- search Site and in the diet of birds The disseminule form of 108 (80%) of the 140 woody plant species in the Aya Research Site in- volved endozoochory (Appendix 1). High propor- tions of shrubs (93%), lianas (84%), and trees (70%) were endozoochorous, while only three (7%) herb species were endozoochorous. Thirty-seven endozoochorous species fruiting in the overstory included 36 species of canopy and sub- canopy trees and one woody epiphyte species (Ap- pendix 2). There were 53 endozoochorous understory species: 23 small trees, 27 shrubs, and three herbs. There were also 21 endozoochorous liana species. Thirty-seven bird species were observed in the 1.2- ha plot from 1991 to 1995. We identified 15 common species of frugivorous birds (Table 1). Endozoo- chorous fruits are included in the food habits of 25 of these bird species. Of the frugivorous birds, eight species (Chinese Bamboo Partridge Bambusicola tho- racica, Jungle Crow Corvus macrorhynchos, Blue- and-white Flycatcher Cyanoptila cyanomelana, White-backed Woodpecker Dendrocopos leucotos, Oriental Turtle Dove Streptopelia orientalis, Brown Thrush Turdus chrysolaus, Dusky Thrush T: nau- manni, White’s Thrush Zoothera dauma) were rare in the plot. These were observed only once or twice dur- ing the five year observation. Long-tailed Tit Aegit- halos caudatus and Grey Bunting Emberiza vari- abilis were common in the plot, but each is reported to eat only one endozoochorous fruit (Rhus javanica and Eurya japonica, respectively) (Kiyosu 1978; Nimura 1988). Non-frugivorous birds observed in the plot were Sparrowhawk Accipiter nisus, Jungle Nightjar Caprimulgus indicus, Horsfield’s Hawk Cuckoo Cuculus fugax, Little Cuckoo C. polio- cephalus, Oriental Cuckoo C. saturatus, Ruddy King- fisher Halcyon coromanda, Ashy Minivet Pericroco- tus divaricatus, Fairy Pitta Pitta brachyura, Nuthatch Sitta europaea, Hodgson’s Hawk-eagle Spizaetus ni- palensis, Black Paradise Flycatcher Terpsiphone atro- caudata and Short-tailed Bush Warbler Urosphena Classification of bird-dispersed plants Table 1. Common frugivorous birds in the study site, and number of fruit species in the diets of the bird, identified from the lit- erature (see Methods). No of fruit species* Species Season* Weight O/Us (8) O U L Total Aya Brown-eared Bulbul R 60-75 50 81 16 147 57 0.6 Copper Pheasant R 745-1348 1] 29 8 48 21 0.4 Pale Thrush FW 67-77 11 16 2 29 17 0.7 Japanese Green Woodpecker R 120-136 8 13 5 26 10 0.6 Jay R 122-147 i 13 3 23 12 0.5 Japanese White-eye R 9-11 7 15 1 23 1] 0.5 Japanese Green Pigeon R 217-300 13 8 ] 22 1] 1.6 Narcissus Flycatcher S,F 14-17 5 9 2 16 7 0.6 Red-flanked Bushrobin W 12-17 0 11 2 13 8 0.0 Masked Grosbeak R 63 6 3 7 11 6 = Japanese Pygmy Woodpecker R 12-21 3) 4 3 10 4 - Great Tit R 14-16 2 3 | 6 2 - Eye-browed Thrush lf 50-117 4 0 5 2 = Bush Warbler F, W 14-18 l 3 ] 5 1 - Varied Tit R 14-20 0 2 ] 3 3 - * R, resident observed in all seasons; S, observed in spring and summer; F, observed in fall; W, observed in winter. > Body weight described in Kiyosu (1978). “ O, overstory; U, understory; L, liana; Aya, species in flora of the study site. ‘ Proportion of number of overstory species to understory calculated for bird species that ate at least ten species of overstory and understory fruits in total. squameiceps. Both overstory and understory fruits were eaten by most of the common frugivorous birds (Table 1). The ratio of the number of overstory to understory species in the diet (O/U) was calculated for nine bird species that ate at least ten species of overstory and under- story fruits in total. For seven bird species, O/U ranged from 0.4 to 0.7, however, the differences be- tween the bird species were not significant (chi- square test, P>0.05). O/U exceeded 1.0 only for the Japanese Green Pigeon Sphenurus sieboldii (1.6). It was significantly different from the O/U of Brown- eared Bulbul Hypsipetes amaurotis (chi-square test, wv =4.20, P=0.040), Japanese White-eye Zosterops japonicus (y7°=3.90, P=0.048), Copper Pheasant Syrmaticus soemmerringii (7°=6.83, P=0.009), and Red-flanked Bushrobin Tarsiger cyanurus (Fisher’s exact probability test, P<0.001), but not significantly different from the O/U of Narcissus Flycatcher Ficedula narcissina, Jay Garrulus glandarius, Japan- ese Green Woodpecker Picus awokera, or Pale Thrush Turdus pallidus (chi-square test, P>0.05). Overstory fruits were not found in the diet of the Red-flanked Bushrobin. The O/U of this species dif- fered significantly from those of the six other bird species (Fisher’s exact probability test, P<0.05), ex- cept Japanese White-eye (P=0.067) and Copper Pheasant (P=0.093). The fruits of lianas were also eaten by every common bird species except Eye- browed Thrush Turdus obscurus. 2) Phenological pattern of fruits and seasonal changes of birds Depending on the phenological patterns of fruiting and dispersal at the study site, the fruiting species were divided into 10 “summer species”, 24 “fall species”, and 22 “persistent species”, according to the temporal pattern of the seed rain. Seed rain was ob- served for six species for one to three months from April to June, and for four species mainly from June to August (Fig. 1). These species were considered to be of the “summer” type. Summer fruits were pro- duced during the breeding season of birds, and seed dispersal by birds coincided with fruit ripening. Seed rain for seven species was observed for one to four months from June to October, and for 17 species for two to seven months from June to December (Fig. 2). These species were considered to be of the “fall” type. Fall fruits were presented during the fall migra- tion season of birds, and most of the fruits either were Y. KOMINAMT et al. Taxillus yadoriki 1993 ” Qa o 0) © Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar —_— Sp) oO ; fa Prunus jamasakura 40 1994 40 |= oan; 20 20 Cc @ 0 0 w Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar — 500 To ® 400 ® Y 60 300 hd Picrasma quassioides Oo 40 1992 200 oD 20 100 ne} = 0 0 = Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar —_ 200 oO Rubus crataegifolius + 100 1992 [e) Fk 0 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Fig. 1. Elaeagnus glabra 1993 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar 150 Actinodaphne longifolia 100 1995 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Sambucus sieboldiana 1994 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Broussonetia kaempferi 1994 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Persea japonica Premna japonica 1992 1995 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Seasonal patterns of seed rain for summer species. Solid triangles indicate total numbers of bird-dis- persed seeds in 263 traps, and open circles indicate total numbers of all seeds (dispersed by birds+fallen under own weight). The year for the data follows the name of each species. eaten or fell before January. In the fall type, the bird- dispersed seed rain peaked in early (August and Sep- tember) or late (October to December, often in No- vember) fall. Seed rain for 22 species was observed for three to eight months, from August until the fol- lowing April (Fig. 3). These species were considered to be of the “persistent” type. Persistent fruits re- mained on the tree from fall into winter (and some- times until the following spring), and the seeds were dispersed by birds mainly during late fall or winter. Of the species for which there were few seed rain data, six were classified as follows: Myrica rubra and Persea thunbergii (summer type), Daphniphyllum tei- jsmannii and Syzygium buxifolium (fall type), and Myrsine seguinii and Vaccinium bracteatum (persist- ent type) (Noma & Yumoto 1997). Ten common frugivorous bird species were ob- served in the study site in all seasons, while five oc- curred in specific seasons (Table 1). The Narcissus Flycatcher was a summer visitor that was observed until November. The Eye-browed Thrush was ob- served only in the late fall migration season (October to December), and the Pale Thrush was observed from late fall through winter (November to March). Although the Bush Warbler Cettia diphone was a res- ident of the forest edge 200m from the study site, it was observed under the closed forest only from fall to winter. The Red-flanked Bushrobin was observed only in winter (December to March). 3) Fruit size Fruits were classified in terms of size as “small” (fruits eaten by birds with a wide range of body sizes), and “large” (fruits that are difficult for small birds to eat). The fruits eaten by seven of the com- mon small frugivorous bird species (i.e., body weight <30g: Bush Warbler, Japanese Pygmy Woodpecker Dendrocopos kizuki, Narcissus Flycatcher, Great Tit Parus major, Varied Tit P. varius, Red-flanked Bushrobin, and Japanese White-eye) (Table 1) were mainly those with diameters <10mm (Fig. 4), al- though one extremely large fruit (fruit diameter 45 mm; but seed 4mm: Akebia quinata) was eaten by three of the small bird species. Seed size was also re- lated to the fruit composition in the food habits of the birds; few of the small birds ate fruits containing seeds that had a shorter diameter exceeding 5mm. The same tendency for fruit and seed sizes was found in the food habits of both the Japanese Green Wood- pecker and the Japanese Green Pigeon (body weight >100 g each). Over 20% of the fruit species eaten by six medium or large bird species, with body weights >50g (Masked Grosbeak Eophona personata, Jay, Brown-eared Bulbul, Copper Pheasant, Pale Thrush, Classification of bird-dispersed plants 2000 Ficus erecta Cleyera japonica 1993 1995 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Podocarpus macrophyllus Actinidia arguta 1992 1993 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Cornus controversa 1993 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Daphniphyllum macropodum 300 > 1992 1000 200 200 500 100 100 0 0 0 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr MayJun Jul Aug Sep Oct NovDec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar 300 > Cinnamomum japonicum 300 Ampelopsis cantoniensis 300 — Parthenocissus tricuspidata 1994 1994 1992 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar ” a © —_ isp) ice) N Cc aa 150 : 150 (oI : : Temstroemia gymnanthera bi cp ite @ Neolitsea aciculata 19 Akebia trifoliata = 100-1992 100 100 | 4994 © ay £7) To ® (ab) 0 2 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar e) _ 60 Actinodaphne lancifolia Cornus brachypoda 60 Symplocos myrtacea ® 1995 1993 1993 r= 40. 40 40 S 20 20 20 = w 0 0 0 io} AprMay Jun Jul AugSep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar = 40 Lonicara hypoglauca Actinidia polygama 301992 68 1994 20 20 20 Chloranthus glaber 10 10 10 0 0 0 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar 20— Stauntonia hexaphylla 20—, Dendropanax trifidus 20—, Elaeocarpus japonicus 92 1994 1992-94 10 10 10 0 0 0 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar 10 10 10 Menispermum dauricum 1992 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Fig. 2. and Eye-browed Thrush), had a fruit size of at least 10mm or a seed size of at least 5mm. The diet of the Brown-eared Bulbul included fruits with a wide range of fruit and seed sizes. Therefore, fruits with a fruit size <10 mm, and a seed size <5 mm, were con- sidered as small, and the other fruits were considered as large. Of the 111 endozoochorous species at the study site, the fruits of 67 species (60%) were classi- Diospyros japonica 1991 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Lonicera japonica 1991-94 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Seasonal patterns of seed rain for fall species. The symbols are the same as in Fig. 1. fied as “small”, those of 40 species (36%) were clas- sified as “large”; and four species were unknown. 4) Classification by all three traits Endozoochorous species occurring in the study site were classified using the three traits related to avian frugivory (Table 2). Of the eighteen possible types (3 phenological patterns 3 growth forms X2 sizes), Y. KOMINAML et al. Idesia polycarpa Eurya japonica 1993 1992 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Ficus sarmentosa var. nipponica 1992 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Zanthoxylum ailanthoides 1992 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Symplocos prunifolia 1991 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Evodia meliaefolia 1994 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Total number of seed fallen in 263 traps 30 Meliosma rigida 1991 20 10 0 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Xylosma congestum 1994 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Neolitsea sericea 1992-94 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Fig. 3. only one of the expected combinations was not found (summer-liana-large). For the overstory type, large differences in the numbers of species and families were not found among phenological patterns or fruit sizes. For the understory type, 68% of the species and 75% of the families produced small fruits. Four fruit types (fall-overstory-small (FOS), fall- summer-overstory-large (SOL), and persistent-understory-small (PUS)) were (FOL), overstory-large Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Celastrus orbiculatus 1993 100 0 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar 80 Maesa japonica 1992 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Ilex integra 1994 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Lasianthus japonicus 1991 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Callicarpa mollis 1994 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar 20 Kadsura japonica 1991-94 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Seasonal patterns of seed rain for persistent species. The symbols are the same as in Fig. 1. 90%). 10 relatively abundant in the study site (Fig. 5). In the overstory, the total number of individuals with dbh >5cm in the 4-ha plot producing FOS, FOL, and SOL fruits was 1104 (51% of all overstory individu- als), 328 (15%), and 650 (30%), respectively. In the understory, 140 individual (84%) small trees pro- duced PUS fruits. PUS fruits of shrub and herb were found in the greatest total number of quadrats (451, Ardisia crenata 1994 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Diospyros morrisiana 1994 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Irex goshiensis 60 1991 40 20 0 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar 60 llex rotunda 1992 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Symplocos theophrastaefolia 1991 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Paederia scandens var. mairei 1994 20 10 0 Fr Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Damnacanthus indicus 1991-94 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Diameter of seed (mm) Classification of bird-dispersed plants 6 Varied Tit 4 2 {0} 0 2 4 6 8 10 12 14 16 2 Great Tit 4 2 e (0) 0 2 4 6 8 10 12 14 16 6 Red-flanked Bushrobin Narcissus Flycatcher o 8 oO N 0 2 4 6 8 10 12 14 16 Japanese Green Pigeon e 0 2 4 6 8 10 12 14 16 Masked Grosbeak 0 2 4 6 8 10 12 14 16 Eye-browed Thrush 10 o MD OF DD 0 2 4 6 8 10 12 14 16 Bush Warbler Japanese White-eye Ar 10 12 14 16 18 20 2236 38 40 42 44 Japanese Pygmy Woodpecker AT ae 10 12 14 16 18 20 2236 38 40 42 44 Japanese Green Woodpecker 10 12 14 16 18 20 2236 38 40 42 44 0 2 4 6 8 10 12 14 16 18 20 22 36 38 40 42 44 Copper Pheasant 0 2 4 6 8 10 12 14 16 18 20 2236 38 40 42 44 Diameter of fruit (mm) Fig. 4. Size distribution of fruits in the diets of common frugivorous birds. Each dot indicates the fruit and seed size of one plant species. Vertical and horizontal lines indicate the threshold between large and small fruits. 1] Y. KOMINAMT et al. Table 2. phenology) related to frugivory by birds. Number of species and families (in parentheses) in the classification by three traits (growth form, fruit size, fruiting Fruiting phenology Growth Fruit form size 2 Uo Summer Fall Persistent Unknown Overstory Large 5( 3) 7( 4) 2( 2) 5( 5) 19 (11) Small 2( 2) 4( 2) 6( 5) 6( 4) 18 (12) Total Tale) 11( 6) 8( 6) 11( 8) 37 (20) Understory Large 1( 1) 3( 3) 3( 2) 8( 5) 15 (10) Small 2; (2) 3( 3) 10( 8) 21 (17) 36 (21) Unknown 0( 0) 0( 0) 0( 0) 2( 2) 2( 2) Total 3( 3) 6( 6) 13 (10) 31 (21) 53 (28) Liana Large 0( 0) 3( 2) 1( 1) 2( 2) 6( 4) Small 1( 1) 5( 3) 2( 2) PG) 13 ( 9) Unknown 1( 1) 1( 1) 0( 0) 0( 0) A) Total 2( 2) 9( 5) 3( 3) ra @e)) 21 (13) All Large 6( 3) 13 ( 9) 6( 5) 15 (12) 40 (22) Small 5( 5) 12( 8) 18 (12) 32 (22) 67 (32) Unknown a) 1( 1) 0( 0) 2( 2) 4( 4) Total 12( 9) 26 (17) 24 (15) 49 (30) 111 (46) Twenty-one plant species were common, and nine of the common species were abundant (Fig. 5). In the overstory, 14 species of five types (SOL, FOS, FOL, persistent-overstory-large (POL), and persistent-over- story-small (POS)) were represented by more than 10 individuals each (Fig. 5a). Of these common species, Cleyera japonica, Persea japonica, P. thunbergii, Cinnamomum japonicum, Actinodaphne longifolia, and Neolitsea aciculata were abundant (>100 indi- viduals). Actinidia arguta, a fall-liana-large type (FLL), and Ficus sarmentosa var. nipponica, a per- sistent-liana-large type (PLL), were common lianas with more than 10 individuals each (Fig. 5b). None of the lianas was abundant. In the understory, Eurya Japonica, a PUS type, was an abundant small tree with more than 100 individuals (Fig, 5c). None of the other small trees were represented by more than 10 individuals each. Two species of PUS shrubs (Maesa japonica and Damnacanthus indicus) were observed in more than 110 quadrats (>25% of all quadrats, Fig. 5d). Lasianthus japonicus, a PUS type, and Ar- disia crenata, a persistent-understory-large (PUL) type, were also common in the understory, and were observed in 43 quadrats (10%) and 39 (9%), respec- tively. The other species of shrubs and herbs were ob- served in fewer than 16 quadrats each (<4%). Many species provided considerable numbers of fruits (Fig. 5). In the overstory, 11 of 14 common species contributed more than 100 fruits in 54- months seed rain (Fig. 5a). Six overstory species with fewer than 10 individuals each also contributed more than 100 fruits each. Four species of liana contributed more than 100 fruits each, although the populations were of fewer than 10 individuals each (Fig. 5b). In the understory, only three species (Ardisia crenata, Eurya japonica, and Ficus erecta) contributed more than 100 fruits (Fig. 5c, d). DISCUSSION 1) Major and complementary species of fruits At the Aya Research Site, plant species provided birds with various fruits of different fruiting phenol- ogy, fruit size, and growth form. To understand the pattern of fruit-bird interactions, we defined major and complementary fruit species and distinguished the plant species. A major species was defined as one that produced abundant fruit that was important for a certain bird species. A complementary species was defined as one producing fruit that compensated for low diversity or for a temporal lack of a major species in a particular season. Classification of bird-dispersed plants No of fruits No of fruits 10 100 1000 2000 1 10 100 1000 2000 No of individuals+1 No of individuals+1 = sooo, © LUNderstory, small tree) 1000 = = : 100 FE No of fruits ii No of fruits 10 ; atte 100 1000 2000 1 10 100 1000 No of individuals+1 No of quadrats+1 = Fig. 5. Abundance of each plant species, classified by traits related to frugivory by birds. The number of indi- viduals with dbh >5cm in the 4-ha plot, and the numbers of fruits in 54-months seed rain are shown for each species of overstory (a), liana (b), and small tree (c). The number of quadrats in which each species was observed, out of the 441 quadrats in the 4-ha plot, and the number of fruits are shown for each species of shrub and herb (d). To indicate species that were not counted in the tree census or in the quadrats, one is added to the number of indi- viduals and that of quadrats. Plant types are shown as follows: shape of symbol indicates phenological type (trian- gle, summer; circle, fall; square, persistent), open symbols indicate small fruits, and solid symbols indicate large fruits. Species names are shown for major (roman letter) and complementary species (italic letter) as follows: AcA, Actinidia arguta; ALa, Actinodaphne lancifolia; ALo, A. longifolia; AmC, Ampelopsis cantoniensis; ArC, Ardisia crenata; CeO, Celastrus orbiculatus; CiJ, Cinnamomum japonicum; ClJ, Cleyera japonica; CoB, Cornus brachypoda; CoC, C. controversa; DaM, Daphniphyllum macropodum; D1, Damnacanthus indicus; DiM, Diospy- ros morrisiana; E\G, Elaeagnus glabra; EuJ, Eurya japonica; FE, Ficus erecta; FS, F sarmentosa vat. nipponica; IdP, Idesia polycarpa; IG, Ilex goshiensis; Ill, I. integra; LJ, Lasianthus japonicus; MaJ, Maesa japonica; MeD, Menispermum dauricum; MR, Meliosma rigida; NA, Neolitsea aciculata; PaT, Parthenocissus tricuspidata; Pe, Persea japonica and P. thunbergii; PoM, Podocarpus macrophyllus; PrJ, Prunus jamasakura; SP, Symplocos pruni- folia; TaY, Taxillus yadoriki; TeG, Ternstroemia gymnanthera; ZA, Zanthoxylum ailanthoides. Data for P japonica and P thunbergii was combined. medium and large bird fall & winter A Pale Thrush B Brown-eared Bulbul Jay (Copper Pheasant) spring - Summer early-fall resident Y. KOMINAMT et al. Bird Cc Japanese Green Japanese White-eye Japanese Green Pigeon Woodpecker Fruit late-fall Cinnamomum japonicum Neolitsea aciculata Daphniphyllum macropodum Podocarpus macrophyllus FOS Cleyera japonica small bird summer & fall winter D Narcissus Flycatcher E Red-flanked Bushrobin winter POL Diospyros morrisiana PUL Ardisia crenata POS Meliosma rigida Symplocos prunifolia PUS Eurya japonica Ternstroemia gymnanthera PUS Lasianthus japonicus Maesa japonica PUS Damnacanthus indicus SOL FOL Actinodaphne longifolia Persea thunbergii = Persea japonica — = Cc FOS Cornus controversa SOL Prunus jamasakura FOS Cornus brachypoda FOS Pal SOS Taxillus yadoriki FLS @ | SLS Elaeagnus glabra Parthenocissus tricuspidata | FLS = Ampelopsis cantoniensis E 2 E fo) FOL Actinodaphne lancifolia 1S) FLL Actinidia arguta FUL Ficus erecta Fig. 6. Menispermum dauricum PLS Celastrus orbiculatus POS Idesia polycarpa Zanthoxylum ailanthoides POS Ilex goshiensis POL Ilex integra PLL Ficus sarmentosa var. nipponica Relationship patterns between fruit and bird species. Letters A to E indicate the ranges of fruits that are considered to be eaten by the birds with the same letter. Major and complementary fruit species are arranged verti- cally, and phenological patterns are arranged horizontally. Plant types are represented by three letters as follows: the first letter indicates phenological type (S, summer; F, fall; P, persistent), the second letter indicates growth form (O, overstory; U, understory; L, liana), and the last letter indicates fruit size (L, large; S, small). For example, SOL indicates summer-overstory-large type. Of the birds, Copper Pheasant (in parenthesis) is a seed predator. Three tree species producing SOL fruits were major species in spring and summer (Fig. 6). Two dominant canopy species (Persea thunbergii and P. japonica) and a subcanopy species (Actinodaphne longifolia) showed abundant population (Fig. Sa). Most frugivorous birds eat mainly insects during the spring and summer, while Brown-eared Bulbul and Japanese Green Pigeon frequently eat fruits in these seasons (Haneda & Kobayashi 1967; Kiyosu 1978; Moriyama et al. 1985). Growth form and fruit sizes of SOL species are available for Brown-eared Bulbul and Japanese Green Pigeon. Both species ate many species of overstory fruits (Table 1). The Brown- eared Bulbul was a generalist with respect to fruit size (Fig. 4). Although 81% of the fruit species in the diet of Japanese Green Pigeon were small ones, this pigeon is known to eat the SOL fruits of Prunus ja- masakura (Kiyosu 1978). P. jamasakura, a deciduous canopy species with a low population density (0.25 individuals ha '), was also considered an important food for birds, because the number of fruits in the 54- months seed rain equalled that of A. longifolia (Fig. 5a), and the bird-dispersed seed rain was observed every year. A summer-overstory-small (SOS) type woody epiphyte Taxillus yadoriki and a summer- liana-small (SLS) type Elaeagnus glabra provided the fruits from April to June that compensated for the lack of SOL fruit in early spring (Fig. 1). P ja- masakura, T. yadoriki and E. glabra were considered complementary species (Fig. 6). Four FOL, two FOS, and one PUS fruits were major species in fall (Fig. 6). Two subcanopy species (Cinnamomum japonicum and Neolitsea aciculata) were abundant FOL species (Fig. 5a). A canopy tree Classification of bird-dispersed plants Daphniphyllum macropodum and a conifer Podocar- pus macrophyllus were common FOL species that contributed large number of fruits in the seed rain (Fig. 5a). A dominant subcanopy tree Cleyera japon- ica was the most abundant FOS species, both in pop- ulation and in seed rain (Fig. 5a). A deciduous tree Cornus controversa was a common FOS species that contributed more than 6,000 fruits in the seed rain (Fig. 5a). A small tree Eurya japonica was an abun- dant PUS species (Fig. 5c). The demand for fruits in the fall is great, because this is the season when many bird species in the temperate zone shift their food habits from animal food to vegetable matter (Sorensen 1981; Herrera 1984). Due to the abun- dance of FOL, FOS, and PUS species in the study site (Fig. 5), fruits are abundant and available for var- ious birds in the fall. FOL fruits are consumed mainly by birds that eat a wide range of fruit sizes, such as Brown-eared Bulbul and Pale Thrush. Bird species with a wide range of body size consume FOS and PUS fruits. Of seven major plant species in fall (Fig. 6), only one FOS species Cornus controversa provided a con- siderable quantity of fruit in early fall (Fig. 2). Large numbers of C. controversa seeds were collected in traps (>500 for all seeds, >200 for bird-dispersed seeds) in every year from 1992 to 1995. C. contro- versa fruits were available in early fall in 1992, 1993, and 1995; however, most were eaten by birds before August in 1994. The fluctuation in the availability of C. controversa fruits indicates that complementary species are necessary to maintain a stable supply of | fruit for birds in early fall. Six species of five types _ were considered as complementary species in early fall (Fig. 6). A subcanopy tree Actinodaphne lancifo- lia and a liana Actinidia arguta were common FOL _ and FLL species fruiting in early fall, respectively (Fig. 2, Fig. 5a, b). A deciduous tree producing FOS fruit Cornus brachypoda, a small tree Ficus erecta (fall-understory-large (FUL) fruit), and two species _ of liana, Parthenocissus tricuspidata and Ampelopsis, | (fall-liana-small (FLS) fruits), contributed more than 100 fruits in the seed rain, although their populations were small (Fig. 5a, b, c). Large numbers of seeds of | these four species were dispersed by birds in early fall (Fig. 2). _ Many persistent species fruit in late fall, and 35 species (17 fall species and 18 persistent species) contributed to a peak in the bird-dispersed seed rain. Of seven major plant species in fall (Fig. 6), seed rain of six species peaked in late fall (Fig. 2). Therefore, the number of major species is relatively large in late fall, as compared to the number in early fall. Coinci- dentally, the number of frugivorous birds in lucido- phyllous forests peaks in late fall (Noma & Yumoto 1997). Complementary fruit species are also avail- able at that time. Eight species of seven types were considered as complementary species in late fall (Fig. 6). A subcanopy FOS species Ternstroemia gymnanthera, a liana PLL species Ficus sarmentosa var. nipponica, and two shrub PUS species (Maesa Japonica and Lasianthus japonicus) were common species fruiting in late fall (Fig. 2, Fig. 3, Fig. Sa, b, d). A subcanopy POL species Ilex integra, a sub- canopy POS species /. goshiensis, a persistent-liana- small (PLS) species Celastrus orbiculatus, and a liana FLS species Menispermum dauricum con- tributed more than 100 fruits in the seed rain, al- though their populations were small (Fig. 5a, b). Bird-dispersed seed rain of these four species peaked in late fall (Fig. 2, Fig. 3). Five species of persistent fruits were major species in winter (Fig. 6). Eurya japonica, a major species in fall, had a second peak of bird-dispersed seed rain in winter (Fig. 3). Two subcanopy POS species (Me- liosma rigida and Symplocos prunifolia), a canopy POL species Diospyros morrisiana, and a shrub PUL species Ardisia crenata were common species which contributed more than 100 fruits in the seed rain (Fig. 5a, d). Persistent fruits that have not been consumed in late-fall are an important food resource for frugiv- orous birds in winter. PUS fruits, in particular, are important for Red-flanked Bushrobin, a small bird that visits in winter and prefers fruits in the under- story. PUL and POL fruits are consumed mainly by medium and large birds, such as Brown-eared Bulbul and Pale Thrush. Seven complementary species of persistent fruits in late fall (Fig. 6) may improve the total fruit availability in winter, as considerable amounts of these fruits remain during this season. In addition, three species were considered complemen- tary species in winter. These were a PUS species Damnacanthus indicus, an abundant shrub (Fig. 5d), and two deciduous POS species (Zanthoxylum ailan- thoides and Idesia polycarpa) both of which con- tributed more than 100 fruits in the seed rain, al- though their populations were small (Fig. 5a). 2) Overview of fruit-bird interactions We classified the relationships of fruits and birds into five patterns according to the composition of major and complementary fruits in the study site (Fig. Y. KOMINAMT et al. 6). Among the resident birds recorded, three species are considered to be seed dispersal agents of most fruit types (B in Fig. 6). The Brown-eared Bulbul is a major dispersal agent in lucidophyllous forests, where it is abundant throughout the seasons (Sako et al. 1971; Kawaji 1988; Eguchi et al. 1989; 1992; Noma & Yumoto 1997) and eats fruits of a wide range of sizes (Fig. 4). Fukui (1995) confirmed the efficiency of the Brown-eared Bulbul as a seed dis- persal agent. In the study site, 57 (51%) of 111 endo- zoochorous species and 14 (82%) of 17 fruit types were found in the Brown-eared Bulbul’s diet. Thus, it appears that most types of fruit are eaten and dis- persed by this species. Both the Japanese Green Pi- geon and the Jay are resident frugivorous birds. These species are possible be important dispersal agents in all seasons, however, the fruit size and growth form preference of the Japanese Green Pigeon may restrict the range of fruits eaten and dispersed by this species compared with the Brown-eared Bulbul. Fruits of 21 plant species in the study site were found in the diet of the Copper Pheasant, giving it the second most diverse fruit diet after the Brown-eared Bulbul (Table 1). The Copper Pheasant is, however, capable of digesting many different kinds of seeds. Ogasawara (1968) found intact seeds in its crop, but not in the gizzard. Thus, the Copper Pheasant is pre- sumably a seed predator rather than a seed disperser, especially, of poorly protected seeds such as those of the Lauraceae. Five other species of birds are also considered as important seed dispersal agents, although the fruits that they eat are limited by growth form, fruiting sea- son, or fruit size. Pale Thrush (Turdidae) is one of the most important dispersal agents in late fall and win- ter. The Turdidae includes many seed-dispersing species found in various vegetation types in the tem- perate zone (Karasawa 1978; Thomson & Willson 1979: Sorensen 1981; Herrera 1984; Skeate 1987; Debussche & Isenmann 1989; Jordano 1993). The Pale Thrush population peaks in November and many individuals overwinter in lucidophyllous forests (Noma & Yumoto 1997). This study showed that the Pale Thrush is a generalist concerning fruit size and growth form (Table 1, Fig. 4). Therefore, the Pale Thrush is thought to disperse seed of most fruit types available in late fall and in winter (A in Fig. 6). Small birds, such as Japanese White-eye, Narcissus Fly- catcher, and Red-flanked Bushrobin, eat small fruits primarily, and the medium-sized Japanese Green Woodpecker also tends to prefer small fruits (C, D, E 16 in Fig. 6). These birds seemed to be important disper- sal agents for small fruits. Since Narcissus Flycatcher is insectivorous in summer (Chiba et al. 1972; Kiyosu 1978) and absent from the study site in winter, this species eats fruits mainly in fall (D in Fig. 6). The Red-flanked Bushrobin, which eats understory small fruits in winter (E in Fig. 6), is the most limited agent among the major bird species contributing seed dis- persal. Therefore, eight species of birds were considered to be major species for seed dispersal in the study site. An overview of the relationships between these birds and various types of fruits revealed five patterns (Fig. 6). The importance of other common frugivo- rous birds in the study site (Table 1) as dispersal agents remains unknown. They may contribute to seed dispersal to some extent. The Masked Grosbeak, however, is a seed predator rather than a dispersal agent, because it is a species of finch (Fringillidae), which are mashers of fruits and crush seeds in their bills (Herrera 1984; Levey 1987). 3) Key species This study described interspecific relation patterns of fruit-bird interactions in the Aya lucidophyllous forest (Fig. 6). We considered an irreplaceable major species to be a key species. Two species of Theaceae (Eurya japonica and Cleyera japonica) were consid- ered to be key fruit species. Of the major fruit species, E. japonica was the only fruit found in all five patterns of fruit-bird relationship (A-E in Fig. 6), and C. japonica was the only fruit found in four pat- terns (A-D). The populations of those two species were the largest, and the amounts of seed rain from these species were the greatest for the persistent and fall types, respectively (Fig. 5). The annual fluctua- tion in fruit production by E. japonica and C. japon- ica 1s small in lucidophyllous forest (Noma 1997). Thus E. japonica and C. japonica have abundant populations, bear considerable numbers of fruit, pro- duce fruit constantly, and are associated with a wide range of birds. No other species has such a complete set of these traits. Therefore, the fruits of E. japonica and C. japonica are considered as important food re- sources for most frugivorous birds. Any reduction in the number of fruiting individuals might seriously re- duce the overall availability of fruits in the study site. There was no key plant species in spring and sum- mer. The SOL species were important food resources for Brown-eared Bulbul and Japanese Green Pigeon. However, it is still not known whether any of the Classification of bird-dispersed plants SOL species is replaceable. The fruiting periods within summer differed between two species of Persea and the other two species (Actinodaphne longifolia and Prunus jamasakura). The fruit produc- tion by these species showed considerable annual fluctuation and was not synchronized among species (Kominami unpubl.). Therefore, the SOL species pro- vide fruits in turn. In spring and summer, each SOL species might be replaceable in the short term, but the complete set of these species might be necessary in the long term. Cornus controversa was considered a key species because its fruit was the only major species in early fall. It produced the largest amount of fruits in early fall, and these were constantly available for birds. It was also included in the food of at least six species of common frugivorous birds. Therefore, C. controversa fruit is expected to be a popular food resource for many birds in early fall. We did not find any key fruit species as component species in a strong relationship. Fruit-bird interac- tions in the study site overlapped in various ways (Fig. 6). Hence, there was no strong relationship in- volving birds and fruits that were independent of all other relationships as one might expect in a tropical forest area. In the diffuse pattern of relationships seen at the study site, the abundance and generality of key species may improve the stability of food resources and of dispersal success. Some studies of temperate forests have found that a few abundant species are “key species”, playing crucial roles in fruit-bird inter- actions (Skeate 1987; Debussche & Isenmann 1989). In our study site, an important species set composed of three key fruit species and a group of summer fruits provided continuous and familiar food for many bird species. The key fruit species in this study site are also common species in other lucidophyllous forests (Noma & Yumoto 1997: Hattori & Mi- namiyama 2001), and may play similar roles there. 4) Conservation of fruit-bird interactions How do the details of the relationship patterns de- tected in this study contribute to the conservation of fruit-bird interactions in lucidophyllous forests? To | diagnose the stability of fruit-bird interactions, it is _ useful to determine whether the composition of fruit types in a forest is similar to that in the study site. Observed relationship patterns in a primary lucido- phyllous forest represent one fundamental pattern for lucidophyllous forests in Japan. Since there are vari- ous lucidophyllous forest communities (Sato 1983), the species composition of each relationship, e.g., A-E in Fig. 6, is likely to differ somewhat among communities. It is important to detect differences in the overall composition of fruit types rather than in species composition. If a fruit type is absent from a lucidophyllous forest, the cause of this absence should be determined. If the absence of a type is caused by natural variation in the vegetation, man- agement is not necessary; however, if the absence is caused by human impact, the type should be restored. In particular, a sudden decrease in key species might seriously affect fruit-bird interactions, causing a food scarcity for birds, and preventing adequate seed dis- persal for plants. Therefore, monitoring the popula- tions of key species is essential for conserving fruit- bird interactions in lucidophyllous forests. The results of this study suggest that conservation of the overall composition of fruit types improves the stability of food resources for birds and facilitates dispersal success for the plants themselves. The fru- givorous bird population and the degree of frugivory in temperate forests undergo large annual fluctua- tions, especially in fall and winter (Stapanian 1982; Malmborg & Willson 1988; Jordano 1993; Herrera 1998). Fruit abundance also varies markedly among years (Herrera 1984, 1998; Wheelwright 1986; Jor- dano 1993; Noma 1997). These fluctuations in birds and fruits may lead to an annual, unpredictable vari- ability in the available species of the fruit-bird inter- action. Insofar as endozoochorous plants are rich in fruit types, some of the types may compensate for any sudden fluctuation in the key or major species. ACKNOWLEDGMENTS The fieldwork was possible due to the long-term eco- logical research collaboration between Drs. Kaoru Ni- iyama, Satoshi Saito, Hiroyuki Tanouchi and Shin-Ichi Yamamoto. The Kyushu Regional Forestry Office gave approval and provided facilities for the research in Aya lucidophyllous forest. The Forestry and Forest Product Research Institute supported the study overall. The study was funded in part by a Grant-in-Aid for the Bio Cosmos Project from the Ministry of Agriculture, Forestry and Fisheries, Japan. Data analysis was sup- ported by the Database Development Program from Japan Science and Technology Corporation. We are grateful to all of these organizations. We also thank Dr. Shin-Ichi Seki and two anonymous reviewers for their helpful comments on an earlier draft. Special thanks are Y. KOMINAMT et al. due to Dr. Keisuke Ueda for inviting us to prepare this paper. REFERENCES Chiba S (1969) Stomach analysis of Japanese wood- peckers. J Yamashina Inst Ornithol 5: 487-505 (in Japanese with English summary). Chiba S, Asahi M, Matsumoto F, Nishihiro S, Asanaka K & Sakagami E (1972) Stomach analysis of Japan- ese Brambling, Shrike, Thrush, Swallow, Skylark, Flycatcher, Warbler, Wagtail, Nuthatch, Tree-creeper and Tit. Sci Rept Takao Mus Nat Hist 4: 43-77 (in Japanese with English summary). Debussche M & Isenmann P (1989) Fleshy fruit charac- ters and the choices of bird and mammal seed dis- persers in a Mediterranean region. Oikos 56: 327-338. Eguchi K, Takeishi M, Nagata H, Henmi Y & Kawayi N (1989) Altitudinal distribution of forest birds in the Yaku-shima Island, Kagoshima prefecture, Japan. I. Breeding season. Jpn J Ecol 39: 53-65 (in Japanese with English summary). Eguchi K, Takeishi M, Nagata H & Henmi Y (1992) Al- titudinal distribution of forest birds in the Yaku-shima Island, Kagoshima prefecture, Japan. II]. Non-breed- ing season. Jpn J Ecol 42: 107-113 (in Japanese with English summary). Fennell GN (1965) Stomach analyses of Korean birds. J Yamashina Inst Ornithol 4: 24-35. Fukui A (1995) The role of the Brown-Eared Bulbul Hypsipetes amaurotis as a seed dispersal agent. Res Popul Ecol. 37: 211-218. Haneda K & Kobayashi T (1967) Life history of the Brown-eared Bulbul. 1. Breeding biology. J Ya- mashina Inst Ornithol 5: 61—71 (in Japanese with English summary). Hattori T & Asami K (1998) Shouyoujurin no shizen hogo (Conservation of lucidophyllous forest). In: Nu- mata M (ed) Shizen hogo handbook. pp 371-382. Asakura Shoten, Tokyo (in Japanese). Hattori T & Minamiyama N (2001) Flora of the lucido- phyllous forest in Kyushu and Northward in Japan. Hum Nat 12: 91—104 (in Japanese with English sum- mary). Hawthorne WD & Parren MPE (2000) How important are forest elephants to the survival of woody plant species in Upper Guinean forest? J Trop Ecol 16: 133-150. Hegde SG, Ganeshaiah KN & Uma Shaanker R (1991) Fruit preference criteria by avian frugivores: their im- plications for the evolution of clutch size in Solanum pubescens. Oikos 60: 20-26. Herrera CM (1984) A study of avian frugivores, bird- dispersed plants and their interaction in Mediter- ranean scrublands. Ecol Monogr 54: 1-23. Herrera CM (1985) Habitat-consumer interactions in frugivorous birds. In: Cody ML (ed) Habitat selection in birds. pp 341-365. Academic Press, Orlando. Herrera CM (1998) Long-term dynamics of Mediter- ranean frugivorous birds and fleshy fruits: a 12-year study. Ecol Monogr 68: 511-538. Herrera CM (2002) Seed dispersal by vertebrates. In: Herrera CM & Pellmyr O (eds) Plant-animal interac- tions An evolutionary approach. pp 185-208. Black- well Science, Oxford. Hoppes WG (1987) Pre- and post-foraging movements of frugivorous birds in an eastern deciduous forest woodland, USA. Oikos 49: 281-290. Howe HF (1977) Bird activity and seed dispersal of a tropical wet forest tree. Ecology 58: 539-550. Howe HF (1986) Seed dispersal by fruit-eating birds and mammals. In: Murray DR (ed) Seed dispersal. pp 123-189. Academic Press, Sydney. Jordano P (1987) Frugivory, external morphology and digestive system in Mediterranean sylviid warblers Sylvia spp. Ibis 129: 175-189. Jordano P (1993) Geographical ecology and variation of plant-seed disperser interactions: southern Spanish ju- nipers and frugivorous thrushes. Vegetatio 170/108: 85-104. Kannan R & James DA (1999) Fruiting phenology and the conservation of the Great Pied Hornbill (Buceros bicornis) in the Western Ghats of southern India. Biotropica 31: 167-177. Karasawa K (1978) Relationships between fruit-eating birds and seed dispersal in urbanized areas. Tori 27: 1-20 (in Japanese with English summary). Kawaji N (1988) Breeding bird communities of the lower montane zone in southern Kyushu. Jpn J Or- nithol 36: 153-158. Kitamura S & Murata G (1981) Coloured illustrations of woody plants of Japan Vol. 1. Hoikusha, Osaka (in Japanese). Kiyosu Y (1978) The birds of Japan Vol. 1-2. Kodan- sha, Tokyo (in Japanese). Kominami Y, Manabe T, Tanouchi H, Sato T & Niiyama K (1995) Fruit type composition of seed rain in Aya evergreen broad-leaved forest. Bull Kyushu Branch Jpn For Soc 48: 111-112 (in Japanese). Levey DJ (1987) Seed size and fruit-handling tech- niques of avian frugivores. Am Nat 129: 471-485. Loiselle BA & Blake JG (2002) Potential consequences of extinction of frugivorous birds for shrubs of a trop- ical wet forest. In: Levey DJ, Silva WR & Galetti M (eds) Seed dispersal and frugivory: ecology, evolution Classification of bird-dispersed plants and conservation. pp 397-406. CABI Publishing, Wallingford. Malmborg PK & Willson MF (1988) Foraging ecology of avian frugivores and some consequences for seed dispersal in an Illinois woodlot. Condor 90: 173-186. McConkey KR & Drake DR (2002) Extinct pigeons and declining bat populations: are large seeds still being dispersed in the tropical pacific? In: Levey DJ, Silva WR & Galetti M (eds) Seed dispersal and frugivory: ecology, evolution and conservation. pp 381-395. CABI Publishing, Wallingford. Mizobuchi H (1958) On the food habit of the Brown- eared Bulbul, Microscelis amaurotis in the island of Shikoku, Japan. Choju-shuhou 16: 229-240 (in Japanese). Moriyama H, Yamaoka K, Shigematsu H, Harada N & Enomoto S (1985) Creation of the green in cities, VI. The minimum size and structure of urban forest is- lands which carry a reproductive pair of Brown-eared Bulbul, Aypsipetes amaurotis amaurotis (TEM- MINCK). Ningen to Kankyou 11: 23-30 (in Japan- ese). Nagamatsu D, Kominami Y, Sato T & Saito S (2002) Population structure of dominant tree species in Aya evergreen broad-leaved forest. Kyushu J For Res 55: 50-53 (in Japanese with English summary). Nakagoshi N (1982) Seed dispersal by birds in the Hi- gashi-senda-cho campus of the Hiroshima University, Hiroshima. Seed Ecology 13: 1-6 (in Japanese). Nakamura T (1970) A study of Paridae community in Japan. II. Ecological separation of feeding sites and foods. J Yamashina Inst Ornithol 6: 141-169 (in Japanese with English summary). Nakanishi H (1991) Annual, monthly and daily varia- tions of avian seed dispersal in an urban area. Hikobia 11: 73-83. Nakanishi H (1994) Dispersal types of the component species of natural forests in Japan. Nagasaki Women’s Junior Coll Ann Rep Stu 18: 1-5 (in Japan- ese with English summary). Nakanishi H (1996) Fruit color and fruit size of bird- disseminated plants in Japan. Vegetatio 123: 207-218. Nakayama S, Inokuchi M & Minamitani T (2000) Seeds of wild plants in Japan. Tohoku University Press, Sendai (in Japanese). Narita T, Hijikata Y & Seki M (1989) Birds which used feed trees in Tama Experimental Station. Trans Mtg Kanto Br Jpn For Soc 41: 149-152 (in Japanese). Nimura K (1988) Ashiu-enshurin no chourui sou no kisetsu henka (Seasonal variations of avifauna in Kyoto University forest in Ashiu). Rep Kyoto Univ For 18: 1-16 (in Japanese). Noma N (1997) Annual fluctuations of sapfruits produc- tion and synchronization within and inter species in a warm temperate forest on Yakushima Island. Tropics 6: 441-449. Noma N & Yumoto T (1997) Fruiting phenology of ani- mal-dispersed plants in response to winter migration of frugivores in a warm temperate forest on Yakushima Island, Japan. Ecol Res 12: 119-129. Ogasawara K (1968) Winter habitats and food habits of the Green and Copper Pheasants. J Yamashina Inst Ornithol 5: 351-362 (in Japanese with English sum- mary). Ohwi J & Kitagawa M (1983) New flora of Japan. Shi- bundo, Tokyo (in Japanese). Peres CA & Roosmalen M van (2002) Primate fru- givory in two species-rich neotropical forests: impli- cations for the demography of large-seeded plants in overhunted areas. In: Levey DJ, Silva WR & Galetti M (eds) Seed dispersal and frugivory: ecology, evolu- tion and conservation. pp 407-421. CABI Publishing, Wallingford. Pijl L van der (1982) Principles of dispersal in higher plants. 3rd ed. Springer-Verlag, Berlin. Sako S, Kusumoto T & Yokoyama E (1971) Avifauna in the Takakuma experimental forest of Kagoshima Uni- versity. Bull Kagoshima Univ For 5: 1-31 (in Japan- ese with English summary). Satake Y, Hara H, Watari S & Tominari T (1989) Wild flowers of Japan. Woody plants. Vol 1—2. Heibonsha, Tokyo (in Japanese). Satake Y, Ohwi J, Kitamura S, Watari S & Tominari T (1982) Wild flowers of Japan. Herbaceous plants (in- cluding dwarf subshrubs). Vol 2. Heibonsha, Tokyo (in Japanese). Sato T, Kominami Y, Saito S, Niiyama K, Manabe T, Tanouchi H, Noma N & Yamamoto S (1999) An in- troduction to the Aya Research Site, a Long-term Ecological Research site, in a warm temperate ever- green broad-leaved forest ecosystem in southwestern Japan: Research topics and design. Bull Kitakyushu Mus Nat Hist. 18: 157-180. Sato T (1983) Temperate broad-leaved evergreen forests of Japan. In: Ovington JD (ed) Temperate broad- leaved evergreen forests. pp 169-189. Elsevier, Ams- terdam. Stapanian MA (1982) Evolution of fruiting strategies among fleshy-fruited plant species of eastern Kansas. Ecology 63: 1422-1431. Skeate ST (1987) Interactions between birds and fruits in a northern Florida hammock community. Ecology 68: 297-309. Sorensen AE (1981) Interactions between birds and fruit in a temperate woodland. Oecologia 50: 242-249. Y. KOMINAMT et al. Stiles WS (1980) Patterns of fruit presentation and seed dispersal in bird-disseminated woody plants in the eastern deciduous forest. Am Nat 116: 670-688. Takushi A (1983) Ki no mi ki no tane (Fruits and seeds of trees). Shinsei-Tosho Shuppan, Naha (in Japanese). Tanouchi H & Yamamoto S (1995) Structure and regen- eration of canopy species in an old-growth evergreen broad-leaved forest in Aya district, southwestern Japan. Vegetatio 117: 51-60. Terborgh JT, Pitman N, Silman M, Schichter H & Nunez V P (2002) Maintenance of tree diversity in tropical forests. In: Levey DJ, Silva WR & Galetti M (eds) Seed dispersal and frugivory: ecology, evolution and conservation. pp 1-17. CABI Publishing, Wallingford. The Ornithological Society of Japan (2000) Check-list of Japanese birds. 6th ed. OSJ, Obihiro. Thompson JN & Willson MF (1979) Evolution of tem- perate fruit/bird interactions: phenological strategies. Evolution 33: 973-982. Ueda K & Fukui N (1992) Fruit food of crows Corvus corone and C. macrorhynchos: analysis of seeds col- lected at the pre-roosting sites. Jpn J Ornithol 40: 67-74 (in Japanese with English summary). Wheelwright NT (1985) Fruit size, gape width, and the diets of fruit-eating birds. Ecology 66: 808-818. Wheelwright NT (1986) A seven-year study of individ- ual variation in fruit production in tropical bird-dis- persed tree species in the family Lauraceae. In: Estrada A & Fleming TH (eds) Frugivores and Seed Dispersal. pp 19-35, Dr W Junk Publishers, Dor- drecht. Wheelwright NT (1988) Fruit-eating birds and bird-dis- persed plants in the tropics and temperate zone. Tree 3: 270-274. Willson MF (1992) The ecology of seed dispersal. In: Fenner M (ed) Seeds The ecology of regeneration in plant communities. pp 61-85. CABI, Wallingford. MY AN AU BA SP Total Number of species 0 11 0 1 0 84 0 11 0 1 0 59 0 0 0 0 0 25 0 0 0 2 0 29 0 4 0 0 0 25 0 0 1 0 0 2 0 15 1 3 0 140 1 36 0 2 0 43 0 0 0 0 20 20 | 51 | 5 20 203 Appendix 1. Disseminule form composition of higher plants in the Aya Research Site. Growth form EN SY EP Tree 59 13 0 Canopy and subcanopy 36 1] 0 Small tree 73} 2 0 Shrub Da 0 0 Liana 21 0 0 Woody epiphyte 1 0 0 All woody plants 108 13 0 Herb 3 0 1 Fern 0 0 0 All plants 111 13 1 EN, endozoochore dispersed by frugivores; SY, synzoochore dispersed by food hoarding; EP, epizoochore carried on the outside of animals; MY, myrmecochore dispersed by ants; AN, anemochore dispersed by wind; AU, autochore dispersed by the plant it- self; BA, barochore dispersed by weight only; SP, spore. 20 Classification of bird-dispersed plants Appendix 2. Traits and abundance of 111 endozoochorous plant species in the Aya Research Site. E : , Growth’ .. 4 Seed® Fruit? Noof! Noof® No of? Pamey, ppecies | form ~~" size size individuals quadrats fruits Role. Actinidiaceae _ Actinidia arguta FE L L 1.3 19 98 C Actinidiaceae Actinidia polygama F L JL. ol 2 5 1 Anacardiaceae Rhus ambigua u IL S 3.6 4 2 Anacardiaceae Rhus javanica u U S 2:5 3.8 3 Anacardiaceae Rhus sylvestris u U S 4.6 Tell 1 Anacardiaceae Rhus trichocarpa u U S) 4.2 Aquifoliaceae = I/ex buergeri u O S Del 6.8 1 1 Aquifoliaceae = [lex integra P O Ib Al velox 6 166 Aquifoliaceae = /ex latifolia u O S 24 6.8 Aquifoliaceae [lex macropoda u O S 1.9 Aquifoliaceae = [lex rotunda P O S ite} 5.6 ! 17 Aquifoliaceae = _//ex goshiensis P O S 1.4 1 259 C Araceae Arisaema sp. u U u Araliaceae Aralia elata u U S 3 4.0 Araliaceae Dendropanax trifidus F U S 1.8 6.2 1 6 Caprifoliaceae Lonicara hypoglauca F L S 2.6 1 15 Caprifoliaceae Lonicera japonica F L S 2.0 1 2 Caprifoliaceae Sambucus sieboldiana S U S LES 3.7 1 29 Caprifoliaceae = Viburnum awabuki u O S Bhi 5.0 1 1 Caprifoliaceae Viburnum dilatatum u U S 3.9 5).5) 2 Celastraceae Celastrus orbiculatus lr L S ral 4.9 3} 138 C Cephalotaxaceae Cephalotaxus harringtonia u U L 9 Chloranthaceae Chloranthus glaber F U S BQ) 12 15 Cornaceae Aucuba japonica u U IL, 96 11.7 1 Cornaceae Cornus brachypoda F O S 3) 43 6 281 Cornaceae Cornus controversa F O S 4.9 14 6835 K Ebenaceae Diospyros japonica F O L 6.2 7 46 Ebenaceae Diospyros morrisiana P O IL 5.1 25 247 M Elaeagnaceae Elaeagnus glabra S Ib S 4.4 79 C Elaeagnaceae Elaeagnus pungens u U S Elaeocarpaceae Elaeocarpus japonicus F U L 3.7) 28 Ericaceae Vaccinium bracteatum p> U S 1.0 4.5 Euphorbiaceae Antidesma japonicum u U S De. 1 10 Euphorbiaceae Daphniphyllum macropodum F O L 5.8 9.8 26 946 +M Euphorbiaceae Daphniphyllum teijsmanni F° O L 3/0) eal 8 Euphorbiaceae Mallotus japonicus u O S 3.2 2 2 Flacourtiaceae /desia polycarpa P O S 1.3 7.9 4 136 M Flacourtiaceae Xylosma congestum P U S 2.6 1 20 Lardizabalaceae Akebia trifoliata F L u 3.9 1 Lardizabalaceae Stauntonia hexaphylla F L IL 4.7 44.2 Lauraceae Actinodaphne lancifolia F O L 5.1 20 68 C Lauraceae Actinodaphne longifolia S O L 7.7 10.8 128 76 M Lauraceae Cinnamomum camphora u O L Sys) 1)! l 8 Lauraceae Cinnamomum japonicum F O L 6.6 7.6 149 645 M Lauraceae Lindera erythrocarpa u O S) 4.5 Sail 4 Lauraceae Neolitsea aciculata F O L 5.8 105 399 M Lauraceae Neolitsea sericea P U L 7.6 6 21 Lauraceae Persea japonica Ss O 4) 9:3 356 DNB tare Lauraceae Persea thunbergii s? O L 94 10.6 165 - M Leguminosae Euchresta japonica u U L 7.1 9.1 15 21 Appendix 2. (Continued) Y. KOMINAMT et al. Family Species Liliaceae Smilax china Loganiaceae Gardneria nutans Loranthaceae Taxillus yadoriki Magnoliaceae -Michelia compressa Meliaceae Melia azedarach Menispermaceae Menispermum dauricum Menispermaceae Stephania japonica Moraceae Broussonetia kaempferi Moraceae Ficus erecta Moraceae Ficus sarmentosa vat. nipponica Moraceae Maclura cochinchinensis vat. gerontogea Moraceae Morus bombycis Myricaceae Myrica rubra Myrsinaceae Ardisia crenata Myrsinaceae Ardisia japonica Myrsinaceae Maesa japonica Myrsinaceae Myrsine seguinii Myrtaceae Syzygium buxifolium Oleaceae Ligustrum japonicum Oleaceae Osmanthus insularis Piperaceae Piper kadzura Podocarpaceae Podocarpus macrophyllus Rhamnaceae Berchemia racemosa Rosaceae Prunus jamasakura Rosaceae Prunus spinulosa Rosaceae Rubus buergeri Rosaceae Rubus crataegifolius Rosaceae Rubus hirsutus Rosaceae Rubus minusculus Rosaceae Rubus palmatus Rosaceae Rubus sieboldii Rubiaceae Damnacanthus indicus Rubiaceae Lasianthus japonicus Rubiaceae Paederia scandens vat. mairei Rubiaceae Randia cochinchinensis Rutaceae Evodia meliaefolia Rutaceae Skimmia japonica Rutaceae Zanthoxylum ailanthoides Sabiaceae Meliosma rigida Schisandraceae Simaroubaceae Symplocaceae Symplocaceae Symplocaceae Symplocaceae Symplocaceae Symplocaceae Kadsura japonica Picrasma quassioides Symplocos glauca Symplocos lancifolia Symplocos lucida Symplocos myrtacea Symplocos prunifolia Symplocos theophrastaefolia Phenology* Sm™mwAemMeEe ME & MMUMUWM~VEFE fF fF EME SHAE Meee yume ££ 8 wm eg es ye re Growth‘ rm Be ES EvOr Oro re a Cn GheKeVeetene, je eugioig! ©) Grice tee terenie| Go ee e¢ eon ooee « Size! Seed® Fruit® eS OA) ee ea aa ica 72) Wa) tek [eh oh (=) et eel tee) te! leh ay Ie! Wa el at Io et ak a le! le! NANNNNATIANURNUNANRANURURAR size 3.7 6.6 7.4 4.6 4.4 1.4 1.1 1.0 1.1 6.4 5.0 4.3 0.5 5.4 32 2.6 9.6 2.8 5.5 ee 1.5 1.2 1.0 0.9 12 1.2 We) 1.6 3.4 2.0 4.3 2.8 4.9 3.5 Shai 6.2 3.4 4.7 3.0 2.8 3.9 No of! No of® No of" size individuals quadrats fruits 8.4 7.6 13.0 14.5 7.6 13.5 6.9 5.0 8.0 3) 2.0 8.4 133) 3) Wh 5.0 2 39 270 131 22 1836 Role! Classification of bird-dispersed plants Appendix 2. (Continued) Growth’ .. 4 Seed® Fruit? Noof! Noof® No of® Baril apes Ppeaelbey form BIZe size size individuals quadrats fruits Role Theaceae Cleyera japonica F O S 2.1 7.3 1069 7706 «3K Theaceae Eurya japonica IP U S 1.2 5.8 139 1115 K Theaceae Ternstroemia gymnanthera F O S 15 41 C Thymelaeaceae Daphne Kiusiana u U S 3.8 6.6 3 1 Ulmaceae Aphananthe aspera u O L WM Op 1 Urticaceae Villebrunea frutescens u U S 3 Verbenaceae Callicarpa japonica u U S 1.5 35 1 Verbenaceae Callicarpa mollis le U S 1.7 6 22 Verbenaceae Clerodendron trichotomum u U S 4.4 6.4 Verbenaceae Premna japonica S U S 2.9 8 28 Vitaceae Ampelopsis brevipedunculata u U S 3.4 5.8 Vitaceae Ampelopsis cantoniensis F L S 2.8 359 C Vitaceae Parthenocissus tricuspidata F L S 3.3 2 537 C Vitaceae Vitis flexuosa u L S 1 * Phenological pattern of fruiting and dispersal: S, summer type; F, fall type; P, persistent type; u, unknown. > The type is considered by the phenological patterns reported in Noma & Yumoto (1997). © O, overstory; U, understory; L, liana. 4 Fruit size class: L, large; S, small; u, unknown. © Mean shorter diameters (mm) of seed and fruit. Only original data are shown. Number of individuals with DBH of >5 cm in the 4-ha plot. A blank indicates no record in the tree census of this study. ® Number of quadrats where the shrub and herb species was observed among 441 quadrats in the 4-ha plot. A blank indicates no record in the quadrats of this study. " Total number of fruits fallen into 263 traps (total cover 153 m?) in 54 months, calculated from the total number of seeds fallen into traps and the number of seeds per fruit. ' The data of Persea japonica and P. thunbergii are combined. } Role in fruit-bird interaction: K, key species; M, major species; C, complementary species. 23 Ornithol. Sci. 2: 25-32 (2003) SPECIAL FEATURE _ Ecology of seed dispersal Fruiting of fleshy-fruited plants and abundance of frugivorous birds: Phenological correspondence in a temperate forest in central Japan Yoichiro TAKANOSE!* and Tomohiko KAMITANI’ ' Graduate School of Science and Technology, Niigata University, Niigata 950-2181, Japan ? Faculty of Agriculture, Niigata University, Niigata 950-2181, Japan Abstract We investigated the seasonal correspondence between the fruiting phenol- ogy of fleshy-fruited plants and the abundance of frugivorous birds, in a temperate forest, in central Japan. The majority of fleshy fruits ripened in the fall when frugivo- rous birds were most abundant. This correspondence occurred earlier than in a warm temperate forest, located in southern Japan; these relationships in East Asia coincide with those of the temperate regions of North America and Europe. We also examined whether the abundance of frugivorous birds led to profitable effects for seed dispersal among fleshy-fruited plants. The Brown-eared Bulbul Hypsipetes amaurotis was con- sidered to be an important seed disperser for many fleshy-fruited plants because of its high frequency of occurrence throughout the fruiting season, and its large gape size, which allowed it to swallow all sizes of fruits found at the study site. Although nu- merous Brown-eared Bulbuls and other frugivorous birds were present in the fall, fruit removal rates in fall-fruiting species were not always higher than in summer- fruiting species. The abundance of frugivorous birds alone could therefore not ade- ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2003 quately explain the concentrated fruiting phenology among fleshy-fruited plants. Key words nology The relationship between the fruiting phenology of fleshy-fruited plants and their dispersal agents has been described at the community level in various re- gions (e.g., Thompson & Willson 1979; Stiles 1980; Sorensen 1981; Herrera 1984; Izhaki & Safriel 1985; Skeate 1987; Loiselle & Blake 1991; Debussche & Isenmann 1992; French 1992; Noma & Yumoto 1997). Fruiting periods that correspond to seasonal variation in the abundance of dispersal agents have been identified as one of the characteristic events of temperate regions. In North America, this correspon- dence has been examined in different climatic re- gions. The fruiting periods of many fleshy-fruited plants in cool temperate forests are generally concen- trated in the fall, when a large number of frugivorous birds migrate south (Thompson & Willson 1979; Stiles 1980; Stapanian 1982). Fruiting periods in a warm southern temperate forest, however, were post- (Received 9 September 2002; Accepted 29 October 2002) * Corresponding author, E-mail: f98t868g @mail.cc.niigata-u.ac.jp 25 Fleshy-fruited plant, Frugivorous bird, Fruit removal rate, Fruiting phe- poned to the winter, when large populations of fru- givorous birds arrived to overwinter (Skeate 1987). Similar relationships have also been reported in Eu- rope (Herrera 1984), and variations in the dates of maximum fruit availability among communities situ- ated along latitudinal gradients coincide with those of maximum abundance of frugivorous birds (Fuentes 1992). In East Asia, Noma and Yumoto (1997) found a close correspondence between the fruiting periods of many fleshy-fruited plants and the fluctuation of frugivorous birds in a warm temperate forest, in southern Japan. This correspondence was similar to previous reports from lower latitudes in North Amer- ica and Europe, however, no detailed comparison with other regions has been conducted in East Asia. We examined the possibility that efficient seed dis- persal may cause the fruiting period of a fleshy- fruited plant to coincide with the fall migration and overwintering of many frugivorous birds. Abundant dispersal agents may cause higher rates of fruit re- moval and constitute, therefore, a selective influence Y. TAKANOSE and T. KAMITANI on the fruiting phenology of fleshy-fruited plants. Contrary to this hypothesis, however, not all species, whose fruits ripen when frugivorous birds are plenti- ful, show evidence of higher fruit removal rates. In fleshy-fruited plants that have low quality fruits, the fruits persist on the trees for a long time and are dis- persed less efficiently (Stiles 1980; Kominami 1987). The objective of our research was to examine the correspondence between the fruiting periods of vari- ous fleshy-fruited plants and seasonal variation in the abundance of frugivorous birds in a temperate forest, in central Japan. In particular, it was thought that a comparison with the findings from a warm temperate forest in southern Japan (Noma & Yumoto 1997) would help to explain the mechanisms of seed disper- sal by frugivorous birds in East Asia from a pheno- logical point of view. We also examined whether fru- givorous birds present during the fruiting periods of fleshy-fruited plants would have a large gape size enough to swallow fruits whole. Frugivorous birds with a small gape can swallow only small fruits, while those with a large one can swallow fruits of various sizes (Wheelwright 1985; Noma & Yumoto 1997). Finally, we tested the hypothesis that fruiting during a period when frugivorous birds are most abundant leads to stable and higher fruit removal. We also discuss profitable fruiting periods for seed dis- persal among fleshy-fruited plants. METHODS 1) Study site The study site was located at the foot of Mt. Kakuda (482 m) in Niigata Prefecture, central Japan (37°46'N, 138°50’E). According to the Niigata Mete- orological Observatory, the mean annual temperature and precipitation are 13.5°C and 1,776mm, respec- tively. The elevation is approximately 100m az.s.l. The major canopy species are Quercus serrata and Pinus densiflora, whereas the shrub layer is com- posed mainly of Eurya japonica, Viburnum dilatatum, and Aucuba japonica var. borealis. Both cool and warm temperate forest species are found at the study site. 2) Frugivorous birds The population and species composition of birds were recorded by the line-census method from May 1995 to April 1997. A census route approximately 1.5km long was established along a path through the study site. We walked along the census route for three hours just after sunrise and recorded all birds ob- served within 25 m of both sides of the line. The cen- sus was usually conducted once a week, but in the fall, when bird abundance fluctuated dramatically, it was conducted three times a week. Due to inclement weather, the frequency of observations was reduced in the winter. The literature (Kiyosu 1978) and direct observations served to establish whether observed bird species were frugivorous (see Appendix 1). 3) Fleshy-fruited plants We monitored the flowering and fruiting phenol- ogy of 51 woody plant species that bore fleshy fruits (see Appendix 2). During the same period that birds were censused, five individual plants of each species were tagged along the census route, and their pheno- logical traits, such as flowering, fruiting, and fruit disappearance, were recorded once a week. The flow- ering and fruiting periods of each species were de- fined in terms of the number of days from the first flowering and fruiting to the disappearance of all flowers and fruits on the tagged plants. From these observations of fruiting phenology, the 51 species were Classified into three fruiting types; summer-, fall-, or spring-fruiting species (see Appendix 2). Fruit removal from the six principal species of each fruiting type was examined for three consecu- tive years from 1996 to 1998. Broussonetia kazinoki (Moraceae), Prunus apetala ssp. pilosa (Rosaceae), and Sambucus racemosa ssp. sieboldiana (Caprifoli- aceae) were summer-fruiting species and Eurya Japonica (Theaceae), Viburnum dilatatum, and V. wrightii (Caprifoliaceae) were fall-fruiting species. Fruit removal rates were examined in the following two ways. In 1996, four to 12 individuals of each species were randomly selected and their fruits were counted once or twice every two weeks under natural conditions. We also selected several individuals from each species and covered their inflorescences with nylon nets to eliminate fruit consumption by birds. Fruits that dropped off naturally were caught within the nets and counted. The difference between the number of fruits that disappeared under natural con- ditions and the numbers on the netted plants was re- garded as the number of fruits removed by birds. In 1997 and 1998, we again selected three to five indi- viduals of the same plant species and set up seed traps underneath them. The number of fruits on the plants and the number of fruits that fell into the seed traps were recorded once or twice every two weeks. The difference between the two fruit counts was con- Phenological Correspondence sidered to be the number of fruits removed by birds. We measured the size of 30 fruits selected ran- domly from the sample species. Gape sizes of frugiv- orous birds were recorded from the literature (Noma & Yumoto 1997). 4) Data analysis Seasonal differences in the mean number of frugiv- orous bird species counted in each census were ana- lyzed using Tamhane’s multiple comparison, and dif- ferences in the mean species numbers were tested by season using Bonferroni’s multiple comparison. The relationship between frugivorous bird abundance and the number of fleshy-fruited plant species with ripe fruits was investigated using Kendall’s rank correla- tion test. For the above analyses, census data from two consecutive years were combined. Fruit removal rates in summer-fruiting species were compared using a t-test to detect any differences between study years, whereas removal rates in fall-fruiting species were tested with a Kruskal-Wallis test. RESULTS 1) Seasonal variation in the abundance of frugivo- rous birds Of the 63 bird species that we observed during the monitoring periods, 24 were frugivorous. Of these frugivorous species, nine were residents while the other 15 were migrants. The population of frugivo- rous birds was stable throughout the summer months (Fig. 1). Brown-eared Bulbuls Hypsipetes amaurotis were observed most frequently, while the second most observed species was the Japanese White-eyes Zosterops japonicus. These two species accounted for >80% of the total population of frugivorous birds recorded during the summer. The population of fru- givorous birds increased rapidly in the early fall. The first peak in numbers occurred in mid-September in 1995, and in early October in 1996, and this peak was due mainly to the increase of Japanese White-eyes (Fig. 2). The second peak in both study years oc- curred around late October, when migratory winter birds such as Dusky Thrushes Turdus naumanni and Pale Thrushes T. pallidus began to reach the study site (Fig. 2). The Brown-eared Bulbul was observed frequently from September to November (Fig. 2). The total population of the Brown-eared Bulbul and _ the Japanese White-eye reached over 70% of the total population of frugivorous birds recorded during the fall. The mean number of frugivorous birds counted during each census was higher in the fall than in the other seasons (Fig. 3). The population of frugivorous birds decreased in December and remained low rim 80 oy oS Number of birds — i=) Nn Number of bird species (=) MJJASONDJFMAMJJASONDJFMA 1995 1996 1997 Month of the year Fig. 1. Seasonal variation in the number of individuals and species of frugivorous birds recorded along a census route be- tween May 1995 and April 1997. Hypsipetes amaurotis Nn Uv = aa) Gey io) o a) S| =) Zz Zosterops japonicus 40 inpera Le aw! \ N\A— IIS pate MJJASONDJFMAMJJASONDJFMA 1995 1996 1997 Month of the year Fig. 2. Seasonal variation in the number of individuals of the principal frugivorous birds recorded along a census route between May 1995 and April 1997. Y. TAKANOSE and T. KAMITANI 40 ”n = 30 : *** kK 5 oH <) 20 = 3 10 Z 0 8 10 3 ‘i eee | Toy I n HS ee 3} o 2 5 iw O Spring Summer Fall Winter Fig. 3. Seasonal differences in the mean number of individ- uals and species of frugivorous birds recorded during each census (+SE). The combined census data for two consecutive years were divided into four categories: Spring (March—May), Summer (June—August), Fall (September—November), Winter (December—February). The number of individuals and species of birds were compared between the categories using Tamhane’s and Bonferroni’s multiple comparison test, respec- tively. *P<0.05; **P<0.01; ***P<0.001. throughout the winter months. One notable exception was a large flock of Dusky and Pale Thrushes ob- served in the study site in late January 1997. In mid April, the number of birds began to increase again. The dominant species were the Brown-eared Bulbul and the Japanese White-eye, which accounted for 78% of the total population of frugivorous birds ob- served during the spring. Seasonal variation in the number of frugivorous bird species showed a similar pattern to that of individuals (Fig. 1). The mean num- ber of species of frugivorous birds counted during each census was higher in the fall than in the other seasons (Fig. 3). 2) Phenology of fleshy-fruited plants The flowering periods of fleshy-fruited plants were concentrated in the period from May to July, and fruiting periods also showed a notable seasonality (Fig. 4). The number of fleshy-fruited plant species with ripe fruits fluctuated slightly during the summer, however, it increased rapidly in early September and reached a peak during late October and mid-Novem- ber (Fig. 4). Thereafter, the number of fruiting species decreased gradually in late November and re- 2 30 Q (a) ‘eS s 20 BS 10 oO a 2 0 2 20 3 (b) ° GS ie = 10 RS ) or E Z, MJJASONDJFMAMJJASONDJFMA 1995 1996 1997 Month of the year Fig. 4. Seasonal variation in the number of fleshy-fruited plant species that bore ripe fruits (a) and flowers (b) between May 1995 and April 1997. mained low from late winter to spring. We found that the number of fruiting species and the number of frugivorous birds were significantly correlated (Kendall’s rank correlation test, T= 0.19, N= 75, P= ().02); numbers of bird species were also correlated to the number of fruiting species (t= 0.39, N=75, P=0.00). Of the 51 plant species investigated, 18 were de- fined as summer-fruiting species. Fruits of the sum- mer-fruiting species ripened soon after flowering, and almost all such fruits disappeared before the fall. Fruits of several species (Prunus apetala ssp. pilosa, Rubus palmatus var. coptophyllus, and R. hirsutus) began to ripen around late May. Fruits of the 32 fall- fruiting species ripened from September to Novem- ber, when frugivorous birds were abundant. More than 60% of the fleshy-fruited plant species in the study site belonged to this fruiting type. Several species maintained intact fruits during midwinter. The only spring-fruiting species was Aucuba japon- ica var. borealis. Although most fleshy-fruited plants produced fruits in the fall or summer, a few fruits of A. japonica var. borealis began to ripen in late No- vember, as the abundance of frugivorous birds de- creased gradually. The majority of the fruits ripened in late March, and almost all fruits had disappeared Phenological Correspondence Table 1. Fruit removal rates in six fleshy-fruited plant species. 1996 1997 1998 P Summer-fruiting species Prunus apetala ssp. pilosa (JSWSSE3) 7 65.0+13.4 NS (N= 10) €N=5) Sambucus racemosa ssp. sieboldiana 81.5+£7.3 NOW == NS (N=6) (Ne 5) Broussonetia kazinoki SDESES 31.1£11.4 NS (N=12) (N=5S) Fall-fruiting species Eurya japonica 75.8+10.6 16.4+10.0 71.4+£8.8 <0.05 (N=10) (N=4) (N=5) Viburnum dilatatum 63.9+6.1 19.2+9.9 20.1+8.5 <0.01 (N=9) (N=4) (NESS) Ve wrightii 49.9+18.0 32.8+3.4 31.4+2.5 NS (N=4) «N=3) (N=5) ' Figures are means +SE. Fruit removal rates of the summer- and fall-fruiting species were compared with a t-test and a Kruskal—Wallis test, respectively. 3 NS=no significant difference (P>0.05); -: =no observations. prior to fruit production of the summer-fruiting species. 3) Fruit removal by frugivorous birds Mean fruit sizes of the summer- and fall-fruiting species were 8.4+2.1 mm (N=11) and 5.9+1.8mm (N=26), respectively. The gape size of the Brown- eared Bulbul, which was the frugivorous bird most often observed in summer and the second most often observed in fall, is 13.3 mm (Noma & Yumoto 1997), i.e. large enough to swallow the fruits of both fruiting types whole. The gape size of the Japanese White- eye, the second most frequently observed frugivore in _ summer, and the most frequent in fall, is 6.1mm ' (Noma & Yumoto 1997), which is only large enough to swallow the smaller fruits of both fruiting types. We found no significant differences in fruit re- | moval rates among summer-fruiting species (B. kazi- noki, P. apetala ssp. pilosa, and S. racemosa ssp. | sieboldiana) between the study years (Table 1). Not | all fruits were removed to the same degree by frugiv- orous birds. P. apetala ssp. pilosa and S. racemosa _ ssp. sieboldiana experienced high fruit removal rates | (65-86%), whereas those of B. kazinoki were low | (31-32%) in both years. In contrast, the fall-fruiting species showed different tendencies over the study years (Table 1). Fruit removal rates in E. japonica and V. dilatatum were significantly different over three consecutive years, and fluctuated widely from 29 16 to 76%. The fruits of V. wrightii were consistently removed in three consecutive years. No conspicuous differences in fruit removal rates were noticed be- tween summer- and fall-fruiting species. DISCUSSION The fruits of more than 60% of fleshy-fruited plants in a temperate forest in central Japan ripened in the fall, when frugivorous birds were most abun- dant. Noma and Yumoto (1997) found a similar cor- respondence during the winter in a warm temperate forest in southern Japan. These findings in latitudi- nally diverse regions, imply a phenological corre- spondence during different seasons in East Asia, a correspondence that is similar to phenomena that have been reported in North America (Thompson & Willson 1979; Skeate 1987) and Europe (Fuentes 1992). We attempted to explain these phenological corre- spondences by focusing on several key seed dis- persers and fleshy-fruited plants. The Brown-eared Bulbul and the Japanese White-eye are probably the main seed dispersers, as they account for 44-86% of the total population of frugivorous birds during each season. The increased population of frugivorous birds recorded in the fall can be attributed to migratory flocks of Brown-eared Bulbuls and Japanese White- eyes that breed further north. In a warm temperate Y. TAKANOSE and T. KAMITANI forest, in southern Japan, the winter expansion of the frugivorous bird population is primarily due to the migration of wintering populations of these two bird species (Noma & Yumoto 1997). This suggests that regardless of latitude, many fleshy-fruited plants may depend on particular bird species, such as the Brown- eared Bulbul and the Japanese White-eye, for seed dispersal. Moreover, the bulbul is a very important dispersal agent qualitatively because its gape is large enough to easily swallow fruits of various sizes. Fukui (1995) found that the Brown-eared Bulbul con- sumed fruits of 53 species from 24 plant families. The Japanese White-eye is probably a major seed dis- perser for plants that bear smaller fruits. Large thrushes, such as the Dusky Thrush and the Pale Thrush, are also important bird species that are usu- ally observed around late October at the study site. Although populations of these thrushes are not as large as those of the Brown-eared Bulbul or the Japanese White-eye, they contribute to seed dispersal due to their timely arrival and large gape size (>10.0mm; Karasawa 1978). Kominami (1987) in- ferred that the Dusky Thrush contributed much to fruit removal in V. dilatatum in a cool temperate for- est, in northern Japan, despite the presence of other dominant species such as the Brown-eared Bulbul. We investigated whether fruit removal was prof- itable in the fall when frugivorous birds were most abundant at our study site. Fruit removal rates from fleshy-fruited European plants that ripen during the migratory seasons of frugivorous birds exceeded 80% (Jordano 1982; Herrera 1984), which might attest to the effect of frugivorous bird abundance on seed dis- persal. High fruit removal rates at our study site were recorded in two summer-fruiting species, P. apetala ssp. pilosa and S. racemosa ssp. sieboldiana, rather than in fall-fruiting species. Fruit removal rates in the fall-fruiting species differed greatly among species and study years. Such instability may be caused by keener competition among plant species for seed dis- persers due to the concentrated fruiting of fleshy- fruited plants in the fall. The peculiar position of A. japonica var. borealis, which was the only spring- fruiting species, is very interesting in our phenologi- cal study. A few frugivorous bird species, such as the Brown-eared Bulbul, were comparatively abundant in the spring, whereas fleshy-fruited plant species with ripe fruit were scarce. The spring fruiting of A. japon- ica var. borealis may therefore be a phenological Strategy to escape interspecific competition. Thus, fruiting in the fall when frugivorous birds 30 are most abundant would not necessarily lead to prof- itable dispersal. Our results imply that the fruits of fleshy-fruited plants are likely to be removed effi- ciently during each fruiting season. A long-term and continuous study (e.g., Herrera 1998) will be required to further explain the phenological correspondence between fleshy-fruited plants and the abundance of frugivorous birds in temperate regions. ACKNOWLEDGMENTS We are grateful to K. Takahashi for his valuable sug- gestions, in relation to our research, and we thank grad- uate students of Niigata University for their helpful as- sistance during the study. This study was partly sup- ported by a grant from the Ministry of Education, Sci- ence and Culture of Japan (no. 14608022). REFERENCES Debussche M & Isenmann P (1992) A mediterranean bird disperser assemblage: composition and phenol- ogy in relation to fruit availability. Rev Ecol (Terre Vie) 47: 411-432. French K (1992) Phenology of fleshy fruits in a wet sclerophyll forest in southeastern Australia: are birds an important influence? Oecologia 90: 366-373. Fuentes M (1992) Latitudinal and elevational variation in fruiting phenology among western European bird- dispersed plants. Ecography 15: 177-183. Fukui AW (1995) The role of the brown-eared bulbul Hypsypetes amaurotis as a seed dispersal agent. Res Popul Ecol 37: 211-218. Herrera CM (1984) A study of avian frugivores, bird- dispersed plants, and their interaction in Mediter- ranean scrublands. Ecol Monogr 54: 1-23. Herrera CM (1998) Long-term dynamics of Mediter- ranean frugivorous birds and fleshy fruits: a 12-year study. Ecol Monogr 68: 511-538. Izhaki I & Safriel UN (1985) Why do fleshy-fruit plants of the mediterranean scrub intercept fall- but not spring-passage of seed-dispersing migratory birds? Oecologia 67: 40-43. Jordano P (1982) Migrant birds are the main seed dis- persers of blackberries in southern Spain. Oikos 38: 183-193. Karasawa K (1978) Relationships between fruit-eating birds and seed dispersal in urbanized areas. Tori 27: 1-20 (in Japanese). Kiyosu Y (1978) The birds of Japan. Kodansha, Tokyo (in Japanese). Kominami Y (1987) Removal of Viburnum dilatatum Phenological Correspondence fruit by avian frugivores. Ecol Rev 21: 101-106. Loiselle BA & Blake JG (1991) Temporal variation in birds and fruits along an elevational gradient in Costa Rica. Ecology 72: 180-193. Noma N & Yumoto T (1997) Fruiting phenology of animal-dispersed plants in response to winter migra- tion of frugivores in a warm temperate forest on Yakushima Island, Japan. Ecol Res 12: 119-129. Skeate ST (1987) Interactions between birds and fruits in a northern Florida hammock community. Ecology 68: 297-309. Sorensen AE (1981) Interactions between birds and fruit in a temperate woodland. Oecologia 50: 242-249. Stapanian MA (1982) Evolution of fruiting strategies among fleshy-fruited plant species of eastern Kansas. Ecology 63: 1422-1431. Stiles EW (1980) Patterns of fruit presentation and seed dispersal in bird-disseminated woody plants in the eastern deciduous forest. Am Nat 116: 670-688. Thompson JN & Willson MF (1979) Evolution of tem- perate fruit/bird interactions: phenological strategies. Evolution 33: 973-982. Wheelwright NT (1985) Fruit size, gape width, and the diets of fruit-eating birds. Ecology 66: 808-818. 31 Appendix 1. Frugivorous bird species observed from May 1995 to April 1997 at the study site. White-bellied Green Pigeon (Sphenurus sieboldii) Japanese Green Woodpecker (Picus awokera) Great Spotted Woodpecker (Dendrocopos major) White-backed Woodpecker (Dendrocopos leucotos ) Japanese Pygmy Woodpecker (Dendrocopos kizuki) Brown-eared Bulbul (Hypsipetes amaurotis ) Red-flanked Bluetail (Zarsiger cyanurus ) Daurian Redstart (Phoenicurus auroreus ) Scaly Thrush (Zoothera dauma) Japanese Thrush (T7urdus cardis) Brown-headed Thrush (Turdus chrysolaus) Pale Thrush (7urdus pallidus) Eyebrowed Thrush (T7urdus obscurus) Dusky Thrush (Turdus naumanni) Narcissus Flycatcher (Ficedula narcissina) Blue-and-white Flycatcher (Cyanoptila cyanomelana) Dark-sided Flycatcher (Muscicapa sibirica) Grey-streaked Flycatcher (Muscicapa griseisticta) Asian Brown Flycatcher (Muscicapa dauurica) Japanese White-eye (Zosterops japonicus) White-cheeked Starling (Sturnus cineraceus) Eurasian Jay (Garrulus glandarius) Carrion Crow (Corvus corone) Large-billed Crow (Corvus macrorhynchos) Y. TAKANOSE and T. KAMITANI Appendix 2. Flowering and fruiting periods of fleshy-fruited plants observed at the study site. 1995-1996 1996-1997 Spence Fruiting . Flowering Fruiting Flowering Fruiting pattern period period period period | Morus australis Apr.—May Jun.—Aug. May Jun.—Jul. S Broussonetia kazinoki May Jun.—Aug. May —Jun. Jul. S Prunus verecunda Apr. Jun. May Jun.—Jul. S P. jamasakura Apr. Jun.—Jul. Apr.—May Jun.—Jul. S P. apetala ssp. pilosa Apr. May Mar.—Apr. May —Jul. S P. grayana May Jul.—Oct. May Aug.—Sep. S Rubus crataegifolius May Jun.—Jul. May —Jun. Jun.—Jul. S R. microphyllus Apr.—May Jun—Jul. May Jun.—Jul. S R. palmatus var. coptophyllus Apr. May —Jul. Apr.—May Jun.—Jul. S R. phoenicolasius Jun. Jul—Aug. Jun. Jul. S R. parvifolius Jun. Jul. Jun. Jul. S R. hirsutus May May —Jun. May —Jun. Jun. S Coriaria japonica Apr.—May Jun.—Jul. May Jul. S Elaeagnus multiflora var. hortensis Apr.—May Jun. May Jun. S Swida controversa May Jul.—Oct. May —Jun. Aug. S Acanthopanax spinosus May Jul.—Aug. Jun. Jul.—Aug. S Sambucus racemosa ssp. sieboldiana Apr. Jun.—Jul. Apr.—May Jun.—Jul. S Viburnum plicatum f. glabrum May Jul. —Aug. May —Jun. Jul.—Aug. S Celtis sinensis var. japonica Apr.—May Oct.—Jan. May Oct.—Dec. F Magnolia salicifolia Sa Sep. Be ice F Neolitsea sericea oF Oct.—Feb. Oct. Oct.—Nov. F Eurya japonica ane Oct.—Jan. Apr. Sep.—Dec. F Sorbus commixta May Oct.—Jan. May Eee F S. alnifolia May Oct.—Feb. rae Sec F Pourthiaea villosa May Oct.—Feb. May —Jun. Nov.—Dec. F Rosa multiflora May —Jun. Oct.—Mar. Jun. Oct.—Jan. F Mallotus japonicus Jun.—Jul. Sep.—Jan. Jul. Sep.—Dec. F Zanthoxylum piperitum May Sep.—Nov. May Sep.—Nov. F Z. schinifolium Aug. Oct.—Jan. Aug. Oct.—Dec. F Z. ailanthoides Jul.—Aug. Oct.—Jan. Aug. Oct.—Dec. F Rhus trichocarpa May Aug.—Aug. Jun. Sep.—Dec. F R. ambigua May Aug.—Jul. May —Jun. he F R. javanica var. roxburghii Aug.—Sep. Oct.—Jul. Aug. Oct.—Feb. F Meliosma myriantha Jun. Oct._Nov. one he F Ilex macropoda pa Sep.—Dec. F I. crenata var. paludosa Jun. Oct.—Dec. Jun. Oct. F Euonymus sieboldianus May —Jun. Oct.—Nov. Jun. ses If E. alatus f. stiatus May Oct.—Jan. May —Jun. Oct.—Nov. F E. oxyphyllus var. magnus Apr.—May Aug.—Oct. May —Jun. Sep. F Vitis flexuosa Jun. Sep.—Dec. Jun. Sep.—Nov. F Kalopanax pictus Jul. Sep.—Oct. bbe bbc Ie Aralia elata pia Oct. Sep. Oct.—Nov. F Vaccinium oldhamii a8 Sep.—Dec. Jun. Sep.—Dec. F Ardisia japonica as Oct.—Mar. a Oct.—Apr. F Ligustrum tschonoskii Jun. Oct.—May Jun.—Jul. Oct.—Jan. ie Clerodendrum trichotomum Aug. Sep.—Nov. Aug. Sep.—Dec. F Callicarpa japonica Jun.—Jul. Sep.—Jan. Jul. Oct.—Dec. F Lonicera japonica Jun. Oct.—Feb. Jun. Oct.—Feb. F Viburnum dilatatum May —Jun. Sep.—Feb. May —Jun. Sep.—Jan. F Ve wrightii May Sep.—Nov. May Sep.—Jan. F Aucuba japonica var. borealis Apr.—May Nov.—May Mar.—May Nov.—May SP S. F and SP represent the summer-, fall- and spring-fruiting species, respectively. 32 Ornithol. Sci. 2: 33-40 (2003) SPECIAL FEATURE _ Ecology of seed dispersal Seed dispersal of Japanese stone pine by the Eurasian Nut- cracker Mitsuhiro HAYASHIDA Faculty of Agriculture, Yamagata University, Tsuruoka 997-8555, Japan Abstract Seed dispersal of Japanese stone pine Pinus pumila by the Eurasian Nut- cracker Nucifraga caryocatactes was studied at Mt. Apoi in Hokkaido, northern Japan. The seed foraging and caching behavior of diurnal birds and mammals was ob- served, and the relative importance of each species for pine seed dispersal was exam- ined. All mature cones disappeared from the pine shrubs by mid-October each year regardless of the cone crop size. Eurasian Nutcrackers, Varied Tit Parus varius, Eurasian Nuthatch Sitta europaea, Red Squirrel Sciurus vulgaris, and Siberian Chip- munk Tamias sibiricus were all potential seed dispersal agents, however, observations revealed that nutcrackers carried 96% of all seeds transported from the pine trees. The nutcracker carried 142 seeds on average (max. 209) in one trip. Nutcrackers mainly carried pine seeds into their mixed coniferous forest, breeding habitat, where stone pines cannot normally become established and cached them in the soil there. Nut- cracker caches averaged 12 seeds with a maximum of 51 seeds. Pine seedlings were found growing in clusters (in groups of two or more trees). The number of seedlings per cluster closely resembled the number of seeds in nutcracker caches. Results sug- gest that most stone pine seedlings originated from nutcracker caches. Eurasian Nut- crackers thus play an important role in the regeneration of Japanese stone pine despite their small number of caches. ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2003 Key words Caching behavior, Nucifraga caryocatactes, Parus varius, Pinus pumila, Scatter-hoard, Sitta europaea Several members of the family Corvidae are known to store seeds and to be important seed disper- sal agents (Turcek & Kelso 1968; Vander Wall & Balda 1981; Tomback & Linhart 1990; Vander Wall 1990). The interaction between Clark’s Nutcracker Nucifraga colombiana and several pine species has been well studied in North America for Pinus albi- caulis (Hutchins & Lanner 1982: Lanner 1982; Tomback 1982), P. flexilis (Lanner & Vander Wall 1980; Tomback & Kramer 1980), and P. edulis (Van- der Wall & Balda 1977). Those studies documented the central role played by Clark’s Nutcracker in the regeneration of the pines. Similar interactions were recognized between the Eurasian Nutcracker N. cary- ocatactes and P. cembra in Europe (Mattes 1982), P. sibirica in Siberia (Kondratov 1953; Reimers 1959), P. koraiensis in East Asia (Hutchins et al. 1996), and P. pumila (Saito 1983a; Hayashida 1994; Kajimoto et (Received 19 July 2002; Accepted 19 November 2002) E-mail: hayasida@tds1.tr.yamagata-u.ac.jp 33 al. 1998) and P. pariviflora (Hayashida 1989a, 1994) in Japan. Japanese stone pine P. pumila occurs from Siberia and Kamchatka south to Japan and Korea (Critchfield & Little 1966). A vegetation zone dominated by stone pine, known as the P. pumila zone, is present above the forest limit on most high mountains in northern Japan (Okitsu & Ito 1989; Yanagimachi & Ohmori 1991). The Japanese stone pine is a dwarf pine, and its thickets can regenerate vegetatively from adventitious roots (Okitsu & Ito 1984; Kajimoto 1992). P. pumila has large wingless seeds, as do P. albi- caulis and P. cembra, which belong to the Cembrae group within the genus Pinus (Mirov 1967). Several authors have reported that P. pumila seeds are cached by Eurasian Nutcrackers and germinate in their caches, as indicated by fragmentary observations of nutcracker behavior, cone and seed morphology, and seedling clumps (Saito 1983a; Hayashida 1994; Kaji- moto et al. 1998). Previous studies of P. pumila dis- M. HAYASHIDA persal have, however, lacked both quantified data of seed dispersal by animals and detailed behavioral ob- servations of the animals dispersing the seeds. The objective of the research described here was to determine the relative importance of the Eurasian Nutcracker and other potential dispersers for Japan- ese Stone Pine seed dispersal. The seed harvesting, transportation, and caching behavior of nutcrackers and other dispersal agents was studied in order to clarify the effects of animal behavior on pine seed re- generation. STUDY AREA AND METHODS 1) Study area The study was conducted on Mt. Apoi (811m above sea level), which is located in south-central Hokkaido, Japan (42°06'N, 143°02’E). Mt. Apoi is composed of ultramafic rocks (dunite, lherzonite, etc.) of the Horoman peridotite complex (Niida 1984), and is covered with coniferous forests. The P. pumila zone extends from elevations of 500 to 800 m. Japanese white pine Pinus parviflora var. pentaphylla forest occupies the rocky slopes and ridges below 500 m, while the greater part of the mountain is cov- ered by mixed coniferous stands dominated by Picea glehnii, P. parviflora var. pentaphylla, and Abies sachalinensis (Hayashida 1989a). The soil (based on ultramafic rock peridotite) and meteorological conditions (frequent fog in summer and strong winds and little snow in winter), support a unique alpine flora despite the exceptionally low alti- tude (Tatewaki 1952). The height of the P. pumila scrub is dependent on the degree of wind exposure and snow depth (Okitsu & Ito 1984). Pine scrub (less than 100cm tall) and alpine meadows predominate on wind-exposed ridges and slopes in the study area. The native seed predators among birds are the Eurasian Nutcracker, Eurasian Jay Garrulus glandar- ius, Eurasian Nuthatch Sitta europaea, and several tits Parus spp. Nutcrackers can be observed in all seasons on Mt. Apoi, except in years when the Japan- ese stone pine and the Japanese white pine produce few cones. Although nesting by nutcrackers has not been confirmed in the area, there is no doubt that they do breed as they are often found in the mixed conifer- ous forest during the breeding season. 2) Cone fate and seed dispersal by vertebrates Eight patches of P. pumila scrub (each patch had a 4 to 8m? crown) were chosen for cone fate observa- 34 tion at elevations between 520 and 550m. It takes two growing seasons for stone pine cones to mature. Immature second-year cones were mapped in each patch in June, and these cones were counted at inter- vals of two to four weeks until the cones were de- pleted. Data were collected from 1985 to 1990. I gathered information on diurnal animal pine-seed harvesting rates, seed transportation, and seed caching behavior using 7X binoculars and a 25 tel- escope from an observation site at an elevation of 550 m. The observations were made within 900m’, in- cluding the eight patches for the cone fate studies. A total of 74.5 hours of observation were logged at the site from 21 August to 14 September 1985 (29.5h), and from 12 August to 11 September 1987 (45h). I usually observed from dawn till noon once or twice a week. Cone harvesting, cone transporting, feeding, seed harvesting, seed transporting, and seed caching were the recognized behaviors of various vertebrates. I recorded the frequency of seed transportation and the number of seeds carried each trip. These data were used to calculate the number of seeds trans- ported per hour and to determine the relative number of seeds transported by each species. Observations of seed storage behavior by birds and mammals at any time during the study periods were recorded. As nutcrackers flew to their caching sites after seed harvesting, it was difficult to follow them to observe caching behavior directly. Therefore I recorded the flight directions of the birds and the places where I lost sight of them, so as to be able to estimate where the caching sites were. 3) Seed caching by Eurasian Nutcrackers Stored seeds are retrieved by nutcrackers and by other animals. Nutcrackers, however, probe into caches by thrusting their bills into the soil or forest litter, thus visible prod holes in the ground remain as signs of nutcrackers having dug up seed caches. Such excavations are easily distinguished from those of ro- dents. Rodent holes are three or more times larger than those made by nutcrackers, moreover, nutcrack- ers usually open the seeds at the recovery site, leav- ing behind a pile of seed coats (Tomback 1980), whereas rodents carry away seeds to consume them elsewhere. I followed such nutcracker holes on the ground and counted the number of seeds around or in caches along a 3.3 km long hiking trail (2.2 km in the mixed coniferous forest and 1.1 km in the P. pumila zone). This survey was conducted eight times in June and July 1986. Seed dispersal by Eurasian Nutcracker 4) Clusters of Japanese stone pine trees The seedlings of bird-dispersed pines have often been found in clusters consisting of several seedlings of the same age (Tomback 1982; Kajimoto et al. 1998). I counted P pumila trees in two 5mxX5m quadrats (Q1, Q2) and in a 10mX10m quadrat (Q3) in the P. pumila zone, in order to acquire information on the clumping frequency and age distribution of this species. Ages were determined by counting nodes and bud scars on trunks. I also recorded the numbers of seedlings and saplings on barren slopes along the hiking trail in the mixed coniferous forest (a 1.2 km line census, 5 m in width). RESULTS 1) Cone fate All second year cones of Japanese stone pine dis- appeared from the pine shrubs by mid-October every year, regardless of the size of the cone crop, and even though the crop size varied from year to year (Fig. 1). The cones began to disappear in early August in 1985, 1987, 1988, and 1990, and decreased rapidly from late August to mid-September. They had begun to disappear by mid-July in 1986 and 1989, and were depleted by late July. The pedicels of mature cones were loose enough to be detached easily from the branch, indicating that they might fall off under their own weight, however, I observed no intact cones falling to the ground beneath pine crowns. These findings indicate that the pine cones were harvested by animals. 2) Seed harvest and transportation by vertebrates I observed twelve bird species and three mammal species in the study area during the study periods. These included the Eurasian Nutcracker, Eurasian Jay, Jungle Crow Corvus macrorhynchos, Oriental Greenfinch Carduelis sinica, Siberian Meadow Bunting Emberiza cioides, Eurasian Nuthatch Sitta europaea, Willow Tit Parus montanus, Coal Tit P. ater, Great Tit P. major, Varied Tit P. varius, Indian Tree Pipit Anthus hodgsoni, Great Spotted Wood- pecker Dendrocopos major, Red Squirrel Sciurus vul- garis, Siberian Chipmunk Tamias_ sibiricus, and Snowshoe Hare Lepus timidus. Six species of these species (Eurasian Nutcracker, Oriental Greenfinch, Eurasian Nuthatch, Varied Tit, Red Squirrel, and Siberian Chipmunk) foraged on the pine seeds. The Oriental Greenfinch only ate pine seeds in situ, whereas the other five species harvested and trans- 35 500 400 300 Number of cones Sep Oct Fig. 1. Annual cone crop size and number of cones remain- ing on Pinus pumila trees 1985-1990. Data were collected from eight patches of P pumila shrubs. ported cones or seeds. The nutcrackers began to forage on the stone pine cones as early as mid-July in 1986 and 1989, and in early August in 1985, 1987, 1988, and 1990. These foraging times coincided with the timing of cone dis- appearance (Fig. 1). During these periods, the cone scales were tight and the seed coats were thin and fragile. Consequently, nutcrackers and other animals were unable to pick up whole seeds and acquired only seed kernels directly from the cones. Some cones turned brown in mid-August and their seeds matured. Nutcrackers harvested the brown cones se- lectively and frequently transported the seeds. Nut- crackers were usually seen in pairs, and sometimes in small flocks of up to six birds. Observations of nutcrackers foraging for cones fol- lowed a general pattern during which a bird first se- lected a cone then detached it by gripping it in the bill. Brown (mature) cones were easily detached from their subtending branches, and the detached cone was then carried in the bill to a perch on a rock, on the bough of a tall tree, or on the ground, where the bird extracted seeds from the cone. Following seed extrac- tion, some seeds were squeezed in the bill until they cracked, then they were eaten. Most seeds derived from cones, however, were stored in the sublingual pouch rather than eaten. Each seed was apparently tested for edibility by rattling between the mandibles. Forty-three (95.6%) of the 45 pine seeds discarded by Nutcrackers were aborted (empty). Forty-two (32.3%) of the 130 seeds left behind on perches by nutcrackers after harvesting cones were aborted seeds. These facts demonstrate that Nutcrackers se- lectively transported only edible seeds. Nutcrackers tip their heads upwards in an easily recognised motion, when pouching seeds for trans- M. HAYASHIDA Table 1. Comparative transport of Pinus pumila seeds by five vertebrates during August and September in 1985 and 1987. Number of Number of Seeds Relative number of transporting seeds per transporting seeds transported per hour* transport** per hour (seeds/10,000) Nutcracker 2.07 142 293.94 9556 Varied Tit 1.54 ] 1.54 50 Nuthatch 0.12 1 0.12 4 Squirrel 0.13 43 S).)) 182 Chipmunk 0.16 40 6.40 208 * Total obserbation periods were 74.5 hours. ** Refer into text. portation. Nutcrackers extracted 9-50 seeds tained during this study. (mean+SD=29.2+10.4, N=39) from each cone, and required 15-78 seconds (mean+SD=35.7+22.5, N=7) to do so. In each session, nutcrackers harvested 2-10 cones (mean+SD=4.95+1.88, N=20). There- fore, nutcrackers carried 67-209 seeds (mean+ SD=142.2+48.2, N=6) in their sublingual pouches when flying to caching sites. Varied Tits and Eurasian Nuthatches visited P. pumila shrubs in small mixed flocks or in pairs. They were unable to open the immature cones themselves and depended primarily on nutcrackers to expose the seeds for them. They were, however, able to extract seeds from more mature cones when seeds were slightly loose. When they successfully harvested a seed, Varied Tits and nuthatches flew 5-30 m to a tree in the coniferous forest. There they either ate or cached the seed and then returned to harvest another. Both Varied Tits and Eurasian Nuthatches transported seeds singly in the bill (Varied Tit N=115; Eurasian Nuthatch N=9). Both species usually stored seeds in tree branches or under bark, but occasionally cached seeds in the soil on the ground. Red Squirrels were common inhabitants of the mixed coniferous forests on Mt. Apoi, and sometimes visited pine shrubs near the forest edges. Squirrels peeled off the cone scales to extract seeds, ate some seeds and then carried other intact cones in their mouths to the coniferous forest (N=10). As cones contained an average of 43 seeds (Hayashida 1994), Red Squirrels carried an average of 43 seeds each trip. Siberian Chipmunks, as well as Red Squirrels, were found near the forest edges. Chipmunks har- vested about 40 seeds at a time in their cheek pouches, and then returned to the forest (N=12). No data on Red Squirrel or chipmunk caching were ob- 36 Of the five vertebrates observed transporting seeds Eurasian Nutcrackers and Varied Tits transported seeds most frequently (see Table 1). The relative number of seeds transported by each vertebrate, was calculated by multiplying the mean number of seeds per trip by the number of trips per hour. These calcu- lations indicated that 96% of all the seeds trans- ported, were carried by Eurasian Nutcrackers (Table by): 3) Seed caching behavior by Eurasian Nutcrackers Eurasian Nutcrackers transported seeds in all di- rections in both 1985 and 1987 (see Fig. 2). Most of the observed transport flights ranged from 100 to 1,000 m, with a few birds transporting seeds farther than 1,000m. Transportation could be divided into two periods. In the first (mainly in August), they flew down into the mixed coniferous forest on all trips. During the second period (in September), the nut- crackers often carried seeds into the P. pumila zone (53% in 1985, 60% in 1987). Thus, 83% (79% in 1985, 86% in 1987) of seed transports by nutcrackers were into the mixed coniferous forests and 17% (21% in 1985, 14% in 1987) were into the P. pumila zone. Nutcrackers stored P. pumila seeds in many differ- ent cache sites. I directly observed and confirmed nutcrackers making two caches consisting of two and twelve pine seeds in soil at depths of 1 to 2.cm. In ad- ditional caches (N=26) that I found near nutcracker excavations, the number of seeds per cache ranged from | to 51 (mean+SD=12.0+10.1). These caches were | to 3cm deep in soil. Seven (27%) of the 26 caches were found in the PR. pumila zone, and 19 (73%) in the mixed coniferous forest. The density of caches in P. pumila zone and the coniferous forest were 6.36 and 8.64 per km, respectively. These densi- Seed dispersal by Eurasian Nutcracker Fig. 2. The direction and frequency of seed transport by Nutcrackers in 1985 and 1987. Open arrows and solid arrows show the transporting directions in the first and second half of the observation periods, respectively. Figures show the fre- quency of transporting seeds in each direction. Small arrows indicate one trip in each direction. The hatched area is the Pinus pumila zone. ties did not differ significantly (y’=0.5, ns). One cache was found 2.8 km from the nearest seed source of P. pumila. Nutcracker not only predated seeds prior to germination, but also took emergent seeds (with a germinated but not expanded cotyledon). Nut- crackers predated 34% of the germinated seeds in 26 caches. 37 Table 2. Frequency of stem clustering in Pinus pumila in three quadrats and the line census. No. of individuals 10) No. of cluster (%) per cluster single =? trees max. mean+SD Ql 29 (55) 24 (45) 16 DS 2e]. Q2 15 (34) 29 (66) i7/ 4.3+3.8 Q3 19 (49) 20 (51) 5 1.8+£1.1 Line 9 (16) 47 (84) 26 SO2e 56) 4) Clusters of Japanese stone pine trees In the three quadrats studied in the P. pumila zone, the overall number of pine trees per cluster ranged from 1 to 17, and 53% of the clusters consisted of groups of two or more trees (Table 2). In the mixed coniferous forest, 84% of the 56 clusters of pine seedlings along the hiking trail were in groups of two or more trees. In particular, all seedlings under five years old were in groups of two or more trees. The mean number of individuals per cluster was 5.6, and the maximum number was 26 seedlings. DISCUSSION 1) Seed dispersal agents of the Japanese stone pine All Japanese stone pine cones in the Mt. Apoi study area were carried away by animals by late au- tumn, even in years with bumper crops. Five species of vertebrates (Eurasian Nutcracker, Varied Tit, Eurasian Nuthatch, Red Squirrel, and Siberian Chip- munk) were found to be potential seed dispersal agents of the Japanese stone pine, and their contribu- tions to seed dispersal were examined. My observations indicated that the Eurasian Nut- cracker is the most frequent harvester of pine cones. Nutcrackers in this study carried up to 209 pine seeds in their sublingual pouches, close to the maximum capacity of 218 seeds recorded in previous studies (Turcek & Kelso 1968). Mattes (1982) observed that Eurasian Nutcrackers carried up to 134 Cembra pine seeds, which were estimated to be 42.4 ml in capac- ity, based on comparison with the similarly sized Pinon pine seed (Vander Wall & Balda 1981). The volume of 209 Japanese stone pine seeds was meas- ured using a graduated cylinder, and found to be 42.1 ml (Hayashida, unpublished data), consistent with the capacity reported by Mattes (1982). The mean number of P. pumila seeds transported M. HAYASHIDA by Eurasian Nutcrackers in their sublingual pouches was 142 in the Mt. Apoi study. Bergmann et al. (2001) reported that nutcrackers in the Russian Far East carried far fewer P. pumila seeds (80 on average) per trip, which represented the harvestable contents of 2.8 cones. In the Mt. Apoi study, however, the mean number of seeds counted is consistent with cal- culations of the number of seeds carried per trip (144.5), obtained by multiplying the average number of seeds extracted from each cone (29.2) by the aver- age number of cones harvested per trip (4.95). Varied Tits and Eurasian Nuthatches often har- vested pine seeds and transported them singly to cache them under tree bark or in the ground. The Var- ied Tit has long been known to cache seeds (Higuchi 1977), storing as many as 63% of cached seeds of Taxus cuspidata on the ground (Sakakibara 1989), and thus playing an important role in forest regenera- tion. Nuthatches have also been observed to cache seeds (Hendricks 1995; Hutchins et al. 1996; Hardling et al. 1997). Based on their occasional ground caching behavior, both species may poten- tially help Japanese stone pine seedlings become es- tablished, however, as they only carry seeds one at a time their roles are probably not as important as that of the Eurasian Nutcracker. Red Squirrels scatter-hoard Korean pine seeds in soil and contribute to the regeneration of Korean pine forest (Miyaki 1987; Hayashida 1989b). In contrast, they store whole cones of Japanese white pine (P. parviflora var. pentaphylla, the seeds of which are smaller than those of the Korean pine) in the ground (Hayashida 1988; 1989a). Although Red Squirrels were not observed caching cones during this study, they are suspected of storing whole Japanese stone pine cones because stone pine cones and seeds are similar in size to those of the white pine. An experi- ment by Saito (1983b) demonstrated that Japanese stone pine seeds in a cone buried in the soil are able to germinate, but are unable to become established as seedlings. These findings indicate that Red Squirrels do not play a role in the effective dispersal of the Japanese stone pine. Siberian Chipmunks usually scatter and larder hoard acorns (Kawamichi 1980). They also harvest stone pine seeds and carry them in their cheek pouches. Hayashida (1989a) observed a chipmunk caching 33 Japanese white pine seeds in the ground at a depth of 3cm, confirming that chipmunks can be seed dispersal agents. The frequency and quantity of seeds that they harvest are insignificant, however, in 38 comparison with the Eurasian Nutcracker. In addition to the diurnal vertebrates, two noctur- nal rodents are also potential seed dispersal agents in the study area. These are both wood mice Apodemus speciosus and A. argenteus, both of which scatter- hoard seeds (Imaizumi 1979; Miyaki & Kikuzawa 1988). Although these two mice are abundant in vari- ous forest types, they are uncommon in Japanese stone pine scrub (Ota 1968), thus nocturnal rodents are unlikely to be significant seed dispersers of P. pumila. In northeast Siberia, the cones of P. pumila are a primary food source of Brown Bears Ursus arctos during the fall (Krechmar 1995). Although no Brown Bears were observed foraging in the Mt. Apoi study area, recent activity was evident. The large amount of fecal material left behind contained several thousand ingested cones but no intact seeds (Hayashida, un- published data), suggesting that the brown bear is a seed predator and not a potential seed dispersal agent. It is clear from various aspects of this study that the Eurasian Nutcracker transports most Japanese stone pine seeds, and that the Japanese stone pine de- pends mostly on nutcrackers for seed dispersal. 2) The behavior of the Eurasian Nutcracker as a seed dispersal agent of P. pumila Eurasian Nutcrackers usually store more than one Japanese stone pine seed in a cache, as they do with the seeds of other stone pine species (Vander Wall & Balda 1977; Hutchins & Lanner 1982; Mattes 1982; Tomback 1982; Hutchins et al. 1996; Bergmann et al. 2001). Though they sometimes cache seeds under tree bark (Hayashida 1989a), most seeds are cached at depths of 1—3.cm in soil, in a microhabitat that is favorable for germination (Saito 1983b). The number of Japanese stone pine seeds in a single cache varied more widely than for other stone pines (Hutchins & Lanner, 1982; Tomback 1982; Hutchins et al. 1996), probably because the seed of P. pumila is smaller than that of other stone pine species. Some experiments suggest that nutcrackers find most of their caches by memory (Balda 1980; Vander Wall 1982; Balda & Kamil 1989; Kamil et al. 1993; Kamil & Jones 1997), however, since germinated but not expanded cotyledon seeds were eaten along with ungerminated seeds from caches, nutcrackers may also find caches by using emergent seedlings as land- marks. If so, then many emergent pine seedlings may be damaged when birds dig up their caches. Though nutcrackers are principle predators of pine seeds, the Seed dispersal by Eurasian Nutcracker results of this study (cf. Table 2) and Kajimoto (2002) suggest that nutcrackers also play an important role in seed regeneration of P. pumila. My observations indicated that nutcrackers trans- ported most Japanese stone pine seeds to a mixed coniferous forest beyond the stone pine habitat. It is difficult for P. pumila seedlings to grow and become established on the dark floors of dense coniferous forests (Kajimoto 1995) even if cached seeds survive and germinate. Nutcrackers usually store seeds in the habitat of each pine species, as with most bird-dis- persed pine species (Lanner & Vander Wall 1980; Hutchins & Lanner 1982; Mattes 1982; Tomback 1982), with the exceptions of P. edulis (Vander Wall & Balda 1977) and P. monophylla (Vander Wall 1988). Stands of these two pine species and those of Japanese stone pine do not provide suitable breeding habitat for nutcrackers, thus Eurasian Nutcrackers breed in coniferous forests (Yamashina 1934; Mattes 1982), and recover cached seeds to carry to their nestlings (Swanberg 1956). Transporting seeds into mixed coniferous forest (breeding habitat) allows for more efficient exploitation of the cached food during the breeding season. Nutcrackers cached seeds in the coniferous forest in August and often cached seeds in the P. pumila zone in September. The reason for this shift in caching area is not clear, but it may be due to lower levels of cache robbing by rodents (e.g. squirrels, chipmunks, mice and voles) in the P pumila zone than is likely to occur in the coniferous forests. Eurasian Nutcrackers disperse small amounts of Japanese stone pine seeds in the P. pumila zone, but their rate of seed retrieval is low and consequently seedlings become established especially during mast crop years. ACKNOWLEDGMENTS I thank D. Tomback for valuable comments on an ear- lier version of the manuscript. I also acknowledge the helpful support of the staff of Samani town office during the field survey. This study was partly supported by Grant-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology, Japan (No. 14608022). REFERENCES Balda RP (1980) Recovery of cached seeds by a captive Nucifraga caryocatactes. Z Tierpsychol 52: 331-346. Balda RP & Kamil AC (1989) A comparative study of 39 cache recovery by three corvid species. Anim Behav 38: 486-495. Bergmann HH, Englander W & Arkhipov VU (2001) Harvesting and caching capacities of thin-billed Nut- crackers in the Russian Far East. J Ornithol 142: 129-143 (in German with English summary). Critchfield WB & Little EL Jr (1966) Geographic distri- bution of the pines of the world. USDA For Serv Misc Publ 991. Hayashida M (1988) The influence of social interactions on the pattern of scatterhoarding in red squirrels. Res Bull Exp For Hokkaido Univ 45: 267-278 (in Japan- ese with English summary). Hayashida M (1989a) Seed dispersal and regeneration patterns of Pinus parviflora var. pentaphylla on Mt. Apoi in Hokkaido. Res Bull Exp For Hokkaido Univ 46: 177-190 (in Japanese with English summary). Hayashida M (1989b) Seed dispersal by red squirrels and subsequent establishment of Korean pine. For Ecol Manage 28: 115-129. Hayashida M (1994) Role of nutcrackers on seed disper- sal and establishment of Pinus pumila and P. penta- phylla. In: Schmidt WC & Holtmeier FK (eds) Pro- ceedings-International Workshop on Subalpine Stone Pines and Their Environment: the Status of Our Knowledge. St. Moritz, Switzerland. USDA For Serv Gen Tech Rep INT-309, 159-162. Hardling R, Kallander H & Nilsson JA (1997) Memory for hoarded food: An aviary study of the European Nuthatch. Condor 99: 526-529. Hendricks P (1995) Ground-caching and covering of food by a Red-breasted Nuthatch. J Field Ornithol 66: 370-372. Higuchi H (1977) Stored nuts Castanopsis cuspidata as a food resource of nestling Varied Tits Parus varius. Tori 26: 9-12. Hutchins HE & Lanner RM (1982) The central role of Clark’s Nutcracker in the dispersal and establishment of whitebark pine. Oecologia 55: 192-201. Hutchins HE, Hutchins SA & Liu BW (1996) The role of birds and mammals in Korean pine (Pinus koraien- sis) regeneration dynamics. Oecologia 107: 120-130. Imaizumi Y (1979) Seed storing behavior of Apodemus speciosus and Apodemus argenteus. Zool Mag 88: 43-49. Kajimoto T (1992) Dynamics and dry matter production of below ground woody organs of Pinus pumila trees growing on the Kiso mountain range in central Japan. Ecol Res 7: 333-339. Kajimoto T (1995) Ecological characteristics of Pinus pumila with special reference to dry matter produc- tion and regeneration. Jpn J Ecol 45: 57—72 (in Japan- ese with English summary). M. HAYASHIDA Kajimoto T (2002) Factors affecting seedling recruit- ment and survivorship of the Japanese subalpine stone pine, Pinus pumila, after seed dispersal by nut- crackers. Ecol Res 17: 481-491. Kajimoto T, Onodera H, Ikeda S, Daimaru H & Seki T (1998) Seedling establishment of subalpine stone pine (Pinus pumila) by nutcracker (Nucifraga) seed dis- persal on Mt. Yumori, northern Japan. Arc Alp Res 30: 408-417. Kamil AC & Jones JE (1997) The seed-storing corvid Clark’s nutcracker learns geometric relationships among landmarks. Nature 390: 276-279. Kamil AC, Balda RP, Olson DJ & Good S (1993) Re- turns to emptied cache sites by Clark’s nutcrackers, Nucifraga columbiana: a puzzle revisited. Anim Behav 45: 241-252. Kawamichi M (1980) Food, food hoarding and seasonal changes of Siberian chipmunks. Jpn J Ecol 30: 211-220. Kondratov AV (1953) On the restoration of the Siberian Cedar-pine in the wild by nest sowing. Agrobiologiia 3: 161—164 (in Russian). Krechmar MA (1995) Geographical aspects of the feed- ing of the brown bear (Ursus arctos L.) in the ex- treme northeast of Siberia. Russ J Ecol 26: 436-443. Lanner RM (1982) Adaptation of whitebark pine for seed dispersal by Clark’s Nutcracker. Can J For Res 12: 391-402. Lanner RM & Vander Wall SB (1980) Dispersal of lim- ber pine seed by Clark’s Nutcracker. J For 78: 637-639. Mattes H (1982) The coadaptive system of the nut- cracker, Nucifraga caryocatactes (L.), and the Swiss stone pine, Pinus cembra L., and their importance to forestry in the upper mountain region. Swiss Federal Institute of Forest Research, 241, Birmersdorf (in German with English summary). Mirov NT (1967) The genus Pinus. Ronald Press, New York. Miyaki M (1987) Seed dispersal of the Korean pine (Pinus koraiensis) by red squirrels (Sciurus vulgaris). Ecol Res 2: 147-157. Niida K (1984) Petrology of the Horoman ultamafic rocks in the Hidaka Metamorphic Belt, Hokkaido, Japan. J Fac Sci Hokkaido Univ Ser IV 21: 197-250. Okitsu S & Ito K (1984) Vegetation dynamics of the Siberian dwarf pine (Pinus pumila Regal) in the Taisetsu mountain range, Hokkaido, Japan. Vegetatio 58: 105-113. Okitsu S & Ito K (1989) Conditions for the develop- ment of the Pinus pumila zone of Hokkaido, northern Japan. Vegetatio 84: 127-132. Ota K (1968) Studies on the ecological distribution of 40 the murid rodents in Hokkaido. Res Bull Exp For Hokkaido Univ 26: 223-295. Reimers NF (1959) Birds of the Cedar-pine forests of south-central Siberia and their role in the life of the cedar-pine. Trudy Biol Inst Sibirsk Otdel Akad Nauk USSR 5: 121—166 (in Russian). Saito S (1983a) Caching of Japanese stone pine seeds by nutcrackers at the Shiretoko Peninsula, Hokkaido. Tori 32: 13—20 (in Japanese with English summary). Saito S (1983b) On relations of the caching by animals on the seed germination of Japanese stone pine, Pinus pumila Regel. Bull Shiretoko Mus 5: 23-40. Sakakibara S (1989) The role of the varied tit, Parus varius T. and S., in the seed dispersal of Japanese yew, Taxus cuspidata S. and Z. J Jpn For Soc 71: 41-49 (in Japanese with English summary). Swanberg PO (1956) Territory in the thick-billed nut- cracker, Nucifraga caryocatactes. Ibis 98: 412-419. Tatewaki M (1952) Apoi dake to kaigan chitai no syoku- sei (The vegetation on Mt. Apoi and the coastal area). Hayashi (8): 54-59 (in Japanese). Tomback DF (1980) How nutcrackers find their seed stores. Condor 82: 10-19. Tomback DF (1982) Dispersal of whitebark pine seeds by Clark’s nutcracker: A mutualism hypothesis. J Anim Ecol 51: 451-467. Tomback DF & Kramer KA (1980) Limber pine seed harvest by Clark’s Nutcracker in the Sierra Nevada: Timing and foraging behavior. Condor 82: 467-468. Tomback DF & Linhart YB (1990) The evolution of bird-dispersed pines. Evol Ecol 4: 185-219. Turcek FJ & Kelso L (1968) Ecological aspects of food transportation and storage in the Corvidae. Commun Behav Biol A 1: 277-297. Vander Wall SB (1982) An experimental study of cache recovery in Clark’s nutcracker. Anim Behav 30: 84-94. Vander Wall SB (1988) Foraging of Clark’s nutcrackers on rapidly changing pine seed. Condor 90: 621-631. Vander Wall SB (1990) Food hoarding in animals. The University of Chicago Press, Chicago. Vander Wall SB & Balda RP (1977) Coadaptations of the Clark’s Nutcracker and the pinon pine for efficient seed harvest and dispersal. Ecol Monogr 47: 89-111. Vander Wall SB & Balda RP (1981) Ecology and evolu- tion of food-storage behavior in conifer-seed caching Corvids. Z Tierpsychol 56: 217-242. Yamashina Y (1934) Nihon no chorui to sono seitai (The birds and their ecology in Japan). Azusa Syobo, Tokyo. Yanagimachi O & Ohmori H (1991) Ecological status of Pinus pumila scrub and the lower boundary of the Japanese alpine zone. Arc Alp Res 23: 424—435. | Ornithol. Sci. 2: 41-48 (2003) SPECIAL FEATURE _ Ecology of seed dispersal Relationship between seed retention time in bird’s gut and fruit characteristics Akiko FUKUI** Graduate School of Environmental Sciences, University of Tsukuba, Tsukuba, Ibaraki 305-8572, Japan ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2003 Abstract Seed retention time (SRT) of 16 fruit species in the guts of the Brown- eared Bulbul (Hypsipetes amaurotis), a major fruit consumer in central Japan, was studied to examine the relationship between SRT and fruit characteristics, i.e. fruit size, seed size, seed weight, and water content. Caged bulbuls were videotaped after feeding on fruits, and the time of defecation of each seed was recorded. Most seeds were always defecated in fecal pellets, with the exception of Aucuba japonica (the largest of the seeds studied), a seed of which was regurgitated on one occasion. Bul- buls defecate large seeds more rapidly than small seeds. The SRT of the last defecated seed, mean SRT, and standard deviation of SRT were significantly negatively corre- lated with seed size, fruit size, and seed weight, while SRT of the first defecated seed and water content were not correlated with any of the fruit characteristics examined. This suggests that Brown-eared Bulbuls are somehow able selectively to eliminate bulky seeds from the gut rapidly in order to overcome digestive limitations. If birds would prefer fruit species with large seeds that they can regurgitate and with short seed retention times in the gut, the results suggest that large seeds have the advantage of quantity of seed dispersed. Small seeds retained in the gut for longer have the ad- vantage of being carried further and thus can achieve greater dispersal distances and more diverse destinations. The evolutionary interaction between fruiting plants and avian seed dispersers, may affect the diversity of fruit characteristics mediated by the length of retention time in a bird’s gut. Key words Frugivorous birds, Gut limitation, Hypsipetes amaurotis, Seed retention time, Seed size Mutual interactions between birds and_ plants species have been emphasized in a range of studies of endozoochory, during which frugivorous birds con- sume fruit and subsequently disperse plant seeds (Snow 1971; Herrera 1985, 1995; Jordano 1995). There are, however, various conflicts between fruit- ing plants and frugivorous birds, for example, fruit pulp production is costly for plants, but provides fru- givores with energy (Sorensen 1984; Fukui 1996). Fruiting plants and frugivorous birds are also in con- flict over seed retention time (SRT), which is defined here as: the time from ingesting a fruit to the time when its seed(s) are eliminated, corresponding to the time spent passing through the bird’s digestive sys- tem. (Received 28 August 2002; Accepted 12 December 2002) * E-mail: aki@fieldnote.com * Present address: WBSJ Center for Wild birds & Nature of the Globe, Minamidaira 2—35-2, Hino, Tokyo, 191-0041, Japan 4] Several studies have suggested that seed dispersal distance is a function of SRT, however, small seeds tend to be dispersed further (Hoppes 1988; Murray et al. 1994), indicating a relationship between seed size and SRT. Small seeds tend to remain in the gut longer than large seeds. Furthermore, long SRT in the avian gut apparently enhances seed germination (Barnea et al. 1991). Seeds ingested by blackbirds Turdus merula usually had a higher germination rate than those ingested by bulbuls Pycnonotus xanthopygos. This differential rate in germination has been ex- plained by longer SRT in blackbirds than in bulbuls. Longer SRT may increase abrasion of the seed coat by the avian digestive system and thus improve ger- mination rate. SRT represents part of the handling cost of food for birds. Ingested seeds also represent a cost to frugivorous birds because the seeds displace gut volume that could otherwise be filled with di- gestible fruit pulp. Furthermore, the seed mass in- A. FUKUI creases the energy demands for locomotion. Thus whereas long SRT is advantageous for fruiting plants, it is disadvantageous for frugivorous birds. There are several ways to overcome gut limita- tions. One is enlargement of gut volume, as seen in large herbivorous mammals (Westby 1974). Gut vol- ume enlargement is a less helpful solution for frugiv- orous birds as they must be highly mobile to visit ephemeral fruiting sources; increased gut volume brings the disadvantage of increased body weight and thus a severe restriction for actively flying birds. An alternative solution is more rapid food processing. To reduce the ingestion cost of bulky seeds, frugiv- orous birds have several mechanisms for processing fruit quickly (Countney & Sallabanks 1992; Levey & Duke 1992). Levey (1987) identified two types of frugivorous birds in terms of their fruit processing mechanisms: mashers, and gulpers. Mashers include the tanagers Thraupidae and the finches Fringillidae, which crush fruits, separating the pulp from the seeds with their mandibles before ingesting only the pulp. Mashers solve the gut limitation problem by not in- gesting seeds, but, as a result, their handling time costs are higher. Gulpers, including the manakins Pipridae, waxwings Bombycillidae, and thrushes Tur- didae, ingest whole fruit, and separate the pulp from the seeds in their digestive system. Gulpers must off- set their gut limitation by rapid elimination of seeds. Gulpers ingest all seeds, regardless of seed size, then eliminate the seeds by either regurgitation or defecation (Levey 1987, 1992). Levey (1987) showed that seed size influenced processing methods and time. Frugivorous birds more frequently regurgitate large seeds (>4 mm) than small seeds (<2 mm). Me- dian regurgitation times for two manakin species Pipra mentalis and Manacus candei were signifi- cantly less than that of defecation (Levey 1992). Re- gurgitated seeds only pass through part of a bird’s di- gestive system, whereas defecated seeds pass through the whole process, and this may explain why regurgi- tation time is usually shorter than defecation time. Size is important in determining how quickly a seed can be voided with fecal matter (Levey 1992). Large seeds had a significantly shorter SRT in the Cedar Waxwing Bombycilla cedrorum than did small seeds. Rapid elimination of large bulky seeds may be one way for gulpers to overcome their gut limitations. Understanding the relationship between seed size and seed retention time is essential when studying in- teractions between frugivorous birds and _ fruiting plants. Seed size is an important factor determining the foraging efficiency of frugivorous birds. Seed size affects SRT, which is important for fruiting plants as it affects survival into the next generation. A number of studies have examined SRT in relation to seed size for various plant species (Sorensen 1984; Johnson et al. 1985; Levey 1987; Levey & Grajal 1991). In this study, I examined fruit size and water con- tent as well as seed size, since these characteristics appear to influence the foraging behaviour, digestion, and SRT of the Brown-eared Bulbul Hypsipetes amaurotis. The Brown-eared Bulbul is one of the major fruit consumers in central Japan, and an effi- cient seed disperser (Fukui 1995). When feeding, bul- buls usually gulp whole fruit, except when larger fruits are close in size to the gape width. Sixteen fruit species with a wide range of fruit properties were used in feeding experiments with Brown-eared Bulbuls, and the relationship between fruit characteristics and SRT were analyzed statisti- cally. The questions raised during this research, and discussed here are as follows: 1) Does the Brown- eared Bulbul regurgitate large seeds more frequently than small seeds?, 2) Does the bulbul eliminate large bulky seeds more rapidly than small seeds?, and 3) What properties of fruits affect seed retention time in the gut? MATERIALS AND METHODS 1) Study area This study was conducted on the University of Tsukuba campus, in southern Ibaraki Prefecture (36°06'N, 140°06’E), from 1989 to 1991. Six species of frugivorous birds were observed in the study area. These were the Brown-eared Bulbul, Gray Starling Sturnus cineraceus, Azure-winged Magpie Cyanopica cyana, Dusky Thrush Turdus naumanni, Pale Thrush 7: pallidus, and Daurian Redstart Phoenicurus auroreus. The Brown-eared Bulbul was the most abundant of the six, and the only one resi- dent in the study area throughout the year. 2) Fruit characteristics The Brown-eared Bulbul is a gulper known to feed on the fruits of 53 plant species from 24 families oc- curring in this study area (Fukui 1995). These fruits vary in diameter from 3.5 mm Callicarpa mollis, Ver- benaceae to 12.0mm Aucuba japonica, Cornaceae, and all were smaller than the maximum width of the bulbul’s gape (17.0+1.7mm [mean+SE]). Sixteen common fruiting plants preferred by wild Brown- Seed retention time in bird’s gut Table 1. Characteristics of fruits examined. Bins ehecies Seed size Fruit size Seed weight Water content (mm) (mm) (mg) (%) Aucuba japonica 16.1 16.5 0.69 74.36 Daphniphyllum macropodum 9.3 12.0 0.23 68.23 Elaeagnus umbellata 6.5 10.0 0.10 68.08 Cinnamomum camphora 6.3 8.5 0.12 56.29 Paederia scandens 6.2 6.2 0.029 65.68 Ligustrum japonicum 5.4 6.5 0.046 66.43 Cornus florida 5.3 11.0 0.091 64.9 Nandina domestica Srl 6.5 0.10 66.98 Viburnum dilatum 5.0 a 0.025 74.15 Tlex crenata 4.8 6.0 0.058 56.54 Celastrus orbiculatus 3.8 7.5 0.023 71.12 Phytolacca americana 3.0 8.0 0.0076 73.29 Ilex serrata 2.9 5.0 0.008 78.57 Pyracantha angustifolia 2.8 5.5 0.025 73.00 Callicarpa mollis 1.9 3.5 0.0031 79.35 Eurya japonica 1.8 6.0 0.0023 70.87 eared Bulbuls were chosen for this study (see Table 1). All 16 species produce fruit in autumn and winter. To minimize intraspecific variation in fruit features, I collected fruits for the experiments from a single mother tree within a five-day period. For each of the 16 fruit species, four characteristics (Table 1) were recorded: fruit size, seed size, seed weight, and water content. Twenty or more fruits of each species were measured. Fruit and seed sizes, were the diameters of the longest axis measured using a pair of vernier callipers. Forty fresh fruits of each species were weighed individually with a Met- tler balance. Seeds were removed from half of the 40 fruits and weighed. Another 20 fruits were dried in an oven at 60°C until their weights became constant. Water content was defined as the difference between | dry weight and fresh fruit weight. 3) Seed retention times Seven wild Brown-eared Bulbuls were captured with mist nets between September 1988 and March 1990. When captured they weighed 78.7+9.6¢ (mean+SE). These seven birds were housed indoors in separate cages (1.0X1.0m’X0.5m high) and maintained on diets of wild fruits, pieces of apple, /commercially-available food for fruit-eating birds, and water. All birds were fed commercially-available food overnight between the days of experiments and they were deprived of food for an hour prior to the onset of the experiments. Each bird was kept for a 43 week then released. All birds remained in good health throughout the duration of the experiments. The SRT for the bulbuls was determined as fol- lows. During experiments, each bulbul was fed on a single species of fruit and provided with plenty of water. From | to 20 fruits were supplied depending on fruit size (Table 1). Ten fruits of each of 16 plant species with intermediate sized fruits were given dur- ing experiments. There were three exceptions. Twenty fruits of Callicarpa mollis were supplied as this species bears very small fruit. Five fruits of the large fruit bearing Daphniphyllum macropodum were given, and just one fruit of the very large fruited Au- cuba japonica was given. The number of experimen- tal replications for each fruit ranged from 3 to 32, be- cause the number of bulbuls available for testing var- ied from time to time. Several experiments were carried out sequentially in a single day. One experiment consisted of supply- ing fruit followed by about two-hours of observation. Different fruit species were fed to a bird in successive experiments so as to be able to distinguish the seeds from each experiment. In each experiment the bul- buls consumed all the fruits within 20seconds of them being provided. Each experiment was video- taped, and a sheet of graph paper on the floor of the cage made it possible to identify exactly the positions of feces, the condition of each fecal pellet, and the number of seeds in each fecal pellet. The time at which seeds were either defecated or regurgitated A. FUKUI was available in seconds from the video record. The SRT of each type of seed was calculated as the period between feeding and elimination. Four indices were used to define the relationship between SRT and fruit characteristics: mean SRT, first SRT, last SRT, and standard deviation of SRT. Mean SRT was defined as the overall mean of the mean retention time for all defecated seeds in an ex- periment using a single fruit species. First SRT was defined as the overall mean SRT for the first defe- cated seed of each experiment. Last SRT was defined as the overall mean SRT for the last defecated seed of each experiment. The standard deviation of SRT was defined as the overall mean of the standard deviations of SRT for all defecated seeds in an experiment. The first three indices represent the SRTs of first, mean, and last defecated seeds of a given species, and thus they provide some indication of the expected disper- sal distance and dispersal range of seeds by wild bul- buls. The standard deviation of SRT indicates the de- gree of variation in seed destination dispersal by birds, which, in addition to seed dispersal distance, is an important parameter of seed dispersal success for plants. 4) Analysis of seed retention time and fruit charac- teristics Relationships between each index of seed retention time and fruit characteristics were analyzed using Pearson’s correlation coefficient. All the data, except water content, were log-transformed before analysis. Water content data was transformed to arcsine square root in order to satisfy the requirement of normality. Table 2. Comparison of retention time between individual birds. RESULTS 1) Treatments of seeds All but one of the seeds of all of the plant species used in the experiments were defecated by captive bulbuls. The exception was a single seed of A. japon- ica, which was regurgitated. The retention time of the regurgitated A. japonica seed was 14.2 minutes (N=1), but it was excluded from the calculation of retention time. All of the observed feces contained only seeds of a single species. Individual difference between bulbuls It was not practical to examine every fruit species with every bird, because the fruiting periods of the plants examined varied and each bulbul was retained for experimentation for just one week. One-way ANOVA was applied to the five fruit species for which SRTs were measured for two or three individ- ual bulbuls. No significant differences were found among individual bulbuls in any of the experiments (Table 2). 2) 3) Seed retention times In most cases, seeds were defecated within one hour (Table 3). The mean SRT for all seed species was 20.8 minutes. The shortest SRT recorded was 2.4 for Pyracantha angustifolia, and the longest SRT recorded was 123.0 minutes for C. mollis. 4) Relationships between fruit characteristics and seed retention times Last SRT and mean SRT were significantly nega- tively correlated with seed size, fruit size and seed weight (last SRT vs. seed size: r=—0.68, N=16, p=0.0028; mean SRT vs. seed size: r=—0.56, Individuals birds Plant species Geaoeredeer ee) df MS F-ratio P Ligustrum japonicum Avs Bvs C 2 317.63 Hil 50 (10, 10, 10) Pyracantha angustifolia Avs B vs C 2 20.83 1.34 .26 (80, 83, 91) Ilex crenata Cvs DvsE 2 267.83 2.65 07 (20, 57, 51) Aucuba japonica E vs F vs G 2 6.15 .20 .82 (4, 13, 12) Cornus florida Avs B | 4.9] ll 74 (46, 4) 44 Seed retention time in bird’s gut Table 3. Seed Retention Times by Brown-eared Bulbul. Plant name ING Mean of mean Mean of first Mean of last Aucuba japonica 32 11.6+4.9 6.0 17.6 Daphniphyllum macropodum 17 IS) H/S55)35) 9.5 30.0 Elaeagnus umbellata 4 19.8+7.2 7.0 31.8 Cinnamomum camphora 9 13.1+4.8 6.2 22S Paederia scandens 3) 24.2+3.6 16.0 37.0 Ligustrum japonicum 4 2 eee: 3.0 26.5 Cornus florida 6 13.1+6.0 6.8 29.8 Nandina domestica 6 20.8+5.4 15.3 26.0 Viburnum dilatum 4 28.0+£15.5 11.6 58), I Ilex crenata 4 IM S38 i210) DES 53.6 Celastrus orbiculatus 4 26.9+10.6 14.5 42.5 Phytolacca americana 11 21.1£16.7 52 45.3 Ilex serrata 4 16.8+£11.5 9.5 43.0 Pyracantha angustifolia 8 Gy se7-2. 2.8 45.0 Callicarpa mollis 4 41.5+14.4 26.3 76.3 Eurya japonica 6 2117.9 6.8 40.6 N* shows replications of trials of each fruit species. Log (time of seed retention in gut) Log (time of seed retention in gut) mote «5 G.6i,.7 628.984. A.401.2.1:3 Log (seed size) 3 -2.5 -2 -1.5 -1 -.5 0 Log (time of seed retention in gut) Log (time of seed retention in gut) Tee Ont a ee oN At AA 1S 55 ‘6 565 "7 Wo, ?-o foo 6.9.90 Log (fruit size) Arcsine(,/water content ) Fig. 1. Correlations between three indices of seed retention time (first SRT, mean SRT, and last SRT) and four fruit features (seed size, fruit size, seed weight and water content). Open squares=first SRT; solid circles=mean SRT, and open circles=last SRT. Unbroken lines indicate interactions between last SRT and each fruit feature. Broken lines indicate interactions between mean SRT and each fruit feature. 45 A. FUKUI N=16, p=0.022; last SRT vs. fruit size: r=—0.71, N=16, p=0.0014; mean SRT vs. fruit size: r= —0.68, N=16, p=0.0031; last SRT vs. seed weight: r=—0.71, N=16, p=0.0014; mean SRT vs. seed weight: r=—0.63, N=16, p=0.0077) (see Table 1, Fig. 1). The first SRT did not have significant correla- tion with any fruit characters (seed size: r=—0.12, N= 16, p=0.66; fruit size: r=—0.24, N=16, p=0.38; seed weight: r=—0.23, N=16, p=0.41). Water con- tent was not significantly correlated with SRT (first SRT: r=0.41, N=16, p=0.12; mean SRT: r=0.37, N=16, p=0.17; last SRT: r=0.28, N=16, p=0.30). The standard deviation of SRT was significantly negatively correlated with seed size (r=—0.65, N=16, p=0.0053), fruit size (r=—0.57, N=16, p=0.021), and seed weight (r=—0.60, N=16, p=0.013), but was not correlated with water content (r=0.49, N=16, p=0.054) (see Fig. 2). Log (the standard deviation RT) 12) 63 14) {SU AGS Oe) allel 20 ded Log (seed size) ae. 41:5 oc = 2 o 3) Tv Z © z 5 5 ” o & Ss O = A rca: Smee dram, «OIE ly PDs J Log (fruit size) Fig. 2. DISCUSSION As the key aspect of endozoochory is that plant seeds must pass through a bird’s gut, seed retention time is a significant factor for both plants and birds. Brown-eared Bulbuls are gulpers, and so might be expected to overcome any gut limitation by rapid voiding of seeds, such as by regurgitation. During my experiments, a bulbul only once regurgitated a seed, and that was of A. japonica, which produces the largest seed of the fruits examined. Studies of man- akins indicate that seed regurgitation is one way of voiding seeds rapidly in order to offset gut limitation because regurgitated seeds do not pass through the gut (Levey 1992). Levey (1992) found that manakins regurgitated large seeds more frequently than small seeds, whereas small seeds were always defecated. Like manakins, Brown-eared Bulbuls defecated small seeds and regurgitated a very large seed, although in the bulbul the average length of time to regurgitation fe 1.5 c Ao) ic) 3 1 me} 2 Oo no) § 5 ” ® £ 8 oO -3 -2.5 -2 -1.5 -1 -.5 0 Log (seed weight) Fae et a) o = Ao) 2 4 ® no) J (3) z is 5 (7) ® £ 3 0 55 6 65 .7 .75 86 .85 9 .95 Arcsine(,/water content ) The correlation between the standard deviation of SRT and four fruit features (seed size, fruit size, seed weight and water content). Unbroken lines indicate interactions between the standard deviation of SRT and each fruit feature (seed size, fruit size and seed weight). 46 Seed retention time in bird’s gut is not known because regurgitation was observed only once. The Brown-eared Bulbul defecated large bulky seeds more rapidly than small seeds (last SRT and mean SRT were both significantly negatively corre- lated with seed size, fruit size, and seed weight; see Fig. 1). These three fruit characteristics, representing bulk to frugivorous birds, being negatively correlated with SRT demonstrate that fruits and seeds with bulky characteristics are processed more quickly. In contrast, water content was not correlated with any indices of retention time, perhaps representing nutri- ent reward to frugivorous birds rather than bulk. It is concluded that it is important for bulbuls to process bulky seeds quickly. The first (shortest) SRT was not correlated with any of the four fruit characteristics examined (Fig. 1), indicating that gut limitations do not apply to first SRT. For fruiting plants, rapid seed processing by birds is disadvantageous for their seed dispersal success. Since last SRT and mean SRT are considered to rep- resent seed dispersal distances, the results imply that small seeds, which are retained in the gut longer, have longer dispersal distances than large seeds. A large standard deviation of SRT is likely to cor- respond to high diversity of seed destination. If so, small seeds have a higher diversity of seed destina- tion than large seeds because there was a negative correlation between seed size and the standard devia- , tion of SRT (Fig. 2). The standard deviation of SRT was also negatively correlated with fruit size and seed weight. Thus, it is concluded that seeds within a small fruit or light seeds are dispersed at various des- tinations as well as at greater distances than larger, bulkier seeds. Endozoochorous plants have evolved special fea- tures of their fruit to allow their dispersal agents to consume fruit easily (Gautier-Hion et al. 1985; Will- son & Whelan 1990; Jordano 1995). Fruiting plants depending on frugivorous birds for their seed disper- sal, are characterized by vivid colors (Willson & Whelan 1990). Among bird-dispersed fruits, nutri- tional characteristics differ between summer and win- ter fruit, coinciding with seasonal differences in the ‘nutrient requirements of birds (Herrera 1982). Al- though such fruit characteristics are advantageous for _frugivores, other characteristics, e.g. seed size, are disadvantageous (Sorensen 1984). Thus there exists a conflict between fruiting plants and frugivorous birds, although seed dispersal interactions have been dis- cussed from the viewpoint of mutualism. 47 How quickly and how thoroughly seeds are processed has important evolutionary consequences for both fruiting plants and frugivores (Herrera 1995; Murray et al. 1994). Fruits are generally considered to be a food resource high in bulk (Herrera 1987), and bulk overloads the gut thereby reducing ingestion rate. Rapid seed processing may clear the gut and thus allow an increased ingestion rate (Levey & Gra- jal 1991; Levey 1992). Waxwings fed two types of artificial fruit (agar-based sugar solution as pulp and plastic beads as seeds) with equal seed loads but with different seed sizes consumed significantly more of the larger-seeded fruits. Because large seeds were defecated more quickly than small seeds, the increase in fruit consumption indicated that waxwings were process rate limited (Levey & Grajal 1991). Levey (1992) reported that freely feeding manakins took natural and artificial fruits very quickly after regurgi- tation. Evidence from various birds suggests that the seeds already in a bird’s gut limit the rate of ingestion of more fruit. If rapid seed processing results in in- creased nutritional gain, then development of such an advantageous mechanism would spread through a population of frugivorous birds. From the viewpoint of fruiting plants, seed disper- sal success is determined by two factors: the quantity of seeds dispersed, and the distance over which seeds are dispersed. There is evidence that fruit species with seeds that are regurgitated, or with seeds that have short seed retention times (e.g. hawthorn Crataegus monogyna, sloe Prunus spinosa, and ivy Hedera helix) were the fruits most preferred by birds (Sorensen 1984). By implication, such fruit species have better chances of achieving higher removal rates of seeds, than other less-favored species. Considering that large seeds tend to have short retention times (Fig. 1), plant species producing large seeds benefit from the food preferences of birds and are able to dis- perse large numbers of their seeds, though their seed destinations may be limited because of the short throughput time. In contrast, plants producing small seeds (tending to have longer retention times) are more likely to achieve longer dispersal distances (corresponding to longer SRTs), which results in a greater diversity of seed destinations. Therefore, plants producing large seeds have the advantage of greater quantity of their seeds dispersed, whereas plants producing small seeds have the advantage of greater dispersal distances of their seeds. The results from this study of the Brown-eared Bulbul imply that there are two types of fruit charac- A. FUKUI teristic associated with frugivorous birds in term of seed dispersal success. This evolutionary interaction between fruit plants and seed dispersers may affect the diversity of fruit features, e.g. seed size, fruit size and seed weight, mediated by the length of time the seeds are retained in a bird’s gut. ACKNOWLEDGMENTS I am very grateful to: T. Saitou and H. Banno, who gave valuable suggestions during the experiments; Y. Kominami, S. Matsuoka and N. Noma, who provided helpful discussions; M. Yasuda, who gave numerous valuable comments on early versions of the manuscript, and finally, members of the Laboratory of Animal Ecol- ogy, University of Tsukuba, who gave valuable advice and encouragement during this study. REFERENCES Barnea A, Yom-Tov Y & Friedman J (1991) Does in- gestion by birds affect seed germination? Func Ecol 5: 394-402. Countney SC & Sallabanks R (1992) It takes guts to handle fruits. Oikos 65: 163-166. Fukui A (1995) The Role of Brown-eared Bulbuls as a Seed Dispersal Agent. Res Popul Ecol 37: 211-218. Fukui A (1996) Retention Time of Seeds in Bird Guts: Costs and Benefits for Fruiting Plants and Frugivo- rous Birds. Plant Sp Biol 11: 141-147. Gautier-Hion A, Duplantier J-M, Quris R, Feer F, Sourd C, Decoux J-P, Dubost G, Emmons L, Erard C, Heck- estsweiler P, Moungazi A, Russilhon C & Thiollay J- M (1985) Fruit characters as a basis of fruit choice and seed dispersal in a tropical forest vertebrate com- munity. Oecologia 65: 324-337. Herrera CM (1982) Seasonal variation in the quality of fruits and defuse coevolution between plants and avian dispersers. Ecology 63: 773-785. Herrera CM (1985) Determinants of plant-animal co- evolution: the case of mutualistic dispersal of seeds by vertebrates. Oikos 44: 132-141. Herrera CM (1987) Vertebrate-dispersed plants of the 48 Iberian peninsula: A study of fruit characteristics. Ecol Monogr 57: 305-331. Herrera CM (1995) Plant-vertebrate seed dispersal sys- tems in the Mediterranean: Ecological, evolutionary, and historical determinants. Ann Rev Ecol Syst 26: 705-727. Hoppes WG (1988) Seedfall pattern of several species of bird-dispersed plants in an Illinois woodland. Ecol- ogy 69: 320-392. Johnson RA, Willson MF & Thompson JN (1985) Nu- tritional values of wild fruits and consumption by mi- grant frugivorous birds. Ecology 66: 819-927. Jordano P (1995) Angiosperm flesh fruits and seed dis- persers: a comparative analysis of adaptation and constraints in plant-animal interactions. Am Nat 145: 163-191. Levey DJ (1987) Seed size and fruit-handling tech- niques of avian frugivorous. Am Nat 129: 471-485. Levey DJ (1992) Digestive processing of fruits and its consequences for fruit-frugivore coevolution. Acta XX Congresses Internationalis Ornithologici 1624— 1629. Levey DJ & Duke GE (1992) How do frugivores process fruit?: Radiographic studies of gastrointesti- nal transit in cedar waxwings (Bombycilla cedrorum). Auk 109: 722-730. Levey DJ & Grajal A (1991) Evolutionary implication of fruit-processing limitations in Cedar waxwings. Am Nat 138: 171-189. Murray KG, Russell S, Picone CM, Winnett-Murray K, Sherwood W & Kuhlmann ML (1994) Fruit laxatives and seed passage rates in frugivores: consequences for plant reproductive success. Ecology 75: 989-994, Snow DW (1971) Evolutionary aspect of fruit-eating by birds. Ibis 113: 194-202. Sorensen AE (1984) Nutrition, energy and passage time: Experiments with fruit preference in European black- birds. J Anim Ecol 53: 545-557. Westby M (1974) An analysis of diet selection by large generalist herbivores. Am Nat 108: 290-304. Willson MF & Whelan CJ (1990) The evolution of fruit color in flesh-fruited plants. Am Nat 136: 790-809. Ornithol. Sci. 2: 49-58 (2003) SPECIAL FEATURE Ecology of seed dispersal The role of the Velvet Asity Philepitta castanea in regenera- tion of understory shrubs in Madagascan rainforest Hajanirina RAKOTOMANANA'™, Teruaki HINO”*, Mamoru KANZAKFP and Hiroyuki MORIOKA‘4 ' Department of Zoology, Faculty of Science, Kyoto University, Kitashirakawa-Oiwakecho, Sakyo-ku, Kyoto 606-8502, Japan * Kansai Research Center, Forestry and Forest Products Research Institute, Kyoto 612-0855, Japan 3 Division of Forest and Biomaterials Science, Graduate School of Agriculture, Kyoto University, Kyoto 606-8502, Japan * Department of Zoology, National Science Museum, Tokyo 169-0073, Japan ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2003 Abstract In the Madagascan rainforest, the role of the Velvet Asity Philepitta cas- tanea, an endemic frugivorous bird, in the regeneration of five understory shrub species (Myrsinaceae and Rubiaceae) was examined during the dry season (August to October). Effective dispersal distance was 33.3 m/h. Based on seed retention time in captivity, more than 85.7% of regurgitated seeds and all defecated seeds were esti- mated to be transported outside the crowns of mother plants. Seeds passed by the Vel- vet Asity germinated less successfully than unmanipulated (control) seeds in four out of the five species of shrubs. The reduced germination rate of processed seeds was partly due to the non-adapted morphology of the Velvet Asity as a seed disperser, in particular its voluminous, thick-walled, muscular gizzard. The narrow, slightly de- curved bill and the semi-tubular tongue with vibrissae at the tip of this bird are nor- mally features of insect- and/or nectar-eaters. Moreover, since manual removal of fruit pulp decreased the germination rate of seeds, the shrub species studied may not have developed adaptations for seed dispersal by animals. The most probable expla- nation for this situation is that the Velvet Asity has shifted relatively recently to oc- cupy the niche of a fruit-eater of the understory and as yet insufficient time has passed for a sophisticated relationship with fruiting plants to have coevolved. Key words In tropical forests, many trees bear fleshy fruits adapted for animal dispersal (Howe & Smallwood 1982; Janzen 1983), and the role played by animals, particularly by birds, is well documented (e.g., Howe 1977, 1981, 1986; Howe & Estabrook 1977; Murray 1988). That is, birds provide a survival advantage to fruiting plants by allowing them to escape seed pre- _ dation near conspecific parent plants (Janzen 1970; - Howe 1977; Regal 1977; Hubbel 1980; Gorchov et al. 1993). Most of the plants adapted to bird-dispersal have: red or black fruits to attract birds visually _ (Willson & Melampy 1983); relatively small fruits that are easily swallowed by birds (Wheelright 1985); and thin-husked fruit, which are quickly processed by birds (Leighton & Leighton 1983). Specialized fruit- (Received 27 September 2002; Accepted 17 December 2002) * Corresponding Author, E-mail: tkpk @affre.go.jp * Present address: Department of Animal Zoology, Faculty of Sci- ence, University of Antananarivo, Antananarivo 101, Madagas- car 49 Coevolution, Frugivory, Madagascar, Seed dispersal, Velvet Asity eating bird species involved in mutualistic interac- tions with such plants, typically have: a wide gape for swallowing large fruits whole (Snow & Snow 1988); a short intestine and a thin-walled, non-muscular giz- zard for processing seeds gently and evacuating them quickly (Walsberg 1975); and a large liver for detoxi- fying fruits (Pulliainen et al. 1983). The island of Madagascar, because of its ancient isolation from Africa, has a unique flora and fauna with a high proportion of endemic species, many of which are threatened by the rapid habitat destruction and forest fragmentation that has occurred in the last few decades and which continues today (Green & Sussman 1990; Ganzhorn et al. 1997). Understanding the relationships between frugivores and fruiting plants is extremely important for the preservation of these forest fragments because such interactions may influence plant distribution and floristic heterogeneity (Howe 1977). Belher and Bohning-Gaese (in press) have shown that frugivorous bird species are fewer, H. RAKOTOMANANA et al. and their seed dispersal is far less efficient in Mada- gascar than in South Africa, while the seeds of Mada- gascan plants are less well adapted to bird-dispersal than those of South Africa. The Velvet Asity Philepitta castanea is endemic to Madagascar where it is the only regularly fruit-eating bird species occurring among rainforest understory shrubs. This species consumes a wide variety of fruits throughout the year, and also feeds fruits to its young as an important dietary component during the breed- ing season (Rakotomanana & Hino 1998; Rako- tomanana & René de Roland, in press). Nevertheless, this bird may not be fully adapted to frugivory as in- dicated by its narrow, slightly decurved-bill, a bill structure more typical of nectar- or insect-eaters (Langrand 1990; Yamagishi et al. 1997). In this paper we describe research into the role of the Velvet Asity in the regeneration of five understory shrub species (Oncostemon leprosum, Psychotria sp. 1, sp. 6, sp. 8 and Saldinia sp.) in a southeastern Madagascan rainforest. We aimed to answer the fol- lowing questions: (1) Whether the internal and exter- nal morphological characteristics of the Velvet Asity are adapted for seed dispersal, or not; (2) How far the birds disperse seeds by regurgitation and by defeca- tion; (3) How successfully the seeds passed by the birds germinate and grow. Based on the results, we discuss the relationship that has coevolved between the Velvet Asity and the fruiting plants of Madagas- car. METHODS 1) Study area The study was carried out in the Ranomafana trop- ical rainforest, about 365 km southeast of the capital Antananarivo, southeastern Madagascar (21°16’S— 47°28’E), at an altitude of 800 to 1200m above sea level. The study area was ca. 12.3 ha with many trails and paths. Two main seasons occur in the area: a rainy season (November—April) and a dry season (May—October). The annual precipitation was 2,600 mm (Nicoll & Langrand 1989) and the mean annual temperature was about 21.4°C (Ranomafana-Ifanadi- ana Station). The vegetation was characterized by discontinuous canopy layers composed of scattered trees 20-30 m high. Trees over 30m tall were rare in the forest, whereas the middle-canopy layer (4-10 m) was quite dense. Trees and shrubs consisted mainly of Psycho- (ria spp. (Rubiaceae), Oncostemon spp. (Myrsi- 50 naceae), Aphloia theaformis (Flacourtiaceae), Euge- nia spp. (Myrtaceae), Ficus spp. (Moraceae), Dombeya_ spp. (Sterculiaceae), Tambourissa spp. (Monimiaceae), Ravensara spp. (Lauraceae), Poly- scias spp. (Araliaceae), and Weinmannia spp. (Cunoniaceae). Large areas of the forest floor are covered with an introduced exotic plant Psidium catt- leyanum. The most common epiphyte was Asplenium nidus (Aspleniaceae) and the most common orchids were Bulbophyllum spp. and Eulophiella spp. (Orchi- daceae). 2) Bird measurements External morphometric measurements of Velvet Asity were obtained from museum specimens (Cor- nell University Museum, USA, the Peabody Mu- seum, USA, and the Tsimbazaza Botanical and Zoo- logical Park or PBZT, Madagascar) and from mist- net-captured birds in the field. Their bill, wing, tarsus and tail lengths were measured to the nearest 0.1 mm using callipers and birds were weighed to the nearest 0.1 g using a pesola balance. Captured birds were marked with colored plastic bands and released un- harmed after being measured. The tongue and diges- tive tract of one individual (taken as a specimen for PBZT), were observed under a light microscope. Wing areas were also measured and wing loading was calculated as: body weight (g)/wing area (cm’), according to Pennycuick (1975) and Greenwalt (9S): 3) Field observations of foraging behavior A total of 96 hours of field observations were made between 06:00 and 12:00 or between 14:00 and 17:00 from August to October in 1995 and 1996. Once a bird was found, we followed and recorded its forag- ing behavior until it was out of sight (using 1040 binoculars and a tape recorder), using a continuous recording method (Martin & Bateson 1986). Both un- marked and marked birds (4 males and 11 females) were followed. One sequence of continuous record- ing averaged 69.8 min (range: 30-201 min, N=40 se- quences). During each observation period, the times when birds visited and left each plant, and the forag- ing behavior of the focal bird, were recorded. The shrubs and trees that the birds visited were tagged with plastic tapes and marked on a map. For each movement, the distance in a straight line from the ini- tial location where the bird was encountered was measured on this map. Median distances were calcu- lated for every five-minute interval spent in move- The Velvet Asity and regeneration of shrubs ment from the initial locations. A quadratic equation was devised to express the relationship between dis- tance traveled and time elapsed during foraging movements. 4) Experimental determination of seed retention times From August to October 1996, five Velvet Asity (2 males and 3 females) were mist-netted, and kept for experimental determination of seed retention time (SRT) of five common fruiting species: O. leprosum, Psychotria sp. 1, sp. 6, sp. 8 and Saldinia sp. (Psy- chotria species were identified by number, as their species names are not known; see Appendix 1). In the field, each captured individual was _ temporarily housed in a small cage (1 mX1mX1m) for a maxi- mum of five hours, and supplied with ripe fruits of the target plant. One trial was conducted for each in- dividual, and water was available during each trial. The number of fruits consumed was counted and the time from consumption to evacuation of seeds by re- gurgitation and defecation was measured using a stopwatch. Since uneaten fruits were removed five minutes after the first fruit was consumed, the num- ber of ingested seeds differed among species (Appen- dix 2). Since we did not know exactly when the seeds evacuated were ingested, we used, as the time of in- gestion, the midpoint of the interval during which fruits was consumed. Unclear data were removed from the analysis because determination of SRT is critical to the estimation of effective seed dispersal. After each trial, the bird was released unharmed. Processed seeds were classified as regurgitated (clean) and defecated (mixed with feces). Some SRT followed by regurgitation data were also obtained through direct observation in the field. The distribution of the seeds dispersed (i.e. the seed shadow) by the Velvet Asity was estimated based on SRT followed by regurgitation or by defeca- tion, for each shrub species, using the movement dis- tance-time equation. Data relating to Psychotria sp. 1 and sp. 6, which have similar-sized seeds (Appendix 2), were combined owing to the small SRT sample sizes for each species. 5) Seed-germination experiment Processed seeds (regurgitated and defecated by the Velvet Asity) and non-processed seeds (naturally- fallen fruits) of the five shrub species were collected from the forest floor and dried in a cool place over several weeks. Non-processed seeds were divided 51 into two groups; in one group (manipulated seeds) the pulp was manually removed, in the other group (control seeds) the pulp remained intact. Processed and non-processed seeds (excluding insect-damaged seeds) were used for germination experiments. Be- cause all of the seeds of O. leprosum went rotten after defecation by the Velvet Asity, and because very few Psychotria sp. 8 seeds were defecated, defecated seeds could not be investigated in these two species. Forty to two hundred seeds were used in germina- tion experiments of each seed treatment of each shrub species. Seeds were planted in sand in small contain- ers (28.5cm lengthx24.5cm widthx5cm height). The experiments were carried out in a glasshouse under constant luminosity and at a constant tempera- ture of 24°C. Watering was twice a week. The germi- nation of seeds, seedling survival, and seedling stem lengths, were all monitored at weekly intervals for 100 days after sowing. RESULTS 1) Morphological characteristics The average body weight and external morpho- metric measurements (accompanied with standard deviation and sample size in parentheses) of the adults were: body weight=8.9+3.2¢ (19); bill length=18.1+1.9mm (55), bill width=6.0+0.9 mm (55); bill depth=5.1+0.4mm (55); tarsus length= 24.3+2.4mm (53); wing length=82.1+0.3 mm (55); tail length=42.5+0.5 mm (55). The bill was narrow (width/length=0.33+0.05 and depth/length=0.28+ 0.04), and slightly decurved. The wing loading averaged 0.53 g/cm? (2), and the wing tip has poorly developed slots (Rakotomanana unpublished data). The tongue was semi-tubular with a deep cleft (5 mm in length) anteriorly with vibrissae at the tip (Fig. 1). The intestine was about 20 cm long and there was a muscular thick-walled (ca. 4mm), voluminous gizzard (averaging 3% of body weight). 2) Foraging movements and behavior Among foraging observations of the Velvet Asity 91.1% of items were ripe or partially-ripe fruits taken from understory shrubs (<7.5 m in height) belonging to the Rubiaceae and Myrsinaceae. The size and shape of fruits and the size and number of seeds var- ied among shrub species (Appendix 1). During field observations, birds were observed to take from | to 16 fruits from one shrub, depending on the size or abundance of fruits. The seeds were either regurgi- H. RAKOTOMANANA et al. tated or defecated away from the parent plants. The seeds of almost all of the larger fruits, such as Psy- chotria sp. 8, were regurgitated, not defecated. The median distances of movement, measured in a straight line from the initial encounter point, in- creased almost linearly with increasing time and to 33.3m in one hour (Fig. 2). The maximum distance was 58.4 m. 3) Seed retention time and seed-shadow The Velvet Asity was observed to pluck fruit and swallow them whole without manipulating them with its bill. All the seeds appeared to be voided intact through regurgitation and defecation. More than 60% of regurgitation of all five shrub species’ seeds, was observed within 5-10 min after ingestion (Fig. 3). SRT data were obtained for O. leprosum in the field and experimentally, but there was no significant dif- ference between them (7 test=2.5, df=4, P>0.05). No seeds of Psychotria sp. 8 were observed to be defecated either in the field or in captivity (Appendix 2). SRT followed by defecation was more variable than SRT followed by regurgitation, with the excep- tion of Saldinia sp. SRTs of the seeds of Psychotria sp. 1 and sp. 6 were longer than those of O. leprosum and Saldinia sp. (Fig. 4; Mann Whitney’s U test, Psy- chotria sp. 1, sp. 6/O. leprosum, z=—4.7, P<0.001, Psychotria sp. 1, sp. 6/Saldinia sp., z=—5.7, P< 0.001 and O. leprosum/Saldinia sp., z=—1.7, NS). Brushy tip 2 : : ; The estimated maximum distances of dispersal by Groove regurgitation were 18.2m in O. leprosum, 5.4m in Saldinia sp, 6.7m in Psychotria sp. 1, 18.7 m in Psy- Shallow chotria sp. 6, and 13.8m in Psychotria sp. 8 (Fig. 5). Defecated seeds were transported farther than regur- Deep cleft gitated seeds (Fig. 4). The dispersal distances by defecation were estimated to range from 8.5m to 18.7 m in O. leprosum, from 9.1m to 15.4m in Sal- dinia sp, and from 14.4m to 29.0 m in Psychotria sp. ’ b 1 and sp. 6. The average crown diameters of the un- x6 derstory shrubs were 1.6+0.8m in O. leprosum, 1.9 x25 m+0.6 m in Psychotria sp. 1, 1.6+0.5m in Psycho- Fig. 1. (a) The whole tongue and (b) tongue tip of P cas- tria sp. 6, 2.8+0.9m in Psychotria sp. 8, and tanea. A t 50 d = 1.622 + 0.647 t - 0.002 t r = 0.82 40 E 2 30 =| e Se igze S | Y = 10 0 0 10 20 30 50 60 70 80 90 Time (min) Fig. 2. Relationship between time and median distance of movement in a straight line from the initial encounter point. 100 100 Oncostemon leprosum N=599 50 50 ~-~ SS _— : s hn 3 0 32 Bo 2 100 100 3 a Psychotria sp.6 N=42 50 50 0 0 Fig. 3. 100 | Oncostemon leprosum N=14 50 Seed defecation (%) sage times. 1.7£0.6m in Saldinia sp. (N=20 for each species). Thus for all of the five shrub species more than 85.7% of the regurgitated seeds and all of the defe- cated seeds were estimated to be transported beyond the crown of the mother plant. i 4) Germination ratios and growth rates of processed _and non-processed seeds Seed-germination experiments revealed that the The Velvet Asity and regeneration of shrubs Psychotria sp.1 N=24 C] experimental data || field data 100 Psychotria sp.8 N=14 Saldinia sp. N=30 50 Time after ingestion (min) Psychotria sp.1 and sp.6 N=20 20 53 SRT of seeds regurgitated by P castanea for each of five shrub species. Saldinia sp. N=50 30 40 50 60 Time after ingestion (min) Fig. 4. SRT of seeds defecated by P castanea for each of five shrub species. Arrows correspond to median pas- Velvet Asity has a negative effect on the germination of the seeds of four of the five shrub species (Psycho- tria sp. 8 was the exception) (Table 1). The processed and manipulated seeds of Saldinia sp., Psychotria sp. 1 and Psychotria sp. 6 germinated less successfully than their control seeds. The processed seeds of O. leprosum all went rotten before germination, although the germination ratio of manipulated seeds was higher than the control seeds. No differences in ger- H. RAKOTOMANANA et al. Oncostemon leprosum Estimated seed Shadow (%) Psychotria sp.8 10 15 20 25 30 Psychotria sp.1 and sp.6 Saldinia sp. ——= : regurgitated seeds axa : defecated seeds Distance from origin (m) Fig. 5. Table 1. Estimated seed shadow produced by regurgitation and defecation for each of five shrub species. Cumulative germination ratios (%) of seeds at 100 days in four treatments for five shrub species. Sample sizes are shown in parentheses. The same letters indicate significant differences between treatments for each species, with Fisher’s exact test (Bonferroni’s method, P<0.05). ce. Psychotria sp. 1 Psychotria sp. 6 Psychotria sp. 8 Saldinia sp. Control 11.7 (60)° 37-5. (200)° 55.0 (200)**° 14.5 (200) 66.5 (200) Manipulated 35.0 (60) 17.0 (100)"4 25.0 (100)? 39.0 (100) 30.5 (200) Regurgitated 0.0 (24)? 6.0 (200) 11.0 (50)? 39.0 (100)° 40.0 (80)° Defecated - 7.0 (100)° 8.0 (50)° ~ 32.0 (50)° b 17.4 Ti 148.7 65.9 58.3 df. p 3 3 2 3 P <.001 <.0001 <.0001 <.0001 <.0001 : Not investigated. mination rates were found between regurgitated and defecated seeds in four species. For Psychotria sp. 8, however, regurgitated and manipulated seeds germi- nated more successfully than the control seeds. On the other hand, the Velvet Asity had a positive effect on the growth rates of seedlings after germina- tion. In Psychotria sp. 1 and sp. 6 and Saldinia sp., 54 the seedlings that germinated from defecated seeds were significantly taller than, or not different from, those from other treatments, although those from re- gurgitated seeds were shorter than their control seeds. In O. leprosum and Psychotria sp. 8, for which defe- cated seeds were not investigated, the seedlings that germinated from manipulated or regurgitated seeds The Velvet Asity and regeneration of shrubs Table 2. Stem lengths (mean+SD) of seedlings at 100 days after sowing in four treatments for five shrub species. Sample sizes are shown in parentheses. The same letters indicate significant differences between treatments for each species, with post-hoc F- test (Bonferroni’s method, P<0.05). Psychotria sp. 6 Psychotria sp. 8 Saldinia sp. Creesanan Psychotria sp. | leprosum Control 20.0#1.2 (7)* 29.0+0.7 (70)*** Manipulated 30.0+0.9 (14)? —23.0+0.6 (12)*** Regurgitated - PAOHOT U2)" Defecated = 30.0+0.8 (7)°* F 492.6 358.6 d.f. 1 3 P<.0001 <.0001 25.0+0.8 (105) 26.0+1.0 (12)*° PUOHTS 6) 25.0+0.8 (4) 15.0+0.8 (26)*° 35.0+0.8 (39)*° 40.0+1.1 (33) 31.0+1.1 (130) 270018 (Si) 28.0+0.8 (30)> 35.0+0.8 (14) 5746.1 357.6 357.6 3 2 3 <.0001 <.0001 <.0001 —: Not investigated. were significantly taller than those from the control seeds (Table 2). DISCUSSION The seed shadow produced by the Velvet Asity in- dicates some contribution of this bird to the seed dis- persal of shrubs in the Madagascan rainforest, al- though the maximum dispersal distance (58.4 m) was much shorter than that of Neotropical frugivorous birds (ca. 220-510m, Murray 1988). The short dis- persal distances recorded in Madagascar might be due to the Velvet Asity’s aerodynamically non-ad- vanced wing features, such as its short wings, high wing loading, and poorly slotted wing, which would produce poor lift and acceleration during flight (Sav- ile 1957; Pennycuick 1969, 1975; Norberg 1981; Rayner 1981; Rakotomanana 1998). The Velvet Asity has some of the general features of specialized fruit-eaters, including, for example, a diet consisting mostly of fruits (Snow 1980; Morton 1973), great attentiveness and faithfulness to fruiting plants, the habit of regurgitating seeds, rapid passage of seeds through the gut (McKey 1975), and posses- sion of a short digestive tract (Desselberger 1931; Docters Van Leeuwen 1954; Walsberg 1975). The de- creased germination ratios of seeds regurgitated or defecated by the Velvet Asity demonstrate, however, that this species may not be adapted for dispersing seeds of fruiting plants. The rough chemical and physical treatment experienced inside the bird’s mus- cular, thick-walled, voluminous gizzard, which is twice the size of that of a specialized frugivorous bird such as Phainopepla nitens (Walsberg 1975), may cause seed damage, indicating that the Velvet Asity is 3) partially a seed predator. Although the Velvet Asity’s narrow, slightly decurved bill, and its semi-tubular tongue with vibrissae at the tip, may help it obtain in- sects (Gardner 1925) and/or nectar, these characteris- tics are unlikely to be useful for seed-eating (Amadon 1950; Richard & Bock 1973). Only Psychotria sp. 8 which produces seeds with a thick, hard seed coat, benefited from processing in the Velvet Asity’s gut, and from artificial manipula- tion. This species of shrub may have adapted to seed dispersal by the Velvet Asity, although its foraging preference was not high when compared with other species owing to the large size of its fruits (Rako- tomanana & Hino 1998). Since the control seeds, with pulp, germinated more successfully than manip- ulated seeds, without pulp, however, the shrub species other than Psychotria sp. 8 may not have de- veloped adaptations for seed dispersal by animals in- cluding birds. How do we explain the current relationship be- tween the Velvet Asity and the understory shrub species, in which the asity appears to be a partial seed predator, but one that carries seeds some distance from the mother plant? The most probable explana- tion is that the Velvet Asity has shifted relatively re- cently from being an insect- and/or nectar-eater to being a frugivore in the understory and as yet insuffi- cient time has passed for a sophisticated relationship with fruiting plants to have coevolved. That a niche shift may have occurred, is indicated by the morpho- logical structure of the Velvet Asity’s bill and tongue, and by observations that show that it takes arthropods and nectar on rare occasions (Prum & Razafindratsita 1997; Rakotomanana et al. in press). In addition, two species of Sunbird-Asity Neodrepanis coruscans and H. RAKOTOMANANA et al. N. hypoxantha, which are classified in the same en- demic family, Philepittidae, as the Velvet Asity, and considered to have the same ancestral origins as the Velvet Asity, both feed mainly on nectar and insects and have long, thin, decurved bills (Langrand 1990; Morris & Hawkins 1998) and tubular tongues (Morioka, unpublished). In Madagascar, lemurs are also important seed-dis- persers. In the same forest in which our study site was located, Dew and Wright (1998) have shown that seeds passed by lemurs germinate more successfully, and grow faster than those not passed by lemurs, in- dicating mutualistic adaptations between lemurs and fruiting plants. Dew and Wright (1998) also sug- gested that fruits more than 10mm in diameter are most likely to be dispersed by lemurs, while those less than 10 mm in diameter are most likely to be dis- persed by birds. The fruits of four of the five shrub species we studied were less than 10 mm in diameter, only Psychotria sp. 8 fruits were larger. Evidently, based on fruit size, seed dispersal of these fruiting plants should depend primarily on birds. If they are indeed dependent on the Velvet Asity for seed disper- sal, how have these fruiting plants maintained their populations in spite of their low germination success rate? In order to understand the dynamics of under- story shrub survival it is necessary to examine plant seed dispersal in more detail, examining not only the contribution of the Velvet Asity, but also that of occa- sional frugivorous visitors such as the Madagascar Bulbul Hypsipetes madagascariensis, the Red- fronted Brown Lemur Eulemur fulvus rufus and the Red-bellied Lemur Eulemur rubriventer, and bats such as Pteropus rufus and Rousettus madagas- cariensis. ACKNOWLEDGMENTS We are grateful to S. Yamagishi, M. Imafuku and A. Mori for their critical reading and comments on the pre- liminary draft of this paper and to M. Brazil for lan- guage editing. We are also indebted to A. Randrianjafy and other staff of the PBZT, R. Rakotonindrina and other staff of Association Nationale pour la Gestion des Aires Protégées for facilitating the administration, P. Wright and J. Rakotomalala for arranging the use of pre- liminary research facilities, and R. L. Ramamonjisoa and H. Razafiarisoa for collaborating in the germination experiments. This study would have not been possible without the essential help of S. Rakotomanana, D. Ran- drianantenaina and P. Rasabo in the field. Our apprecia- 56 tion also goes to the following institutions: Direction Générale des Eaux et Foréts, Ranomafana National Park Project, Silo National des Graines Forestiéres, Douroucouli Foundation, Yale University, Cornell Uni- versity and World Wide Fund for Nature, the Laboratory of Animal Sociology and Laboratory of Plant Ecology (Osaka City University) and Laboratory of Animal Be- haviour (Kyoto University) for their cooperation in the realization of the present study. This study was sup- ported by the Japanese Government and a Grant under the Monbusho International Scientific Research Pro- gram (Field Research No. 06041093). REFERENCES Amadon D (1950) The Hawaiian honeycreepers (Aves: Drepaniidae). Bull Am Mus Nat Hist 95: 151-262. Belher B & Bohning-Gaese K (in press) The role of birds in seed dispersal and its consequences for forest ecosystem. Acta Zool Sinica (Suppl). Desselberger H (1931) Der Verdauungskanal der Dicae- den nach Gestalt und Funktion. J Ornithol 79: 353-370. Dew JL & Wright P (1998) Frugivory and seed disper- sal by four species of primates in Madagascar’s East- ern rain forest. Biotropica 30: 425-437. Docters van Leeuwen WM (1954) On the biology of some Javanese Loranthaceae and the role birds play in their life histories. Beaufortia 4: 105-205. Gardner LL (1925) The adaptive modifications and tax- onomic value of the tongue in birds. Proc US Nat Mus 67: 1-49. Ganzhorn JU, Langrand O, Wright PC, O’Connor S, Rakotosamimanana B, Feistner ATC & Rumpler Y (1997) The state of Lemur conservation in Madagas- car. Prim Conser 8: 1-12. Green G & Sussman RW (1990) Deforestation history of the eastern rain forest of Madagascar from satellite images. Science 248: 212-215. Greenwalt CH (1975) The flight of birds. Trans Am Phil Soc New Ser 65: 1-67. Gorchov DL, Cornejo F, Ascorra C & Jaramillo M (1993) The role of seed dispersal in the natural regen- eration of rain forest after strip-cutting in the Peru- vian Amazon. Vegetatio 107/108: 337-349. Howe HF (1977) Bird activity and seed dispersal of a tropical wet forest tree. Ecology 58: 539-550. Howe HF (1981) Dispersal of a neotropical nutmeg (Vi- rola sebifera) by birds. Auk 98: 88-98. Howe HF (1986) Seed dispersal by fruit-eating birds and mammals. In: Murray DR (eds) Seed Dispersal. pp 123-189. Academic Press, Sydney. Howe HF & Estabrook GF (1977) On intraspecific com- The Velvet Asity and regeneration of shrubs petition for avian dispersers in tropical trees. Am Nat 111: 817-832. Howe HF & Smallwood J (1982) Ecology of seed dis- persal. Ann Rev Ecol Syst 13: 201-228. Hubbell SP (1980) Seed predation and the coexistence of tree species in tropical forests. Oikos 35: 214-229. Janzen DH (1970) Herbivores and the number of tree species in tropical forests. Am Nat 104: 502-528. Janzen DH (1983) Dispersal of seeds by Vertebrates guts. In: Futuyama DJ & Slatkin M (eds) Coevo- lution. pp 232-262. Sinauer Associates, Sunderland. Langrand O (1990) Guide to the birds of Madagascar. Yale University Press, New Haven. Leighton M & Leighton DR (1983) Vertebrate re- sponses to fruiting seasonality within a Bornean rain forest. In: Sutton SL, Whitmore TC & Chadwick AC (eds) Tropical rain forest. pp 181-196. Blackwell, Oxford. McKey D (1975) The ecology of coevolved seed disper- sal systems. In: Gilbert LE & Raven PH (eds) Coevo- lution of animals and plants. University of Texas Press, Austin. Martin P & Bateson P (1986) Measuring behaviour: an introductory guide. Cambridge University Press, Cambridge. Morris P & Hawkins F (1998) Birds of Madagascar: a photographic guide. Pica press, London. Morton ES (1973) On the evolutionary advantages: dis- advantages of the fruit-eating habit in tropical birds. Am Nat 107: 8-22. Murray KG (1988) Avian seed dispersal of three neotropical gap-dependent plants. Ecol Monogr 58: 271-298. Nicoll M & Langrand O (1989) Madagascar: Revue de la conservation et des aires protégées. WWF, Gland Suisse. Norberg UM (1981) Flight, morphology and ecological niche in some birds and bats. Symp Zool Soc Lond 48: 173-197. Pennycuick CJ (1969) The mechanics of bird migration. Ibis 111: 525-556. | Pennycuick CJ (1975) Mechanics of flight. In: Farner DS, King JR & Parkes KC (eds) Avian Biol Vol 5. pp 1-75. Sf Prum RO, Razafindratsita VR (1997) Lek behavior and natural history of the Velvet Asity (Philepitta cas- tanea: Eurylaimidae). Wilson Bull 109: 371-392. Pulliainen E, Helle P & Tunkari P (1983) Adaptive radi- ation of the digestive system, heart and wings of Tur- dus pilaris, Bombycilla garrulus, Sturnus vulgaris, Pyrrhula pyrrhula, Pinicola enucleator and Loxia py- tyopsittacus. Orn Fenn 58: 21-28. Rakotomanana H (1998) Negative relationship between relative tarsus and wing lengths in Malagasy rain for- est birds. Jpn J Ornithol 47: 1-9. Rakotomanana H & Hino T (1998) Fruit preference in the Velvet Asity Philepitta castanea in a rain forest of Madagascar. Jpn J Ornithol 47: 11-19. Rakotomanana H & René de Roland L-A (in press) Breeding ecology of the Velvet Asity Philepitta cas- tanea, an endemic species to Madagascar. Ostrich. Rayner JMV (1981) Flight adaptation in vertebrates. Symp Zool Soc Lond 48: 137-172. Regal PJ (1977) Ecology and evolution of flowering plant dominance. Science 196: 622-629. Richard LP & Bock WJ (1973) Functional anatomy and adaptive evolution of the feeding apparatus in the Hawaiian honeycreeper genus Loxops (Drepanidi- dae). Ornithol Monogr 15: 1-173. Savile DBO (1957) Adaptive evolution in the avian wing. Evolution 11: 212-224. Snow DW (1980) Regional differences between tropical floras and the evolution of frugivory. Acta XVII Congr Int Ornithol 2: 1192-1198. Snow B & Snow D (1988) Birds and berries. T & AD Poyser, Calton. Walsberg GE (1975) Digestive adaptations of Phainope- pla nitens associated with the eating of Mistletoe berries. Condor 77: 169-174. Wheelwright NT (1985) Fruit size, gape width, and the diets of fruit eating birds. Ecology 66: 808-818. Willson MF & Melampy MN (1983) The effect of bi- colored fruit displays on fruit removal by avian frugi- vores. Oikos 41: 27-31. Yamagishi S, Masuda T & Rakotomanana H (1997) A field guide to the birds of Madagascar. Kaiyusha, Tokyo (in Malagasy and Japanese). H. RAKOTOMANANA et al. Appendix 1. Fruit characteristics of five srub species. Oncostemon Psychotria Psychotria Psychotria Saldinia leprosum sp. 1 sp. 6 sp. 8 sp. Plant family Myrsinaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Fruit colour red red blue red blue Fruit type berry berry berry berry berry Fruit shape round ovoid ovoid ovoid round Seeds number 1 1 or 2 1 or 2 1 or 2 1 or 2 Average fruit size (mm)* 7.1 5.9 73 11.9 8.2 Average seed size (mm)? 49 4.7 4.8 9.6 2.8 a: Number of sample is 60 for each species. b: Samples were 122, 30, 84, 42 and 185 in the order of species from the left. Appendix 2. The number of fruits ingested and seeds processed by the Velvet Asity for each shrub species in the cage experi- ment. Oncostemon Psychotria Psychotria Psychotria Saldinia leprosum sp. l sp. 6 sp. 8 sp. Ingested fruits 139 25 50 7 52 Regurgitated seeds 125 24 42 14 2 Defecated seeds 14 18 2 0 50 58 ORIGINAL ARTICLE Ornithol. Sci. 2: 59-63 (2003) Head-bobbing patterns, while walking, of Black-winged Stilts Himantopus himantopus and various herons Masaki FUJITA'** and Kazuto KAWAKAMI? ' Division of Anthropology, Department of Biological Sciences, the University of Tokyo, 7-3-1, Hongo, Bunkyo-ku, Tokyo, 113-0033, Japan ? Tama Forest Science Garden, Forestry and Forest Product Research Institute, Todori-cho 1833, Hachioji, Tokyo, 193-0843, Japan ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2003 Abstract Head bobbing patterns of walking Black-winged Stilts Himantopus hi- mantopus and eight species of herons were studied. Though several of the species studied had been previously reported as non-bobbing birds, all nine species usually head bobbed while walking during our observations. The head-bobbing pattern most frequently observed was ‘one bob per step’ in which a bird bobs its head once for each step it takes. In several species, one of two other patterns was also observed. The ‘one bob per two steps’ pattern was observed in three species of herons when they were walking slowly, and the ‘two bobs per step’ pattern was observed in Black- winged Stilts. Non-bobbing walking was observed in Japanese Night Herons Gor- sachius goisagi walking at relatively fast speed during foraging, and in two other species of herons when they were not foraging. Head bobbing may be affected by walking speed and by whether birds are foraging or not. | Key words bobbing walking Many birds head bob while walking, and most bob once for each step (Daanje 1951; Bangert 1969). Dur- ing head bobbing, the head is held stable relative to the environment during the hold phase and is thrust forward during the thrust phase (Dunlap & Mowrer 1930; Whiteside 1967). Previous studies have pro- vided evidence that head bobbing has visual func- tions (Friedman 1975; Frost 1978; Pratt 1982: Davies and Green 1988; Wallman & Letelier 1993; Green et al. 1998; Troje & Frost 2000). In addition, a neural constraint (Troje & Frost 2000) and a biomechanical function (Fujita 2002) have been proposed as the fac- tors facilitating synchronization of head and leg /movements of walking birds. Nevertheless, it is not clear why some birds bob their head during walking while others walk without doing it. Dagg (1977), the only author to have made an issue of this question, listed 28 head-bobbing and (Received 29 July 2002; Accepted 18 October 2002) * Corresponding Author, E-mail: mfujita@es.a.u-tokyo.ac.jp * Present address: Laboratory of Biodiversity Science, School of Agriculture and Life Science, The University of Tokyo, Yayoi 1—I-1, Bunkyo-ku, Tokyo 113-8657, Japan 59 Black-winged Stilts, Foraging, Head-bobbing patterns, Herons, Non- 21 non-bobbing species and mentioned that most non-bobbers generally live near water, whereas head- bobbers generally eat seeds or fruits rather than moy- ing prey. Dagg’s (1977) results were based, however, on small numbers of walking steps, and no other in- formation (such as walking speed, ground condition, walking situation) was given. Since several species walk both with and without head bobbing (Dagg 1977), detailed observations are required before a species can be confirmed exactly as only a head-bob- ber or a non-bobber. The present study provides more detailed informa- tion on several species classified as non-bobbers by Dagg (1977) and on several related species. These were the Japanese Night Heron Gorsachius goisagi, Malayan Night Heron G. melanolophus, Cattle Egret Bubulcus ibis, Intermediate Egret Egretta intermedia, Little Egret E. garzetta, Pacific Reef Egret E. sacra, Gray Heron Ardea cinerea, Purple Heron A. purpurea, and Black-winged Stilt Himantopus hi- mantopus. Observation of these species was expected to provide insight into the common characteristics of head-bobbers because all nine species live near water M. FUJITA and K. KAWAKAMI and eat small animal prey. MATERIALS AND METHODS The walking behavior of wild individuals of nine species (eight herons/egrets and one shorebird) was filmed using a DCR-TRV30 digital video camera (SONY, Tokyo, Japan). Video films were analyzed first to show whether birds walked with or without head bobbing. Head- bobbing while walking was defined as walking while alternately extending and flexing the neck, and non- bobbing walking was defined as walking without ob- vious neck extension or flexion. When birds were walking away from, or towards, the camera, it was difficult to distinguish whether head bobbing oc- curred or not, thus those observations were discarded. Subsequent analysis focussed on the number of bobs per step among head-bobbing birds. In the ‘M bob(s) per N step(s)’ pattern, M bob(s) occurred dur- ing N step(s) of walking (M and N are integral num- bers). Body movements of these head-bobbing pat- terns were digitized by Frame Dias software for mo- tion analyses. Digitized body points were the eye, the shoulder, and the tip of the first digit (as previously described by Fujita (2002). The distal end of the ob- servable portion of the tarsometatarsus segment was digitized when the digits were not observable. Ground conditions and foraging/non-foraging be- havior were also observed. Ground conditions were classified as grassland, flatland, and water. Flatland included sandy beaches, dry sandy areas, and open soil. “Water” indicated that the birds were walking with their feet in the water. As an indicator of walking speed, step duration was counted (in terms of video frames; 30 frames/ sec) and was expressed in seconds. Step duration was counted from one “foot-down” to the next. “Foot- down” was defined in this study as the moment when the entire sole touched the ground. When the feet were hidden in water, grass etc., the moment of foot- down was defined as the moment when the down- ward movement of the tarsometatarsus segment stopped around the end of the single support phase. A t-test was used to compare the mean step duration for each head-bobbing pattern to that for the ‘one bob per step’ pattern. RESULTS Purple Herons, Intermediate Egrets and Japanese 60 Night Herons generally walked slowly. The mean values of step duration in these three species were greater than one second (see Table 1). In all nine species, head-bobbing walking was ob- served during both foraging and non-foraging behav- iors. The head-bobbing pattern that was most fre- quently observed was ‘one bob per step’ (Table 1). During ‘one bob per step’, the thrust phase occurred from the double support phase to the beginning of the single support phase (Fig. 1, A—B), and the hold phase was observed from the single support phase to the beginning of the double support phase (Fig. 1, C—D). This pattern was observed in all species. The second pattern was ‘one bob per two steps’ which was observed in Purple Herons, Japanese Night Herons and Malayan Night Herons (Table 1). During ‘one bob per two steps’, the thrust phase started during the double support phase (Fig. 2, A) and continued to the beginning of the single support phase. Then the hold phase started during the single support phase (Fig. 2, B), and continued to the next single support phase (Figs. 2, C-E). Then the thrust phase started again almost simultaneously or immedi- ately after the beginning of the successive double support phase (Fig. 2, F). This pattern was observed when the birds were walking very slowly and looking for prey. The mean step duration of this pattern in Malayan Night Herons was significantly larger than that of the ‘one bob per step’ pattern (P<0.05). In the Purple Herons, and the Japanese Night Herons, the mean step duration during ‘one bob per two steps’ was more than three seconds, but did not differ sig- nificantly from that during ‘one bob per step’. The third pattern was ‘two bobs per step’ which was observed only once in the Black-winged Stilt (Table 1). In this pattern, the head movement was similar to that in ‘one bob per step’, but there was one more short thrust phase during the hold phase (Fig. 3). Thus, two thrust phases and two hold phases occurred during one step. This pattern was observed while the bird walked at normal speed looking for prey. Non-bobbing walking was observed in the Japan- ese Night Herons, Pacific Reef Egrets and Gray Herons (Table 1). Whenever Gray Herons were ob- served non-bobbing walking, individuals were walk- ing on dry sandy ground and were not looking for prey irrespective of their walking speed. The mean step duration of non-bobbing walking was slightly longer than during ‘one bob per step’ head-bobbing walking, but the difference between them was not Head-bobbing in Black-winged Stilts and herons Table 1. pattern. Step numbers, ground conditions, and mean and standard deviation of the step duration (sec) for each head-bobbing 2 bobs Total 1 bob | step 1 bob 2 steps isp Non-bobbing Japanese Night Heron N (nf) 40 (0) 27 (0) 8 (0) 0 5 (0) GC F F F F Mean+SD 3.54+1.92 4.03+1.63 3.70+1.87" - 0.65+0.55** Malayan Night Heron N (nf) 72 (0) 68 (0) 4 (0) 0 0 GC F/G F/G G Mean+SD 0.93 £0.85 0.87+0.85 1.83+0.22* - - Cattle Egret N (nf) 39 (0) 39 (0) 0 0 0 GC G G Mean+SD 0.62+0.32 0.62+0.32 _ - — Intermediate Egret N (nf) 40 (0) 40 (0) 0 0 0 GC G/W G/W Mean+SD 1.54+0.57 1.54+0.57 _ = = Little Egret N (nf) 31 (0) 31 (0) 0 0 0 GC W W Mean+SD 0.54+0.04 0.54+0.04 - _ — Pacific Reef Egret N (nf) 90 (22) 87 (19) 0 0 3 (3) GC F/W F/W F Mean+SD 0.59+0.13 0.60+0.13 - - 0.46+0.04"° Gray Heron N (nf) 60 (52) 18 (10) 0 0 42 (42) GC F/W F/W F Mean+SD 0.76+0.52 0.64+0.42 - — 0.81+0.56"° Purple Heron N (nf) 30 (0) 16 (0) 14 0 0 GC G/W G/W G Mean+SD 3 ASE) 4.00+1.92 3.45+1.60" - - Black-winged Stilt N (nf) 75 (0) 74 (0) 0 1 (0) 0 GC W W WwW Mean+SD 0.76+0.45 0.76+0.45 - 1.1 = | of the ‘1 bob 1 step’ pattern. significant. Japanese Night Herons, however, were observed non-bobbing walking even when walking | quickly while foraging. The average step duration of 'this species while non-bobbing was significantly | shorter than during ‘one bob per step’ (P<0.01). A Pacific Reef Egret was also observed walking without | head bobbing while it walked somewhat quickly while not foraging, though the mean step duration of non-bobbing was not significantly different from that of ‘one bob per step’. DISCUSSION The ‘one bob per step’ pattern was seen in most of our observations (Table 1) of eight species of herons/egrets and one shorebird. The same pattern Asterisks refer to the statistically significant (**: P<0.01, *: P<0.05, ns: not significant) difference from the mean step duration | Ground conditions (GC) were exhibited in following abbreviations; F: flatland; G: grassland; W: water. N: step numbers of both foraging and non-foraging steps; nf: numbers of non-foraging steps. has been observed among other species (Daanje 1951; Bangert 1960; Fujita 2002). In contrast to Dagg’s (1977) observations, we found that Pacific Reef Egrets and Black-winged Stilts usually walked in the ‘one bob per step’ pattern. We also observed ‘one bob per two steps’ in several species, and ‘two bobs per one step’ in one species. The ‘one bob per two steps’ pattern observed in several heron species (Table 1) had not previously been reported, perhaps because walking herons had not been sufficiently critically observed before. Herons have a longer neck than other birds and can extend it further than other species. The ‘two bobs per step’ pattern observed in a Black-winged Stilt seems to be a variation of the typ- ical ‘one bob per step’ pattern, with one extra short 61 M. FUJITA and K. ®@ Eye @ Shoulder O First digit hold phase relative body movement in sagittal axis (pixel) 0 a 4 6 8 1 time (sec) Fig. 1. The ‘one bob per step’ head bobbing walk of a Pa- cific Reef Egret. Upper diagrams show body movements traced from a video film of the distinctive phases: A, the thrust phase during the double support phase; B, the hold phase dur- ing the double support phase: C, the hold phase during the sin- gle support phase; D, the thrust phase during the double sup- port phase. The lower diagram shows the relative horizontal movement of the body points, the eye, shoulder, and first digit, plotted against time. The gray boxes indicate the double sup- port phase during which both feet are on the ground. relative body movement in sagital axis (pixel) y 0 | 2 3 4 5 6 Time (sec) Fig. 2. The ‘one bob per two steps’ head bobbing walk of a Japanese Night Heron. Upper diagrams show the body move- ments traced from a video film of the distinctive phases. The first step is from A to C, and the second step is from D to F. The neck is extended during the first double support phase (A-B). The hold phase starts during the first single support phase (B), and continues to the third double support phase (C-—E). Then the thrust phase starts again during this double support phase (F). The lower diagram shows the relative hori- zontal movement of body points, the eye, shoulder, and first digit, plotted against time. The gray boxes indicate the double support phase during which both feet are on the ground. thrust phase during the hold phase. The bird may have altered the direction of its gaze during this short thrust phase. This pattern has previously been re- ported for chickens and starlings (scientific names not given) (Dunlap & Mowrer 1930). KAWAKAMI 2 extra short thrust phase hold phase +4 hold phase a ro) QODDDOONDOGNOOGID OatevareBarteats PELEEE EDGED GEE® wn i) y' @ Shoulder O Distal tarsometatarsus D ConnpconDonooO- relative movement of body points (pixel) i=) 0 2 4 6 8 1 12 time (sec) Fig. 3. Relative horizontal movement of the eye, shoulder, and distal tarsometatarsus, during the ‘two bobs per step’ pat- tern. Distal tarsometatarsus is defined as the distal end of the observable portion of the tarsometatarsus segment. This was digitized instead of the first digit because the individual walked with its feet in water. The gray boxes indicate the dou- ble support phase. Pacific Reef Egrets and Gray Herons walked with- out head bobbing when they were not foraging. As these birds were not looking for prey, they may have not needed strict stabilization of the head. Thus, their non-bobbing walking may be indicative of the impor- tance of the visual function of head bobbing when foraging. In the Japanese Night Herons, however, non-bobbing walking was observed when the bird walked quickly while foraging. The reason for its non-bobbing probably differs from that of Pacific Reef Egrets and Gray Herons, and may have been re- lated to its walking speed. Dagg (1977) reported that White Ibis Threskiornis molucca and Bar-tailed Godwit Limosa lapponica walked both with and without head bobbing. Three species of herons also walked in both manners in our observations. Pacific Reef Egrets and Black-winged Stilts (classified as non-bobbers by Dagg (1977)), usually head bobbed while walking in our observa- tions. One reason for this inconsistency may be that Dagg (1977) focussed on bobbing behavior and did not record walking speed or foraging behavior. Dagg’s (1977) observations of these species may also have been over a small range of walking speeds and under restricted situations. Our observations indicate that non-bobbing walking sometimes occurs when birds are walking at relatively high speeds, or when not foraging. Dagg (1977) mentioned that most non-bobbing birds generally live near water, and that many head- bobbing birds generally eat seeds or fruits rather than moving prey. In our observations, however, herons and Black-winged Stilts, which live near water and Head-bobbing in Black-winged Stilts and herons eat moving prey, usually head bobbed while walking (Fig. 1, Table 1). One of the common characteristics among head-bobbing species seems to be that their foraging behavior consists of looking for food while walking. Comparing many different species of birds that employ different foraging techniques will give more insight into the reasons for head bobbing. ACKNOWLEDGMENT We thank Mr. Seiki Shiratama for helping us to film the walking behavior of Malayan Night Herons. REFERENCES Bangert H (1969) Untersuchungen zur Koordination der Kopf- und Beinbewegungen beim Hausuhn. Z Tierpsychol 17: 143-164. Daanje A (1951) On locomotory movements in birds and the intention movements derived from them. Be- haviour 3: 48-98. Dagg AI (1977) The walk of the Silver gull (Larus no- vaehollandiae) and of other birds. J Zool Lond 182: 529-540. Davies MNO & Green PR (1988) Head-bobbing during 63 walking, running and flying: relative motion percep- tion in the pigeon. J Exp Biol 138: 71-91. Dunlap K & Mowrer OH (1930) Head movements and eye functions of birds. J Comp Psychol 11: 99-113. Friedman MB (1975) Visual control of head movements during avian locomotion. Nature 225: 67-69. Frost BJ (1978) The optokinetic basis of head-bobbing in the pigeon. J Exp Biol 74: 187-195. Fujita M (2002) Head bobbing and the movement of the center of gravity in walking pigeons (Columba livia). J Zool Lond 23: 373-379. Green PR, Davies MNO & Thorpe PH (1998) Head- bobbing and orientation during landing flights of pi- geons. J Comp Physiol A 174: 249-256. Pratt DW (1982) Saccadic eye movements are coordi- nated with head movements in walking chickens. J Exp Biol 97: 217-223. Troje NF & Frost BJ (2000) Head-bobbing in pigeons: How stable is the hold phase? J Exp Biol 203: 935-940. Wallman J & Letelier JC (1993) Eye movements, head movements, and gaze stabilization in birds. In Zeigler HP & Bischof HJ (eds) Vision, Brain, and Behavior in Birds. pp 245-263. The MIT Press, Cambridge. Whiteside TCD (1967) The head movement of walking birds. J Physiol Lond 188: 31. Ornithol. Sci. 2: 65-72 (2003) ORIGINAL ARTICLE Response to manipulation of partner contribution: A handicapping experiment in the Barn Swallow Kazuyoshi TAJIMA* and Masahiko NAKAMURA* Laboratory of Animal Ecology, Department of Biology, Joetsu University of Education, I Yamayashiki-machi, Joetsu-shi, Niigata 943-8512, Japan Abstract In birds with biparental care, two parents cooperate to provide the appro- ORNITHOLOGICAL priate amount of care for the young. Evolutionary stable strategy (ESS) models pre- SCIENCE dict that cooperation can be stabilized when parents respond to reductions in care by their partners by increasing their effort, while not fully compensating for the reduc- tion. To examine whether parents adjust their effort according to their partner’s con- tribution and what cues the parents use in the bargaining process, we manipulated parental care in the Barn Swallow Hirundo rustica. Twenty-eight pairs were ran- domly assigned to three groups: (i) reduced male parental care (7 pairs), (ii) control (13 pairs), (111) reduced female parental care (8 pairs). Parental care was manipulated by attaching small weights to the base of a bird’s tail feathers. The manipulation suc- cessfully reduced parental provisioning in the handicapped birds, while still maintain- ing biparental care. Regardless of sex, however, handicapping of individuals led to no compensatory responses by the mates. The handicapped birds spent more time rest- ing, causing lowered provisioning rates. Males with a handicapped female decreased their provisioning rates to guard the resting females against extra-pair males. Since the provisioning parents in the three groups seldom met at their nest, it is unlikely that parents monitor their partner’s provisioning rate directly. We predicted that parents would adjust their provisioning rates according to the begging behaviors of their nestlings. However, no significant relationship was detected between the begging in- tensity (begging level and calling duration) and travelling time. Moreover, there were no significant differences in either begging level or calling duration among the three groups. Since various factors affect the provisioning rate in a handicapping manipula- tion, our data do not support the prediction derived from ESS models of biparental care. © The Ornithological Society of Japan 2003 Key words Barn Swallow, Biparental care, Hirundo rustica, Parental effort, Weighting manipulation In species with biparental care, two parents coop- erate to provide the appropriate amount of care for pensation is a feature of optimal investment strategy (Winkler 1987; Kacelnik & Cuthill 1990), and this the young. The division of labor between mates is best understood as an evolutionary stable strategy (ESS) whose equilibrium depends on the relative costs and benefits of investment to the male and fe- male (Chase 1980; Houston & Davies 1985). Cooper- ation can be stabilized when parents respond to re- ductions in care by their partners by increasing their effort, while not fully compensating for the reduction _ (Houston & Davies 1985). Incomplete partner com- (Received 5 August 2002; Accepted 16 October 2002) * Corresponding author, E-mail: masahiko @juen.ac.jp * Present address: Omaezaki Junior High School, 800-1 Sagara- machi, Shizuoka 421—0533, Japan 65 prediction is supported by experimental studies in which one partner’s contribution is completely re- moved (Bart & Tornes 1989; Hatchwell & Davies 1990; Markman et al. 1996). According to the ESS models, parental feeding rates are determined by a bargaining process between parents (Chase 1980; Houston & Davies 1985). The best way to test the process is experimental study in which one partner’s contribution is reduced either by clipping a number of feathers (Slagsvold & Lifjeld 1988, 1990; Whittingham et al. 1994; Weimerskirch et al. 1995; Sanz et al. 2000) or by attaching weights to the birds (Wright & Cuthill 1989, 1990a, b; Seether K. TAJIMA and M. NAKAMURA et al. 1993; Markman et al. 1995), and the compensa- tory response of the partners is then measured. Some experimental studies have supported the ESS models, while others have shown a range of responses to re- duced parental care by one partner, from no response through incomplete compensation to complete com- pensation (see Table 4 in Sanz et al. 2000). Thus, the bargaining process is still obscure and more experi- ments are necessary to elucidate it. ESS models of biparental care assume that parents adjust their effort according to their partner’s contri- bution. The bargaining process is possible only if each parent can monitor the behavior of the other. What proximate cues might parents use in the bar- gaining process? At least three are possible. First, parents may monitor their partner’s provisioning rate directly and adjust their effort accordingly. Second, if handicapped partners are spending more time resting near the nest, their mates might monitor resting part- ners and increase their own provisioning rate. Third, the cue might be nestling behavior, especially beg- ging, rather than parental behavior. Studies on a range of bird species suggest that nestling begging signals regulate the provisioning rate of adult birds (Budden & Wright 2001). If handicapped partners de- crease their provisioning rates and nestlings are hun- gry, nestling begging should be intense and the other partner should increase their provisioning rate. Most experiments have examined the response of the birds to handicapping (e.g. it reduced their provisioning rate), but have not considered the proximate cues that affect parental provisioning rates in the bargaining process. Our aims with this study are twofold; first, to in- vestigate the responses of male and female Barn Swallows Hirundo rustica to experimental _ tail- weight manipulation of their partners under the same brood size condition, and second, to examine whether parents adjust their effort according to the partner’s provisioning rate, the behavior of the handicapped partner, or nestling begging. The Barn Swallow is a suitable organism for this study because nestlings are fed by both parents, which contribute approximately equally, to the raising of young (M@gller 1994), and also because it is easy to find their nests and observe nestling behavior. MATERIALS AND METHODS 1) Study area and population This study was conducted in a residential area at 66 Joetsu City, Niigata Prefecture, Japan (37°07'N, 138°15’E, 10m alt.), in 1997 and 1998. Our study season, from May to June, corresponded to the first breeding period of the swallows. Intensive observa- tion in the week following hatching allowed us to pick a sample from 53 pairs, which nested under the eaves of a covered sidewalk. The average distance between nests was 16.2m (+3.4 SE) and both par- ents cared for the young in each nest. Thus, Barn Swallows in our study area bred in monogamous pairs in a loose colony. Of the 53 pairs, we selected 28 with broods of five chicks for our experiment. Be- fore the experiment, all adults were mist netted, color ringed, and marked with a non-toxic paint on the tail feathers, throat, or forehead to allow individual iden- tification. 2) Weighting experiment The manipulation designed to produce the reduc- tion in parental provisioning rate was the attachment of small lead weights (each weighing 0.3 or 0.4 g) to the base of the bird’s three central tail feathers with quick-drying glue (“Aron alfa”, Konishi, Japan). The birds in the experimental groups were weighted with a total of 1.0-1.5g (approximately 5—8% of their body mass). The nestling period, which is the period from hatching of the last egg to fledging of the last nestling, lasts on average 21 days (Mgller 1994). The weighting manipulation was carried out between 0100 and 0300 when chicks were 10 days old. The tail-weighting was temporary, as weights were shed during tail molt at the end of the breeding season. At least 53.3% of the handicapped birds (56.1% for con- trol pairs) started second breeding after the successful fledging of their first broods. Pairs were randomly assigned to three treatment groups: (i) male parental feeding reduced (N=7); (ii) control pairs under natural condition (N= 13); (iii) fe- male parental feeding reduced (N=8). For each pair in the three groups, we recorded the number of provi- sioning visits per hour using video cameras. We videotaped each nest for at least 2 hours a day, and for at least 2 days between 0600 and 1200. The cam- eras were placed so that they were aimed up towards the nest at a 45° angle from a distance of 3—5 m. This did not seem to disturb the parents because feeding activities were resumed immediately after we set the video camera in place. Data from all nests were col- lected using a video cassette recorder, when chick 12-14 days old. When recording parental behavior at nests, we observed their behavior near the nest using Response to Partner Contribution 8X30 binoculars for at least 30 minutes. Barn Swal- lows sometimes perched on electric lines. We re- garded perching within 5 m of the nest as resting, and calculated the time budget for resting (percentage of time that individuals spent resting). Fieldwork was performed in fine weather conditions. 3) Encounters between parents After one parent delivered food to the nestlings, provisioning by the mate usually followed within 1-3 seconds if the mate was ready to visit the nest. In this case, we assumed that the mate could monitor provi- sioning by the other parent. We defined an encounter between parents at the nest as occurring when the parents took turns at provisioning within 3 seconds or when both parents provisioned simultaneously. The degree of encounter between parents was cal- culated using a slight modification of Ekman’s (1979) coherence index. This index was defined as: Encounter index (%)= < 100 G Na+Nb—Ne Where Na and Nb are the numbers of times indi- viduals a and b provisioned, respectively, and Nc is the number of encounters between parents within 3 seconds. If two parents delivered food at the same time in all cases, the encounter index would be 100%. 4) Begging behavior To quantify begging intensity, we utilized a widely used begging index (e.g. Redondo & Castro 1992). The degree of postural intensity of the begging dis- _ play was categorized into five levels: the nestling 1) fails to beg; 2) gapes silently while resting on its belly; 3) gapes calling while resting on its belly; 4) _ gapes calling while elevating its body; and 5) same as 4) plus wing flapping. We estimated the begging level at the time when parents arrived at their nest. Total parental provisioning rates have been clearly demonstrated to depend on whole-brood begging rates or total noise as a signal of brood demand for food (Budden & Wright 2001). Thus, the begging levels of each nestling in a nest were pooled to give the whole-brood begging intensity. For example, in this study, when each nestling in a nest begged at level 5, the whole-brood begging intensity was 25. From the recorded videotapes, we examined the beg- ging level of each nestling and measured the total calling duration (time from first sound to end of last sound) using a stopwatch. 67 Parents flew from their nest to a feeding site and them took food home to the nestlings. We defined the round trip as the feeding trip and measured the time spent in each feeding trip as the travelling time. We predicted that parents responded to intense begging or long calling duration would shorten their travelling time. We were able to measure total calling duration and travelling time in all nests, but begging levels were not examined in 14 out of 28 nests in which we could not monitor the begging behaviors of all five nestlings. 5) Statistical analysis Each nest was assumed to be a statistically inde- pendent observation and parameters (provisioning rate, time spent resting, travelling time and call dura- tion) were normally distributed, so we used a two- way ANOVA to determine the effect of experiment and sex on parameters. When there was a significant interaction in the result of two-way ANOVA, the data were divided into male and female, then a one-way ANOVA was used to compare some parameters among the three treatments. When one-way ANOVA revealed significant differences among treatments, Scheffe’s F-test was used for post-hoc testing. Nestlings begged several times during a continuous observation period. We used the median of the beg- ging intensity and call duration of a nest as a repre- sentative value to minimize bias due to pseudorepli- cation. To examine the relationship between begging intensity (begging level and calling duration) and travelling time, we used Spearman’s correlation coef- ficient. All statistical procedures were done using StatView 5.0 (SAS 1998). RESULTS 1) Provisioning rates The average provisioning rates differed markedly according to the experimental treatment (Fig. 1, two- way ANOVA, F, ;)-=18.25, P<0.0001). There was no sex difference in the provisioning rate (F; ;).=0.001, P=0.98), but the interaction between sex and group was significant (F, 5)=17.83, P<0.0001). One-way ANOVA showed a significant difference in the provisioning rate among the three groups for either male or female (male, F,,,=15.74, P<0.0001; female, F,,;=20.17, P<0.0001). The provisioning rate of handicapped males was significantly lower than that of control males (Scheffe’s F-test, K. TAJIMA and M. NAKAMURA nN w c=) o Provisioning rate per hour —s =) Female hand (8) Experimental group Male hand’ Control (7) (13) Fig. 1. Average (SE) provisioning rates per hour in the three treatment groups (open bars, males; hatched bars, fe- males). The total provisioning rate per nest for males plus fe- males is depicted above. Sample sizes (number of nests) are given in parentheses. P<0.0001) and that of males with handicapped fe- males (P=0.040). Likewise, there was a significant difference in the male provisioning rate between the control and female-handicapped groups (P<0.05). The provisioning rate of handicapped females was lowest in the three groups (Scheffe’s F-test, P<0.0001). However, there was no significant differ- ence in the female provisioning rate between the male-handicapped and control groups (P=0.995). The total provisioning rates per nest (top of Fig. 1) differed significantly among the three groups (one- way ANOVA, F,,,=15.24, P<0.0001). Scheffe’s F- test revealed a significant difference between the male-handicapped and control groups (P<0.01), and between the female-handicapped and control groups (P<0.0001), but no significant difference between the male- and female-handicapped groups (P=0.39). 2) mH Encounters between pairs at the nest The mean encounter index of each group. was below 4% (Fig. 2). There was no significant differ- ence in the encounter index among the three groups (one-way ANOVA, F, ,,=2.49, P=0.10). 68 5 % x @ me) £ = 7) ae e = ie) rz) € WJ 0 Male hand Control Female hand (7) (13) (8) Experimental group Fig. 2. Average (SE) encounter indices (%) in the three treatment groups. Sample sizes (number of nests) are given in parentheses. 30 i) (=) % time spent resting Female hand (8) Experimental group Male hand’ Control (7) (13) Fig. 3. Average (+SE) percentage of time that individuals spent resting within 5m of the nest in the three treatment groups (open bars, males; hatched bars, females). Sample sizes (number of nests) are given in parentheses. 3) Time spent resting Two-way ANOVA showed that the effect of treat- ment on time spent resting was significant (Fig. 3; F, 5).=16.50, P<0.0001), as was the interaction be- tween sex and group (F, ,7=15.22, P<0.0001). How- ever, there was no sex difference in the time budget (F; 59= 3-40, P=0.07). Response to Partner Contribution A significant difference in the time budget was de- tected among the three groups for both sexes (male, F,5;=13.45, P<0.0001; female, F,,;=19.07, P< 0.0001). Handicapped males spent more time resting than control males (Scheffe’s F-test, P<0.001) and than males with handicapped females (P<0.01). Males with handicapped females spent more time resting than control males (Scheffe’s F-test, P<0.05), because they perched close to the resting females. We observed that extra-pair males approached resting fe- males (14 cases); male partners usually chased the extra-pair males away (12 of 14 cases), while we ob- served extra-pair copulations in the other two cases. Handicapped females spent more time resting than control females (Scheffe’s F-test, P<0.001) and than females with handicapped males (P=0.001), but there was no significant difference between the con- trol and male-handicapped groups (P=0.87). Females with handicapped males did not associate with resting males. 4) Response to nestling begging The average (+SE) travelling time of males and females was 277.2 (+12.5, N=757) and 230.0 (+10.1, N=841) seconds, respectively (data were pooled for all individuals). No significant relationship was found between the begging level and travelling time in 14 males (P>0.05 for 4 handicapped males, 4 _ control males and 6 males with handicapped females) and 14 females (P>0.05 for 4 females with handi- _ capped males, 4 control females and 6 handicapped females): male and female parents did not shorten their travelling time as begging level increased. Like- | wise, there was no significant relationship between call duration and travelling time in 28 males (P>0.05 _ for 7 handicapped males, 13 control males and 8 males with handicapped females) and 28 females - (P>0.05 for 7 females with handicapped males, 13 control females and 8 handicapped females). Regardless of treatment or sex, the average beg- ging intensity of the entire brood was ca. 19 (Fig. 4). Two-way ANOVA showed no significant effect of either treatment (F,,,=0.99, P=0.39) or sex (F; ..=0.26, P=0.62) on begging intensity and no significant interaction between the two variables (F,5,=0.29, P=0.75). This was because almost all | the nestlings begged at level 4. We obtained data on 5,180 begging episodes from 70 nestlings. Begging at level 4 accounted for 84.5% (4,377) of the cases (level 1, 2.3%; level 2, 3.7%; level 3, 8.3%; level 5, 1.2%). 69 > = 20 aE F = 2 Y 2 £ re) £10 fe?) aD ® mM 0 Female hand (4) (4) (6) Experimental group Male hand Control Fig. 4. Average (+ SE) whole-brood begging intensity in the three treatment groups (open bars, males; hatched bars, fe- males). Sample sizes (number of nests) are given in parenthe- ses. Oo iY YG AS Total calling duration NO //, Male hand Control Female hand (7) (13) (8) Experimental group (=) Fig. 5. Average (SE) total calling duration (seconds) in the three treatment groups (open bars, males; hatched bars, fe- males). Sample sizes (number of nests) are given in parenthe- ses. Regardless of treatment or sex, nestlings begged for about 5 seconds (Fig. 5). The average total call duration did not differ according to the experimental treatment (two-way ANOVA, F, ,.=0.54, P=0.58). Likewise, there was no sex difference in call duration (F, 597=1.70, P=0.20) and no significant interaction between sex and group (F, ,)=0.26, P=0.77). K. TAJIMA and M. NAKAMURA DISCUSSION In our experiments, manipulation successfully re- duced parental provisioning by handicapped birds, while still maintaining a biparental system (Fig. 1). Regardless of sex, however, handicapping of individ- uals led to no compensatory responses from their mates (Fig. 1). Previous handicapping experiments in passerines have shown compensatory responses by males or females (Sanz et al. 2000), so our results are exceptional. The low encounter index shows that pro- visioning parents seldom meet at the nest. Therefore, it is unlikely that parents monitor their partner’s pro- visioning rate directly. No significant relationship was detected between begging intensity (begging level and calling duration) and travelling time. There were no significant differences in either begging in- tensity or calling duration among groups (Figs. 4 & 5). It is unclear whether parents adjust their provi- sioning rates according to nestling behavior. Handicapped females spent more time resting near the nest, thereby decreasing their provisioning rates (Figs. | & 3). Before and after males visited their nests for provisioning, they were able to see their resting females. However, male swallows with handi- capped females did not increase their provisioning rates. While females were resting, extra-pair males attempted to copulate with them but were chased off by their male partners. Colonial breeding species have a high frequency of extra-pair copulation (Birk- head & Moller 1992). Mgller (1987a, b) found that Barn Swallow females in colonies, where the likeli- hood of extra-pair copulation is greater, tend to be guarded more closely than females of solitary pairs. Barn Swallows in our study area bred in a loose- colony. Thus, it is reasonable to suppose that males decreased their provisioning rates to guard their rest- ing females. Wright and Cuthill (1989, 1990a, b) ma- nipulated the parental care of the European Starling Sturnus vulgaris in a nestbox colony and found that, regardless of sex, the provisioning rate was reduced in handicapped birds with an incomplete compensa- tory increase by their unmanipulated partner. How- ever, they did not report on the resting of handi- capped females or on mate guarding. Females with handicapped males did not increase their provisioning rates as a result of their partners’ lower provisioning rate (Fig. 1). It is unclear whether this is because handicapped females could not work harder due to an upper limit on energetic expenditure (Drent & Daan 1980), or whether handicapped fe- 70 males maintained the frequency of their feeding visits by switching prey types. Parental care in terms of food delivered to the nest depends not only on the provisioning rate, but also on what is delivered (Whittingham et al. 1994). Starling parents main- tained their provisioning rates by switching prey types to those that were more quickly gathered or to smaller prey types (Wright & Cuthill 1989, 1990a, b). Compensation by switching prey types may be true of the male swallows with handicapped females as well as the females with handicapped males. Unfortu- nately, we have no data on food items. The total provisioning rates of the manipulated groups were lower than those of the control groups (Fig. 1). We expected that the low provisioning rates lead to an increase in nestling hunger level and nestling begging intensity. The more frequent beg- ging is or the larger begging call is, the more parents provision their nestlings (Budden & Wright 2001). Our data did not show this trend, however, and par- ents did not shorten their travelling time as begging level increased. Moreover, most nestlings in each group begged at level 4. There are at least two possi- ble reasons for this. First, the nestlings may always be hungry, even if the parents provision at a high rate. Saino et al. (2000) demonstrated in Barn Swallows that begging behavior is a signal of need (the amount of resources requested by an offspring to achieve sati- ation). Begging at level 4 when chicks are 12-14 days old seems to achieve the maximum level and re- flect their need level. Another possible reason is the shortness of the period of observation. We could not collect long-term data because many chicks in our study nests were predated by the Carrion Crows Corvus corone after chicks were 15 days old. Some previous experiments have shown that compensation requires a longer time to be effective. For example, Wright and Cuthill (1989) collected data for starling chicks between 11 and 20 days of age, and their re- sults showed compensatory responses by males and females. Long-term manipulation may lead to an in- crease in nestling hunger level and nestling begging intensity. Although our study presents negative data for ESS models of biparental care, we can not assert that part- ners with handicapped mates do not show any com- pensation. The provisioning rate by each partner is affected by many factors (age, sex and number of chicks, parent-offspring relatedness, conditions of parent and offspring, time of season, and paternity). In this study, we limited ourselves to the response of Response to Partner Contribution a parent to its partner’s contribution. However, many factors affect the provisioning rate in a handicapping manipulation. Is mate-guarding for handicapped fe- males always found? Do parents maintain the fre- quency of their visits by switching prey types? Does long-term manipulation lead to an increase in levels of nestling hunger? Moreover, recent studies indicate that there are sex differences in responsiveness to begging signals (Kilner 2002; Macgregor & Cock- burn 2002). Further research on these points is needed to assess our experimental results more pre- cisely. ACKNOWLEDGMENTS We are grateful to the members of the Laboratory of Animal Ecology of Joetsu University of Education, | Sachie Wakayama, Sachie Shinohara, Masako Yanagi- sawa, Takayoshi Okamiya, Takashi Yamamura, and Hi- romi Fujita, for their help in collecting data and provid- | ing constant advice. Two anonymous referees kindly commented on the manuscript. We thank to the residents of the Takada region of Joetsu City for their kind assis- | tance. We also thank the Shizuoka Prefectural Board of Education for their kindness. REFERENCES | Bart J & Tornes A (1989) Importance of monogamous male birds in determining reproductive success. Evi- dence for house wrens and a review of male-removal studies. Behav Ecol Sociobiol 24: 109-116. Birkhead TR & Mller AP (1992) Sperm competition in birds. Academic Press, New York. | Budden AE & Wright J (2001) Begging in nestling birds. In: Nolan V Jr & Thompson CF (eds) Current Ornithology. Vol 16. pp 83-118. Plenum Publishers, New York. Chase ID (1980) Cooperative and noncooperative be- havior in animals. Am Nat 115: 827-857. | Drent RH & Daan S (1980) The prudent parent: ener- getic adjustments in avian breeding. Ardea 68: 225-252. Ekman J (1979) Coherence, composition and territories of winter social groups of the Willow Tit Parus mon- tanus and the Crested Tit P. cristatus. Ornis Scand 10: 56-68. Hatchwell BJ & Davies NB (1990) Provisioning of | nestlings by Dunnoks, Prunella modularis, in pairs and trios: compensation reactions by males and fe- males. Behav Ecol Sociobiol 27: 199-209. ' Houston AI & Davies NB (1985) The evolution of co- 71 operation and life history in the Dunnock, Prunella modularis. In: Sibly RM & Smith RH (eds) Behav- ioural ecology: the ecological consequences of adap- tive behaviour. pp 471-487. Blackwell, Oxford. Kacelnik A & Cuthill I (1990) Central place foraging in starlings. II. Food allocation to chicks. J Anim Ecol 59: 655-674. Kilner RM (2002) The evolution of complex begging displays. In: Wright J & Leonard ML. (eds) The evo- lution of begging. pp 87-106. Kluwer Academic Pub- lishers, London. Macgregor NA & Cockburn A (2002) Sex differences in parental response to begging nestling in Superb Fairy-wrens. Anim Behav 63: 923-932. Markman S, Yom-Tov Y & Wright J (1995) Male parental care in the Orange-tufted Sunbird: behav- ioural adjustments in provisioning and nest guarding effort. Anim Behav 50: 655-669. Markman S, Yom-Tov Y & Wright J (1996) The effect of male removal of female parental care in the Or- ange-tufted Sunbird. Anim Behav 52: 437-444. Moller AP (1987a) Extent and duration of mate guard- ing in swallows Hirundo rustica. Ornis Scand 18: 95-100. Moller AP (1987b) Advantages and disadvantages of coloniality in the Swallow, Hirundo rustica. Anim Behav 35: 819-832. Moller AP (1994) Sexual selection and the barn swal- low. Oxford University Press, Oxford. Redondo T & Castro F (1992) Signalling of nutritional need by Magpie nestlings. Ethology 92: 193-204. Saino N, Ninni P, Incagli M, Calza S, Sacchi R & Meller AP (2000) Begging and parental care in rela- tion to offspring need and condition in the Barn Swal- low (Hirundo rustica). Am Nat 156: 637-649. Sanz JJ, Kranenbarg S, & Tinbergen JM. (2000) Differ- ential response by males and females to manipulation of partner contribution in the Great Tit (Parus major). J Anim Ecol 69: 74-84. SAS Institute (1998) StatView for Macintosh, Version 5.0, SAS Institute Inc., Cary, North Carolina. Sether BE, Andersen R & Pedersen HC (1993) Regula- tion of parental effort in a long-lived seabird: an ex- perimental manipulation of the cost of reproduction in the Antarctic Petrel, Thalassoica antarctica Behav Ecol Sociobiol 33: 147—150. Slagsvold T & Lifjeld JT (1988) Ultimate adjustment of clutch size to parental feeding capacity in a passerine bird. Ecology 69: 1918-1922. Slagsvold T & Lifjeld JT (1990) Influence of male and female quality on clutch size in tits (Parus spp.). Ecology 71: 1258-1266. Weimerskrich H, Chastel O & Ackermann L (1995) Ad- K. TAJIMA and M. NAKAMURA justment of parental effort to manipulated foraging ability in a pelagic seabird, the Thin-billed Prion Pachyptila belcheri. Behav Ecol Sociobiol 36: 11-16. Whittingham LA, Dunn PO & Robertson RJ (1994) Fe- male response to reduced male parental care in birds: an experiment in Tree Swallows. Ethology 96: 260-269. Winkler DW (1987) A general model for parental care. Am Nat 130: 526-543. Wright J & Cuthill I (1989) Manipulation of sex differ- 72 ences in parental care. Behav Ecol Sociobiol 25: 171-181. Wright J & Cuthill I (1990a) Manipulation of sex differ- ences in parental care: the effect of brood size. Anim Behav 40: 462-471. Wright J & Cuthill I (1990b) Biparental care: short-term manipulation of partner contribution and brood size in the Starling, Sturnus vulgaris. Behav Ecol 1: 116-124. SHORT COMMUNICATION Ornithol. Sci. 2: 73-74 (2003) Low prevalence of blood parasites in five Sylviidae species in Japan ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2003 Hisashi NAGATA!” and Navjot S. SODHI’ ' Laboratory of Wildlife Conservation, National Institute for Environmental Studies, Onogawa 16-2, Tsukuba, Ibaraki 305-8506, Japan ? Department of Biological Sciences, National University of Singapore, Blk S2, 14 Science Drive 4, Singapore 117543, Republic of Singapore About 450 species of blood parasites have been recorded from nearly 4,000 sampled bird species (Bishop & Bennett 1992). Due to their potential detri- mental affects on the reproduction and survival of different bird species (e.g. Sundberg 1995; Dufva 1996), blood parasites are gaining increasing promi- nence (McCallum & Dobson 1995; Sodhi 1995; Dale et al. 1996). Before blood parasites can be used for various rigorous ecological studies, it is an important first step to determine the level of prevalence of blood parasites in various species. We determined the prevalence of five blood parasites (Haemoproteus spp., Trypanosoma spp., Splendidofilaria spp., Plas- modium spp. and Leucocytozoon spp.) in five Sylvi- idae species (Cisticola juncidis, Locustella pryeri, L. pleskei, Acrocephalus arundinaceus, and A. bistrigi- ceps) in Japan. Previously we found that seven Emberiza species have low blood parasite presence in three of the study sites, Kamisu, Ukishima, and Seimei (Sodhi et al. 1996; Sodhi et al. 1999). Only three species, E. rus- tica, E.spodocephala and E. schoeniclus, were in- fected with a blood parasite. We also examined four Sylviidae species inhabiting reedbed and the Styan’s Grasshopper Warbler (L. pleskei) inhabiting dwarf laurel forest that is a congener to the endangered Japanese Marsh Warbler (L. pryeri). We sampled warblers at three sites within Japan, Kamisu (35°51'33”N, 140°38'20"E) along lower Tone River, Ukishima (35°57'N, 140°28’E) and Fu- nako (36°01'N, 140°16’E) along Lake Kasumigaura, and Ohtsukuejima (33°40’N, 130°18’E). At each site, warblers were caught using mist nets in the breeding season of 1996 and 1997. From some of the caught warblers, we collected a small amount of blood by (Received 30 August 2002; Accepted 14 October 2002) * Corresponding Author, E-mail: hnagata@nies.go.jp 73 puncturing the brachial vein, and released them after measuring and taking blood samples. Blood smears were prepared by using the method described by Clayton and Moore (1997) and were air-dried and fixed in 100% ethanol. The blood smears were stained using Giemsa’s stain. The presence of blood parasites was determined for each blood smear by ex- amining 100 randomly selected fields under a 100x oil immersed objective (see Bennett et al. 1995 for details). A total of 252 individuals belonging to five war- bler species were examined for the prevalence of blood parasites. Only 2 (0.8%) individuals of the Great Reed Warbler were infected with Plasmodium sp. (Table 1). Because only early developing gameto- cytes were found, the blood parasites could not be identified to specific level. Between 1965 and 1971, of 425 individuals exam- ined belonging to 27 bird species from Tsunoshima island, Yamaguchi Prefecture (34°21'N, 130°51’E), 53 (12%) were infected with a blood parasite (Mc- Clure et al. 1978). On this island, three species of Phylloscopus warblers were examined. Twenty-two Crowned Willow Warblers (P. coronatus) were free of blood parasites. Only one of 52 individuals exam- ined of the Arctic Warblers (P. borealis) and one of 27 individuals examined of the Pale-legged Willow Warblers (P. borealoides) were infected with a blood parasite from this area. The present study and previous reports (McClure et al. 1978: Sodhi et al. 1996; Sodhi et al. 1999) show that in general the blood parasite prevalence is low in warbler and bunting species in Japan. Both migratory and resident species had low blood parasite preva- lence (Table 1, Sodhi et al. 1999). It is unclear what the possible mechanisms are that help birds to main- tain a relatively low prevalence of blood parasites in this region. Low prevalence of blood parasites found H. NAGATA and N. S. SODHI Table 1. The number of individuals sampled for the presence of blood parasites in five species of warblers in Japan. Sampling sites Kamisu Lake Kasumigaura Ohtsukue-jima Ibaraki Ibaraki Fukuoka M F U M F U M F U Resident species Fan-tailed Warbler, Cisticola juncidis 2 1 - 12 10 2 ~ ~ - Japanese Marsh Warbler, Locustella pryeri 26 4 l 3) 7 8 ~ - — Migratory species (Summer visitor) Styan’s Grasshopper Warbler, L. pleskei 1 ] Great Reed Warbler, Acrocephalus arundinaceus 2 1(1) l 41(1) 21 4 — - Black-browed Reed Warbler, A. bistrigiceps 10 - - 20 30 19 - - = The number of individuals infected is indicated in parentheses. F=female, M=male, and U=unknown sex. by us may be due to several reasons such as the lack of a suitable arthropod vector (e.g. mosquitoes), host specificity in available arthropod vector, and/or lack of blood parasite susceptibility. It is also possible that the examined birds had latent infections that are diffi- cult to detect. Species of avian Plasmodium have much broader host specificity than Haemoproteus or Leucocytozoon and may infect coexisting species be- yond avian families (Atkinson & van Riper 1991). From a conservation perspective, it is probably much easier for maintaining the local population of endan- gered Japanese Marsh Warbler and vulnerable Japan- ese Reed bunting because the coexisting species have less prevalence of blood parasites. This study was partly supported by a Grant-in-Aid from the Japan Society for the Promotion of Science (No. 13480181) and a Global Environment Research Fund from Japan’s Ministry of the Environment. The junior author (NSS) thanks the Japan Research and Development Corporation for granting him a STA postdoctoral fellowship for foreign researchers in col- laboration with the Natural Science and Engineering Research Council of Canada. REFERENCES Atkinson CT & Van Riper III C (1991) Pathogenity and epizootiology of avian haematozoa: Plasmodium, Leucocytozoon, and Haemoproteus. In: Loy, Y.E. & Zuk M. (eds) Bird-Parasite Interaction. pp 19-48. Oxford University Press, New York. Bennett GF, Squires-Parson D, Siikamaki P, Huhta E, Allander K & Hilllstrom L (1995) A comparison of the blood parasites of three Fenno-Scandian popula- 74 tion of the Pied Flycatcher Ficedula hypoleuca. J Avian Biol 26: 33-38. Bishop MA & Bennett GF (1992) Host-parasite cata- logue of the avian Haematozoa, Suppl.1 and Bibliog- raphy of the avian blood-inhabiting Haematozoa. Suppl.2. Memorial University of Newfoundland, Oc- casional Paper in Biology 15. Clayton DH & Moore J (1997) Host-parasite evolution: General principles and avian models. Oxford Univer- sity Press, Oxford. Dale S, Kruszewicz A & Slagsvold T (1996) Effects of blood parasites on sexual selection and natural selec- tion in the pied flycatcher. J Zool 238: 373-393. Dufva R (1996) Blood parasites, health, reproductive success, and egg volume in female Great Tits Parus major. J Avian Biol 27: 83-87. McCallum H & Dobson A (1995) Detecting disease and parasite threats to endangered species and ecosys- tems. Trends Ecol Evol 10: 190-194. McClure HE, Poonswad P, Greiner EC & Laird M (1978) Haematozoa in the birds of eastern and south- ern Asia. Memorial University of Newfoundland, Newfoundland. Sodhi NS (1995) Parasitic infection of migratory species. Trends Ecol Evol 10: 417-418. Sodhi NS, Bennett GF & Nagata H (1996) Absence of blood parasites in the Japanese Reed Bunting Ember- iza yessoensis. Jpn J Ornithol 45: 115-117. Sodhi NS, Adlard RD, Nagata H & Kara AU (1999) Low prevalence of blood parasites in six Emberiza species in Japan. Jpn J Ornithol 47: 65-67. Sundberg J (1995) Parasites, plumage coloration and re- productive success in the yellowhammer, Emberiza citrinella. Oikos 74: 331-339. A List of Referees 2002 The editors of Ornithological Science extend hearty gratitude to the referees, who kindly and voluntarily helped the advancement of ornithology. The figures in parentheses denote the number of papers they refereed. Mitsuhiko ASAKAWA, Kazuhiro EGUCHI, Yuzo FUJIMAKI, Masahiro FUJIOKA, Shoji HAMAO, Mit- suhiro HAYASHIDA, Teruaki HINO (5), Masanobu HOTTA (2), Sadao IMANISHI, Ken ISHIDA, Akiko KATO, Noritomo KAWAJI, Kazuto KAWAKAMI, Kimiya KOGA, Yohsuke KOMINAMI, Reiko KURO- SAWA, Hajime MATSUBARA, Hiroshi MOMOSE, Masashi MURAKAMI, Hiroshi NAKAMURA, Kazuo NAKAMURA, Masahiko NAKAMURA, Sumio NAKAMURA, Isao NISHIUMI, Nariko OKA, Yuichi OSA, Yasuyuki SAITO, Hidetsugu SAKAI, Shigeho SATO, Tetsuro SHIMADA, Masaoki TAKAGI (2), Hitoshi TOJO, Yukihiko TOKUNAGA, Keisuke UEDA (3), Eiichiro URANO, Takeshi WADA, Yutaka WATANUKI, Ken YODA, Hoshiko YOSHIDA Rewriter: Mark BRAZIL Abstracts of the Japanese Journal of Ornithology, Volume 51 Noinber 1 held between 1998 and 2001 during the annual meet- ing of the Ornithological Society of Japan. The meet- ARTICLES ings were steered by cormorant researchers, in which the coexistence of people and Great Cormorants was Aims of the basic researches of Great Cormorant discussed. The special issues in this volume sum- (Phalacrocorax carbo), assessment of its impact marise the results of these meetings. and the damage control. Jpn J Ornithol 51: 1-3. 2002. Masae NARUSUE and Hisashi SUGAWA Changes in the distribution and abundance of the Great Cormorant Phalacrocorax carbo in Japan. Michio FUKUDA, Masae NARUSUE and Nanae KATO The Great Cormorant (Phalacrocorax carbo), a high- trophic level consumer in the wetland ecosystem, nests in the colonies sometimes in urban areas. Re- cent expansion of its distribution and the increase of _ the population size are presumably attributed to the The distribution and abundance of the Great Cor- _ changes both of the human activities and the wetland morant Phalacrocorax carbo in Japan have changed ecosystems in Japan. The Great Cormorants occurred markedly. Before 1920 cormorants were widely dis- widely throughout Japan before 1940, then both their tributed throughout Japan south of Hokkaido. From ranges and numbers decreased, so that by the begin- the end of 19th century to 1940s, cormorants de- | ning of the 1970s this species was considered as creased rapidly because of illegal hunting. After threatened. Subsequently, however, the population 1945, the expansion of human activities, develop- _ size has increased gradually, then during the 1990s ment, and water pollution were the causes of further _ the population underwent rapid increase in numbers decreases in the cormorant population and range. In and spread widely. As a consequence, they gave im- 1971, fewer than 3,000 cormorants bred in just three pacts on forests where they nest and inland fisheries. colonies in Japan. From the late 1970s onwards, num- In order to assess the extent of the current impacts bers of cormorants gradually began to increase; they and to control the damage, the four meetings were formed sub-colonies or roosts in areas close to these 75 three colonies. The main reasons for this increase are assumed to have included improvements to freshwa- ter quality, progress in freshwater purification (which led to increased fish stocks), and reduced disturbance of cormorants by people. During 1980s cormorants began to disperse widely due to the culling for pest control in Aichi, Gifu and Mie prefectures. At pres- ent, there are an estimated 50,000-—60,000 Great Cor- morants in Japan, occurring from Oita prefecture in the south, to Aomori prefecture in the north. Jpn J Ornithol 51: 4-11. 2002. Diet and foraging site selection of the Great Cor- morant in Japan. Kayoko KAMEDA, Takeshi MATSUBARA, Hiroshi MIZUTANI and Yoshihiro YAMADA During the 1990s, the number of Great Cormorant (Phalacrocorax carbo) in Japan has been increasing. As a consequence, there have been increasing con- flicts between cormorants and fresh water fisheries. We reviewed the species and size range of fish eaten by cormorants, examined their food requirements, and foraging site selection in the Kanto, Tokai, and Kansai areas of Honshu. Cormorants eat various fish species from fresh, brackish, or marine waters, de- pending on seasonal changes in food availability in each habitat. Cormorants generally eat fish measuring 3-30cm in length. Each individual cormorant re- quires approximately 500 g of fish each day. Individ- ual differences in foraging site selection were indi- cated by stable isotope analysis. A greater under- standing of the foraging ecology of cormorants is necessary in order to manage the population and/or behaviour of this piscivorous bird effectively. Jpn J Ornithol 51: 12-28. 2002. A review of studies on effects of the Great Cor- morant (Phalacrocorax carbo hanedae) colonies and roosts on forest ecosystem. Akira ISHIDA The present paper reviews current knowledge on the effects of Great Cormorant (Phalacrocorax carbo hanedae) nesting and roosting on forest ecosystems. The intensity of their effects was considered depend- ent on the duration of their residence and their activ- ity levels in forests. Field surveys and experimental studies showed that both plants and soils beneath colonies and roosts are affected by the deposition of feces, and damage to twigs and foliage (caused by ‘lapping and trampling, and by collection for nest ma- 76 terials). These activities may change the interactions among plants and between plants and soil, and may affect the community structure of soil animals and fungi. This change of the interactions among biologi- cal/chemical and physical factors resulted in the changes of the succession of the plant communities. The mechanism causing the decline and death of plants, the succession of forest vegetation, the dy- namics of plant community structure and biological interactions in the cormorant colonies and roosts are still unclear. A greater understanding of the basic ecological implications of cormorant nesting and roosting is essential in order to have management strategies for reducing damage on forests by cor- morants. Jpn J Ornithol 51: 29-36. 2002. The current status of dioxin pollution and its in- trinsic effects on Great Cormorants (Phalacroco- rax carbo) in Japan: an overview. Naomasa ISEKI, Jun HASEGAWA, HAYAMA and Shigeki MASUNAGA Shin-ichi In this paper we outiline the history of toxic contami- nants in wild birds in Japan. Pollution by dioxin and dioxin-like compounds has become a common issue in recent decades. As such pollution poses a consider- able health probem, countermeasures and technology to reduce the impacts are important. Very few papers have so fare focussed on the effects of dioxin and dioxin-like compounds on wild life in Japan. For the purposes of our research, we selected the Great Cor- morant (Phalacrocorax carbo). This fish-eating species nests colonially, and can be regarded as an in- dicator species of the effects of dioxins and dioxin- like compounds. We monitored cormorant health and compared it with published information on other. The cormorant residue levels were found to be higher than among other birds. The residue of PCDD/Fs consit- sted mainly of 2,3,7,8-substitution, in which 1,2,3,7,8- PeCDD and 2,3,4,7,8-PeCDF were the greatest con- tribution to toxic equivalency (TEQ). These com- pounds are accumulated more in the liver than egg and muscle. Based on the half-lives of dioxin and dioxin-like compound in the body of the cormorants, a decadal change of pollutant levels of their eggs was calculated using that of the environmental. It seems likely that embryo mortality, caused by dioxins, was the main toxic effect during the 1970s, but this declined dramatically over the following decades. We conclude that the estimated embryo mortality caused by PCDD/Fs and co-PCBs pollution (27%) was so small and would not impact population status. How- ever, studies of the other end points such as LOEL of enzyme activity and immunotoxicity are still needed. Our sample size was small and it is desirable to moni- tor large number of birds with unlethal techniques. Jpn J Ornithol 51: 37-55. 2002. Policy for the management of the Great Cor- morant in Japan. Shin-ichi HAYAMA The Great Cormorants (Phalacrocorax carbo) have recently gave impacts on forestry and fishery in Japan. To decrease the population of this species, culling was operated in many locations, which ap- peared not to be so effective. These human-cormorant conflicts have not been mitigate easily because so many factors are contributed. The special animal management planning system will be applied in the future under the Wildlife Protection and Hunting Law. Jpn J Ornithol 51: 56-61. 2002. The home range and flock size of the Azure- winged Magpie Cyanopica cyana during the non- breeding season in Nagano Prefecture. Sadao IMANISHI The home ranges and flock sizes of the Azure-winged Magpie Cyanopica cyana were studied during the non-breeding season in the Ina area (ca. 800 m alt.) from 1977 to 1981 and in the Nobeyama area (ca. 1350 m alt.) from 1980 to 1983 in Nagano Prefecture, central Honshu, Japan. The climate of the Nobeyama area was more severe in winter than in the Ina area. In the Ina area, the average home range size of ten flocks was 135.1 ha, while in the Nobeyama area five flocks average home ranges of 287.6ha. The fall, winter and spring ranges of one flock in Ina area Xal- most completely overlapped, but the range of the sin- | gle flock at Nobeyama area expanded during winter and contracted in spring. The home ranges of neigh- boring flocks partly, or largely overlapped. The aver- age size of ten flocks at Ina was 28.7 birds, while the average of five flocks at Nobeyama was 16.4 birds. The size of the Ina flock decreased over time, but that _ of Nobeyama remained the same. Population density was 9.6 birds/at Ina and 4.2 birds/at Nobeyama. Flock | sizes in both areas were unstable during October, and decreased gradually from November to April. The flock size was reduced from October to April by the 77 rate of 33.5% at Ina and 31.4% at Nobeyama, respec- tively. The difference in the home ranges of the flocks in the two areas may result from environmental fac- tors such as weather conditions and food availability during winter. Jpn J Ornithol 51: 62-73. 2002. The biology of Hazel Grouse Bonasa bonasia. Yuzo FUJIMAKI The Hazel Grouse (Bonasa bonasia) is a small forest grouse occurring in temperate and boreal forests from Scandinavia to the Far East. The species is assumed to have reached Hokkaido, northern Japan, via Sakhalin Island, during the last ice age about 40,000 years ago. The subspecies occurring in Hokkaido is now recognisably distinct as B. b. vicinitas. Pairs are formed from late March to early May. During this pe- riod males whistle actively. Six to ten eggs are laid in early or mid-May and hatch in early June after incu- bation of 23 to 25 days. Young attain adult size by late August and have adult plumage by mid-Septem- ber. The main diet consists of the leaves and seeds of herbaceous plants and trees and arthropods during late spring and summer, the buds of broad-leaved trees and vine fruits during autumn and winter, and buds and catkins in early spring. The Hazel Grouse has two large caeca supporting effective digestion of the plant fibers comprising their main diet. Hazel Grouse prefer broad-leaved and mixed forests with relatively dense undergrowth, and they avoid larch plantations in Hokkaido. Recently, the Hazel Grouse population has decreased in Hokkaido, the main cause of which is considered to be predation by the red fox (Vulpes vulpes), which increased in numbers from the early 1970s until the 1990s. Brood sizes were smaller during low population periods than dur- ing normal population periods. In order to maintain, or increase, Hazel Grouse population levels, habitat management and predator control is considered nec- essary. Jpn J Ornithol 51: 74-86. 2002. SHORT NOTE The first captured record of Willow Warbler Phyl- loscopus trochilus from Japan. Hisashi NAGATA, Hisahiro TORIKAI and Takema SAITOH Jpn J Ornithol 51: 87-91. 2002. Number 2 ARTICLES Seed dispersal of Styrax japonica by Varied Tits Parus varius on Miyake-jima, Izu Islands. Hiroshi HASHIMOTO, Takashi KAMIJYO and Hiroyoshi HIGUCHI Dispersal of Styrax japonica (a deciduous tree) seeds by Varied Tits Parus varius was observed on Miyake- jima, Izu Islands. Seven sample trees were monitored for 65-780 min. to examine seed consumption by birds. Varied Tits visited all the sample trees and car- ried off some of the fruits. The tits removed the toxic pulp and testae from the S. japonica fruits using their bills, and ate only the seed albumen. The tits stored some of the seeds close to the trees, but carried many others out of sight. Half of the stored seeds were placed on the ground, and these seeds presumably had a chance of germination, thus the Varied Tit may be an important seed disperser of S. japonica on Miyake-jima. Jpn J Ornithol 51: 101-107. 2002. Effects of brood size on chick growth and survival in early and late breeding Black-tailed Gulls in two years of different environmental conditions. Kenichi ISEKI and Yutaka WATANUKI To examine the effects of brood size on chick growth and survival, we manipulated the brood size of early and late breeding Black-tailed Gulls (Larus cras- sirostris) that usually lay two eggs, in two years of different environmental conditions at Rishiri Island, Hokkaido. Early breeders produced more chicks that grew faster than late ones, though there were no sig- nificant interactions between the effects of laying pe- riod and brood size. In 2000, the parents of enlarged 3-chick-broods produced greater number of 25-day- old chicks than those of reduced 1-chick-broods. In 2001 with heavy rain, however, the effects of brood size on the number of 25-day-old chicks were not sig- 78 nificant because the survival of 3-chick-broods was small. This indicated that the parents might not be able to care three chicks during bad weather. Body mass of chicks at 25-day-old was smaller for larger broods in both early and late breeders and in both years. In our study, parents might not have enough ability to provision three chicks. Jpn J Ornithol 51: 108-115. 2002. Breeding biology of the Great Cormorant Pha- lacrocorax carbo in Japan. Michio FUKUDA Great Cormorants Phalacrocorax carbo breed year round in Japan, though there are variations between colonies in the timing and duration of breeding. Males bring more nesting material than females do. The clutch consists of 2-3 eggs and both male and fe- male parents alternate incubation duties for a month. At the Shinobazu Pond colony, Tokyo, the chicks of nests on trees fledged in 31-59 days (average 45 days). Although essentially monogamous the cor- morants often mate with other partners even during the same breeding season. At the Shinobazu Pond colony, males start to breed at a younger age (1-6 years-old, average 2.1 years) than females (1-8 years- old, average 2.6 years). Jpn J Ornithol 51: 116-121. 2002. SHORT NOTE The first record of the Rose-coloured Starling Sturnus roseus from Okinawa Prefecture. Satoshi TOKOROZAKI, Kaori TOKOROZAKI and Eiki SUNAGAWA Jpn J Ornithol 51: 122-124. 2002. TECHNICAL REPORT A device to inspect woodpecker cavities. Shigeru MATSUOKA Jpn J Ornithol 51: 125-128. 2002. Instructions to Authors Ornithological Science is the English journal published by the Ornithological Society of Japan (OSJ) twice a year. Member- ship to OSJ is not required for contributors. The contribution can be classified into four categories: Reviews, Original NArticles, Short Communications and Comments. Comments are only limited to the papers published on Ornithological Science. Accepted papers automatically become the copyright property of OSJ and may not be reproduced, in whole or part, without a written permission from OSJ. Submission of Manuscripts Manuscripts have to be original and so far unpublished. The contributors are requested to present 3 complete photocopies (the original not required at the time of reception) of manu- script, tables and figures. Floppy disks (3.5-inch) containing a text file of the manuscript are also required after acceptance. Papers prepared by Microsoft WORD or PDF files, from coun- tries other than Japan, may be submitted via e-mails. Manuscripts are peer-reviewed by two anonymous referees. In case the contributors are requested to revise their manuscripts, they are supposed to resubmit then in consideration of the comments by reviewers within a month with a note describing how revision has been carried out. The final acceptance or re- jection for publication is decided by the Editor-in-Chief. The address for submission is: Dr. Teruaki Hino, Associate Editor Kansai Research Center, Forestry and Forest Products Re- search Institute, Kyoto 612—0855, JAPAN. E-mail: tkpk @affre.go.jp TEL: +81-75-611-1201, FAX: +81-75-611-1207 Organization of Manuscripts Manuscripts should be printed, double-spaced, on one side only of A4 sheets (210297 mm), with one-inch margins and 20 lines per sheet. Original Articles should be presented in the following order: (1) Title page (2) Abstract (3) Introduction (4) Materials and Methods (5) Results (6) Discussion (7) Ac- knowledgements (8) References (9) Tables (10) Figures and Legends. Short Communication is required not to exceed four printed pages and no abstract is needed. All manuscript sheets should be consecutively numbered. The use of footnotes is not recommended. A charge of 2,500 yen will be added to every page that ex- ceeds 20 printed pages. Each color page will be charged 50,000 yen. General Instructions 1) Cover page The title page should contain the title, author names, affilia- tions, addresses, contact name and address (phone and/or fac- simile number together with an e-mail address) for correspon- dence. Authors with different affiliations should be identified with superscripts. On the same page, a running title (less than 40 letters) and the number of tables and figures are also re- quired. To discriminate the family name(s) from the first and/or the middle owns, contributors are encouraged to spell their family names in all capitals. 2) Abstract and key words Abstracts should be less than 300 words, followed five or less key words in alphabetical order. 3) Names of animals and plants Scientific names of species should be given both in the ab- stract and in the article at the first mention. Scientific and Eng- lish names of birds should follow an authoritative source. Cap- itals should be used for the initial letters for each word of Eng- lish names (e.g., Black-billed Magpie) but not for a group name (e.g., Crows). 4) References References in the text: Tokugawa (1995) or (Tokugawa 1995) Oda and Smith (1996) or (Opel & Smith 1996) Mori et al. (1997) or (Mori et al. 1997; Richard et al.1999a) Journals: Yamaguchi N & Kawano KK (2001) Effect of body size on the resource holding potential of male varied tits Parus varius. Jpn J Ornithol 50: 65-70. Books: Cambell RC (1974) Statistics for biologists. 2nd ed. Cam- bridge Univ Press, London. Chapters in a book: Dawson WH (1996) Energetic features of avian thermoreg- ulatory responses. In: Carey C (ed) Avian energetics and nutritional ecology. pp 85-124. Capman & Hall, New York. Papers written in a language other than English: Yamagishi S (1981) Mozu no yomeiri — toshikouen no mozu noseitai wo saguru (The bridal of shrikes— ecological research of bull-headed shrikes in an urban park). Dai- Nippon-Tosho, Tokyo (in Japanese). Manuscripts accepted for publication but not yet published should be listed in the reference section as “in press”. All con- tributors are advised to prepare a list of references cited in the text arranged alphabetical by the author names. The sources of the reference should follow the commonly accepted abbrevia- tions for journal titles (refer to the ‘International List of Peri- odical Title Abbreviations’). The use of “in preparation,” “sub- mitted” or “personal communication” is not allowed in the ref- erence list. 5) Units and Abbreviations All measurements should be given in SI units. Time should be shown as 0930 and 1500 and the geographical location as 35°20'25"N, 136°10'20’E. Statistical symbols should be capi- talized except df (e.g., SD, SE, P, N). Italicized expressions are not preferred. 6) Tables and Figures The contributors are supposed to take the size of the printed page into consideration when they prepare their tables and fig- ures: the actual print is on 165223 mm in double columns with 80 mm width. ORNITHOLOGICAL SCIENCE Volume 2 Number1 February 2003 Contents SPECIAL FEATURE ORIGINAL ARTICLES Ecology of seed dispersal Fujita M & Kawakami K Ueda K Head-bobbing patterns, while walking, of Introduction. 1 Black-winged Stilts Himantopus himantopus Kominami Y, Sato T, Takeshita K, sin og a Manabe T, Endo A & Noma N Tajima K & Nakamura M Classification of bird-dispersed plants by Response to manipulation of partner fruiting phenology, fruit size, and growth contribution: A handicapping experiment in form in a primary lucidophyllous forest: the Barn Swallow. 65 an analysis, with implications for the conservation of fruit-bird interactions. bs SHORT COMMUNICATION Takanose Y & Kamitani T Nagata H & Sodhi NS Fruiting of fleshy-fruited plants and Low prevalence of blood parasites in five abundance of frugivorous birds: Phenological Sylviidae species in Japan. 73 correspondence in a temperate forest in central Japan. 25 A List of Referees 2002 is Hayashida M Seed dispersal of Japanese stone pine by the Abstracts of the Japanese Journal of Eurasian Nutcracker. 33 Ornithology, Volume 51 75 Fukui A Relationship between seed retention time in bird’s gut and fruit characteristics. 4] Rakotomanana H, Hino T, Kanzaki M & Morioka H The role of the Velvet Asity Philepitta castanea in regeneration of understory shrubs in Madagascan rainforest. 49 Published by the Ornithological Society of Japan Printed by Kokusai Bunken Insatsusha Co., Ltd. ISSN 1347-0558 ORNITHOLOGICAL SCIENCE Vol.2 No.2 September 2003 ~ The Ornithological Society of Japan ORNITHOLOGICAL SCIENCE Official journal of the Ornithological Society of Japan Editor-in-Chief Keisuke Ueda, Rikkyo University, Tokyo Associate Editors Teruaki Hino, Forestry and Forest Products Research Institute, Kyoto Hidetsugu Sakai, Nihon University, Tokyo Editorial Board Masahiko Nakamura, Joetsu University of Education, Joetsu Isao Nishiumi, National Science Museum, Tokyo Kazuo Okanoya, Chiba University, Chiba Eiichiro Urano, Yamashina Institute for Ornithology, Abiko Advisory Board Alexander V. Andreev, Institute of Biological Problems, Magadan Walter J. Bock, Columbia University, New York Jiro Kikkawa, The University of Queensland, Brisbane Woo-Shin Lee, Seoul National University, Suwon Bernd Leisler, Max-Planck-Gesellschaft, Radolfzell Anders P. Moller, Universite Pierre et Marie Curie, Paris Richard Noske, Northern Territory University, Casuarina Pilai Poonswad, Mahidol University, Bangkok Lucia Liu Severinghaus, Academia Sinica, Taipei Navyjot S. Sodhi, National University of Singapore, Singapore Jeffrey R. Walters, Virginia Polytechnic Institute and State University, Blacksburg John C. Wingfield, University of Washington, Seattle Jeong-Chil Yoo, Kyung-Hee University, Seoul Editorial Policy Ornithological Science publishes reviews, original articles, short communications and comments covering all aspects of ornithology. Manuscripts are judged on the basis of their contribution of original data and ideas or interpretation. All articles are peer-reviewed by at least two researchers expert in the field of the submitted paper. Manuscripts are edited where necessary for clarity and economy. Ornithological Science aims to publish as rapidly as is consistent with the requirements of peer-review and normal publishing constraints. Submission Manuscripts and editorial correspondence should be ad- dressed to: Dr. Teruaki Hino, Associate Editor, Forestry and Forest Products Research Institute, Kyoto 612-0855, Japan. Tel: +81-75-611-1201, Fax: +81-75-611-1207, e-mail: tkpk @affre.go.jp For detailed instructions concerning the submission of man- uscripts, please refer to the Instructions to Authors at the in- side of the back cover of each issue of the journal or visit the OSJ web page: http://wwwsoc.nii.ac.jp/osj/ Subscriptions Membership of the Ornithological Society of Japan (OSJ) is open to anyone interested in ornithology and the aims of the OSJ. The annual fee for an ordinary member is Japanese Yen 5,000. Members are entitled to receive two journals. Ornithological Science (in English; two issues per year) and Japanese Journal of Ornithology (with English summary; two issues per year). To join the OSJ, please apply to: The OSJ Office, Laboratory of Biodiversity Science, School of Agricultural and Life Sciences, The University of Tokyo, Yayoi 1-1-1, Tokyo 113-8657, Japan. Tel: +81-3-5841-7541, Fax: +81-3-5841-8192, e-mail: osj@lagopus.com Copyright Submission of a manuscript implies: that the work de- scribed has not been published before; that is not under con- sideration for publication elsewhere; that its publication has been approved by all co- authors, if any, as well as by the re- sponsible authorities at the institute where the work has been carried out; that, if and when the manuscript is accepted for publication, the authors agree to automatic transfer of the copyright to the OSJ; and that the manuscript will not be pub- lished elsewhere in any language without the consent of the copyright holders. ISSN: 1347-0558. All articles published in the Ornithological Science are pro- tected by copyright, which covers the exclusive rights to repro- duce and distribute the article (e.g. as offprints) as well as all translation rights. No material published in this journal may be reproduced photo-graphically or stored on microfilm, in elec- tronic data bases, video disks, etc., without first obtaining writ- ten permission from the copyright holders. Special requests should be addressed to: Dr. Keisuke Ueda, Editor-in-Chief, Laboratory of Animal Ecology Faculty of Science, Rikkyo University, Ikebukuro, Tokyo 171-0021, Japan, Tel: +81-3-3985-2596, Fax: +81-3-3985-2596, e-mail: keisuke @rikkyo.ac.jp Printer: Kokusai Bunken Insatsusha Co., Ltd., Takada-no-baba 3-8-8, Shinjuku-ku, Tokyo 169-0075, Japan. Tel: +81-3-3362- 9741, Fax: +81-3-3368-2822. Cover design: Eiichiro Urano EA 2004 1A1A8~2005412A 3148 Hifel 20038 710A 13 A8~2003411A1248 fiz A 2003 4 11A148 pain BAAS MSA AGH 1005A 1. BRAA HACK 0k 249%), AKER 248, RHEE ARB {e482 $80 [GA HIGUCHI, Hiroyoshi fet 2438 ® ( fee ) aft 5 2. FMARE HECK 193918 192%), AME 139 RUE AREER (SRAMSAt+SIA) 1. FR dt UEDA, Keisuke 87 4. EK 2. Bee +41 FUJIMAKI, Yuzo 66 4. REED 3. FA ISHIDA, Ken 59 4. REA Sh (UUs Gey YAMAGISHI, Satoshi 59 4. EK 5. He NAKAMURA, Hiroshi 58 4. ER 6. RR iEt8 = FUJIOKA, Masahiro 55 a, REE AKER 7. ER #- KARASAWA, Koichi 5S 4. EKER 8. iL 8 EZAKI, Yasuo 47) - Sah, EK 9. et HEH NAKAMURA, Masahiko 438 42. WEAR 10. JIS BilA = KAWAdI, Noritomo 41 4. WER dite, Sel Ale EGUCHI, Kazuhiro 38-4, AER 11. lB GABA = OZAKI, Kiyoaki 38-4. AER 13. KA mls NAGATA, Hisashi 35 4. EKER 14. pues ABE, Manabu 33 4. ER 15: ene WATANUKI, Yutaka 32 RA mO LA HIGUCHI, Hiroyoshi 19 #k4Az Election Results of President and Executives AAR Fark ABARSGR BARTSRACHEAS RE ABS Ht 2(A,. AR 1) 54( A, AR 20) SBOE 1 WII SRISHMACESICMECN CUETO, 14 2MNARLGVET. ERROIED, MODVROBBA 1 BHVELT. alse ie Fa ETS Oe BARSTSSh FRG AREOPRICLLOE, RMFMAOMCRHLLGE, RMORISRLE THRE ICIS UTFODRMEENEL. FEMA SaTe A CRIC< 2004 £1A18 AS 2005 # 12 A 31 RECT. alak pet oi Hate (ASIA) +H Bit KB eis ef] IE te NAKAMURA, Hiroshi UEDA, Keisuke NAGATA, Hisashi FUJIOKA, Masahiro ~ irebice9 to efluesh noffosla my Bae SFA G Fie a i ay 7 ~— v g : yy Kea * ar i in ” = . ru x = Ornithol. Sci. 2: 79-88 (2003) INVITED ARTICLE Adaptations and maladaptations to island living in ae Seychelles Warbler ces ze OQ ¢ i. N Jan KOMDEUR* ef ; Animal Ecology Group, Zoological Laboratory, University of Groningen, P.O. Box 14, 9750 AA Haren, ny | atlerdan Abstract The Seychelles Warbler (Acrocephalus sechellensis) was an endangered endemic of the Seychelles islands where, until 1988, the entire population of ca. 320 birds was restricted to the one island of Cousin Island (29 ha). Although warblers can breed independently in their first year, some individuals remain in their natal territory as subordinates, and often help by providing nourishment to non-descendent off- spring. The frequency of helping is affected by habitat saturation, variation in terri- tory quality (insect prey availability), and the genetic relatedness between the helper and the offspring. Helping results in indirect benefits from enhancing the reproductive success of close relatives, and direct benefits as improved parental skills and the ac- quisition of parentage. The overall helping benefits are higher for daughters than for sons, and it is therefore no wonder that most helpers are daughters from previous broods. Furthermore, on low-quality territories breeding pairs raising sons gain higher fitness benefits than by raising daughters, and vice versa on high-quality territories. Female breeders adaptively modify the sex of their single-egg clutches according to territory quality: male eggs on low quality and female eggs on high quality. However, despite the saturated nature of the Cousin population, the possibility of obtaining higher reproductive success on new nearby island, and a well developed flight appara- tus, inter-island dispersal by Seychelles Warblers is extremely rare. The Seychelles Warbler is a beautiful example of behavioural and life history adaptations and mal- adaptations to restricted circumstances. ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2003 Key words Co-breeding, Competition, Cooperative breeding, Fitness benefits, Sey- chelles Warbler, Territory inheritance The evolution of flight was a major innovation for birds and insects, permitting them to disperse from inhospitable sites, to exploit habitats unavailable to terrestrial predators and to forage in three dimensions and over very large areas. Some island species, how- ever, refuse to disperse to other relatively close suit- able islands despite the fact that they are capable of sustained flight (Diamond 1981; Raikow 1985). Other island species cannot disperse to other islands, because they have lost their flight capacity. The ca- pacity to fly is an energetically costly affair as the muscles that constitute the flight apparatus of birds represent 17 to 25% of body mass (Greenewalt 1975), and the metabolic demands of these special- (Received 22 August 2003; Accepted 23 September 2003) * E-mail: j.komdeur@biol.rug.nl Proceeding of the Symposium on Behavioral Studies and Conserva- tion Biology. 15 September 2002, Tokyo; Organized by Go Fujita and Hisashi Nagata. WS) ized flight muscles rank among the highest of all tis- sues (e.g., Weber & Piersma 1996). It therefore comes as no great surprise that under particular eco- logical conditions some species of bird and insect have lost their morphological flight capacity (e.g., Di- amond 1981; Livezey 1992a,b; Trewick 1997). The majority of flightless species live on islands with low resource availability and without mammalian preda- tors. As the capacity to fly is costly (Greenewalt 1975; Weber & Piersma 1996), the loss of flight may be a consequence of selection for reduced energy ex- penditure in the face of intra-specific competition for food, or because the benefits of flight (viz. predator escape) no longer apply in these habitats (e.g. Roff 1990, 1994; Wagner & Liebherr 1992; McNab 1994a,b). However, in several isolated island popula- tions that have reached carrying capacity, a consider- able fraction of the reproductively mature birds (‘sur- plus’ birds) is constrained from breeding (e.g., J. KOMDEUR Komdeur 1996a; Holt & Martin 1997; Grant et al. 2000; Kokko & Lundberg 2001) and may never pro- duce offspring over their life time (Komdeur 1992). Mature individuals delaying independent breeding may instead provide care to young that are not their own genetic offspring. The most common form of co- operative breeding involves a breeding pair being as- sisted by offspring from previous broods. Evolution- ary theory is based on the concept that individuals are selected on their ability to efficiently translate re- sources into reproductive success, thereby maximiz- ing their genetic contribution to future generations (Hamilton 1964; Maynard Smith 1964). As such helpers can increase their inclusive fitness by gaining either indirect fitness benefits from enhancing the re- productive success of close relatives, or direct bene- fits from increased opportunity for their own future reproduction (e.g. Emlen & Wrege 1989; Mumme et al. 1989; Emlen 1991, 1995; Mumme 1992; Koenig et al. 1992; Cockburn 1998). If helping results in higher fitness than dispersal, individuals should delay dispersal and help. However, in most species the fit- ness benefit from helping is likely to be considerably less than the potential direct genetic gain from imme- diate independent breeding if a territory and mate could be obtained (Brown 1987). Thus cooperative breeding can be seen as a ‘best-of-a-bad-job’ strategy, adopted when opportunities for independent breeding are limited. However, grown offspring may gain in- clusive fitness benefits by remaining on the natal ter- ritory. To test the fitness benefits of delayed dispersal and helping in the field, detailed knowledge of the fit- ness functions for males and females is required (Leimar 1996; Lessells 1998; Lessells et al. 1998; Koenig & Walters 1999; Pen & Weissing 2000). These data are not available for most species. In this article I will describe the adaptations and maladapta- tions to island living in the Seychelles Warbler (Acro- cephalus sechellensis). The Seychelles Warbler is a rare island endemic, which from 1920 to 1988 occurred only on Cousin Is- land (29ha) in the Seychelles (Fig. la). I will de- scribe the fitness consequences of cooperative breed- ing for the helpers and for their parents, and address some adaptations employed by helpers and parents to enhance their fitness. The Seychelles Warbler is an excellent model species to test the fitness benefits of helping and dispersal behaviour, because of: (1) the ability to measure local dispersal in detail (Komdeur 1992, 2003), (11) the availability of molecular markers lo assess sex, parentage, and coefficients of related- 80 ness between individuals in the breeding group (Richardson et al. 2000, 2001), and (iii) the wealth of long-term data allowing accurate fitness measures for each individual (Komdeur 2003). MATERIALS AND METHODS 1) Study populations and data collection The Seychelles Warbler is a small (15-16 gram) insectivorous bird (Fig. 1b). It is a rare island en- demic, which occurred on several islands in the Sey- chelles (Oustalet 1878; Diamond 1980) until anthro- pogenic disturbance in the 20th century reduced them to just one population on Cousin Island (Fig. 1a; Col- lar & Stuart 1985) where by 1967 only ca. 30 individ- uals remained (Crook 1960; Lousteau-Lalanne 1968). Because the warbler was threatened with extinction, Cousin was purchased in 1968 for Bird Life Interna- tional with the express aim of saving the warbler. The native vegetation was allowed to regenerate resulting in the warbler population growing to its carrying ca- pacity of 320 birds by 1982. The increase in number of territories showed the same trend and reached its saturation level of ca. 115 territories in 1981 (Fig. Ic). The island was completely covered by territorial groups with no empty spaces (Fig. Id). Although warblers can breed independently in their first year, some individuals remain in their natal territories as subordinates, and often assist the breeding pair pro- viding nourishment to offspring (Fig. 1b). Habitat saturation has lead to cooperative breeding in this species (Komdeur 1992). Once paired, warblers re- side permanently in the same territories, sometimes for as long as nine years. On Cousin Island the war- bler usually has a clutch size of one egg only, and adult birds have high annual survival (81%). In 1988 and 1990 respectively, new populations of Seychelles Warblers were established by moving 29 birds to both Aride Island (9km from Cousin) and to Cousine Is- land (1.6 km from Cousin) (Fig. la). These transloca- tions were highly successful, with the founder popu- lations expanding to a combined population of ca. 1,750 individuals by 1996 (Komdeur 2003). The Cousin Island population of Seychelles War- blers (ca. 320 birds) has been intensively studied since 1985, while the populations on the islands of Aride and Cousine have been studied from establish- ment. Since 1985 nearly all birds on Cousin Island have been banded for individual recognition and, since 1993, blood sampled for molecular sex and parentage analyses. Off-island migration by warblers ae Ctiitwwww.... Adaptations and Maladaptations in the Seychelles Warbler Aride td Cousin ** Félicité es e a li. j > ‘ Cousine Praslin Marianne AFRICA +E Seychelles f Fig. 1. first transfer (September 1988), and Cousine as place of second transfer (June 1990). (b) A complete breeding group of Seychelles Warblers on high-quality territory: a breeding pair, three helpers and their offspring (twins) (from left to right: male (1 year, twin), mother (7 years), female helper (4 years), female helper (2 years), female (1 year, twin) father (6 years) and female helper (2 years); photograph: DS Richardson). (c) Increase in Seychelles Warbler numbers and territories on Cousin Island since 1959. (d) Map of Cousin Island with Seychelles Warbler territories divided into three quality categories: low-, medium- and high-quality territory (from Komdeur 1992). is negligible (0.10%; Komdeur et al. in press), so birds that have disappeared are considered dead. Each year (1985-2002) nearly all breeding attempts were monitored and activity by resident birds ob- served during the nest building, incubation and nestling periods (Komdeur 1996b; Richardson et al. 2002). Observations on incubation and food provi- sioning were made at all breeding attempts to deter- mine the status of the birds within each territory. The primary male and female were defined as the domi- nant, pair-bonded male and female in the territory, while the term ‘subordinate included all other birds 81 (c) 350 300 250 120 l 200 on 2 150 80 | a o a) 100 ro) 40 = 50 2 T T T 0 1960 65 70 75 80 85 90 95 2000 year Territory quality: high medium low (a) Map showing the Seychelles islands. Populations of Seychelles Warblers on Cousin, Aride as place of (> eight months old) resident in the territory. Subor- dinates were split into three categories: non-helpers, subordinate helpers (non-parents) and subordinate parents. Playing recorded songs at different locations to provoke territory defence behaviour by the focal birds identified territory boundaries. Territory size was assessed using a compass and aerial photo- graphs. Because the warblers are insectivorous, terri- tory quality was expressed in terms of insect prey availability (Komdeur 1992) a factor that has been assessed on a monthly basis within each breeding season. Territories were classified into three territory- J. KOMDEUR quality categories—low, medium or high (Komdeur 1992). 2) Molecular analyses The sex of each individual within the study popu- lations was determined using a PCR based method (Griffiths et al. 1998). Genotypes were identified for individuals in the Cousin population using 14 mi- crosatellite markers (Richardson et al. 2000). Coeffi- cients of relatedness between individuals were deter- mined using KINSHIP (Goodnight & Queller 1999) and used to determine, in terms of genetic equiva- lents, the exact direct and indirect benefits gained by the subordinates (Richardson et al. 2002). Using CERVUS (Marshall et al. 1998) parentage was as- signed with high confidence (> 90%) to all offspring sampled between 1997-2000 (Richardson et al. 2001). RESULTS AND DISCUSSION 1) Benefits of helping to subordinates: difference between the sexes Indirect fitness benefits have been suggested to be a major selective force behind the evolution of coop- erative breeding (Emlen & Wrege 1989; Mumme et al. 1989; Mumme 1992; Koenig et al. 1992, Emlen 1997). These benefits accrue if helping obeys two re- quirements. First, helping should result in improved survival and reproductive success of the breeding pair. Second, subordinates should preferentially help more closely related kin. In the Seychelles Warbler, removal experiments showed that on high-quality ter- ritories the helping behaviour of subordinates in- creased the reproductive success of the breeding pair (Komdeur 1994a). Furthermore, subordinates were both more likely to become helpers, and provided more help when they were more closely related to the recipient (Komdeur 1994b). However, in this study relatedness was estimated from pedigree data, which was shown to be inadequate when, a decade later, using microsatellite based genotyping we found that complex patterns of shared reproduction and/or extra- group paternity occur in the Seychelles Warbler (Richardson et al. 2001, 2002). Furthermore, it was found that females always become subordinates on their natal territory (N=43), whereas a significant percentage of males (25%, N=20) became subordi- nates on non-natal territories (y7=8.51, P<0.05; Richardson et al. 2002), suggesting that the indirect benefits of helping are more important for females 82 0.40 > 0.30 + 0.20 + 0.10 + 0.00 -0.10 + -0.20 + Subordinate-nestling relatedness -0.30 + -0.40 - Females (o Helpers O Non-helpers | Fig. 2. The relatedness (r) of female and male subordinates to helped or non-helped nestlings. Bars indicate means +SE * P<0.05, ** P<0.01 (from Richardson et al. 2003a,b). than for males. Using microsatellite markers to calcu- late precisely coefficients of genetic relatedness be- tween individuals (see Richardson et al. 2000), we found that female subordinates without parentage ac- curately maximised their indirect benefits by prefer- entially feeding more related offspring, produced by relatives of the female subordinate (Fig. 2; Richard- son et al. 2003a,b). On the other hand, the amount of help provided by male subordinates was low and in- dependent of relatedness to offspring (Fig. 2; Richardson et al. 2003b). In the Seychelles Warbler unrelated subordinates (born in other groups) do sometimes help (DS Richardson, pers. comm.) suggesting that direct bene- fits, such as improved parental skills or gaining parentage (Koenig et al. 1992; Cockburn 1998), or territory inheritance (Emlen 1991; Koenig et al. 1992; Balshine-Earn et al. 1998) are important too. To test whether helping improved the reproductive success of subordinates that later become breeders, we translocated male and female warblers of the same age but with different degrees of previous breeding experience, to the islands of Aride and Cou- sine. Individuals were categorised as: (i) experienced breeders that had fledged young of their own in a pre- vious year; (ii) experienced subordinates with no breeding experience but with helping experience; and (ii1) inexperienced birds that had neither helping nor breeding experience (see Komdeur 1996b). On the new islands birds with helping experience that were paired with an experienced partner produced their first fledgling as fast as experienced breeders, and significantly faster than inexperienced birds paired Adaptations and Maladaptations in the Seychelles Warbler FIRST FLEDGLING [EE female [___] male @ first nest-attempt since pair-formation or independent young 2) x o ® = 2) =< fo) o S inexperienced helping breeding experience experience Fig. 3. (A) The number of weeks between pair formation and the production of the first nest and first fledgling on the is- lands of Aride and Cousine plotted for male and female birds with different past experience. (B) The number of weeks be- tween independence of first fledgling and the production of a new nest and second fledgling on the islands of Aride and Cousine plotted for the same male and female birds as in (A), who have become experienced breeders. In both figures, the birds were between 3 and 7 years of age (i.e., the period during which there are no age effects on reproduction), paired with the same experienced breeding partner, and did not receive as- sistance from helpers. Statistically significant comparisons de- termined by Mann-Whitney U tests. Only significant differ- ences are plotted. Error bars represent one standard error (from Komdeur 1996b). with an experienced partner (Fig. 3; Komdeur 1996b). Females with helping or breeding experience built better nests and spent more time incubating than inexperienced females. During this period no subor- dinates assisted any of the breeding pairs. Pairs com- prising a male with breeding experience and an inex- perienced female, took four times longer to produce their first fledgling than pairs consisting of a female with breeding experience and an inexperienced male (Fig. 3). This is probably because only females build nests and incubate the clutch. Once inexperienced birds had fledged young and had, therefore, acquired 83 [9 direct breeding benefits [] indirect benefits benefits (offspring equivalents) both sexes male subordinates female Fig. 4. The fitness benefits of cooperative breeding gained by female and male subordinates in the Seychelles Warbler (1997-1999). Statistical significance assessed by Mann-Whit- ney Z statistic. Both female (N=43) and male (N=20) subor- dinates gain significantly higher direct breeding benefits (open columns) compared to indirect benefits (filled columns). Di- rect breeding benefits are significantly higher in females than in males, but there is no significant difference between the sexes in indirect breeding benefits. Error bars represent one standard error (from Richardson et al. 2002). breeding experience, they subsequently improved their breeding success by producing a second fledg- ling in the same time interval as birds with either helping or breeding experience (Fig. 3). Another direct benefit gained by subordinate Sey- chelles Warblers is the acquisition of parentage within the breeding group by sneaking their own eggs into their mother’s nest (44% of 43 female subordi- nates; Richardson et al. 2001, 2002). Subordinate males gained significant less often parentage within the group (15% of 20 male subordinates) than fe- males (Fisher’s exact test, P=0.024; Richardson et al. 2002). None of the subordinate females gained repro- ductive success through egg dumping in other territo- ries and none of the male subordinates gained extra- pair fertilisations with females from other groups. Overall direct benefits were significantly higher than indirect benefits, although this difference was more extreme in females (Fig. 4; Richardson et al. 2002). These results show that females remaining on the natal territory as helpers gain greater benefits than male helpers and this may explain why most helpers are female (percentage subordinates that are female: 88% (Komdeur 1996c) and 68% (Richardson et al. J. KOMDEUR eggs nestlings males 1.0 S) ) ° £ Lop) Cc So De} proportion male nestlings A) LOW MEDIUM HIGH 2 ro} LOW MEDIUM’ HIGH 11 14 17 8 7 1 12 14 6 7 6 2 11 3 0 1994 1995 © LOW MEDIUM HIGH territory quality Fig. 5. Sex ratio of nestlings produced by Seychelles Warbler pairs in relation to quality class of breeding territory (tq classes: low-quality territory, medium-quality territory; and high-quality territory; 1993-1995). No additional young were present on the territory. Young were hatched from one-egg clutches only in different years (A, 1993: N=46, G-test of independence: D= 12.23, df=1, P=0.0005, proportion male=1/(1+e 7), z=—2.68+1.36 (tq class); B, 1994: N=45, D=12.03, df=1, P=0.0005, Z=22.68+1.37 (tq class); C, 1995: N=27, D=12.99, df=1, P=0.0003, z=—3.60+2.10 (tq class)). Shaded area represents the maximal and minimal values for the sex ratio assuming that all unhatched eggs were male, or females, respectively (from Komdeur et al. 1997) 2002)). 2) Adaptive manipulation of egg’s sex Given that helpers on high-quality territories in- crease their parents’ reproductive success (Komdeur 1994a), it may be that primary birds on high-quality territories preferentially accept or produce female subordinates to increase their fitness. On the other hand, given that helpers on low-quality territories re- duce their parents’ reproductive success (Komdeur 1994a) females on low-quality territories may prefer- entially produce males, which disperse in order to avoid having future helpers. Helpers are mostly fe- males, and females (in birds females determine the sex of an offspring because they are the heteroga- metic (ZW) sex) may adjust the sex of their single egg to territory quality (Komdeur 1996c). Reviews from the 1980s were unanimous in the belief that fac- ultative adjustments of offspring sex ratio at birth were unlikely to occur in vertebrates (Williams 1979; Charnov 1982; Clutton-Brock 1986; Bull & Charnov 1988). Sex determination is almost ubiquitously asso- ciated with chromosome heterogamety, constraining the physiological or genetic mechanisms for skewing the sex ratio at birth (Williams 1979, 1992; Krackow 1995). The occurrence of adaptive sex ratio manipu- lation at laying in birds has, therefore, been ques- tioned. At the time we were able to demonstrate that 84 Seychelles Warblers adaptively modify the sex of their single egg. The fraction of male eggs produced by unassisted Seychelles Warbler mothers changed significantly with territory quality (Fig. 5; Komdeur 1996c; Komdeur et al. 1997). Unassisted females on high-quality territories produced 88% female eggs (N=32), whereas unassisted females on low-quality territories produced 77% males eggs (N=57), the dis- persing sex (Komdeur et al. 1997). This was con- firmed by experiments: (i) Helper removal experi- ments confirmed that sex ratio bias was for the pur- pose of producing helpers. When females on high- quality territories had their female helpers experi- mentally removed they switched from producing all sons to producing 83% females (Komdeur et al. 1997). (ii) Breeding pairs that were transferred from low- to high-quality territories switched from the pro- duction of male to female eggs (Komdeur et al. 1997). In the short term, the manipulation of the off- spring’s sex is directly adaptive from the perspective of the breeding pair. The next step is to determine whether egg sex-ratio manipulation results in long- term inclusive fitness benefits for the breeding pair. 3) Inter-island dispersal The co-operative breeding and sex-allocation sys- tem of the Seychelles Warbler is a beautiful example of behavioural and life history adaptations to re- Adaptations and Maladaptations in the Seychelles Warbler stricted circumstances. However, a paradox still re- mains. Genetic parentage analysis has demonstrated that a substantial proportion of adult female and male birds died without having produced offspring of their own (12%; D.S. Richardson, unpublished data). These warblers would have done better if they had colonised the suitable islands of Aride and Cousine by themselves, where initially the annual production of yearlings by breeding pairs that were artificially translocated to Aride and Cousine was, on average, 14 times higher than before the translocation (Komdeur 1996d). These islands have been suitable for Seychelles Warblers for at least 25 years. Despite the saturated nature of the Cousin Island population and the possibility of obtaining higher reproductive success on new islands (Komdeur 1996d), inter-is- land dispersal by Seychelles Warblers is extremely rare. During the 20 years of study only two warblers (0.10%, N=1,924) have been observed to fly be- tween these islands (Komdeur et al. in press). Energy conservation is proposed to be the selection pressure behind reduced flight ability and, compared to closely related flying species, flightless birds usually have smaller pectoral muscles, relative to body mass, wing shape and wing loading (Livezey 1990; 1992a,b; McNab 1994b; McCall et al. 1998). The Seychelles Warbler does not fit the predictions of the energy con- servation theory, as neither the average pectoral mass relative to body mass, the skeletal attachment area, the wing shape nor the wing loading is distinct from those of its migratory relatives (Komdeur et al. in press). Seychelles Warblers show the morphological structures required for sustained flight, but still do not manage to disperse successfully to relatively close suitable islands. Seychelles Warblers appear, there- fore, to have a behavioural reluctance to cross what they may regard as extensive bodies of water. This phenomenon of ‘psychological flightlessness’ occurs in other species that are capable of sustained flight (Diamond 1981, Raikow 1985). Given the historical presence of discrete island populations of Seychelles Warbler and the vulnerability to extinction of such populations, we would not expect the Seychelles Warblers to have lost the ability or willingness to dis- perse across water to colonise new islands. The anatomical and behavioural bases of dispersal are subject to natural selection (Diamond 1981). There- fore, one would expect that the ability to disperse over water, and settle on nearby suitable islands (where, as shown by the translocations, individuals could gain higher fitness) should have been main- 85 tained in this species. The reluctance to do so could have developed a long time ago when, perhaps, all Seychelles islands were fully occupied by warblers. Under these conditions, dispersal over water was not an adaptive strategy as all islands may have been sat- urated with warblers. Perhaps it is only now, after 100 years in which warblers have disappeared from all but one island, that flying across water to breed on other islands would be adaptive. The case of the Sey- chelles Warbler may exemplify the inability of natu- ral selection to plan ahead! CONCLUDING REMARKS The Seychelles Warbler is atypical in cooperative breeding birds in that females are more likely than males to become subordinates. Males typically dis- perse. Our results show that female subordinates gain significantly higher inclusive fitness benefits than male subordinates. Female subordinates remain on their natal territory and obtain higher inclusive repro- ductive success by helping closely related relatives, by co-breeding within the group, and through im- proved future parenting ability. Males often become subordinates on non-natal territories and so do not gain indirect reproductive success by helping. The higher inclusive fitness benefits accruing to female subordinates may explain why primary females often skew the sex ratio towards producing female off- spring (Komdeur et al. 1997). With female subordi- nates remaining on the natal territory the primary fe- male gains both an increase in her own productivity and also indirect benefits associated with subordinate females breeding. The presence of several female subordinates in the group may, however, be a disad- vantage to the primary female, because her direct fit- ness declines due to competition for food and in- creased nest failure (mainly egg breakage due to in- creased pressure from simultaneous incubation; Komdeur 1994b). Her indirect fitness may also de- cline because of increased competition over local breeding vacancies between subordinate female rela- tives. At this stage female offspring should refrain from becoming subordinates and disperse. The bal- ance between these forces should determine whether offspring become subordinates or not. There is good evidence that the inclusive fitness consequences are higher for female subordinates than for male subordinates in the Seychelles Warbler. However, the long-term inclusive fitness functions for subordinates and non-subordinates of both sexes J. KOMDEUR should be calculated using molecular parentage analyses and precise coefficients of genetic related- ness, in order to predict under which circumstances males and females should become subordinates. ACKNOWLEDGEMENTS I thank Prof. Hidetsugu Sakai (the organizing com- mittee of 2002 Annual Meeting of Ornithological Soci- ety, Japan), Prof. Hiroyoshi Higuchi and Go Fujita (Uni- versity of Tokyo, Japan) and Dr. Hisashi Nagata (Na- tional Institute for Environmental Studies, Japan) very much for inviting me to present this paper at the 2002 symposium ‘Behavioral Ecology and Conservation Bi- ology’ organized by the Ornithological Society of Japan, in Tokyo. I also want to thank them for their ex- treme kind hospitality during my stay. I thank Mari Takeda very much for doing a fantastic job by translat- ing my talk into Japanese life. Furthermore, I thank the other conference speakers, Drs. Johanna Pierre, Masato Minami, Katsumi Ushiyama and Tatsuya Amano for their nice company throughout. REFERENCES Balshine-Earn S, Neat FC, Reid H & Taborsky M (1998) Paying to stay or paying to breed? Field evi- dence for direct benefits of helping behavior in a co- operatively breeding fish. Behav Ecol 9: 432-438. Brown JL (1987) Helping and communal breeding in birds. Princeton University Press, Princeton. Bull JJ & Charnov EL (1988) How fundamental are Fisherian sex ratios? Oxford Surv Ecol Biol 5: 96-135. Charnov EL (1982) The theory of sex allocation. Prince- ton University Press, Princeton. Clutton-Brock TH (1986) Sex ratio variation in birds. Ibis 128: 317-329. Cockburn A (1998) Evolution of helping behaviour in cooperatively breeding birds. Annu Rev Ecol Syst 29: 141-177. Collar NJ & Stuart SN (1985) Threatened birds of Africa and related islands. ICBP/IUCN Red Data Book, Part 1. International Council for Bird Preserva- tion, Cambridge. Crook JH (1960) The present status of certain rare land- birds of the Seychelles islands. Unnumbered Sey- chelles Government Bulletin. Diamond AW (1980) Seasonality, population structure and breeding ecology of the Seychelles brush warbler Acrocephalus sechellensis. Proc Pan—Afr Orn Cong 4: 253-266. Diamond JM (1981) Flightlessness and fear of flying in 86 island species. Nature 293: 507-508. Emlen ST & Wrege PH (1989) A test of alternate hy- potheses for helping behavior in white-fronted bee- eaters of Kenya. Behav Ecol Sociobiol 25: 303-320. Emlen ST (1991) Evolution of cooperative breeding in birds and mammals. In: JR Krebs & NB Davies (eds) Behavioural Ecology: an evolutionary approach. pp 301-337. Blackwell, Oxford. Emlen ST (1995) An evolutionary theory of the family. Proc Natl Acad Sci USA 92: 8092-8099. Emlen ST (1997) Predicting family dynamics in social vertebrates. In: JR Krebs & NB Davies (eds) Behav- ioural Ecology: an evolutionary approach. pp 228- 253. Blackwell, Oxford. Goodnight KF & Queller DS (1999) Computer software for performing likelihood tests of pedigree relation- ship using genetic markers. Mol Ecol 8: 1231-1234. Grant PR, Curry RL & Grant BR (2000) A remnant population of the Floreana mockingbird on Champion island, Galapagos. Biol Conserv 92: 285-290. Greenewalt CH (1975) The flight of birds. Trans Am Phil Soc 65: 1-67. Griffiths R, Double MC, Orr K & Dawson RJG. (1998) A DNA test to sex most birds. Molec Ecol 7: 1071-1075. Hamilton WD (1964) The genetical evolution of social behaviour I and II. J Theor Biol 7: 1-52. Holt RF & Martin K (1997) Landscape modification and patch selection: The demography of two secondary cavity nesters colonizing clearcuts. Auk 114: 443-455. Koenig WD & Walters JR (1999) Sex-ratio selection in species with helpers at the nest: the repayment model revisited. Am Nat 153: 124-130. Koenig WD, Pitelka FA, Carmen WJ, Mumme RL & Stanback MT (1992) The evolution of delayed disper- sal in cooperative breeders. Quart Rev Biol 67: 111-150. Kokko H & Lundberg P (2001) Dispersal, migration, and offspring retention in saturated habitats. Am Nat 157: 188-202. Komdeur J (1992) Importance of habitat saturation and territory quality for evolution of cooperative breeding in the Seychelles Warbler. Nature 358: 493-495. Komdeur J (1994a) Experimental evidence for helping and hindering by previous offspring in the coopera- tive breeding Seychelles Warbler. Behav Ecol Socio- biol 34: 175-186. Komdeur J (1994b) The effect of kinship on helping in the cooperative breeding Seychelles Warbler (Acro- cephalus sechellensis). Proc R Soc London B 256: 47-52. Komdeur J (1996a) Breeding of the Seychelles magpie Adaptations and Maladaptations in the Seychelles Warbler robin Copsychus sechellarum and its implications for conservation. Ibis 138: 485-498. Komdeur J (1996b) Influence of helping and breeding experience on reproductive performance in the Sey- chelles Warbler: a translocation experiment. Behav Ecol 7: 326-333. Komdeur J (1996c) Facultative sex ratio bias in the off- spring of Seychelles Warblers. Proc Roy Soc Lond B 263: 661-666. Komdeur J (1996d) Seasonal timing of reproduction in a tropical bird, the Seychelles Warbler: a field experi- ment uising translocation. Biol Rhythms 11: 333-346. Komdeur J, Daan S, Tinbergen J & Mateman C (1997) Extreme adaptive modification in sex ratio of the Sey- chelles Warbler’s eggs. Nature 385: 522-525. Komdeur J (2003) Daughters on request: about helpers and egg sexes in the Seychelles Warbler. Proc Royal Soc London B 270: 3-11. Komdeur J, Piersma T, Kraaijeveld K, Kraaijeveld-Smit F & Richardson DS (in press) Why Seychelles War- blers fail to recolonize nearby islands: unwilling or unable to fly there? Ibis. Krackow S (1995) Potential mechanisms for sex ratio adjustment in mammals and birds. Biol Rev 70: 225-241. Leimar O (1996) Life-history analysis of the Trivers and Willard sex-ratio problem. Behav Ecol 7: 316-325. Lessells CM (1998) A theoretical framework for sex-bi- ased parental care. Anim Behav 56: 395-407. Lessells CM, Oddie KR & Mateman AC (1998) Parental behaviour is unrelated to experimentally ma- nipulated great tit brood sex ratio. Anim Behav 56: 385-393. Livezey BC (1990) Evolutionary morphology of flight- lessness in the Auckland Islands Teal. Condor 92: 639-673. Livezey BC (1992a) Flightlessness in the Galapagos Cormorant (Compsohalieus [Nannopterum|] harrisi): Heterochrony, giantisms and specialization. Zool J Linn Soc 105: 155-224. Livezey BC (1992b) Morphological corollaries and eco- logical implications of flightlessness in the Kakapo (Psittaciformes: Strigops habroptilus). J Morphol 213: 104-145. Loustau-Lalanne P (1968) The Seychelles, Cousin Is- land nature reserve. Internal report. International Council for Bird Preservation, Cambridge. Marshall T, Slate J, Kruuk L & Pemberton J (1998) Sta- tistical confidence for likelihood-based paternity in- ference in natural conditions. Mol Ecol 7: 639-655. Maynard Smith J (1964) Group selection and kin selec- tion. Nature 201: 1145-1147. McCall RA, Nee S & Harvey PH (1998) The role of wing length in the evolution of avian flightlessness. Evol Ecol 12: 569-580. McNab BK (1994a) Resource use and the survival of land and freshwater vertebrates on oceanic islands. Am Nat 144: 643-660. McNab BK (1994b) Energy conservation and the evolu- tion of flightlessness in birds. Am Nat 144: 628-642. Mumme RL (1992) Delayed dispersal and cooperative breeding in the Seychelles Warbler. Trends Ecol Evol 7: 330-331. Mumme RL, Koenig WD & Ratnieks FLW (1989) Helping behaviour, reproductive value, and the future component of indirect fitness. Anim Behav 38: 331-343. Oustalet ME (1878) Etude sur la faune ornithologique des isles Seychelles (Study of the ornithological fauna of the Seychelles Islands). Bull Soc Philomath Paris 7: 161-206. Pen I & Weissing FJ (2000) Sex-ratio optimization with helpers at the nest. Proc Roy Soc Lond B 267: 539-543. Raikow RJ (1985) Locomotor system. In King A.S. & McLelland J. (eds) Form and function in birds. Vol 3. pp 57-147. Academic Press, London. Richardson DS, Jury FL, Dawson D, Salgueiro P, Komdeur J & Burke T (2000) Fifty Seychelles War- bler (Acrocephalus sechellensis) microsatellite loci polymorphic in Sylviidae species and their cross- species amplification in other passerine birds. Mol Ecol 9: 2226-2231. Richardson DS, Jury FL, Blaakmeer K, Komdeur J & Burke T 2001. Parentage assignment and extra-group paternity in a cooperative breeder: the Seychelles Warbler (Acrocephalus sechellensis). Mol Ecol 10: 2263-2273 Richardson DS, Burke T & Komdeur J (2002) Direct benefits and the evolution of female-biased coopera- tive breeding in Seychelles Warblers. Evolution 56: 2313-2321. Richardson DS, Komdeur J, Burke T (2003a) Altruism and infidelity among warblers. Nature 422: 580. Richardson DS, Burke T, Komdeur J. (2003b) Sex-spe- cific associative learning cues and inclusive fitness benefits in the Seychelles Warbler. J Evol Biol 16: 854-861. Roff DA (1990) The evolution of flightlessness in in- sects. Ecol Monogr 60: 389-442. Roff DA (1994) The evolution of flightlessness: is his- tory important? Evol Ecol 8: 639-657. Trewick SA (1997) Flightlessness and phylogeny amongst endemic rails (Aves: Rallidae) of the New Zealand region. Phil Trans R Soc London B 352: J. KOMDEUR 429-446. Wagner DL & Liebherr JK (1992) Flightlessness in in- sects. Trends Ecol Evol 7: 216-220. Weber TP & Piersma T (1996) Basal metabolic rate and the mass of tissues differing in metabolic scope: mi- gration-related covariation between individual knots 88 Calidris canutus. J Avian Biol 27: 215-224. Williams GC (1979) On the question of adaptive sex ratio in outcrossed vertebrates. Proc R Soc Lond B 205: 567-580. Williams GC (1992) Natural selection: domains, levels, and challenges. Oxford University Press, Oxford. Ornithol. Sci. 2: 89-96 (2003) INVITED ARTICLE Translocations in avian conservation: reintroduction biology of the South Island Saddleback (Philesturnus carunculatus carunculatus) Johanna P. PIERRE* Marine Conservation Unit, Department of Conservation, P. O. Box 10-420, Wellington, New Zealand Abstract ‘Translocation is a commonly used tool in conservation management. However, because post-release monitoring has been infrequent in the past, reasons for the outcomes of translocations have often been unknown. Here, I review the reintro- duction biology (including dispersal patterns, social organisation, survival, habitat use and foraging patterns) of a population of 26 South Island Saddlebacks (Philesturnus carunculatus carunculatus), on Motuara Island, New Zealand. After release on Motu- ara Island, South Island Saddlebacks dispersed widely through forest areas. During their first post-release breeding season, saddlebacks established territories of 1.9 ha- 8.8 ha (X=4.21 ha, SD=2.42) in size, and territorial confrontations were very rare. Saddlebacks bearing both adult and subadult plumage held territories and attempted to breed, and successful breeding produced approximately 10 fledglings. Saddlebacks foraged on a variety of plant species, dead wood and the ground. Except for five-fin- ger (Pseudopanax arboreus), a preferred foraging substrate, birds foraged in all plant material in proportion to its availability. Saddlebacks preferred to forage in the lower levels of the forest. Although vegetation composition differed significantly between territories, all territories contained forest areas, and birds appeared to prefer foraging in larger sized trees. Large territory sizes, breeding attempts by young birds and rarity of territorial confrontations are most likely products of low population density. As density increases, birds are expected to occupy smaller territories, forage more effi- ciently within these smaller areas, start breeding at older ages, and possibly colonise scrub habitats. The translocated group sustained no more than 50% mortality at 8-10 months after release. In the past, translocations of 15-59 South Island Saddlebacks have been successful, suggesting that the relatively small founder group does not threaten the success of the transfer to Motuara Island. Saddlebacks are flexible in their habitat use, appear to readily adapt to ‘new’ environments and have high repro- ductive potential, increasing the likelihood of success of translocations of this species. ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2003 Key words New Zealand, Philesturnus carunculatus carunculatus, Reintroduction, South Island Saddleback, Translocation Translocation is becoming increasingly commonly used for conservation, and can be defined as the in- tentional release of plants or animals to the wild to establish, re-establish or augment a population (IUCN 1987; Griffith et al. 1989). The technique has been applied to the conservation of mammals (e.g. Dutty et al. 1994; Short et al. 1994), invertebrates (e.g. Sherley 1994) and birds (e.g. Atkinson & Bell (Received 4 March 2003; Accepted 23 September 2003) * E-mail: jpierre @doc.govt.nz Proceeding of the Symposium on Behavioral Studies and Conserva- tion Biology. 15 September 2002, Tokyo; Organized by Go Fujita and Hisashi Nagata 89 1973; Merton 1973, 1975; Bell 1978; Butler & Mer- ton 1992). In New Zealand, translocations have often been incorporated into species management programs in the past, and have been successful in saving some New Zealand birds from extinction. Since the 1880s, almost 400 translocations of 50 taxa (42 birds, five reptiles and three invertebrates) have been conducted in New Zealand, often in emergency situations, such as in 1962 when ship rats (Rattus rattus) arrived at the last stronghold of the South Island Saddleback (Philesturnus carunculatus carunculatus) (Atkinson & Bell 1973; Merton 1973, 1975; Bell 1978). Despite the increasingly common application of J. P. PIERRE translocations in conservation management, and wide recognition of the value of post-release monitoring (e.g. Scott & Carpenter 1987; Griffith et al. 1989; Sarrazin & Barbault 1996), such monitoring does not always occur. In fact, Wolf et al. (1996) reported that only 45% of 336 bird and mammal translocation pro- grams included tagging of released individuals or post-release telemetry, and in about 30% of transloca- tions, the causes of 90-100% of animal losses were never identified. In New Zealand, translocation con- tinues to be widely used as a conservation tool. Post- release monitoring of translocated animals may be becoming more common (e.g. Nelson et al. 2002), but such monitoring seldom occurred in the past (De- partment of Conservation 1994; Lovegrove & Veitch 1994; Armstrong & McLean 1995). Similar to post-release monitoring of survival, monitoring the habitat use of newly translocated or- ganisms is important, yet seldom occurs. Globally, translocations are deemed most likely to succeed where animals are released into areas of high habitat quality and/or quantity (Wolf et al. 1996). This em- phasizes the importance of thoroughly investigating the habitats at potential release sites before conduct- ing translocations and monitoring the habitat use of translocated animals after they are released. Post-re- lease monitoring of habitat use can confirm suitability of the new location, and produce valuable informa- tion relevant to future translocations, thereby facili- tating adaptive wildlife management (Sarrazin & Barbault 1996). For example, mobile animals at low densities may occupy habitats on the basis of prefer- ence rather than requirement, and may colonise less ‘preferred’ areas as their population density increases (Jenkins 1976; Armstrong & McLean 1995). There- fore, monitoring spatiotemporal patterns of habitat use together with population growth can be informa- tive in determining habitat requirements and prefer- ences, and the ability of translocated organisms to colonise ‘novel’ environments. The South Island Saddleback is a forest passerine endemic to the islands of New Zealand. It is an en- dangered subspecies of the near threatened New Zealand Saddleback (IUCN 2000). Diurnal, territorial and largely insectivorous, the South Island Saddle- back forages on foliage, live and dead wood, and the ground. It uses cavities for roosting and is monoga- mous with apparently flexible nest site requirements (Guthrie-Smith 1925; Pierre 1995). Although the South Island Saddleback was formerly widespread over New Zealand’s South Island and southern off- 90 shore islands, it is unable to coexist with introduced predators, and consequently, was virtually extinct by 1900 (Oliver 1955; Roberts 1991). The total popula- tion of South Island Saddlebacks is currently about 650, with birds occurring on 11 islands, all of which are free of introduced predators. Following the South Island Saddleback Recovery Plan (Roberts 1991) administered by the New Zealand Department of Conservation, 26 South Is- land Saddlebacks were translocated to Motuara Is- land in the Marlborough Sounds from the Titi Islands near Stewart Island, New Zealand, in March 1994. (See Figure | for the locations of islands mentioned in the text). Seven adult males, 11 adult females, 5 subadult males, one subadult female and one adult and subadult of unknown sex were included in the translocated founder population. Motuara Island (59 ha) was selected as the release site for translocated South Island Saddlebacks for three main reasons. First, introduced predators (Rattus exulans) were ex- tirpated from the island in 1993. Second, vegetation on Motuara Island has been regenerating since farm- ing was abandoned in 1926 (W.F. Cash pers. comm.), and finally, the bird was historically resident in the area. Conservation managers considered Motuara Is- land to be sufficiently large and with appropriate Motuara I. 200 km Fig. 1. Locations of islands mentioned in the text. (From Pierre 1999). South Island Saddleback reintroduction habitat to support more South Island Saddlebacks than currently existed on any other island (W.F. Cash pers. comm.). In this paper, I review the reintroduction biology of the South Island Saddleback. I discuss post-release dispersal, social organisation, survival, habitat use and foraging patterns of the newly released popula- tion on Motuara Island, New Zealand. Before the work described in this review was carried out, the South Island Saddleback had not been studied in de- tail. Also, future translocations were planned to in- crease the total population of this bird (Roberts 1991). For these reasons, post-release monitoring of the translocated population was critical. METHODS Saddlebacks were weighed, measured and their plumage examined on capture to determine sex and age (W.F. Cash pers. comm.; Nillson 1978; Jenkins & Veitch 1991). Numbered metal and unique combina- tions of coloured plastic legbands were used to iden- tify individuals. After release on Motuara Island, birds were relocated using a recording of South Is- land Saddleback male territorial song, and via their vocal responses and attraction to disturbances (e.g. logs breaking). The work reviewed here results from post-release monitoring involving intensively search- ing for birds for four days, six months after they were released, and searching for and monitoring birds for 56 days, from eight months after release, during their first breeding season on Motuara Island. Territory boundaries were identified during the breeding season by conducting repeated searches for birds and recording the locations of song posts. I monitored birds for 90 minutes during these tracking episodes; tracking beyond five 90 minute sessions did not alter the perceived location of territory bound- aries (Pierre 1999). As well as recording song posts, I monitored habitat use by birds, including substrates used for foraging and other activities, and heights of the forest that birds occupied (Pierre 2001). I also recorded prey identity when possible, and where prey were caught (Pierre 2000). Saddleback nests were lo- cated by following birds holding nesting materials, and by checking nest boxes. To avoid disturbing nest- ing birds, I did not check nests frequently after locat- ing them. Instead, I monitored breeding activities in- directly by assessing the behaviour of adult birds (Pierre 1999). I conducted analyses of habitat use and foraging data using paired t tests and analysis of variance (ANOVA) conducted in SPSS 8.0 (SPSS Inc. 1997). I used Games-Howell post-hoc tests (Day & Quinn 1989) to identify homogeneous subgroups after sig- nificant ANOVAs, and Bonferroni @ adjustments when appropriate, including in G tests described below (Miller 1981). See Pierre (2001) for a discus- sion of how pseudoreplication and problems of inde- pendence were minimized. I characterised vegetation using a transect-quadrat sampling method, and sampled both saddleback terri- tories and one area uninhabited by saddlebacks (Pierre 2001). I defined plant availability as the cross- sectional area of each plant species at breast height, and the availability of dead wood as the proportion of total quadrat area it covered on each transect (Pierre 2001). Vegetation characteristics were compared be- tween sites using G tests (Zar 1996). RESULTS After release, South Island Saddlebacks ranged widely through the forested areas of Motuara Island. By the start of the first post-release breeding season, both adult and subadult birds appeared to have settled on territories, which varied in size from 1.9-8.8 ha (X =4.21 ha, SD=2.42, N=6, Figure 2, Pierre 1999). Adult and subadult saddlebacks announced their presence vocally from a range of vertical locations in their territories, especially in the first two hours after dawn, and for approximately one hour just before dusk. However, confrontations between neighbours appeared very rare, and I observed only one during the study. This low frequency was despite saddle- backs venturing into each other’s territories, which I observed on five occasions during my research, and which at least sometimes appeared related to searches for water (Pierre 1999). The vegetation composition of saddleback territo- ries on Motuara Island differed significantly, and was different again in areas uninhabited by saddlebacks (G tests, vegetation composition in all territories compared to all other territories and a vegetated area not occupied by saddlebacks: G,>2117.36, P<0.005, Pierre 2001). However, all territories were in forest areas and included plants of similar size distributions (G tests: plant sizes in all territories compared to all other territories, G;=14, P>0.005, NS). In contrast, saddlebacks did not inhabit a locality covered by smaller diameter plants, in this case scrub (G tests: all territories compared to an area uninhabited by saddle- J. P. PIERRE Motuara Island Hippa Island Fig. 2. South Island Saddleback territory boundaries during 11 Nov 94-13 Jan 95, the first breeding season after release. Letters denote legband combinations: B=blue, G=green, M=metal, R=red, W=white, Y=yellow. (From Pierre 1999). % foraging time Ground Psearb Kuneri Dead Melram Macexc backs, all G;=26, P<0.005, Pierre 2001). At least five, and possibly six pairs of saddlebacks were present during the first post-release breeding season on Motuara Island. Pairs were not knowingly transferred together, and after release at least some birds interacted with more than one potential mate before settling on a territory with what appeared to be a stable partner (Pierre 1999). Both adult birds and those in subadult plumage formed pairs and_at- tempted breeding during the first post-release breed- ing season. Pairs were able to form and then initiate breeding with what appeared to be minimal delay; one pair built a nest and reared two offspring no more than 35 days after pairing. This pair then appeared to renest (Pierre 1999). The nests of two other pairs were found in a tree hole and a nest box, and held one and two eggs, respectively. However, these nests were deserted for unknown reasons. In all, an esti- mated 10 fledglings were hatched the first breeding season after saddlebacks were released onto Motuara Island (Pierre 1999; W.F. Cash pers. comm.). Translocated saddlebacks utilised a range of forag- ing substrates, including various plant species, dead wood and the ground. Foraging patterns differed be- tween males and females, with males spending the most foraging time on the ground, whereas females preferred to forage in five-finger (Pseudopanax arboreus) (ANOVA: F; 55=4.59, P=0.003, Figure 3, Pierre 2001). Overall, birds apportioned their forag- ing time differently among different foraging sub- strates, and spent the most time foraging on the Phocoo Corlae Copluc Hedarb Brarep Foraging location Fig. 3. Percent time spent on foraging sites (X +1 SE) by male (open bars) and female (hatched bars) South Is- land Saddlebacks on Motuara Island. Brarep=Brachyglottis repanda, Coplue=Coprosma lucida, Corlae=Coryno- carpus laevigatus, Dead=dead wood, Hedarb=Hedycarya arborea, Kuneri=Kunzea ericoides, Macexc= Viacropiper excelsum, Melram=Melicytus ramifloris, Phocoo=Phormium cookianum, Psearb=Pseudopanax ar- boreus. Numbers above bars represent the number of birds contributing to means. (From Pierre 2001). 92 South Island Saddleback reintroduction ground and _ five-finger (ANOVA: F,;,.=13.08, P<0.001, Figure 3, Pierre 2001). However, birds used most plant species (and dead wood) in propor- tion to their availability (t tests: t, <=—1.14-2.87, P=0.21-0.99), except for five-finger (used more than expected, t test: tj=2.74, P=0.03). Flax (Phormium cookianum) may also have been used more than ex- pected, although small sample size rendered this im- possible to test statistically (Pierre 2001). The num- ber of prey items South Island Saddlebacks obtained from foraging substrates was roughly in proportion to the amount of foraging time spent, although dead wood was a particularly rich foraging substrate. Sad- dlebacks were largely insectivorous, but also fed on nectar and honeydew (Pierre 2000). Saddlebacks were recorded significantly more fre- quently in the lowest two metres of the forest than at any other level (ANOVA: F,3)=41.31, P<0.001, Pierre 2001). This was the case for both males and fe- males, whose vertical use of the forest did not appear to differ (ANOVA: F, ;)=0.20, P=0.94, however 1- B=0.16 at a=0.1, Pierre 2001). Saddlebacks were observed drinking water fre- quently, both from natural sources and troughs con- structed on the island before birds were released. They also bathed in these human-made water troughs frequently (Pierre 2001). Mortality of South Island Saddlebacks was 35% (9/26 birds dead) at six months after release. Maxi- mum mortality was 50% 8-10 months after release. Of the surviving 13/26 translocated saddlebacks, there were 3/7 adult males, 4/11 adult females and 6/7 subadults remaining (Pierre 1999). DISCUSSION After release on Motuara Island, South Island Sad- dlebacks traversed forested areas widely, before set- tling on territories after about eight months. Translo- cations of North Island Saddlebacks (P. c. rufusater) also result in wide post-release dispersal, suggesting that this is the norm for the species (Jenkins 1976; Armstrong & Craig 1995). South Island Saddlebacks settled on territories that were unusually large and variable in size. There is little data on territory size in saddlebacks, but on Cuvier Island (170 ha), two pairs of North Island Saddlebacks held territories of 0.89 ha and 1.22 ha in November/December, the height of the breeding season. The smallest territory on Motu- ara Island was about |.5—2 times this size, suggesting that the low density of saddlebacks on Motuara Is- 93 land had strong effects on the area of territories. The size of North Island Saddleback territories has been recorded to change temporally, with territories being largest during breeding (O’ Callaghan 1980). South Island Saddlebacks on Motuara Island were much less vocal than North Island Saddlebacks oc- curring in higher density populations. However, tem- poral singing patterns appear similar for both sub- species (pers. obs.; Jenkins 1976; Murphy 1989). Ter- ritorial confrontations occurred rarely on Motuara Is- land, relative to North Island Saddleback populations (Jenkins 1976; O'Callaghan 1980). This is probably because saddleback territories on Motuara Island were sufficiently large that neighbours would seldom meet, and may have been out of earshot of each other often, if not most of the time. The utility and impor- tance of song in maintaining territorial integrity might be expected to increase with increasing popula- tion densities, however Murphy (1989) showed that the number of neighbours around North Island Sad- dleback territories did not affect singing rates. The frequency of territorial displays, however, is known to increase with population densities in North Island Saddlebacks (Jenkins 1976; O’Callaghan 1980). South Island Saddleback territories on Motuara Island were not areas of completely exclusive use by territory ‘owners’. This may not be unusual for the species. In a relatively high density population, O’Callaghan (1980) documented areas of overlap between North Island Saddleback territories, as well as subadults and non-territorial adults moving through the territories of others. Vegetation composition of all saddleback territo- ries differed on Motuara Island. Further, saddlebacks seemed to prefer areas with larger trees; all territories were at least partially forested, compared to uninhab- ited areas in which the dominant vegetation type was scrub. The species composition of forest however, did not seem important. Habitat characteristics, e.g. the availability of food, affect the size of North Island Saddleback territories (Blackburn 1964; O’Callaghan 1980). Despite this, South Island Saddleback territo- ries were probably much larger than the area required to support a breeding pair (also see above) and as previously mentioned, the large size is almost cer- tainly a product of low population density. In low density North Island Saddleback popula- tions, birds also preferentially occupy forest habitat. However, with increasing density, birds will colonise scrub areas, and can breed very successfully in this habitat type (Jenkins 1976; Craig 1994; B. Walter J. P. PIERRE pers. comm.). Thus, it is expected that as the South Island Saddleback population density increases on Motuara Island, birds will colonise scrub areas. On Tiritiri Matangi Island, the suitability of scrub may have been increased by the erection of roost and nest boxes (B. Walter pers. comm.). Thus, the addition of roost and nest boxes may enhance the suitability of scrub areas for saddlebacks on Motuara Island. In any case, the colonisation of scrub demonstrates the be- havioural plasticity of saddlebacks. Both for saddle- backs and other organisms, behavioural plasticity can be an important factor increasing the success of translocations. After release on Motuara Island, but before settling on a territory in a stable pair, some South Island Sad- dlebacks were observed with more than one bird of the opposite sex. This has also been reported in newly translocated North Island Saddleback popula- tions (Armstrong & Craig 1995). Further, pairs of North Island Saddlebacks translocated together did not maintain their pair bond after release (Armstrong & Craig 1995). Also, similar to my results for South Island Saddlebacks, North Island Saddlebacks in low density populations are recorded breeding at one year old (Craig 1994). However, in high density popula- tions, North Island Saddlebacks breed at two or more years of age (Lovegrove 1980). Further, South Island Saddlebacks were able to breed successfully with one pair taking a maximum of only 35 days between pair- ing and nest building. This pair fledged two chicks and probably renested in the first breeding season after translocation. Although there are no published records of the length of time taken from pairing to nesting, newly released North Island Saddlebacks have also been reported to raise more than one brood per breeding season (Jenkins 1976; Craig 1994; Arm- strong & Craig 1995). Like South Island Saddlebacks on Motuara Island, North Island Saddlebacks appear to be flexible in terms of foraging substrates they can use (Atkinson 1964, 1966; Lovegrove 1980). Male and female South Island Saddlebacks had slightly different forag- ing strategies in terms of substrate used and vertical location. Similarly, albeit at high density, North Is- land Saddleback males spent more time foraging on the ground than their female counterparts (Blackburn 1964; Lovegrove 1980; O’Callaghan 1980), however, whether there are intersexual differences in their for- aging patterns at low density is unknown. Out of all plant species (including dead wood) that saddlebacks on Motuara Island foraged on, five-finger and possi- 94 bly flax were used more than expected. Foraging se- lectivity of high or low density North Island Saddle- back populations cannot be compared, due to lack of research. However, I expect that as population den- sity increases, South Island Saddlebacks may use a wider spectrum of plants to increase the efficiency of foraging in smaller territories. Published data report- ing relative richness of locations of prey capture are nonexistent for North Island Saddleback populations. However, as saddlebacks appear flexible with respect to foraging strategies and diet, these are expected to vary with habitat type as well as season, as are the importance of different foraging substrates and prey types (Atkinson 1964, 1966; Blackburn 1964, 1967; Lovegrove 1980, 1992; O’Callaghan 1980; Pierre L995): South Island Saddlebacks on Motuara Island for- aged mostly in lower levels of the forest, like North Island Saddlebacks at high density (Lovegrove 1980; O’Callaghan 1980). However, high density popula- tions of North Island Saddlebacks also display verti- cal stratification within pairs when foraging (Love- grove 1980; O’Callaghan 1980). This may function to reduce intersexual competition within pairs, and improve the efficiency of resource use in territories. As increasing population density causes territory sizes to decrease on Motuara Island, birds may de- velop vertically stratified foraging behaviour. Coinci- dent with this speculation, an increased degree of ver- tical stratification in North Island Saddleback forag- ing behaviour has been related to decreases in terri- tory size (Lovegrove 1980). Although the diet of saddlebacks contains some water, most of Motuara Island is very dry, and birds used both natural and human-constructed water sources frequently. That saddlebacks used water sources outside their own territories suggests a very strong need for water, and as predicted prior to the translocation, the instalment of water troughs may have increased the quality of saddleback habitat on Motuara Island, possibly increasing the likelihood of success of the translocation. Mortality rate is one of the most important factors determining the size of founder groups, making it a key consideration when planning translocations. On Motuara Island, maximum mortality was 50% 8-10 months after saddlebacks were released. Published records of North Island Saddleback mortality range from 8% at six months to 52% at two years after re- lease (Jenkins 1976; Armstrong & Craig 1995). Natu- rally, mortality will vary due to site-specific charac- South Island Saddleback reintroduction teristics as well as the ability of birds to deal with the stresses of capture and translocation, and their flexi- bility in adapting to new environments. Weather con- ditions after release and the abundance of natural predators are other important considerations. In general, for a translocation into excellent quality habitat to have a 60% chance of success, the recom- mended size of the founder group is fifty individuals (Griffith et al. 1989). Only 26 South Island Saddle- backs were released onto Motuara Island, but in the past, translocations of 15-59 South Island Saddle- backs have been successful (Nillson 1978; Roberts 1991). Thus, from the outset, the success of the trans- fer to Motuara Island may not have been jeopardized by the relatively small founder group. In combination with the predator-free environment of Motuara Is- land, the saddlebacks’ flexible habitat requirements and foraging strategies, ability to readily adapt to ‘new’ habitats, and potential for high reproductive output increased the likelihood that translocations to this island would be successful. In the eight years since translocation, estimates of the number of saddlebacks Motuara Island can sup- port have ranged from about 70 (in years of cold, wet climatic conditions, when breeding success is low) to 150-200 (in the first years after release, with high reproductive output, an abundance of prey and favourable climate) (Pierre 1995; W. F. Cash pers. comm.). Social organisation, survival, foraging ecol- ogy and diet have not been investigated in detail since my work was completed. However, due to the success of this translocation over the first eight years at least, the Motuara Island population is now being used as a source population for other South Island Saddleback translocations. ACKNOWLEDGMENTS I wish to thank the Ornithological Society of Japan for the invitation to present a symposium paper at their 2002 Annual Meeting. I also extend thanks to the Royal Forest and Bird Protection Society Waikato Branch, the Pacific Development and Conservation Trust, the Or- nithological Society of New Zealand and the University of Canterbury for funding my study, and Dr. I. McLean who proposed work on Motuara Island initially. The De- partment of Conservation Permission granted me per- mission to stay on the island, and staff of the Depart- ment of Conservation (especially Bill Cash), Cougar Line, Dolphin Watch Marlborough, and the Sandpiper were a great help with field work, as well as valued con- 95 nections to the world outside Motuara. I would like to thank my field assistants for their hard work, especially Sally Truman and Maureen and Richard Pierre for help both in the field and beyond. Figures were reproduced with permission from Elsevier Science. REFERENCES Armstrong DP & Craig JL (1995) Effects of familiarity on the outcome of translocations, I. A test using sad- dlebacks Philesturnus carunculatus rufusater. Biol Conserv 71: 133-141. Armstrong DP & McLean IG (1995) New Zealand translocations: theory and practice. Pacific Conserv Biol 2: 39-54. Atkinson IAE (1964) Feeding stations and food of North Island saddleback in August. Notornis 11: 93- OT. Atkinson IAE (1966) Feeding stations and food of the North Island saddleback in May. Notornis 13: 7-11. Atkinson IAE & Bell BD (1973) Offshore and outlying islands. In: Williams GR (ed) The Natural History of New Zealand. pp 372-392. AH & AW Reed, Welling- ton. Bell BD (1978) The Big South Cape Islands rat irrup- tion. In: Dingwall PR, Atkinson IAE & Hay C (eds) The Ecology and Control of Rodents in New Zealand Nature Reserves. pp 33-40. Department of Lands and Survey, Wellington. Blackburn A (1964) Some observations on behaviour of the North Island saddleback in August. Notornis 11: 87-92. Blackburn A (1967) Feeding stations and food of the North Island saddleback in November. Notornis 14: 67-70. Butler D & Merton D (1992) The Black Robin: Saving the Worlds’ Most Endangered Bird. Oxford Univer- sity Press, Auckland. Craig JL (1994) Meta-populations: is management as flexible as nature? In: Olney PJS, Mace GM & Feist- ner ATC (eds) Creative Conservation: Interactive Management of Wild and Captive Animals. pp 50-66. Chapman and Hall, London. Day RW & Quinn GP (1989) Comparisons of treat- ments after an analysis of variance in ecology. Ecol Monogr 59: 433-463. Department of Conservation (1994) New Zealand trans- fer guidelines for indigenous terrestrial flora and fauna. In: Serena M (ed) Reintroduction Biology of Australian and New Zealand Fauna. pp 253-255. Surrey Beatty and Sons, Chipping Norton. Dufty AC, Seebeck JH, McKay J & Watson AJ (1994) Reintroduction of the eastern barred bandicoot Per- J. P. PIERRE ameles gunnii at Gellibrand Hill Park, Victoria. In: Serena M (ed) Reintroduction Biology of Australian and New Zealand Fauna. pp 219-225. Surrey Beatty and Sons, Chipping Norton. Griffith B, Scott JM, Carpenter JW & Reed C (1989) Translocation as a species conservation tool: status and strategy. Science 245: 477-480. Guthrie-Smith H (1925) Bird Life on Island and Shore. William Blackwood and Sons, London. IUCN (1987) The IUCN position statement on translo- cation of living organisms: Introductions, re-intro- ductions and re-stocking. \UCN, Gland. IUCN (2000) 2000 IUCN Red List of Threatened Species. IUCN, Gland and Cambridge. Jenkins PF (1976) Social organisation and vocal behay- iour of the saddleback Philesturnus carunculatus ru- fusater (Aves). Ph D Thesis, University of Auckland, Auckland. Jenkins PF & Veitch CR (1991) Sexual dimorphism and age determination in the North Island saddleback (Philesturnus carunculatus rufusater). N Z J Zool 18: 445-450. Lovegrove TG (1980) The saddleback pair bond. M Sc Thesis, University of Auckland, Auckland. Lovegrove TG (1992) The effects of introduced preda- tors on the saddleback (Philesturnus carunculatus), and implications for management. Ph D Thesis, Uni- versity of Auckland, Auckland. Lovegrove TG & Veitch CR (1994) Translocating wild forest birds. Ecological Management 2: 23-35. Merton D (1973) Conservation of the saddleback. Wildlife: A Review 4: 13-23. Merton D (1975) The saddleback: its status and conser- vation. In: Martin RD (ed) Breeding Endangered Species in Captivity. pp 61-74. Academic Press, Lon- don. Miller RG (1981) Simultaneous statistical inference. McGraw Hill, New York. Murphy SM (1989) The interaction of social organisa- tion and parentage with vocalisation in the North Is- land saddleback. M Sc Thesis, University of Auck- land, Auckland. Nelson NJ, Keall SN, Brown D & Daugherty CH (2002) Establishing a new wild population of tuatara (Sphen- odon guntheri). Consery Biol 16: 887-894. Nillson RJ (1978) The South Island saddleback. Wildlife: A Review 9: 32-36. O’Callaghan AP (1980) Use of space by the North Is- land saddleback Philesturnus carunculatus rufusater 96 (Aves). M.Sc. Thesis, University of Auckland, Auck- land. Oliver WRB (1955) New Zealand Birds. 2nd ed. AH & AW Reed, Wellington. Pierre JP (1995) Behaviour, ecology and reintroduction biology of the South Island saddleback Philesturnus carunculatus carunculatus. B Sc (Hons) Thesis, Uni- versity of Canterbury, Christchurch. Pierre JP (2000) Foraging behaviour and diet of a rein- troduced population of the South Island Saddleback (Philesturnus carunculatus carunculatus). Notornis 47: 7-12. Pierre JP (1999) Reintroduction of the South Island sad- dleback (Philesturnus carunculatus carunculatus): dispersal, social organisation and survival. Biol. Con- serv. 89: 153-159. Pierre JP (2001) Reintroduction of the South Island sad- dleback (Philesturnus carunculatus carunculatus): habitat use including foraging patterns. Notornis 48: 63-71. Roberts A (1991) A Recovery Plan for the South Island Saddleback. Department of Conservation, Southland. Sarrazin F & Barbault R (1996) Reintroduction: chal- lenges and lessons for basic ecology. TREE 11: 474— 478. Scott JM & Carpenter JW (1987) Release of captive- reared or translocated endangered birds: What do we need to know? Auk 104: 544-545. Sherley G (1994) Translocations of the Mahoenui giant weta Deinacrida n. sp. and Placostylus land snails in New Zealand: What have we learnt? In: Serena M (ed) Reintroduction Biology of Australian and New Zealand Fauna. pp 57-63. Surrey Beatty and Sons, Chipping Norton. Short J, Turner B, Parker S & Twiss J (1994) Reintro- duction of endangered mammals to mainland Shark Bay: a progress report. In: Serena M (ed) Reintroduc- tion Biology of Australian and New Zealand Fauna. pp 183-188. Surrey Beatty and Sons, Chipping Nor- ton. SPSS Inc. (1997) SPSS for Windows. SPSS _Inc., Chicago. Wolf CM, Griffith B, Reed C & Temple SA (1996) Avian and mammalian translocations: update and re- analysis of 1987 survey data. Conserv Biol 10: 1142- 1154. Zar JH (1996) Biostatistical Analysis. 3rd ed Prentice Hall, Upper Saddle River. Ornithol. Sci. 2: 97-102 (2003) ORIGINAL ARTICLE Sexing White-rumped Munias in Taiwan, using morphology, DNA and distance calls Taku MIZUTA'*, Hiroko YAMADA?, Ruey-Shing LIN*, Yuki YODOGAWA* and Kazuo OKANOYA**5* ' Laboratory of Ethology, Graduate School of Science, Kyoto University, Kitashirakawa-Oiwake-cho, Sakyo-ku, Kyoto 606-8502, Japan ? Graduate School of Science and Technology, Chiba University, 1-33 Yayoi-cho, Inage-ku, Chiba 263-8522, Japan 3 Taiwan Endemic Species Research Institute, 1, Ming-shen East Road, Chichi Township, Nantou County, Taiwan 552, R.O.C. * Faculty of Letters, Chiba University, 1-33 Yayoi-cho, Inage-ku, Chiba 263-8522, Japan ° PRESTO, Japan Science and Technology Agency, Kawaguchi, Saitama, Japan Abstract Methods of sexing the White-rumped Munia Lonchura striata phaethon- ORNITHOLOGICAL toptila were investigated in eastern Taiwan. Twenty-six individuals were captured and SCIENCE their physical parameters were measured in the hand. Distance calls emitted when the birds were released were also recorded. The sex of each bird was confirmed using a DNA sexing method in the laboratory. Among the morphological traits measured, the tails and wings of males were significantly longer than those of females. An increas- ing stepwise discriminant analysis was performed to determine sex on the basis of morphological characters, however, only 84.0% of individuals were sexed correctly by such analysis. Distance calls of the White-rumped Munia were very similar to those of the Bengalese Finch, Lonchura striata var. domestica, the domesticated strain of the White-rumped Munia. Two distinct distance calls were recorded from birds on release, corresponding to the sexual difference confirmed by DNA testing. It is concluded, therefore, that the difference in distance calls is a useful trait that facili- tates separation of the sexes in the field. © The Ornithological Society of Japan 2003 Key words CHD gene, Distance call, Morphology, Lonchura striata phaethon- toptila, Lonchura striata var. domestica, Sexing method Many species of munias (Estrildidae) are sexually monomorphic (Restall 1996), making it impossible to sex in the field. The White-rumped Munia Lonchura striata is one such monomorphic estrildid species that is widely distributed in South- and Southeast Asia (Paynter 1968; Restall 1996). Six (Paynter 1968) to eight (Restall 1996) subspecies occur in this region. They inhabit in a wide range of habitats, from open grassland and cultivated land, to human residential area such as urban gardens and parks (Restall 1996). Although the White-rumped Munia is a common bird in Asia, little fieldwork has been done on this species (but see Avery 1978, 1980a, b; Chattopad- hyay 1980; Young 1989). The Bengalese Finch, Lonchura striata var. domestica, is a domesticated (Received 26 August 2002; Accepted 21 February 2003) * Corresponding author, Email: okanoya@cogsci.L.chiba-u.ac.jp * Present address: Laboratory of Geographical Ecology, Faculty of Science, Toho University, Miyama 2-2-1, Funabashi, Chiba 274-8510, Japan OF strain of the White-rumped Munia (Restall 1996; Honda & Okanoya 1999), and during the process of domestication, several modifications in coloration and behavior have occurred. In particular, males of the Bengalese Finch sing more complex songs than those of wild White-rumped Munias (Honda & Okanoya 1999). The Bengalese Finch has been de- scribed as having sexually dimorphic distance calls (Yoneda & Okanoya 1991; Okanoya & Kimura 1993), as has the White-rumped Munia (Restall 1996). Okanoya et al. (1995) observed a small popu- lation of wild White-rumped Munias in Okinawa, Japan, and recorded two types of distance calls that may correspond to the sexually dimorphic calls of the Bengalese Finch. However, neither Restall (1996) nor Okanoya et al. (1995) described the details of sexual differences in distance calls; Restall (1996) did not show sonograms of the calls, and Okanoya et al. (1995) did not confirm the sex of the calling individu- als. T. MIZUTA et al. In the present study, we investigated the White- rumped Munia L. s. phaethontoptila in Taiwan, in order to establish a sexing method that can easily be applied in the field. We measured morphological characteristics and examined distance calls by ear in the field so as to sex individuals. The sexual differ- ence in distance calls was confirmed by sonogram. Correct sex was determined in the laboratory (CHD gene analysis), and the accuracy of the two sexing methods (morphology and distance calls) was com- pared. Since monomorphic White-rumped Munias live in flocks throughout the year (Restall 1996), being able to sex individuals in the field will be indis- pensable for researchers investigating their ecology, social system, life-history, and breeding behavior. METHODS 1) tics Fieldwork was conducted at a site near Antong hot spring (23°16'32.3"N, 121°20'40.0"E) in Hualien County, eastern Taiwan, from 16 to 18 February 2002. Antong hot spring is a hilly area, around 260 m above sea level. Mist nets were set in the morning (from 06:00 to 11:00), at a grassland area where a group of White-rumped Munias regularly foraged. The following characters of the birds captured were measured using digital caliipers and a ruler: 1) flat- tened wing length (FWL), 2) longest tail length (LTL), 3) tarsus length (TSL), 4) entire culmen length (ECL), 5) bill width (BLW), and 6) bill height (BLH). BLW and BLH were measured at the anterior edge of the nostril. Each bird was also weighed (body weight BDW) using a digital balance. The differences in morphological measurements between males and females were examined using the t-test. In order to establish the probability of success- ful sex determination using morphological measure- ments, both discriminant, and stepwise discriminant analyses were performed. SPSS statistical packages (SPSS 1993, 1994) were used for the statistical analy- ses. Field measurement of morphological characteris- 2) Blood sampling and genomic sex determination Small blood samples (20-60 11) were obtained from each bird for subsequent laboratory analysis to determine sex using the CHD genes of DNA (Grif- fiths et al. 1996). In brief, PCR reaction volumes of 50 ul were made up of taq polymerase (Takara), 200 um of each dNTP, P2, P3 primers, 100 mg of ge- 98 nomic DNA and 0.15 units of tag polymerase. After PCR amplification, the restriction enzyme Haell] was used to cut 8 ul of PCR product in Promega buffer 3 and 50 ng/ul bovine serum albumin in a total volume of 10 yl. Electrophoresis was then used to detect sex- ually dimorphic bands (see Fig. 2). 3) Sex determination by distance calls Field determination When the birds were released, they always gave distance calls. We recorded these calls using a direc- tional gun microphone (AUDIO-TECHNICA AT815b) and a DAT recorder (TASCAM DA-P1). Distance calls were first examined in real-time (i.e., when the birds were released) by ear, by an experi- enced observer, Hiroko Yamada, who determined the sex of the caller in the field without consulting the DNA or morphological data. Spectrographic determination Sex determination, based on hearing calls, was fur- ther confirmed by spectrographic observation. Sono- grams (sound spectrograms) of the calls were ob- tained using a digital signal processor (Avisoft ver- sion 3.93, SASLab Pro 2002) at a sampling rate of 48 kHz and a Fast Fourier Transform (FFT) size of 512 points (see Fig. 3). We counted the number of el- ements in each call and compared those of males and females determined by the DNA sexing data (see Fig. Py Individual variation Individual variation in distance calls by each sex was also analyzed. We were able to record more than one call from several individuals. By taking a Pear- son correlation coefficient between two matrices each derived from the sound spectrograph of a distance call, we were able to measure the similarity of the two calls (Avisoft version 3.93, SASLab Pro 2002). Eighteen calls recorded from six males and twelve calls recorded from four females were separately sub- mitted to a furthest neighbor cluster analysis (Stat- Partner version 2.0, NEC software 1995) and a den- drogram reflecting the similarity among calls was cal- culated. RESULTS 1) Morphological analyses About 50 White-rumped Munias were found feed- ing on the ground at the study site, where they fed on Sexing White-rumped Munias in Taiwan Table 1. Measurements of morphological characters (mm) and body weight (g) of male and female White-rumped Munias cap- tured in eastern Taiwan in February 2002. Male: N=13, Female: N= 12. Male Female Measurements t-value Mean+SD Range Mean+SD Range Wing length 50.52+0.91 47.1-51.7 49.35+1.15 47.0-51.7 2.84* Tail length 42.94+1.95 36.80—46.15 40.19+2.42 35.36-45.36 Bala Tarsus length 13.36+0.43 12.36-13.86 13.20+0.45 12.76-14.20 0.89 Bill length 10.53+£0.37 9.94-11.12 10.55+0.31 9.86-11.19 0.13 Bill width 6.66+0.40 6.24-7.84 6.47+0.27 5.95-6.85 1.33 Bill height 7.17+£0.30 6.66—-7.73 7.05+0.18 6.68-7.54 1.31 Body weight 9.94+0.48 8.8-10.6 9.80+0.45 9.5-10.4 0.74 *P<0.01, ** P<0.005 the seed of silver grass Miscanthus sp. In total, 26 birds were captured, measured, and blood samples taken, during three days in February 2002. From the DNA analysis, 13 birds were identified as males, and 13 as females. As one female was molting her tail feathers, measurements of 13 males and 12 females were used for the following morphological analyses. Males had significantly longer tails and wings than females, although considerable overlaps were ob- served between the sexes Fig. 1. The probability of correct discrimination of the preliminary discriminant analysis was 76.0% (19/25). To simplify the function and to distinguish between the sexes more precisely, the increasing stepwise dis- criminant analysis was performed. Two variables, tail length (LTL) and wing length (FWL), were included in the analysis. The resulting discriminant function is as follows: D=0.32 LTL+0.60 FWL—43.41 (Wilks’ A=0.59, 73=11.55, P<0.005; Fig. 1). The two variables contributed to the function sig- nificantly (LTL: Wilks’ A=0.70, F,,,=9.92, P< 0.005, FWL: Wilks’ A=0.59, F,,,=7.60, P<0.005). D>0 indicates that the individual is male and D grass land forested area blank : urban area X= nest OQ Territory Territory of pairs excluded from analysis Broken line in J1, J2 and J4; Home range Kamo-River 4 500m Fig. 1. Territories of Jungle Crows and Carrion Crows at the Shimogamo-shrine study site in the western part of the study area. Broken lines connecting the split territories of J2 and J4 show the home ranges of the pairs. J8 and C13 were not in- cluded in further analysis. Table 2. rence freqencies. Note that attacks to the owner is 0. between the J2 and CS pairs (Table 2) whose home ranges also overlapped. As in the case of J4 and C7, J2 foraged on the opposite side of the river. The fre- quency of Carrion Crow attacks on Jungle Crows crossing the river, where the Carrion Crows held ter- ritories, did not differ significantly between pairs J2 and J4 (Table 2, Fisher’s exact test, two-tailed, N=24 (J2), 66 (J4), P=0.40, NS), although the sample size is small. Four garbage collection points (“garbage stations”) were located in the foraging area of J4 on the eastern side of the river (Figs. | and 3). The garbage stations were uncovered, thus the garbage in plastic bags was easily accessible for the crows. J2 also visited three to four garbage stations. Compared with the western side, there were more apartment houses on the eastern side of the Takano River where garbage was placed in the garbage stations everyday. Although J3 did not fly across the river, the pair flew Y : forested area IIlll : grass land J5~J7, J9~J11: Jungle Crows C10~12, C15~C17: Carrion Crows blank : urban area X: nest OQ Territory Territory of pairs excluded from analysis 500m “aaa I) Wy Fig. 2. Territories of Jungle Crows and Carrion Crows in the Kyoto University study site in the eastern part of the study area. J9~J11, and Cl4~C17 were not included in further analysis. Observed interactions when Jungle Crows passed across a Carrion Crow’s territory. Figures in table show the occur- Attack to Attack to Attacked by Intruder No interaction eins Black Kite tee Total i fy 22 0 0 2: 24 J4 53 0 3 10 66 106 Comparative study on syntopic corvids Table 3. Time spent on the ground (min). Number of touchdowns indicates the frequency that crows flew down to the ground. Duration on the ground Shobies Observation Number P duration touchdowns* Total % per Mean Min. Max. SD duration* observation duration* duration duration Jungle Crow 1800 96 120 6.6 1.3 0.5 5 0.88 Carrion Crow 1908 181 693 36.3 3.8 0.5 23 3.5) * Indicates a siginificant difference between the species (Mann & Whitney’s U test, P<0.05). over C6’s territory to visit garbage stations on the southwestern side of the J3 territory. The C6 pair at- tacked the J3 pair when they flew over C6’s territory. The J3 pair seemed to mostly attack the C6 pair at the territory border as did the J2 and J4 pairs. However, the J3 and C6 pairs were sensitive to the observer and nested in dense woods where observation from a dis- tance was impossible, thus further interactions be- tween the pairs could not be investigated. This type of split territory was not observed in Car- rion Crows. All Carrion Crow territories were contin- uous, and they stayed inside the areas they defended. Carrion Crows’ home ranges coincided with their ter- ritories (Figs. | to 2). 2) Time spent on the ground A prominent behavioral difference between the two species was the use of different vertical positions within the habitat. Jungle Crows mainly stayed on trees, buildings, utility poles, or power lines, that is, places higher than the ground level. I classified the perching positions into two levels, ground level, and higher level. Jungle Crows used the ground level in only 6.6% of the observation time, whereas Carrion Crows were on the ground during 36.3% of the ob- servation time (Table 3). Jungle Crows stayed on the ground for Smin at the longest, whereas Carrion Crows stayed for up to 23 min. The mean duration was 1.3 and 3.8min for Jungle and Carrion Crows, respectively, which is a significant difference (Table 3, Mann-Whitney U-test, N=96 (Jungle Crow) and 181 (Carrion Crow), U=3253, P<0.001). The fre- quency of flying down to the ground was also signifi- cantly lower in Jungle Crows (Table 3, Mann-Whit- ney U-test, N=96 (Jungle Crow) and 181 (Carrion Crow), U=10, P<0.001). 3) Comparison of feeding sites Mean distances between feeding quadrates were significantly larger in Jungle Crows than in Carrion 107 1 Jungle Crow fed [ Carrion Crow fed x: Garbage Station or Restaulant A: Forest v: Glassland -: River -: Open land Blanc: Urban area Wie initm aity Vin daeisity Fig. 3. 20X20m mesh quadrats were laid over Fig. 1 for quantitative estimation of areas of feeding sites and dispersal. Solid quadrats show the feeding sites used by the crows. Poly- gons show the territories. Crows (Figs. 3 and 4, Tables 4 and 5, Mann-Whitney U-test, N=45 (Jungle Crow) and 317 (Carrion Crow), U=0, P<0.001). In Carrion Crows, the mean distance was nearly 1, indicating that Carrion Crow feeding sites are continuous and form a small number of large clusters. On the other hand, in Jungle Crows H. MATSUBARA AANAA? >>> >> AAyvvy >>> x: Garbage Station or Restaulant (0 Jungle Crow fed Carrion Crow fed A: Forest v: Glassland -: River -: Open land Blanc: Urban area Fig. 4. 20X20m mesh quadrats were laid over Fig. 2 for quantitative estimation of areas of feeding sites and dispersal. Solid quadrats show the feeding sites used by the crows. Poly- gons show the territories. the distances ranged from | to 15, indicating that Jun- gle Crow feeding sites are dispersed, and form iso- lated, small clusters (or consist of isolated single quadrats). Jungle Crows used a small number of feeding quadrats. The number of feeding quadrats in Carrion Crows was significantly larger than in Jungle Crows (Tables 4 and 5, Mann-Whitney U-test, N=7 (Jungle Crow) and 12 (Carrion Crow), U=0, P<0.001). Jun- gle Crows had larger territories than Carrion Crows (Tables 4 and 5, Mann-Whitney U-test, N=7 (Jungle Crow) and 12 (Carrion Crow), U=10.5, P<0.01), and accordingly the feeding site density was signifi- cantly lower in the former than in the latter (Table 4, Table 5, Mann-Whitney U-test, N=7 (Jungle Crow) and 12 (Carrion Crow), U=0, P<0.001). These re- sults indicate that Jungle Crows use a smaller number of more dispersed feeding sites in larger territories than Carrion Crows. In Jungle Crow territories, the frequency of quadrats occupied by urban area was higher than in those of Carrion Crows (Fig. 5, Mann-Whitney U- test, N=7 (Jungle Crow) and 12 (Carrion Crow), U=12, P<0.05). The location of large garbage sta- tions (more than 22m in size) and restaurants 108 Table 4. Characteristics of feeding site distributions in Jun- gle Crows. See also Figs. 3 and 4 No. pe Territory Quadrate : i distance , F Pair feeding size density quadrats No (ha) (quadrat/ha) quadrats) Jl 5 5.20 2.07 2.42 J2 7 4.29 6.32 1.11 J3 6 4.67 6.18 0.97 J4 10 4.50 5.64 7 J5 9 3.80 6.12 1.47 J6 8 4.00 6.38 1.25 J7 6 2.63 357) 1.68 Mean 7.29 4.16 5.18 1eS2 Min 5 2.63 2.07 0.97 Max 10 5.20 6.38 2.42 SD 1.80 0.82 1.69 0.49 Table 5. Characteristics of feeding site distributions in Car- rion Crows. See also Figs. 3 and 4 No. ate Territory | Quadrate ! ; distance : ‘ Pair feeding aN} size density quadrats cada) (ha) (quadrat/ha) Cl 19 1.37 2.07 9.18 C2 24 1.38 3.43 7.00 C3 31 1.03 1.59 19.50 C4 py 1.30 2.38 11.30 C5 37 1.11 4.05 9.14 C6 13 1.15 1.13 11.50 C7 23 1.41 3.51 6.55 C8 29 1.14 1.96 14.80 C9 43 1.13 4.11 10.50 C10 29 1.11 Sul) 9.15 Cll 21 1.10 3.37 6.23 C12 13 1.15 2.10 6.19 Mean DAS 1.20 2.74 10.09 Min 13 1.03 1.13 6.19 Max 43 1.41 4.11 19.50 SD 8.94 0.13 0.99 3.92 (garbage was available everyday) where the crows can feed on garbage are shown in Figs. 3 and 4. The frequency of feeding at garbage stations among all feeding quadrats was significantly higher in Jungle Crows than in Carrion Crows (Fig. 6, Mann-Whitney U-test, N=7 (Jungle Crow) and 12 (Carrion Crow), U=0, P<0.001). In the former case it accounted for 81.2%, but in the latter only for 6.4%. However, the extent to which the species depended on garbage Comparative study on syntopic corvids area in the territories (%) Frequencies of urban Carrion Crow Jungle Crow Fig. 5. The proportion of urban areas in the territories of Jungle Crows and Carrion Crows calculated from the number of quadrats. The proportion is larger in Jungle Crows than in Carrion Crows (Mann-Whitney U-test, P<0.05). The boxes, the lines inside the boxes and the bars indicate 50% of sam- ples, median and standard error, respectively. 100 OS as S| °°) Ee og — OS 60 o5 nn 2 w s Se)) (= os 20 = ov oS 0 ME econo lacs Jungle Crow Carrion Crow Fig. 6. The frequencies of feeding in quadrats containing garbage stations among all feeding quadrats. The frequency is higher in Jungle Crows than in Carrion Crows (Mann-Whitney U-test, U=0, P<0.001). Boxes, lines inside boxes and bars in- dicate 50% of samples, medians and standard errors, respec- tively. could not be estimated because the type and amount of food items taken up by the crows were unknown in many cases. DISCUSSION Jungle and Carrion Crows defended their territo- ries both intra- and interspecifically. This territoriality indicates that intrusions both of members of the same species and of different species have negative conse- quences for the territory owners. Crows attacked many species, but the extent of aggressiveness against each attacked species is uncertain. Because the territories of the two species of crows tended to be distributed alternately, the frequency of antagonis- tic behavior between them could be expected to be high. Each territory of each species was adjacent to 109 territories of the other species, which meant that con- flicts could more easily occur between the species than among members of the same species. Carrion Crows attacked Black Kites more frequently than Jungle Crows did, possibly because Black Kites were usually observed flying over the Kamo River and the Takano River, where Carrion Crow territories were situated. In the cases when Black Kites intruded into Jungle Crow territories, the Jungle Crows also at- tacked the kites aggressively. The mean duration of terrestrial feeding is signifi- cantly smaller in the Jungle Crows. The frequency of flying down to the ground is also significantly smaller, and therefore the total duration of terrestrial feeding is significantly smaller in the Jungle Crows. This difference in feeding behavior between the two species has not been quantitatively estimated in pre- vious studies. The difference in feeding behavior co- incides with their microhabitat preferences; i.e. the longer time spent on the ground by Carrion Crows coincides with their utilization of natural ground sur- faces as feeding sites. The microhabitat composition of the territories dif- fers between the species. The proportion of urban areas 1s significantly higher in the territories of Jungle Crows than in those of Carrion Crows. The same ten- dency was also found in previous studies (Higuchi 1979; Nakamura 1998). The use of feeding habitats differs between the species. Jungle Crows use isolated, small sites within larger feeding areas. Carrion Crows, in contrast, use large, continuous areas within feeding areas. This dif- ference relates to the differing amounts of time spent foraging on the ground between the two species, and their foraging habits (Jungle Crows commonly drop to take food spotted from a vantage point, while Car- rion Crows commonly forage on foot). The feeding sites of Jungle Crows are mainly garbage stations, but those of Carrion Crows are located in various envi- ronments, such as forest floors, grasslands, and river floodplains in addition to a small number of garbage stations. It is suggested that the difference in feeding microhabitats results from an inverse relationship be- tween territory size and the number of feeding quadrats. Jungle Crows feed mainly on human food scrap at a small number of good feeding sites where they can easily access food and feed, such as at large garbage stations. In contrast, Carrion Crows used var- ious microhabitats and spend longer on the ground in each, therefore Carrion Crow territories are smaller but the number of feeding quadrats is larger than in H. MATSUBARA Jungle Crows; Carrion Crows, it seems, are able to find food resources in any microhabitat. Jungle Crows in urban areas have been suggested to depend on organic waste (e.g. Higuchi 1979). Car- rion Crows also forage at garbage stations, although this behaviour is less frequent. It appears that waste is not the main food resource for Carrion Crows in the study area. Small garbage stations, which are not shown in Figs. 3 and 4, are situated in urban areas. However, the territories of Carrion Crows contain less urban area than those of Jungle Crows, therefore the number of such small garbage stations is probably lower in Carrion Crow territories. The different pro- portions of urban areas in the territories of the two species are most likely related to differences in mi- crohabitat preferences and feeding behavior. Another possibility is that Jungle Crows include sufficient garbage stations within their territories to provide all their food resources, thus they do not need to forage in other microhabitats as Carrion Crow do. The split territories observed in two pairs of Jungle Crows in this study, were probably caused by the par- ticular pattern of resource distribution and the de- fence cost. Territoriality is adaptive when the benefit by exclusive use of the area is larger than the defence cost (Maynard Smith 1982); the Takano River is a re- source-poor microhabitat for Jungle Crows, and it would be too costly to exclude territorial Carrion Crows. On the other hand, riverbeds are valuable habitats for Carrion Crows, which implies that the same type of environment can have asymmetrical value for the two species. If food resources that Jun- gle Crows prefer, such as food scraps or carcasses, are found inside a Carrion Crow territory where it overlaps with a Jungle Crows’ home range, resource competition between the species may occur. I ob- served that both Jungle Crows and Carrion Crows fed on food scraps at several garbage stations in a com- mercial area in Kyoto City, and the Jungle Crows dominated the food in several cases (Matsubara, un- published data). This suggests that the two species compete for this food resource to a certain extent. As demonstrated above, feeding habitats and feed- ing behavior differ between the two species, but as indicated in a previous study (Ikeda 1957), their food preferences largely overlap. Jungle Crows are likely to be dominant over Carrion Crows because they are larger: the body weight range of the former is 570- 860 g, and that of the latter is 370-730 g (Tamada & Fujimaki 1993). Another potentially important issue is nest defence, because corvids are known to prey on 110 other bird’s nests. Nest predation by conspecific young individuals has been suggested to be the main reason for reproductive failure in Carrion Crows (Yom-Tov 1974). I observed one case of a Jungle Crow attempting to predate a Carrion Crow’s nest (Matsubara, unpublished data). In Tokyo, successful nest predation by Jungle Crows upon Carrion Crows’ nests has been reported (Hitoshi Fujimura, personal communication). Thus, interspecific territoriality be- tween Jungle Crows and Carrion Crows may partly be due to food resource defence and partly to nest de- fence. The differences in feeding behavior and habitat preference probably reduce resource competition, but apparently they are insufficient to avoid competition for food entirely. It is suggested that Carrion Crows in the study area are excluded from zones with abun- dant supply of organic waste because these are occu- pied by the dominant Jungle Crows; Carrion Crow territories probably do not offer enough garbage as a food resource for Jungle Crows to exclude Carrion Crows. However, Jungle Crows can be strong com- petitors for Carrion Crows if food that is attractive for Jungle Crows is found inside Carrion Crow territo- ries. Interspecific territoriality and overlapping home ranges are likely to be maintained by the differences of feeding behavior, food habits and the distribution of food resources between microhabitats. ACKNOWLEDGEMENTS I am grateful to M. Imafuku and A. Mori for their valuable comments on the manuscript and suggestions about this study. I also thank S. Yamagishi for many helpful remarks and suggestions, and S. Nakamura and H. Fujimura for providing important information. REFERENCES Goodwin D (1982) Crows of the world. 2nd ed. British Museum Natural History, London. Haneda K & lida Y (1966) Life history of eastern car- rion crow (Corvus corone orientalis) | breeding sea- son (I). Jpn J Ecol 16: 97-1059 (in Japanese with English summary). Heinz R (1989) Habitat-specific growth and fitness in Carrion Crows (Corvus corone corone). J. Anim. Ecol. 58: 427-440. Higuchi H (1979) Habitat segregation between the Jun- gle Crow and the Carrion Crow (Corvus macrorhyn- chos, C. corone), in Japan. Jpn J Ecol 29: 353-358. Ikeda S (1957) On the habits of some birds belong to the family corvidae. Ornithological and mammalogical Comparative study on syntopic corvids report 16 Wildlife Section, Forestry Agency, Tokyo (in Japanese). Kuroda N (1974) Some behavior and vocalization of jungle crow, illustrated by photos and sketches. J Ya- mashina Inst Ornithol 7: 427-438. Kuroda N (1981) Observations on a schizochroismic Corvus macrorhynchos, with notes on its territorial life. J Yamashina Inst Ornithol 13: 215-227 (in Japanese with English summary). Kuroda N (1990) Jungle crows of Tokyo. Yamashina Inst Ornithol, Abiko. Loyn RH (2002) Patterns of ecological segregation among forest and woodland birds in south-eastern Australia. Ornithological Science 1: 7—27. Maynard SJ (1982) Evolution and the theory of games. Cambridge Univ Press, Cambridge. Nakamura S (1998) The Territorial behavior of the car- rion crows Corvus corone corone in Japan. Jpn J Or- nithol 46: 213-223 (in Japanese with English sum- mary). Nakamura S (2000) Nest site comparisons between the 111 carrion crow Corvus corone and jungle crow Corvus macrorhynchos in Takatsuki city. Jpn J Ornithol 49: 39-50 (in Japanese with English summary). Nakamura T (1978) A study of paridae community in Japan. IV. Ecological segregation of species of use bill in space and technique. Misc. Rep. Yamashina Inst. Ornithol. 10: 94-118. Takahata Y, Sato K, Niki Y, Mori Y, Komemoto K, Ma- suda Y, Harada T, Simizu K & Koshio C (1996) Pre- liminary report on the distribution of two crow species around Naruto, Tokushima. Bull Naruto Univ Educ. 11: 21-27 (in Japanese with English summary). Tamada K & Fujimaki Y (1993) Breeding biology of Corvus corone and C. macrorhynchos in central Hokkaido. Jpn J Ornithol 42: 9-20 (in Japanese with English summary). Yom-Tov Y (1974) The effect of food and predation on breeding density and success, clutch size and laying date of the crow (Corvus corone L.). An experimental approach. J Anim Ecol 43: 479-498. ha nei ihbey bal! Dada FaatfS at Cee RN = Sef Ne jets Lari i esi et ivr wepered edo wbiteeei Se) AER ate Gui to cebichietusheteattieal . veyed oo polos Mi oant ung condi veaticals bas mae cee al Thee eeanngph tee dihes - Raaipanisilirney test eth jn: in ny tine A HRSA T pence ms «iad hore erent wareulh enc + Wino All omtiken Cow wterpung 4 pnishe dod daaten ut nel nee PA Peat: Abid Cre ey i be j [hp io Ported 9h. wear feat yh One eee eae Ae Cbelhat lied “Ni NEE fibh 3 Sin fe ih PP ar B “ot oh, A ISIGHE T sharelh aD DHA, ; napeintesl stl). oe RET, Geni, ~ : v ae”, Oy Rep. wt tues) a | : MeHg AAD Mes Mico aly acoaneygod RRS rates: Ts » fred comme défenicd a NY ‘ a ba ej } woe tiny Wes rn ah wh poi ": te 2 f ay Moy 2 1, in ‘ ot. sf ban (0 fy Y Bhaen ener ee pe No ae : ; i at Telcdv iene sh tte, wyios 3 are ; rp a Sulldel hi) ULE Wit havo th te i flat may *s cy 5 Anes pene rye y ried ae yb a ey wie hii, du) Miuhl” en sot ‘BAR iP Da spetitiiovtt . wet mye nempst "te SR WESIE patos aye We OY hie PV GS SAT WY i ale erg HW an RE Hi Cees | Ob Ae PIS YOSee AMNLANNTAE Tg vs Tooke y Go ant offge ees! ve Commie Crow & itr hore andabucdnndd alter weseT ib ef int a7 Paula Crewe, ‘lent © eY, wright rowg cand ehelty Sas Br pa wats Ae ar vo Cpe JF food What is ince disknitivn aanthe\ de —fungle Crows 4° found Josie Cameras uyiaiity eb oddbtive ewe the bevels es. Syterapctiic -fertitorialicy dnl OVE, | be lenges Gian We defeats get 2 tiki ly ty te shnintained Dy Tl ; q « Inher River *& re f fieextine behav. ford habits and r views And 4 f [ood maemo hetweer thicashytit, ees - 3 ee , - “ae id) iio ; | Ue hand, iw a ine alan ': ay KNOWLEDGES on ; ‘uch voplhes eat Ce , yn hae saiehals P : ~— ~ heattuku arcade fount iiries th ¢ CootTiwiaedn the MINER DE Ri ‘ - f aoc thank %. Younes if of dnatiane nt sogsestione, And S We . va” hose sance ' Vebangte Tou peo iding lrapewtane mule An (yell Tes AY. bs aan is ; $ Cantina ( REFERENCES = ataieon An cu ry 19KS) Cane té of rhe anil , apa he aa an " or Aarurnl Mas story: London, a eaiatieei : & rid / 7 (}O66T Life bis ri od _ an cre Ere Cope orlentaligy ed - fou J Fetal tr U7. ened (in é a sonriiry). va ficite 2 TSR Piabital-speeitic aa) a bh M : ; afer ys (Gres psy corm \ teases é27 pd) . MS Pipes OF COTY Hottie aegreganon ie ta Jie 4% piezo we @fel the Carnen Crowitt; slam : via & * thie, © ccoonet wide apet, Fen J Eee ea iwHe hed $ 219575 On the herbie phir “! yon fitemity ccvr@ldetee Setnitiods Lie , -_ peak 7 _ ORIGINAL ARTICLE Ornithol. Sci. 2: 113-118 (2003) Reduction of cost of polygyny by nest predation in the Black- browed Reed Warbler Shoji HAMAO* Institute for Nature Study, National Science Museum, 5—21—5 Shiroganedai, Minato-ku, Tokyo 108-0071, Japan ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2003 Abstract When female birds choose already-mated males as their mates, they suffer some costs. One major cost is a reduction in male parental care. However, since nest predation disturbs the nesting cycles of polygynously mated females, it might change the allocation of male parental care between the females. I investigated the effect of nest predation on the relationship between female mating order and their brood status during the nestling-rearing period in the Black-browed Reed Warbler Acrocephalus bistrigiceps. Polygynous males did not feed later-hatched broods. The nest predation rate was high (56% of nests), which gives subsequently mated females a chance to re- ceive male assistance. Four subsequently mated females acquired monogamous status by the time their eggs hatched, because the previously mated females had failed in their breeding attempts and disappeared from the territories. In addition, a subse- quently mated female’s nest was preyed upon and her renesting delayed her own nest- ing cycle, which resulted in the disappearance of the previously mated female with her fledglings and gave monogamous status at egg-hatching to the subsequently mated female. Furthermore, a case of inversion in the hatching order in polygynously mated females occurred by the prolonged pre-laying period of the previously mated female, which resulted in the subsequently mated female obtaining primary status at egg-hatching. As a result, 43% of the females that paired with already-mated males acquired monogamous or primary status at egg-hatching, whereas 69% of the females that paired with unpaired males did so. This suggests that nest predation reduces the cost of polygynous mating in this reed warbler population. Key words Acrocephalus bistrigiceps, Cost of polygyny, Male parental care, Mate choice, Nest predation 1993) and changes the allocation of male parental When female birds choose already-mated males as their mates, they suffer some costs (Verner & Willson 1966; Orians 1969). One of the major potential costs is a reduction in male parental care (Verner 1964; Orians 1969). Feeding by polygynous males de- creases in later-hatched broods, or in broods of subse- quently mated females (Martin 1974; Patterson et al. 1980; Alatalo et al. 1982; Dyrez 1986; Muldal et al. 1986; Bruun et al. 1997). When nesting cycles of polygynously mated females proceed smoothly, a fe- male that chooses an already-mated male as her mate will be a parent of the later-hatched brood, and will incur the cost of reduced male assistance. Nest predation is one of the major interruptions of the normal nesting cycle (Ricklefs 1969; Martin (Received 17 March 2003; Accepted 7 August 2003) * E-mail: hamao@kahaku.go.jp 113 care between previously and subsequently mated fe- males (Temrin & Jakobsson 1988; Urano 1990). If a previously mated female fails in her breeding attempt and does not renest within the male’s territory, the subsequently mated female will receive male parental assistance exclusively. Therefore, the frequency of nest predation influences the probability that subse- quently mated females receive male assistance. To evaluate the cost of polygyny in a population, it is important to know both the nest predation rate and the relationship between female mate choice (whether an unpaired male or an already-mated male) and their brood status during the nestling-rearing pe- riod. I studied the effect of nest predation on the alloca- tion of male parental care in the Black-browed Reed Warbler Acrocephalus bistrigiceps. This paper pres- S. HAMAO ents data showing how polygynous males assist in caring offspring. I also report the percentage of fe- males that become mothers of assisted broods when they pair with unpaired or already-mated males. Fi- nally, I discuss the extent of effect of nest predation on the cost of polygyny in this species. METHODS 1) Study area and species This study was conducted during the 1993-1997 breeding seasons in a 42.4-ha area of rice fields along the Arakawa River (36°52'N, 139°35’E; 8m asl) in central Honshu, Japan. The study area included patches of scattered grassland, which consisted mainly of the common reed Phragmites communis and goldenrod Solidago altissima. The grassland is the breeding habitat of Black-browed Reed Warblers. The environment of the study area is described in de- tail in Hamao (2000). The Black-browed Reed Warbler is a small (ca. 10g), migrant passerine bird that is endemic to Northeast Asia, and is a summer resident of Japan (OSJ 2000). Males arrive at the study area from late May to mid-July and attract females by singing long, complex songs (Ueda & Yamaoka 1998; Hamao 2000). Females arrive at the site from early June to mid-July, and soon settle in the males’ territories. Fe- males build nests alone, taking widely varying peri- ods to complete the nests (Hamao 2001). The modal clutch size is five. Although a female can produce fledglings twice in a breeding season, it is very rare. During one breeding season, 25% of the males ac- quired two, or rarely three, females, 56% acquired one female, and 19% remained solitary (Hamao, un- published data). 2) Field observations Birds were captured in mist nets soon after arrival, and were individually color-ringed. In some cases where females were building nests when I detected them, I marked the females after they laid their third egg to avoid making the females desert their nest. Using this marking method, no bird disappeared from the territory or deserted its nest. The study area was visited daily from late May to mid-August, except in cases of heavy rain. To monitor pair formation and nesting stage, I visited each of the territories daily or every two days, between sunrise (0440+15 min dur- ing the study period) and 0800, and observed the be- havior of marked birds from a 1.8-m-high stepladder. 114 Observations were made for 15-40 min in each terri- tory. During my visits, I also inspected the nest con- tents. If all the eggs or nestlings in a nest disap- peared, I assumed that it was due to predation. The nest predation rate was based on observations of all nesting attempts that were identified before egg-laying started. Sixty-five such nesting attempts were identified, including renesting by females within the original territory. Data on mate choice, whether a female formed a pair-bond with an unpaired male or with an already-mated male, were obtained for 49 fe- males. Forty-eight breeding females were observed from the time they settled in their territories, or from the early stage of nest building. I also included in this data set one female (#91) that I found after her clutch was complete. When I found this female, another fe- male (#90) was also nesting within the same terri- tory, and the male was apparently solitary before he paired with female #90. Therefore, female #91 clearly paired with the male who had already mated with female #90. If a female renested with the same mate after nest predation, I used the final nesting attempt as the re- sult of her mate choice. If the first nesting failed dur- ing the incubation period and the next nesting pro- duced fledglings, I assumed that the female had pro- duced eggs that hatched. 3) Parental activities To monitor the incubation and feeding to nestlings by parents, I recorded the birds that visited focal nests with a Sony CCD-G100 video camera and a Sony GV-100 video recorder. The video recording was made for 2.6+0.2h (mean+SE, N=43). To in- vestigate incubation behavior, I videotaped the nests 5.2+0.7 days (N=13) after the last egg was laid, and to examine feeding behavior, I videotaped the nests 6.0+0.3 days (N=30) after the first egg hatched. The individuals recorded in the videotapes were identified by their leg rings, and I checked to see whether they incubated eggs and fed nestlings. To describe parental activities, I divided the broods into three categories according to the situation on the day the nests were videotaped. When there was one nest in a territory, the brood was called a monoga- mous brood. When there were two nesting attempts within a territory, the brood at the more advanced nesting stage was assumed to be the primary brood and the one at the less advanced nesting stage was as- sumed to be the secondary brood. This definition re- duced the number of secondary broods that I could Nest predation and polygyny in warbler videotape, because it was uncommon for both broods of polygynously mated females to survive until the day I made video recordings, due to the high preda- tion rate (see results). The time spent incubating by each parent was measured for nine monogamous, two primary, and two secondary broods. The share of feeding by parents was based on data obtained from 19 monogamous, eight primary, and three secondary broods. The nestling age at the time of video record- ing did not vary among the three brood categories (monogamous: 5.90.3 days (mean+SE), primary: 5.6+0.5, secondary: 7.0+1.2; Kruskal-Wallis test, H=1.36, P=0.51). The number of nestlings in a nest also did not vary among the categories (monogamous: 3.60.2, primary: 4.1+0.3, secondary: 4.0+0.6; H= 2.57, P=0.28). To describe relation between mating and hatching order, I divided the broods into three categories by the same way according to the situation on the day the eggs hatched. 4) Statistical analyses I compared the feeding behavior of different pairs in three brood categories: monogamous, primary, and secondary broods. Since the variance among the cate- gories was unequal, I performed the robust rank- order test (Siegel & Castellan 1988) corrected by the sequential Bonferroni method (Rice 1989). All the statistical tests were two-tailed. Values are presented as means+SE. RESULTS 1) Male parental care Males seldom incubated eggs, irrespective of the brood status. Males did not participate in incubating primary or secondary broods. They incubated only three of nine monogamous broods, and the time that they spent incubating (3.6+1.2min h ', N=3) was much shorter than that by their mates (31.5+2.0 min h7', N=3). The frequency of male food delivery varied be- tween the monogamous, primary, and secondary broods (Fig. la). Males never fed the secondary broods. Male feeding frequency differed significantly between monogamous and secondary broods (U=~, P<0.005). Male feeding frequency also differed be- tween primary and secondary broods, but the differ- ence was not significant (U=5.72, 0.05 Food delivery (times h" young ) = Monog Primary Secondary Primary Secondary Monog Primary Secondary Fig. 1. Feeding frequency by the male (a), female (b), and both parents (c). The lines on the bars indicate the SE. The data were obtained from 19 monogamous, eight primary, and three secondary broods. categories (all P>0.2; Fig. 1b). Therefore, the fre- quency with which both parents fed secondary broods was slightly lower than the frequency with which both parents fed monogamous and primary broods (Fig. 1c), but the differences between any pair of the three brood categories were not significant (all P20.) 2) Nest predation Of 65 breeding attempts, 35 failed because of nest predation. Five of the nests were deserted by both parents. Two of the nests were destroyed when farm- ers cut the grass supporting the nests. Consequently, 23 nests produced fledglings. Excluding the two cases of artificial destruction, 56% of the nests (N=63) were preyed upon in the study area. 3) Relation between mating and hatching order Thirty-five females paired with unpaired males, S. HAMAO Females that disappeared before hatching due to nest predation Females that deserted the nest and disappeared before hatching 9 oa , pe ey Females that paired 1 ' ; SSS with unpaired males eo 25 Primary: females whose mates became polygynous and the subsequently mated female was in a less advanced nesting stage 1* Secondary: females whose mates became polygynous and the subsequently mated female was in a more advanced nesting stage 5 . . . 14 [Females that paired __ | Females that disappeared before hatching due to nest predation wilh aleady-matecumales as Females whose eggs hatched 5* 42 Fig. 2. Monogamous: females not sharing the male with other nesting females due to the disappearance of a previously mated female Primary: females sharing the male with a previously mated female that was in a less advanced nesting stage Secondary: females sharing the male with a previously mated female that was in a more advanced nesting stage Relationship between female mate choice and their status at egg-hatching. Bold face shows the status at egg-hatching. The figures are the numbers of females. The causes of the status changes labeled with asterisks are explained in Fig. 3. and 10 (29%) of them disappeared from the territo- ries before egg-hatching, due to nest predation (nine cases) or nest desertion (one case; Fig. 2). Therefore, 25 (71%) of the females that paired with unpaired males produced eggs that hatched. When the eggs hatched, the mates of 15 females remained monoga- mous (Fig. 2), while the mates of 10 females became polygynous. In the cases of polygyny, nine females were the mothers of primary broods at egg-hatching; i.e. the subsequently mated females were at a less ad- vanced nesting stage (Fig. 2). However, one female became the mother of a secondary brood; i.e. the sub- sequently mated female was at a more advanced nest- ing stage (Fig. 2). This inversion in the order of the nesting cycles of the polygynously mated females oc- curred because the previously mated female built a nest three times before egg-laying, which prolonged her pre-laying period, and by the fact that her clutch was larger than that of the subsequently mated fe- male, which also delayed the start of incubation (Fig. 116 Fig. 3. Causes and patterns of change in female status. Bro- ken lines: pre-laying, open circles: egg-laying, thin lines: incu- bation, thick lines: nestling, P: nest predation. A prolonged pre-laying period of a previously mated female also changed the status of the previously mated female (a). Nest predation of a previously mated female (b) and a subsequently mated fe- male (c) also changed the status of the subsequently mated fe- males. See text for details. 3a). Therefore, 69% (24/35) of the females that paired with unpaired males became the mothers of monogamous or primary broods at egg-hatching. By contrast, 14 females paired with already-mated Nest predation and polygyny in warbler males, and five (36%) of them disappeared from the territories before egg-hatching, due to nest predation (Fig. 2). Therefore, nine (64%) of the females that paired with already-mated males produced eggs that hatched. At egg-hatching, five females became the mothers of monogamous broods (Fig. 2), because the previously mated females had disappeared due to nest predation (four cases; Fig. 3b), or because renesting of the subsequently mated female due to predation delayed her egg-hatching, and the previously mated female and her fledglings disappeared in this interval (one case; Fig. 3c). One subsequently mated female became the mother of a primary brood because of the prolonged pre-laying period of the previously mated female (Fig. 3a). Furthermore, her brood became a monogamous brood during the nestling-rearing pe- riod, because the previously mated female’s nest was preyed upon (Fig. 3a). Three females that paired with already-mated males became the mothers of second- ary broods, so their status did not change throughout the nesting period. Therefore, 43% (6/14) of the fe- males that paired with already-mated males became the mothers of monogamous or primary broods at egg-hatching. This proportion (6/14) did not differ significantly from the proportion of females (24/35) that paired with unpaired males (y’?=2.79, df=1, P=0.10). DISCUSSION Male Black-browed Reed Warblers did not feed later-hatched broods. This can be a potential cost when a female chooses an already-mated male as her mate. In some cases, however, female status changed between the time of pair formation and nestling-rear- ing. In this discussion, I exclude the status change from a monogamous to a primary brood, which occurs when the mate of a female succeeds in mating polygynously, because it does not affect male parental assistance to the female. With one exception, in which the nesting cycle of a polygynously mated female became inverted (Fig. 3a), any change in fe- male status was caused by nest predation. When a previously mated female failed in her breeding at- tempt and disappeared from a territory, the subse- quently mated female acquired monogamous status (Fig. 3b). This type of status change has been pointed out previously (Temrin & Jakobsson 1988; Urano 1990). Temrin and Jakobsson (1988) reported that al- most half of the subsequently mated females of the Wood Warbler Phylloscopus sibilatrix had exclusive 117 male assistance because the nests of the previously mated females were preyed upon. I found another type of female status change: renesting of a subse- quently mated female due to nest predation delayed her nesting cycle and resulted in her gaining monoga- mous status at egg-hatching (Fig. 3c). The proportion of females acquiring monogamous or primary status at egg-hatching was higher for fe- males that paired with unpaired males than it was for females that paired with already-mated males, al- though the difference was not significant. Therefore, although female Black-browed Reed Warblers will not necessarily receive male parental assistance more readily when they pair with already-mated males, nest predation allowed 43% of the females that chose already-mated males as their mates to receive male assistance. This suggests that nest predation reduces the cost of polygynous mating. If females incur a cost due to time constraints, for example, when they choose males, a female might pair with an already- mated male. It is worth noting that when more than half of all nests were preyed upon, 69% of the females that paired with unpaired males and 43% of the females that paired with already-mated males acquired the status of the mothers of assisted broods. These pro- portions indicate the extent of the effect of nest pre- dation on the cost of polygyny in this reed warbler population. ACKNOWLEDGMENTS I am most grateful to Keisuke Ueda for his advice and discussion throughout this study. Many thanks are due to Eiichiro Urano and two anonymous referees for their valuable comments on the early version of this paper, and to Noriyuki Yamaguchi for providing advice on sta- tistics. I also thank Eisuke Katoh and Hiromi Hamao for their help with analyzing the video recordings, and K. Ueda, Hiroshi Uchida and Takashi Matsuda for their field assistance. This study was supported by Grants- in-Aid for Science Research (No. 05918009 and 10918005) from the Japan Ministry of Education, Sci- ence, Sports and Culture. REFERENCES Alatalo RV, Lundberg A & Stahlbrandt K (1982) Why do Pied Flycatcher females mate with already mated males? Anim Behav 30: 585-593. Bruun M, Sandell MI & Smith G (1997) Polygynous S. HAMAO male starlings allocate parental effort according to relative hatching date. Anim Behav 54: 73-79. Dyrez A (1986) Factor affecting facultative polygyny and breeding results in the Great Reed Warbler (Acro- cephalus arundinaceus). J Ornithol 127: 447-461. Hamao S (2000) The cost of mate guarding in the Black-browed Reed Warbler Acrocephalus_bistrigi- ceps: When do males stop guarding their mates? J Ya- mashina Inst Ornithol 32: 1-12. Hamao S (2001) Male reproductive tactics in the Black- browed Reed Warblers, Acrocephalus bistrigiceps: cost and benefit of mate attraction. Ph. D. disserta- tion, Rikkyo University, Tokyo. Martin SG (1974) Adaptations for polygynous breeding in the Bobolink, Dolichonyx oryzivorus. Am Zool 14: 109-119. Martin TE (1993) Nest predation among vegetation lay- ers and habitats: revising the dogmas. Am Nat 141: 897-913. Muldal AM, Moffatt JD & Robertson RJ (1986) Parental care of nestlings by male Red-winged Black- birds. Behav Ecol Sociobiol 19: 105-114. Orians GH (1969) On the evolution of mating systems in birds and mammals. Am Nat 103: 589-603. Ornithological Society of Japan (OSJ) (2000) Check-list of Japanese birds. 6th ed. Ornithological Society of Japan, Obihiro. 118 Patterson CB, Erckmann WJ & Orians GH (1980) An experimental study of parental investment and polyg- yny in male blackbirds. Am Nat 116: 757-769. Ricklefs RE (1969) An analysis of nesting mortality in birds. Smithsonian Contributions to Zoology 9: 1-48. Rice WR (1989) Analyzing tables of statistical tests. Evolution 43: 223-225. Siegel S & Castellan NJ Jr (1988) Nonparametric statis- tics for the behavioral sciences. 2nd ed. McGraw-Hill Book Company, New York. Temrin H & Jakobsson S (1988) Female reproductive success and nest predation in polyterritorial Wood Warbler (Phylloscopus sibilatrix). Behav Ecol Socio- biol 23: 225-231. Ueda K & Yamaoka A (1998) Decrease of song fre- quency after pairing in the polygynous Schrenck’s Reed Warbler Acrocephalus bistrigiceps. J Yamashina Inst Ornithol 30: 53-56. Urano E (1990) Factors affecting the cost of polygynous breeding for female Great Reed Warblers Acro- cephalus arundinaceus. Ibis 132: 584-594. Verner J (1964) Evolution of polygamy in the Long- billed Marsh Wren. Evolution 18: 252-261. Verner J & Willson MF (1966) The influence of habitats on mating systems of North American passerine birds. Ecology 47: 143-147. ORIGINAL ARTICLE Ornithol. Sci. 2: 119-125 (2003) Seed dispersal agents of two Ficus species in a disturbed tropical forest Kelvin S.-H. PEH'** and Fong Lin CHONG! ' Department of Biological Sciences, National University of Singapore, Blk S2, 14 Science Drive 4, Singapore 117543, Republic of Singapore ? Swedish Biodiversity Centre, CBM, Box 7007, 750 07 Uppsala, Sweden ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2003 Abstract Observations were carried out at the Bukit Timah Nature Reserve, Singa- pore on two species of the keystone genus, Ficus fistulosa and F. grossularoides. This study shows that the two species of different morphological characteristics (e.g. plant height, fruit colour and size) attracted different assemblage of avian frugivores. The frequency of visits by the avian frugivores was significantly different between the two Ficus species. The fig-eating frugivore communities feeding on both Ficus species seemed to be comparatively depauperate and a substantial number of members were the non-obligate (i.e. routine) avian frugivores. Mammalian frugivory was also ob- served. The mean number of feedings at F. grossularoides might not correlate with body size of the avian frugivores. Such information may aid the forest conservation and management of the nature reserve and future attempts at forest restoration. Key words Seed dispersal Ficus is one of the largest genus with approxi- mately 350 species in Southeast Asia (Corner 1965). It is widely considered as a keystone mutualist for many vertebrates in the tropical rain forest (e.g. Janzen 1979). Being abundant and always available throughout the year, figs constitute an important diet for many frugivorous animals when other food re- sources (e.g. insects) are scarce (Leighton and Leighton 1983). It comprises the most important class of plant resources as Terborgh (1986) has shown that the species that feed on the figs constitute about 40% of the animal biomass at Cocha cashu, Peru. Lambert and Marshall (1991) identified the characteristics, which make figs the most important keystone plant resources: their large crop sizes, rela- tively short fruiting intervals, intra-crown synchrony of fruit ripening and intra-population asynchrony. The ecological and evolutionary importance of fru- givores as seed dispersers in tropical rain forest is well studied (e.g., Estranda and Coates-Estranda 1986; Fleming 1986). The mutualistic relationships between Ficus and frugivores are also well known. (Received 8 April 2003; Accepted 14 August 2003) * E-mail: kelpeh@ yahoo.com Biodiversity, Conservation, Ficus, Frugivory, Fig-disperser declines, Ficus is heavily dependent on the frugivores to dis- perse their seeds (Lambert 1989; Lambert 1991). Seed dispersal by animals ensures the long-term sur- vival of many Ficus species. Such _ ecological processes provided by the frugivores may also deter- mine the Ficus species and genetic composition in the disturbed landscapes (Corlett 1995; Hamilton 1999). Thus the patterns of visit by frugivores may influence the succession in disturbed areas such as the forest edge or gaps. On the other hand, the pres- ence of fruiting trees in the disturbed landscapes may maintain the frugivorous faunal communities in these areas (da Silva et al. 1996; Restrepo et al. 1999). Un- derstanding such ecological processes is essential for the conservation of biodiversity and restoration of the disturbed landscapes. In this study we attempted to fill the gaps in knowledge on the fig-eating frugivore assemblage in a disturbed forest habitat. First, we recorded the com- position of diurnal vertebrate assembly, feeding on the figs of the two keystone pioneer Ficus species, F fistulosa and F. grossularoides in a highly isolated, disturbed forest reserve. We compared the species as- semblage and visitation pattern in term of frequency and number of fruit consumption, feeding guild and 119 K. S.-H. PEH and F. L. CHONG feeding method of frugivores between the two Ficus species. The comparison of the feeding guild and fruit consumption of the visiting frugivores for each Ficus species could provide information on their fruit usage and this relates to the importance of figs as a food source. We determined if there was a relation between the frugivore’s body size and the number of figs eaten because it is always assumed that the larger frugivores tend to be more efficient fig dispersers as they consume more figs (e.g. Shanahan et al. 2000). We indirectly examined the relative importance of the observed frugivores for both Ficus species based on literature reviews (e.g. Corlett 1998; Shanahan et al. 2000). Lastly, we reviewed the conservation status of resident fig-eating birds of the reserve. METHODS 1) Study site The study was carried out between 20 January and 24 March 1999 at the Bukit Timah Nature Reserve (hereafter Bukit Timah) in the mainland of Singapore (1°20'N 103°50’E), which situated in a typical equa- torial climate. Bukit Timah is a small, isolated forest fragment with the surrounding matrix of urban areas. The 7lha nature reserve comprises primary hill Dipterocarp forest, which has never been logged, and secondary forest patches of various ages (Corlett 1990). Despite its small size and high human distur- bance, Bukit Timah harbours a high floral diversity with more than 800 species of native plants (Corlett 1990). It is also the last primary forest remnant in Singapore where native mammals, birds, reptiles and invertebrates can still be found. The area surveyed was a Stretch of forest of secondary nature in the more disturbed part of the reserve. 5 ) Study subjects We chose Ficus fistulosa and F. grossularoides for this study primarily because of their relative abun- dance. Both members of the Ficus are pioneer Species, commonly found in the secondary patches of Bukit Timah. They are different morphologically in terms of plant height, fruit size, fruit colour and prob- ably also in their palatability and inherent nutrition. F: fistulosa is the taller of the two, reaching a height of 9m. Its fruits are borne in clusters on woody knobs that line the trunk and main branches. The fruit, of the average diameter of 2.5cm, are green with white dots, ripening to pale yellow or greenish yellow. F- grossulariodes is a shrub-like or small tree that 120 reached a maximum of 3 m. Its fruits have an average diameter of 1.25cm. Those are sessile, round and are mostly on twigs just below the leaves. Those ripen from yellow to brownish ochre to dark red. Descrip- tions of the two Ficus species are derived from Ng (1978). 3) Focal observations and statistical analyses We conducted an observation for 12 days between 0700h and 1000h when the frugivores are most ac- tively feeding in the morning. Different fruiting trees of both Ficus fistulosa and F. grossularoides were sampled for each morning session. Thus 12 trees with a total of 18h of observation were performed for each Ficus species. For every session, we chose one fruit- ing tree of each Ficus species, of which at least half of their crown visible from the ground, for focal watch. The observations were carried out in 15 min blocks that alternated between the two species. Binoculars (8X30) were used to aid in identification of the feeding frugivores. Identification of the avian and mammalian frugivores followed King et al. (1989) and Lekagul and McNeely (1988) respec- tively. Frugivores foraging in the fruiting Ficus trees were recorded and grouped accordingly to their feeding guilds: animals that are obligate frugivores feed mainly on fruits and routine frugivores have a wide range of mixed diet. Foraging habits of the frugivores in the fruiting Ficus were also observed and classified into three feeding methods following Trainer and Mill (1984): species that are “swallowers” took the figs as a whole, “marshers” that squashed the figs in the beaks before swallowing the whole or most of the fruit, and “biters” that pecked at the fruit, feed- ing on the figs in small bits. We noted the number of fruit eaten by the visiting frugivores. The taking of a whole fruit or any re- moval of the fruit in small section was counted as a feeding. When it was not obvious if the animal had fed on some fruit, the head-pecking movements were counted as well. We used Spearman-ranked correla- tion test to determine if there was any relation be- tween the mean number of figs eaten by the “swal- lowers” and their body sizes. Body sizes of the frugi- vores were obtained from Robson (2000). All data are analysed using Statistical Analysis System (SAS). Invertebrates, like ants and other insects, were not in- cluded in this study. Another important group of fru- givores, fruit bats, was also not dealt with. Seed dispersal agents of two Ficus species 4) Fig-eating bird assemblage We compiled a checklist of forest resident fig-eat- ing birds of Bukit Timah, from pre-1940 to 1991, based on literature research (e.g. Lim 1992; Shanan- han et al. 2000) and personal observations. The con- servation status of these fig-eating frugivores in Bukit Timah was reviewed according to Lim (1992). The quality of the fig seed dispersers was difficult to judge and was based on a review paper by Corlett (1998). The nomenclature of avian and mammalian frugivores followed Inskipp et al. (1990) and Corbet and Hill (1991), respectively. RESULTS Birds A total of 15 bird species were spotted feeding on figs of both Ficus species (Table 1). Four species of birds were observed feeding on figs of F. fistulosa whereas 14 species were seen on F. grossularoides. We observed eight individuals belonging to three res- ident and one introduced species on F. fistulosa. The most common species seen feeding on the figs of F: fistulosa was the Pink-necked Pigeons. We found 34 individuals belonging to 13 resident and one migrant 1) Table 1. species feeding on the figs of F. grossularoides. The Pink-necked Pigeons were the most common frugi- vores for F: grossularoides, followed by the Yellow- vented Bulbul and White-vented Myna. The common species that fed on both Ficus were the Pink-necked Pigeons, Black-naped Orioles and Asian Glossy Star- lings. More species of avian frugivores were ob- served feeding on F. grossularoides than on F. fistu- losa (¥°=6.20, P<0.05). Similarly, more individuals were seen feeding on F. grossularoides than on F. fis- tulosa (¥?=16.12, P<0.01). The average number of feedings by all avian “biter” species for F: fistulosa ranged from three to five pecks per visitation (Table 1). The only “swal- lower” for F: fistulosa, the Great Hornbill, was ob- served to consume eight figs during its one-timed vis- itation. The biters for F) grossularoides comprised three species, which had an average of two to six mean feedings per visitation. The average number of feedings for F grossularoides by the “swallowers” species (12 species) ranged from two to nine figs per visit. For F. grossularoides, one Asian Fairy Bluebird was observed to swallow the most number of fruits, a total of 9 figs during its visitation. List of frugivores observed feeding on Ficus fistulosa and F: grossularoides between 20 January and 24 March 1999. Terminology for feeding methods follows Trainer & Will (1984). X indicats no data available. Species Feeding guilds Birds Great Hornbill Buceros bicornis Routine Red-crowned Barbet Megalaima rafflesii Obligate Coppersmith Barbet Megalaima haemacephala Obligate Pink-necked Pigeon Treron vernans Obligate House Crow Corvus splendens Routine Scarlet-backed Flowerpecker Dicaewm cruentatum Routine Orange-bellied Flowerpecker Dicaeum trigonostigma Routine Asian Fairy Bluebird /rena puella Obligate Black-naped Oriole Oriolus chinensis Routine Yellow-vented Bulbul Pycnonotus goiavier Routine Olive-winged Bulbul Pycnonotus plumosus Routine White-vented Myna Acridotheres javanicus Routine Asian Glossy Starling Aplonis panayensis Obligate Short-tailed Babbler Trichastoma malaccense Routine Eye-browed Thrush Turdus obscurus Routine Mammals Common Treeshrew Tupaia glis Routine Long-tailed Macaque Macaca fascicularis Obligate Plantain Squirrel Callosciurus notatus Routine Ficus fistulosa Ficus grossularoides Feeding N MeanN Feeding N MeanN Methods visits feedings Methods visits feedings Swallower ] 8.0 Swallower 1 5.0 Swallower 1 2.0 Biter 4 5.0 Swallower 9 3.5 Swallower 1 X Biter 1 6.0 Biter 1 DS Swallower 1 9.0 Biter 2 3.5 Swallower 2 6.0 Swallower 7 3.8 Swallower 1 7.0 Swallower 5 3.4 Biter 1 3.0 Swallower D 4.0 Biter I 2.0 Swallower 1 x Biter 1 x Marsher 6 XxX Biter 5 2.8 Biter 3 3.0 121 K. S.-H. PEH and F. L. CHONG 2) Mammals Three species of mammals were found to be feed- ing on the figs of one or both Ficus species (Table 1). They were the Common Treeshrews, Long-tailed Macaques and Plantain Squirrels. We observed 12 in- dividuals belonging to three mammalian species feeding on the figs of F. fistulosa but recorded only three Plantain Squirrels feeding on F- grossularoides. Thus, more individuals of mammals were observed feeding on F. fistulosa than on F. grossularoides (7 =5.46, P<0.05). The most common mammal eat- ing the figs of F. fistulosa was the Long-tailed Macaques. The average number of feedings by the Plantain Squirrels for F: fistulosa and F: grossularoides are 2.8 and three per visitation, respectively (Table 1). The data for other species was unable to obtain because poor visibility hampered the observation. 3) Feeding guilds, Foraging methods and Body sizes The relative abundance of non-obligate (i.e. rou- tine) frugivore among the vertebrate assembly was substantial for both Ficus species (Fig. 1). Only three species, which comprised 11 individuals, were con- sidered as obligate frugivores among the species as- semblage for F fistulosa. The obligate frugivores feeding at F. grossularoides included four species of a total of 14 individuals. The numbers of obligate fru- givores (including mammals) between both Ficus species are not significantly different (y°=0.4, P>0.01). However, more obligate avian frugivores were observed feeding at F. grossularoides than at F. fistulosa G43 P<0.05). More “swallowers” were observed feeding on F grossularoides whereas the consumers for F. fistulosa were mostly “biters” (Fig. 2). There was an absence of “marshers” on F: grossularoides. For the avian fru- givores that swallowed figs of F. grossularoides, we found that there was no correlation between the num- ber of fig eaten and their body sizes (Spearman- ranked correlation=0.373, P>0.05). 4) Fig-eating bird assemblage 46.2% of the 65 resident avian fig-eating species from 17 families recorded since pre-1940 still persist in Bukit Timah. Of the remaining 30 species, 66.7% are considered to be of high quality fig-seed dispers- ing (HQFD) species which are more efficient at seed dispersal. However only nine species are HQFD for- est specialists and all of them are classified locally as 100% O Routine frugivores @ Obligate frugivores Ficus species Proportion of feeding guilds observed at each F. grossularoides F. fistulosa Fig. 1. Proportion of feeding guilds observed at both Ficus fistulosa and F: grossularoides based on the number of indi- viduals observed. Routine frugivores are species that have a wide range of mixed diet whereas obligate frugivores feed mainly on fruits. 100% » @ Marshers B Biters O Swallowers 50% 0% Proportion of feeding methods observed at each Ficus species F. fistulosa F. grossularoides Fig. 2. Proportion of feeding methods observed at both Ficus fistulosa and F: grossularoides, based on the number of individuals observed. Definitions for foraging methods follow Trainer and Will (1984). threatened species (Table 2). The more important avian fig-eating families at Bukit Timah in term of the number of genera and species eating figs are Pyc- nonotidae, Irenidae, Corvidae, Nectariinidae, Cuculi- dae and Lybiidae. We observed 10% of the total HQED species assemblage feeding on figs of F: fistu- losa and 4% of that on F. grossularoides (excluding winter migrants, reintroduced and _ non-native species). DISCUSSION Comparison between F. fistulosa and F. grossu- laroides shows the difference in frugivore diversity. Both species exhibit different characteristics, which attracted different assemblage of diurnal frugivores. 122 Seed dispersal agents of two Ficus species Table 2. Contributions of families to total number of high quality fig dispersing (HQFD) species in Bukit Timah from pre- 1940 to 1991. The quality of the fig dispersers was based on Corlett (1998). Migrants and non-native species were not included. Proportion of Total fig Extinct Extant HQFD forest Hineec tence HQED in relation af 3 ; ie HQED forest Families dispersers species HQFD specialists A to total extant (N) (N) (N) (N) PSST fig dispersing (N) species (%) Phasianidae 2 2 0 0 0 0 Lybiidae 6 4 2 l 1 6.7 Picidae 2 ] 0 0 0 0 Bucerotidae 3 3 0 0 0 0 Dacelonidae 1 1 0 0 0 0 Trogonidae 1 l 0 0 0 0 Cuculidae 2 0 2 1 1 6.7 Psittacidae 3 0 0 0 0 0 Columbidae 6 3 0 0 0 0 Rallidae 1 0 0 0 0 0 Corvidae 9 5 3) 0 0 10.0 Eurylaimidae 4 4 0 0 0 0 Irenidae 4 0 4 4 4 13.3 Nectariinidae 8 5 3 1 1 10.0 Pycnonotidae 10 6 4 2 2 13.3 Sturnidae 2 0 2 0 0 0 Sylviidae 0 0 0 0 0 The potential primary dispersal agents for F: fistulosa are mainly terrestrial mammals and bats (Lambert 1991). Its figs are too large for most avian frugivores to swallow as a whole. Although the fruits of F: fistu- losa do not show any of the typical bird-dispersal syndromes described by van der Pijl (1972), they caught the attention of some relatively larger birds such as the Great Hornbill and Pink-necked Pigeons and also 10% of the total HQFD species assemblage at Bukit Timah. This showed that F: fistulosa fits into the description for tropical Ficus species associated with generalist nature of the disperser assemblage. The contribution of individual frugivore species var- ied considerably. The HQEFD species observed feed- ing in F- fistulosa were the Black-naped Oriole, Asian Glossy Starling and Long-tailed Macaque. Since half of the recorded avian visits were pink-necked pi- geons, this seed predator exhibited a pattern of domi- nance in avian fig-eating assemblage to F. fistulosa. On the other hand, F: grossularoides were visited by a wider range of fig-eating birds even though Lambert (1989) reported that there was no bird feed- ing on the figs of F. grossularoides at Kuala Lompat, Malaysia. Although F: grossularoides exhibits typical bird dispersing qualities in fruit size, colour and asyn- chrony in ripening, its figs are also readily consumed 123 by the Short-nosed Fruit Bats, Cynopterus spp. (Ling Ong, pers.com). Most of the avian frugivores swal- lowed the figs except the smallest flowerpeckers that pecked the fleshy tissues of the figs. The HQFD species observed on F: grossularoides were the Red- crowned Barbet, Coppersmith Barbet, Scarlet-backed Flowerpecker, Orange-bellied Flowerpecker, Asian Fairy Bluebird, Black-naped Oriole, Yellow-vented Bulbul, Olive-winged Bulbul and Asian Glossy Star- ling. The Pink-necked Pigeons and Yellow-vented Bulbuls dominated the fig-eating assemblage. Most of these species are the smaller frugivores, which are not obligate fig-eaters. Nevertheless, these small gen- eralists are important for the fig seed dispersal as our results suggested that the number of figs eaten might not correlate with body sizes of the birds. Only nine avian HQED species at Bukit Timah are forest specialists that frequent in relatively undis- turbed part of the reserve, and all of them are classi- fied as locally threatened species. They are the Red- crowned Barbet, Drongo Cuckoo, Blue-winged Leaf- bird, Lesser Green Leafbird, Greater Green Leafbird, Asian Fairy Bluebird, Yellow-vented Flowerpecker, Red-eyed Bulbul and Cream-vented Bulbul. The de- clines of the fig-eating frugivores at Bukit Timah since 1800s no doubt limited the number of species K. S.-H. PEH and F. L. CHONG feeding on both F. fistulosa and F. grossularoides. lhe absence of more important fig-seed dispersers at Bukit Timah may lead to a dominance of small gener- alist (routine) frugivores that are more tolerant to fragmentation and able to switch diets opportunisti- cally in their altered landscapes (Table 1; Fig. 2). The only avian HQFD forest specialists contributed to the recorded visits were the Red-crowned Barbet and Asian Fairy Bluebird. The loss of HQFD species in biotic communities may have an irreversible effect on the stability, functioning and sustainability of an ecosystem (Tilman 1997). Bukit Timah serves as a forecast of what may happen to fig-eating frugivores as a result of habitat disturbance and fragmentation. It also reflects a trend of global declines for individual species and populations of fig-eating frugivores. Ex- cluding fig-eating reptiles and fishes, about 18% of all bird and mammalian species known to eat figs are either at risk or near threatened at a global level (Peh K.S.-H., unpublished data). Although habitat destruction and fragmentation have been identified by many studies as root causes of the current global biodiversity crisis and conserva- tion problems (e.g. Turner 1996; Debinski & Holt 2000), serious threats faced by fig-eating frugivores include disturbance of roost sites, poaching, introduc- tion of alien predators, harmful effects of environ- mental pollution and unpredictable natural distur- bances such as hurricanes. All cases of threats have the potential to result in the local loss of fig-eating species. Efforts to address the issues of fig-eating fru- givores declines in disturbed habitats are very much needed. Information on the interaction patterns among fig-seed dispersers and Ficus in rain forests with respect to community structure, degree of gener- alization or specialization of interactions, and ecosys- tems health may be vital for forest conservation and management of small reserves. Such knowledge may also aid in any future attemps in rain forest restora- tion. ACKNOWLEDGMENTS We express our gratitude to I.M. Turner, N.S. Sodhi and Leong Tzi Ming for their invaluable assistance and guidance throughout the study. We thank Johnny de Jong, Torbjorn Ebenhard and Nadejda Andreev for their insightful discussions and constructive criticisms on the earlier draft of this paper. We also thank the National Parks Board, Singapore, which provided the permit for this study to be made possible. REFERENCES Castelletta M, Sodhi NS & Subaraj R (2000) Heavy ex- tinctions of forest avifauna in Singapore: lessons for biodiversity conservation in Southeast Asia. Conserv Biol 14: 1870-1880. Corbet GB & Hill JE (1991) A world list of mammalian species. 3rd ed. Oxford University Press, London. Corlett RT (1990) Flora and reproductive phenology of the rain forest at Bukit Timah, Singapore. J Trop Ecol 6: 55-63. Corlett RT (1995) Tropical secondary forests. Prog Phys Geogr 19: 159-172. Corlett RT (1998) Frugivory and seed dispersal by ver- tebrates in the Oriental (Indomalayan) region. Biol revi 73: 413-448. Corner EJH (1965) Checklist of Ficus in Asia and Aus- tralasia with keys to identification. Gard Bull Singa- pore XXI: 1-186. Debinski DM & Holt RD (2000) A survey and overview of habitat fragmentation experiments. Conserv Biol 14: 342-355. Estrada A & Coates-Estrada R (1986) Frugivory by howling monkeys (Aluoatta palliata) at Los Tuxtlas, Mexico: dispersal and fate of seeds. In: Estrada A & Fleming TH (eds) Frugivores and seed dispersal. pp 93-104. Junk Publisher, Dordrecht. Hamilton MB (1999) Tropical tree gene flow and seed dispersal: deforestation affects the genetic structure of surviving forest fragments. Nature 401: 129-130. Inskipp T, Lindsey N & Duckworth W (1996) An anno- tated checklist of the birds of the Oriental region. Ori- ental Bird Club, Bedfordshire, U.K. Janzen DH (1979) How to be a fig. Annual review of ecology and systematics 10: 13-51. King B, Woodcock M & Dickonson EC (1989) A field guide to the birds of South-east Asia. Collins, Lon- don. Lambert F (1989) Pigeons as seed predators and dis- persers of figs in a Malaysian lowland forest. Ibis 31: 453-458. Lambert F (1991) The conservation of fig-eating birds in Malaysia. Biol Conserv 58: 31—40. Lambert RF & Marshall AG (1991) Keystone character- istics of bird-dispersed Ficus in a Malaysian lowland rain forest. J Ecol 79: 793-809. Leighton & Leighton (1983) Vertebrate responses to fruiting seasonality within a Bornean rain forest. In: Sutton SL, Whitmore TC & Chadwick AC (eds) Tropical rain forests: Ecology and management. pp 181-196. Blackwell, Oxford. Lekagul B & McNeely JA (1988) Mammals of Thai- land. 2nd ed. Association for the Conservation of Seed dispersal agents of two Ficus species Wildlife, Bangkok. Lim KS (1992) Vanishing birds of Singapore. Nature Society (Singapore), Singapore. Ng FSP (1978) Tree flora of Malaya. Vol 3. Longman, Kuala Lumpur. van der Pijl L. (1972) Principles of dispersal in higher plants. p 298. Springer-Verlag, Berlin. Restrepo C, Gomez N & Heredia S (1999) Anthro- pogenic edges, treefall gaps, and fruit-frugivore inter- actions in neotropical montane forest. Ecology 80: 668-685. Robson C (2000) A field guide to the birds of Southeast Asia. New Holland, U.K.. Shanahan M, So S, Compton SG & Corlett R (2000) Fig-eating by vertebrate frugivores: a global review. Biol Rev 76: 529-572. da Silva JMC, Uhl C & Murray G (1996) Plant succes- 125 sion, landscape management, and the ecology of fru- givorous birds in abandoned Amazonian pastures. Conserv Biol 10: 491-503. Terbogh J (1986) Community aspects of frugivory in tropical forests. In: Estrada A & Fleming TH (eds) Frugivores and seed dispersal. pp 371-384. Junk Publishers, Dordrecht. Tilman D (1997) Biodiversity and ecosystem function- ing. In: Daily GC (ed) Nature’s services. Societal de- pendence on natural ecosystems. pp 93-112. Island Press, Washington D.C. Trainer JM & Will TC (1984) Avian methods of feeding on Bursera simaruba (Burseaceae) fruits in Panama. Auk 101: 193-194. Turner IM (1996) Species loss in fragments of tropical rain forest: a review of the evidence. J Appl Ecol 33: 200-209. eonaqaniat qgraebrutaahe comme io eldCaporavig a Tum widqnaie ‘Prk thiG) isifiegepont:- Casidiiony M, Sethi ying? (NBR hAptenrss —~ 04 Fl; shed sbenell WL eR hana. ay. \eengub by vainty zeryvigyripe = C7 er Perce meer ae § “cutee OH & | sted tA igahoves » Dette wigpa elbow) {Ge by. Maeraih dink. Nndlaliberenailiaations wmakriat # EY, (>), ede Slips), al MP Hola Wy - ; 4\ an ALKLO kon, BO, te Vin icp Alreecanta pet Shy > 2-1 90 ra obapribon Naty At Bt): sth) CA oak “lwo i i ina bla AD é cvy fh coh ee gals 6 Joc dA witty conservation ir Senthen Ne MANET eslqrantin’ rab Re Sai eur) (s0ocnoe a) HMDA maine 6D Chee al) a 17? AU tet vay BAppen Welter bo inet CORO Hvar? Braga) acyerod va) Weer gen) Ar? et Vi 1 sae, (a0 FT" 1a ala Aneta: Ania gilt by Er 1 GA CU Satabivs ‘SHS tan i Ve me geet Fade he OE 2 PUTHIONE, ( e-oeteng ER palin — Esomme Tnol( 1060) Deans Ald uh p 1 figt aed hua io al releaw.with keys fo k ert fiat oo rit lin pecs tae ate an “» we XRT A-186, i ¥ : ned Uireal lal o.coreey aeye (Pech eae 1M Flolt RIFC2DO) A yam =— , oa of habital isapugauttion ceperm i i gestiruchon «@ _ ‘Aa » Las — i y IMMA sss sate - Penete A de Coblendisteada 19 7” wy ee he : ‘ag Wedeays tAlarans poiltanya | Vite, §oetie ~' : ished oul tee Of Seeds. Ine : ly Tg nm 7 , cg DM (ecg) Frunivores itmmd week ( i ya Pyebbiqhes, Dordrecht) a iwi. off “ : \enmttine: Ail 1 1904) Tiypacal. he —_ : ‘ a?! -- ey aereal. aofdre sae uticocts therg ~ of (eet fee «= eervieiug forest Irdgmente Nain ue Inoad Joes af eo cag inekion 7, Lines MN & Duckworth W¢ ni the a duc at Pp sre | ted ve blot of the binds of the. Bia Ts ictaptee! kubitee une vere pruch ~~ eral, Gilat Ch, Bedtos alhire, UKs 2 her’ the inieracth atte on DH (10799 How to be li Ade r anf) Sikes. 6 cath Sees Amey and vywicnaatigs 10 TRARL, rect deeercca! weons. | eB Whedcock Ms Dickogsnm re) rather . = ‘ rs 1 Wee Dirals uf Soules anh! | PCOS mbar F 17989) Pigeons as / : nesta : . she . porers of figs ‘ii 4 Malaysta F tawibeet F (1901) Theciiervanny oem * Malayna-Btol Conserv 3891-9079 % DSMENTS ~iwtest RPS Marshal £6 (19917 Kova ) lex of Didi persed fiedy lnk vi HASH ra 6% Sod iin forest. j fin 779: * TGA ; ae ne Ltigiton f Leighton: (1983) Vertebrate” 1 } my de any “Feviting sisnnulity within A Busceii rail py ot “Senton- SE. Whitmore TC Be Chae wid : ‘ +m 2 / Trapyes! rain forenty! Kodtogy. — ~ than the Bidtacaes LAT 106, Plackwelk Crxfark LO Vadeagut Bh & MoNedly. JA 0198 ow ‘oda, Jodeed, Amociaion be nn fan >. SHORT COMMUNICATION Ornithol. Sci. 2: 127-131 (2003) Ants found in scats and pellets taken from the nests of the Japanese Wryneck Jynx torquilla japonica ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2003 Masashi YOSHIMURA", Tomoko HIRATA’, Ayako NAKAJIMA?, and Keiichi ONOYAMA!” ' Course of Biotic Environment, the United Graduate School of Agricultural Sciences, Iwate University, c/o Laboratory of Wildlife Ecology, Obihiro University of Agriculture and Veteri- nary Medicine, Inada-cho, Obihiro, Hokkaido, 080-8555 Japan * Laboratory of Wildlife Ecology, Obihiro University of Agriculture and Veterinary Medicine, Inada-cho, Obihiro, Hokkaido, O80—8555 Japan The Japanese Wryneck Jynx torquilla japonica (Bonaparte 1850) is a subspecies of J. torquilla Lin- naeus, 1758 (Ornithological Society of Japan 2000). J. t. japonica has been recorded breeding only in the southern part of the Kurile Islands, Hokkaido, and the northern part of Honshu, Japan (Fujii 1990; Satoh et al. 1996; Ornithological Society of Japan 2000). Short (1982) and Winkler and Christie (2002) sug- gested that the subspecies J. ¢. japonica is synony- mous with J. t. torquilla. Ecological information is needed to clarify the taxonomic status of the Japanese population. Ants (Formicidae) are considered one of the major food items for J. t. japonica (Yamashina 1941; Kiyosu 1965), and there have been a few reports about ants as prey. Kosugi (1989) observed that chicks were fed larvae and adult ants on Rishiri Is- land, northern Hokkaido. Satoh et al. (1996) also re- ported that ant eggs and pupae were given as food, and Murai and Higuchi (1991) observed ant pupae being carried into a J. t. japonica nest. In his book, Kiyosu (1965) described ant adults and larvae as being a favored food item of J. t. japonica, and Fujii (1990) reported that ant pupae are the most common food for J. t. japonica chicks. Although the ant species recorded by Kiyosu (1965) were Lasius japonicus, and Kuro-ari (possibly Lasius nipponensis or Camponotus (Myrmamblys) itoi) and that recorded in Fujii (1990) was probably Formica sanguinea, the ant species in the reports by Kosugi (1989), Satoh et al. (1996), and Murai and Higuchi (1991) were unidentified. There has not been sufficient analysis of the ants eaten by Jynx torquilla (Received 9 May 2003; Accepted 4 September 2003) # E-mail: myoshimura@ant-database.org 127 Japonica. Detailed identification of the prey ant species will help to clarify the bird’s food habits. In this paper we analyzed ant remains recovered from scats and pellets found in post-breeding nests of Jynx torquilla japonica in Tokachi, Hokkaido, Japan. This is the first detailed report on prey species for J. t. Japonica. We also discuss the food habits of J. ¢. Japonica. MATERIAL AND METHOD Scats and pellets from four nests of Jynx torquilla Japonica were analyzed. Sampling was done immedi- ately after fledging, from June to August in 2002. Scats and pellets were collected using different meth- ods: chips with adhesives (nests I and II), by hand from the bottom of cavities (nest III), or by hand from the bottom of a bird box (nest IV). Of the four sampling nests, I and II were in open land, and III and IV were in humid forests. For nests I, II, and III, J. t. japonica used cavities made by woodpeckers. Nest IV was in a bird box. Measuring the diameters of the entrances suggested that nests I and II were made by Dendrocopos major, and nest II by D. minor, according to Yamauchi et al. (1997). Table 1 shows sampling dates, and describes the nests and their environments. The scats and pellets were broken up in ethanol, and identifiable parts of ants were col- lected with the aid of a stereoscopic microscope. The parts of ants recovered from scats and pellets were treated as samples of each nest. Samples were fixed with 70% ethanol and mounted as dry specimens be- fore identification (Fig. 2). The species in the samples were identified by comparing them with reference specimens collected in Hokkaido. The bodies of the reference specimens were dismembered (head, M. YOSHIMURA et al. fable 1. Sampling information of scats and pellets of the Japanese Wryneck Jynx torquilla japonica. Sampling dates, descrip- tion of the nest surveyed, and the environments surrounding the nest sites are shown. The environments list dominant tree species. Nest No. Sampling date Description of the nest Nest site environment I 2002.vii.3 1 A cavity in a tree Fraxinus mandshurica Grassland, adjoining the forest mainly consisting var. japonica of Salix spp. I] 2002. vili.6 A cavity in a tree Betula platyphylla Row of Betula platyphylla var. japonica var. japonica adjoining open land Ill 2002.vii.15 A cavity ina tree A/nus hirsuta Humid forest, mainly consisting of Alnus hirsuta and Populus maximowiczii IV 2002.vii.16 A nest box Humid forest, consisting of Fraxinus mandshurica var. japonica, Ulmus davidiana var. japonica, and Alnus japonica Table 2. The ant species collected from four post-breeding nests of Jynx torquilla japonica. The appearance of each species are shown as + in each cell of the nest. If only one individual was collected, the caste name is given as a singular form. Developmen- tal stages are given as adult (Ad) or cocoon (Co). Nesting habitats of each ant species are shown as four categories: in the soil of open lands (SO), in the soil of forests (SF), in woods of open lands (WO), and in woods of forests (WF). Breeding nest Scientific name Japanese name Caste Develepmieatat Nesting habitat ee i Formica japonica Kuro-yama-ari workers Ad sr + SO Formica candida Tsuya-kuro-yama-ari workers Ad 36 + SO Formica truncorum Kezune-aka-yama-ari workers Ad oF SO Lasius (Lasius) japonicus Tobiiro-ke-ari workers Ad + + + S0O,SE WO, WF Lasius (Lasius) japonicus Tobiiro-ke-ari queen Ad ete SO, SE WO, WF Lasius (Lasius) japonicus Tobiiro-ke-ari males and Co 5f SO, SFE, WO, WF queens Lasius (Lasius) sakagamii Kawara-ke-ari workers Ad + + SO Lasius (Cautolasius) flavus Kiiro-ke-ari workers Ad aioe wee SORSE Lasius (Dendrolasius) spathepus — Kusa-ari-modoki male Ad sf SE, WF Lasius sp. A Ke-ari-zoku workers Ad + unknown Lasius sp. B Ke-ari-zoku queen Ad AF unknown Lasius sp. C Ke-ari-zoku males and Co + unknown queens Myrmica kotokui Shiwa-kushike-ari workers Ad + WF Myrmica jessensis Ezo-kushike-ari workers Ad a ar ae SO Pheidole fervida Azuma-oozu-ari workers Ad a + SO scapes, pedicels and flagella, mandibles, pronotum, meso-metathorax and propodeum, petiole, postpetiole (if present), gaster, coxae, femora, tibiae, and tarsi) before the comparison (Fig. 1). The Japanese names of the ants follow the Japanese Ant Database Group (2003) and The Myrmecological Society of Japan Ed- itorial Committee (1988). Fig. 1. Reference specimen (Formica japonica) that has RESULTS been taken to pieces. Table 2 shows the names of the species, castes, and developmental stages of the ants from each nest, and 128 Ants found in droppings in the nests of Japanese Wrynecks their nest habitats in Tokachi. A total of 13 ant species from four genera were found in the four nests of J. t. japonica. Ten of the 13 species were identified to species rank: Formica japonica (Fig. 2-J), F. can- dida (Fig. 2-K, P), F- truncorum (Fig. 2-A), Lasius japonicus (Fig. 2-B, F, O, S), L. sakagamii (Fig. 2-E, N), L. spathepus (Fig. 2-M), L. flavus (Fig. 2-C, G), Myrmica kotokui (Fig. 2-l), M. jessensis (Fig. 2-L), and Pheidole fervida (Fig. 2-H). The remaining three were identified to genus rank: belonging to the genus Lasius. Many cocoons of males and queens of L. japonicus were obtained from nest III, those of La- sius sp. C from nest IV, a queen of L. japonicus was found in nest III, and a queen of L. sp. B in nest II. Lasius spathepus from nest II was identified by a male petiole. There were common genera in the fol- lowing nests: the genus Lasius in nests II, III, and IV, the genus Formica in nests II and IV, and the genus Myrmica in nests II, III, and IV. We observed nest habitats in Obihiro for the 10 ant species identified to species rank. Lasius japonicus, L. spathepus, and Myrmica kotokui nest in wood, and the other seven species nest in the soil. M. kotokui and L. spathepus generally nest in forests, L. japoni- cus and L. flavus in both sparse forests and open land, and the remaining six species generally nest in open land. DISCUSSION In previous studies, only three ant species were re- ported as food items for Jynx torquilla japonica (Kiyosu 1965; Fujii 1990). In the present study, 13 species have been recognized, and nine of the records are new. Kiyosu (1965) writes that J. t. japonica for- ages in rotten parts of trees and on the ground, and Yamashina (1941) and Nakamura and Nakamura (1995) also state that the birds commonly forage on the ground in sparse forests. In the present study, seven species of the prey ants nest in the soil on open land in Tokachi. We found not only adult workers but also many cocoons from the reproductive castes among the bird scats and pellets. A forest species, La- sius spathepus, was found, but as the only finding was a male, it is likely that its capture occurred out- side the ant’s nest. Our results suggest that J. t. japon- ica generally forages on the ground, and takes adult ants and their pupae directly from nests in the soil. In addition, because many individuals of Myrmica ko- tokui were found, it is suggested that J. t. japonica also forages where ants make their nests among rot- ten branches or wood on the ground. Because of our analysis of the food habits of J. t. japonica, it 1s prob- able that ant larvae are also one of the important food resources in the breeding season, although we did not find them in the scats and pellets. Reports about the food preferences of the Eurasian Wryneck, J. t. torquilla, come mainly from Europe and Russia. In Europe, most of the food items consist of ants, Lasius niger, L. alienus, L. emarginatus, L. flavus, Formica rufa, F. polyctena, F. fusca, F. cunic- ularia, F. rufibarbis, Myrmica lobicornis, M. sabuleti, M. scabrinodis, M. schencki, M. sulcinodis, M. rubra, Aphaenogaster subterranea, Leptothorax unifascia- tus, Tetramorium caespitum, and Tapinoma erraticum (Cramp 1985; Bitz & Rohe 1993), but other items are common in Russia: grasshoppers, aphids, beetles, craneflies, mayflies, and eggs of the Great Tit Parus major (Cramp 1985). Bitz and Rohe (1993) also re- ported from Germany that species belonging to the genus Lasius are the most common food items for J. t. torquilla through their analysis of 22013 food balls (Nahrungsballen). In our study, the food items from J. t. japonica’s scats and pellets consisted mostly of ants, and rarely included beetles or flies. Lasius flavus is contained as their food item for both J. t. japonica and J. t. torquilla. We also found that ant species be- longing to the genera Lasius, Formica, and Myrmica are common food items for J. t. japonica, which is also the case with the European population. Our study of food habits supports a hypothesis of syn- onymy between J. t. japonica and J. t. torquilla. ACKNOWLEDGMENTS We would like to thank Mr. Yuji Yamaguchi (Eco- system Inc., Obihiro) for providing samples from nest IV, Ms. Keiko Kobayashi (Laboratory of Environmental Entomology, Obihiro University of Agriculture and Vet- erinary Medicine) and Mr. Takumi Akasaka (Laboratory of Wildlife Ecology, Obihiro University of Agriculture and Veterinary Medicine) for providing reference ant collections, Ms. Anne McLellan Howard (Obihiro Uni- versity of Agriculture and Veterinary Medicine) for im- proving the English, and Ms. Takako Ikeda, Ms. Naoko Muraki (Laboratory of Wildlife Ecology, Obihiro Uni- versity of Agriculture and Veterinary Medicine) and Mr. Momoki Kawabe (Higashi Taisetsu Museum of Natural History) for their useful advice. This study was partially supported by a Grant-in-Aid for Scientific Research (C) (No. 15510190 to K. Onoyama) from the Japan Society for the Promotion of Science. 129 M. YOSHIMURA et al. Fig. 2. Identifiable body parts of ants found in scats and pellets from J t. japonica. F: from nest I; D, G, H, L, M: from nest II; A-C, E, I, N, O, S: from nest III; J, K, P-R: from nest IV. A: Formica truncorum; B, F, O, S: La- sius japonicus; C, G: Lasius flavus; D: Lasius sp. B; E, N: Lasius sakagamii; H: Pheidole fervida; 1: Myrmica ko- tokut; J: Formica japonica; K, P: Formica candida; L: Myrmica jessensis; M: Lasius spathepus; Q, R: Lasius sp. C. A-C, E-L, N, P: worker; D, O: queen; M: male; Q—S: cocoon. A—D: head; E-H, J, K: meso-metathorax and propodeum; I: Mesosoma and petiole; L: scape; M—P: petiole. 130 Ants found in droppings in the nests of Japanese Wrynecks REFERENCES Bitz A & Rohe W (1993) Nahrungsokologische Unter- suchungen am Wendehals (Jynx torquilla) in Rhein- land-Pfalz. Beih Veroff Naturschutz Landschaft- spflege Baden-Wiirttemb 67: 83-100 (in German). Cramp S (ed) (1985) Handbook of the Birds of Europe the Middle East and North Africa. The Birds of the Western Palearctic. Vol IV. Terns to Woodpeckers. Oxford University Press, Oxford. Fujii T (1990) Iwate-ken Iwaizumi-cho Hittori ni okeru Arisui no eisou kiroku (The breeding record of the Wryneck Jynx torquilla in Hittori [waizumi, Iwate). Strix 9: 63-70 (in Japanese). Japanese Ant Database Group (2003) Ant Image Database 2003, CD-ROM. Idengaku-Fukyukai (The Association for Propagation of the Knowledge of Genetics), Mishima. (Also available at http: //ant.edb.miyakyo-u.ac.jp/index.html) Kiyosu Y (1965) Nihon chourui dai-zukan I (The birds of Japan I). Kodansha, Tokyo (in Japanese). Kosugi K (1989) Arisui no dochu de no eisou kiroku (A nesting record of the Wryneck Jynx torquilla in clay cliff). Strix 8: 290-291 (in Japanese). Myrmecological Society of Japan Editorial Committee (1988) A list of the ants of Japan with common Japanese names. The Myrmecological Society of Japan, Tokyo. Murai M & Higuchi H (1991) Akagera to Arisui ni yoru 131 happo-suchiroru-sei_ giboku no riyou (Nesting of Great Spotted Woodpeckers and Wrynecks in artifi- cial cylinders). Strix 10: 285-290 (in Japanese). Nakamura T & Nakamura M (1995) Genshoku Nihon yachou seitai zukan, riku-dori-hen (Bird’s life in Japan with color pictures—Birds of mountain, wood- land, and field—). Hoikusha, Osaka (in Japanese). Ornithological Society of Japan (2000) Check-list of Japanese birds. 6th ed. Ornithological Society of Japan, Obihiro (in Japanese). Satoh K, Kikuchi N & Nishide T (1996) Akita-ken Hachiro-gata kantakuchi ni okeru Arisui no hanshoku kiroku (Wryneck Jynx torquilla bred in Hachirogata reclaimed land, Akita Prefecture). Strix 14: 135-141 (in Japanese). Short LL (1982) Woodpeckers of the World. Foris Publi- cations, Cinnaminson, New Jersey. Winkler H & Christie D (2002) Subfamily Jynginae. In: Hoyo JD, Elliott A & Sargatal J (eds) Handbook of the Birds of the World. Vol 7. Jacamars to Woodpeck- ers. pp 420-421. Lynx Editions, Barcelona. Yamashina Y (1941) Nihon no chourui to sono seitai, I (A natural history of Japanese birds II). Iwanami- shoten, Tokyo (in Japanese). Yamauchi K, Yamazaki S & Fujimaki Y (1997) Noukou/juutaku chiiki ni okeru Akagera to Koaka- gera no eisou jouken (Breeding Habitats of Dendro- copos major and D. minor in Urban and Rural Areas). Jpn J Ornithol 46: 121-131 (in Japanese). SHORT COMMUNICATION Ornithol. Sci. 2: 132-134 (2003) Interspecific learning by the Ogasawara Islands Honeyeater Apalopteron familiare from the Japanese White-eye Zosterops japonicus on Hahajima, the Bonin Islands, southern Japan ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2003 Kazuto KAWAKAMI'* and Hiroyoshi HIGUCHI’ ' Tama Forest Science Garden, Forestry and Forest Products Research Institute, Todori 1833, Hachioji, Tokyo 193-0843, Japan ? Laboratory of Biodiversity Science, School of Agriculture and Life Sciences, The University of Tokyo, Yayoi 1—I—1, Bunkyo-ku, Tokyo 113-8657, Japan The Ogasawara Islands Honeyeater Apalopteron familiare is endemic to the Bonin Islands, and classi- fied as vulnerable by BirdLife International (2001). The Japanese White-eye Zosterops japonicus was in- troduced there in the early 1900s as a cage bird (Momiyama 1930), and is now one of the most com- mon bird species on the archipelago (Suzuki 1991). The former prefers primary forests and the latter prefers open habitats (Kawakami & Higuchi in press). The two species are similar in size and forag- ing habits, and their distributions overlap greatly in secondary forests, which constitute a large area of Hahajima (Kawakami & Higuchi in press). In various studies, introduced species are consid- ered to increase the costs or decrease the benefits to native species (e.g. Long 1981, Eguchi & Amano 1999), and the Japanese White-eye is considered to have that kind of impact on the Ogasawara Islands Honeyeater (Morioka & Sakane 1978, Department of Labor and Economy of Tokyo Metropolitan Govern- ment 1985). But in this case, there is also a possibil- ity that the Japanese White-eye benefits the Oga- sawara Islands Honeyeater. The Japanese White-eye has a wide home range and high mobility (Kawakami & Higuchi unpub- lished data), and often feeds on a variety of insects and fruits. Many of the fruits are introduced species such as Carica papaya and Morus australis (Kawakami & Higuchi 2003). Conversely, the Oga- sawara Islands Honeyeater has a smaller home range and its diet consists mainly of native foods such as small arthropods (Kawakami & Higuchi 2003). How- (Received 6 June 2003; Accepted 9 August 2003) E-mail: kazzto@ffpri.affre.go.jp ever, this species also forages on introduced fruits, and it is possible that Ogasawara Islands Honeyeaters have learned to eat such unfamiliar foods from Japan- ese White-eyes. Laboratory experiments with the House Sparrow Passer domesticus (Fryday & Greig 1994), the Red Junglefowl Gallus gallus (McQuoid & Galef 1994), and the Black-capped Chickadee Parus artricapillus (Sherry and Galef 1990), showed that demonstrators that ate unfamiliar food influenced the food prefer- ences of observers. The same phenomenon might be occurring with the two bird species in the Bonin Is- lands. The purpose of the present paper is to show whether interspecific learning exists between the two species. In order to find out if the birds were learning from each other, we experimented in the field to see how Ogasawara Islands Honeyeaters responded to unfa- miliar food in two different situations: with and with- out the presence of Japanese White-eyes. Japanese White-eyes are not evenly distributed across the is- land, and there are almost none in some areas inhab- ited by Ogasawara Islands Honeyeaters. Therefore, we could compare the behavior of Ogasawara Islands Honeyeaters in allopatric and sympatric areas. The population densities of Ogasawara Islands Hon- eyeaters in the allopatric and sympatric areas were about seven and four individuals per hectare, respec- tively. The population density of Japanese White- eyes in the sympatric area was 30-40 individuals per hectare. In forests on Hahajima Island 25 feeding stands were set up. Eleven were set up in the allopatric area of Ogasawara Islands Honeyeaters, and the remaining 14 in the sympatric area shared by the two species. Interspecific learning by the Ogasawara Islands Honeyeater Apalopteron familiare from the Japanese On the stands we used peaches in syrup as unfamiliar food, since we believed that the Ogasawara Islands Honeyeater had never eaten these before. The peaches were cut in half and about 8 cm in diameter. The feeding stands, made of plastic, were fixed to naturally growing trees with steel wire, about 1m above ground level, close to thin branches suitable for birds to perch on (Fig. 1). We selected trees that had at least one side clear of heavy foliage so that the birds could find the feeding stands. The survey was carried out in November and December 1996. During the experiment, the food was replaced daily. And each day the food was checked for evi- dence of pecking. When pecking marks were found, we observed the behavior of the birds that came to the stands. The responses of the two species were recorded from 8:00 am to 12:00 noon on the first and 15th days of the experiment. Responses were classi- fied as one of three types: “no approach’, “‘approach only”, and “approach and ingest”. “Approach” was defined as a situation where an individual came within | m of the feeding stand. On the first day of the experiment, the response of Ogasawara Islands Honeyeaters was not significantly different in the allopatric and sympatric areas (Fig. 2, Fisher’s exact probability test, P=0.6232). No Oga- sawara Islands Honeyeaters ate the food in either area, but individuals did approach it. In the sympatric area, Japanese White-eyes ate the food from the first day on all the stands except one, where food was eaten from the second day onwards. Japanese White- eyes approached the stands every day and always ate the food. We often observed Ogasawara Islands Hon- eyeaters watching the Japanese White-eyes foraging. On the 15th day, Ogasawara Islands Honeyeaters ate a significant amount of the previously unfamiliar food in the sympatric area (Fig. 2, Fisher’s exact probability test, P=0.0001). The test was between the sympatric and allopatric areas. In the allopatric area, the honeyeaters often approached the feeding stands but never ate the bait. These results suggest that Ogasawara Islands Hon- eyeaters learned what food to eat from the behavior of Japanese White-eyes. It is known that the two species sometimes form mixed-species flocks during the non-breeding season (Ueda 1990), suggesting that at least one species should enjoy a benefit from the behavior. Some bird species are known to increase food acquisition by participating in mixed-species flocks (Vijayan 1989, Sasvari 1992, Valburg 1992). We have observed Japanese White-eyes foraging on 133 Fig. 1. A feeding stand on a tree trunk. a) plastic feeding stand, b) peaches in syrup, c) steel wire, d) naturally growing tree. Approach and ingest Approach only No Approach 100 ©O (eo) 3 > 60 (= o Sy 8 40 — LL 20 1st day 15th day 1st day 15th day Sympatric area Allopatric area Fig. 2. Responses of Ogasawara Islands Honeyeaters to un- familiar food in allopatric and sympatric areas on the first and 15th day of the experiment. The y-axis shows the frequency of feeding stands classified to each response type of all feeding stands in each of the two areas. The sample sizes for the sym- patric and allopatric areas are 14 and 11, respectively. introduced fruit, and Ogasawara Islands Honeyeaters watching their behavior in mixed-species flocks. There is a possibility that Ogasawara Islands Hon- eyeaters have started eating unfamiliar foods by learning from the behavior of Japanese White-eyes in such circumstances. K. KAWAKAMI and H. HIGUCHI ACKNOWLEDGMENTS We thank Hayato Chiba, Takaya Yasui, Hajime Takano, Kiyoko Yokota, and Wakako Taniguchi for their helpful comments on the drafts, and Takashi Ishii, Osamu Tsunashima, Ryuji Tokiwa and Seiji Tazawa for accommodation during our field study. REFERENCES BirdLife International (2001) Threatened birds of Asia: the BirdLife International Red Data Book Part B. BirdLife International, Cambridge. Department of Labor and Economy of Tokyo Metropoli- tan Government (1985) A report on the ecology and conservation of Apalopteron familiare (in Japanese). Eguchi K & Amano H (1999) Naturalization of exotic birds in Japan. Jpn J Ornithol 47: 97-114 (in Japanese with English summary). Fryday SL & Greig SPW (1994) The effects of social learning on the food choice of the house sparrow (Passer domesticus). Behaviour 128: 281—300. Kawakami K & Higuchi H (2003) Interspecific interac- tions between the native and introduced white-eyes in the Bonin Islands. Ibis 145: 583-592. Long JL (1981) Introduced birds of the world. Reed, Wellington. McQuoid LM & Galef Jr BG (1994) Effects of access to food during training on social learning by Burmese red junglefowl. Anim Behav 48: 737-739. Momiyama T (1930) On the birds of Bonin and Iwo-Is- 134 lands. Bull Biogeogr Soc Jpn 1: 89-186 (in Japan- ese). Morioka H & Sakane T (1978) Observations of the ecology and behavior of Apalopteron familiare (Aves, Meliphagidae). Mem Natl Sci Mus, Tokyo 11: 169- 188. Sasvari L (1992) Great tits benefit from feeding in mixed-species flocks: a field experiment. Anim Behav 43: 289-296. Sherry DH & Galef Jr BG (1990) Social learning with- out imitation: more about milk bottle opening by birds. Anim Behav 40: 987-989. Suzuki T (1991) Status of the landbirds on the satellite islands of Hahajima, the Ogasawara Islands, with special reference to Common Buzzards, Oriental Greenfinches and Bonin Islands Honeyeaters. In: Ono M, Kimura M, Miyashita K & Nogami M (eds.) Re- ports of the second general survey of natural environ- ment of the Ogasawara (Bonin) Islands. pp 148-157. Tokyo Metropolitan Univ, Tokyo (in Japanese). Ueda K (1990) Tori wa naze atsumaru? (Why do birds make flocks?) Tokyo Kagaku Dojin Co Ltd, Tokyo (in Japanese). Valburg LK (1992) Flocking and frugivory: the effect of social groupings on resource use in the common bush-tanager. Condor 94: 358-363. Vijayan L (1989) Feeding behavior of the Malabar wood shrike Tephrodornis virgatus sylvicola Jerdon at Thekkady, Kerala. J Bombay Nat Hist Soc 86: 396-399. TECHNICAL NOTE Ornithol. Sci. 2: 135-137 (2003) Molecular sexing of individual Ryukyu Robins Erithacus komadori using buccal cells as a non-invasive source of DNA Wit Anata Bien ORNITHOLOGICAL = 9 1chi SEKI SCIENCE © The Ornithological Society of Japan 2003 Kumamoto 860-0862, Japan Buccal cells have been used as a good and non-in- vasive source of DNA for humans (Feigelson et al. 2001), experimental animals (Zimmerman et al. 2000), livestock (Kwon et al. 1993), and wild pri- mates (Hashimoto et al. 1996), but have not been used for wild birds. Blood is the most commonly used source of DNA for avian molecular ecological studies because safe and effective sampling tech- niques have been well-developed (reviewed in Wing- field 1999), hence other non-invasive methods have not been deemed necessary. In the case of the Ryukyu Robin Erithacus ko- madori, endemic only to southwestern Japan (Kawaji & Higuchi 1989), however, blood sampling is diffi- cult. Because of its limited distribution and the steep population declines on some islands, it has been des- ignated as a “national natural treasure” and “national endangered species of wild fauna and flora” of Japan. Consequently, both the Kagoshima prefectural gov- ernment and the national agency for cultural affairs are against taking blood samples from robins, espe- cially from nestlings, for political and cultural rea- sons. However, in order to reveal sex-related mortal- ity and dispersal distances of sexually monomorphic fledglings, molecular sexing is needed (Seki 2002). Here I report the results of molecular sexing of adult robins using DNA extracted from buccal cells, com- paring them with those from blood, and examine the reliability of buccal cells as a source of DNA for avian sexing. Samples of both buccal cells and blood were ob- tained from 20 male and 20 female adult Ryukyu Robins, that had been previously sexed based on morphological characteristics. Robins were captured (under license by mist-net) during the breeding sea- son, from 15 April to 30 June 2002, on Nakanoshima (Received 2 April 2003; Accepted 14 July 2003) * E-mail: seki@ffpri.affre.go.jp 135 Kyushu Research Center, Forestry and Forest Products Research Institute, 4—11—16 Kurokami, Island, Kagoshima, Japan. Buccal cells were collected by rolling cotton swabs against the inside of each robins’ mouth and throat five times. Each swab was put into a microtube and soaked in | ml of preservation buffer (150 mM NaCl, 10mM Tris-HCl pH 8.0, 10 mM EDTA), and kept at room temperature during the one week it took to transport them to the laboratory, where they were stored at 5°C for up to one month (see also Appendix 1 for longer preservation). After adding 10 ul of 10%-SDS, the samples were digested with proteinase K (5 ul of 20mg/ml) at 55°C for one hour and an ad- ditional 16 hours at 37°C. DNA was then extracted by the conventional phenol/chloroform method. Up to 30 ul of blood samples were obtained by brachial vein puncture and no ill effects were recorded. Blood samples were stored in GenTLE solution I (Takara) at room temperature and DNA was extracted using an extraction kit (QlAamp DNA Blood Mini Kit, QTA- GEN) within 10 days of sampling. The sex of each individual was determined by length polymorphism at the chromo-helicase-DNA- binding (CHD) gene, using Ellegren & Fridolfsson’s (1997) primer set 3007F (5'-TACATACAGGCTC- TACTCCT-3’) and 3112R (5'-CCCCTTCAGGT- TCTTTAAAA-3’), following the methods used by Fridolfsson & Ellegren (1999). All PCR reactions were performed in 25 ul volumes on a Perkin Elmer 9600 Thermal Cycler, using Tag DNA polymerase (Takara Tag, Takara). Each reaction mixture con- tained 0.625 unit of Tag, 200 uM dNTPs, 20mM Tris-HCl pH 8.0, 50mM KCl, 2mM MgCl, 5 pmol of primers. The thermal profile comprised an initial denaturing step of 94°C for two minutes, followed by a “touch-down” scheme where the annealing temper- ature was lowered by 1°C per cycle, starting from 60°C until a temperature of 50°C was reached. Then 35 additional cycles were run at a constant annealing temperature of 50°C. Denaturation was at 94°C for S.-I. SEKI 40s, annealing for 30s, and extension at 72°C for 40 s, and a final extension step of five minutes was added after the last cycle. PCR products were sepa- rated in 2.5% agarose gels (NuSieve GTG, BMA), run in standard TBE buffer and made visible using ethidium bromide staining. PCR products from blood DNA gave clear bands for all 40 individuals, which could be sexed on the basis of distinct size differences between the CHD-W and CHD-Z gene copies, a single band for males and double bands for females (Fig. 1). Band patterns from buccal cells were the same as those from blood sam- ples, and all the individuals were correctly sexed, using the DNA extracted by the conventional method. Thus buccal cells have been confirmed as a reliable source of DNA from wild birds. Buccal cell sampling is particularly suitable for the molecular sexing of nestlings as sampling takes less time than drawing blood, and: because there is no need for dis- infection and homeostasis. Safe sampling using this technique does not require extensive training, and only cotton swabs are required. It should be noted, however, that young, blind nestlings misidentify the cotton swabs as food being provided by the parents and so try to swallow them. Blood sampling from young nestlings has been preferred in studies of avian sex allocation (e.g. Komdeur et al. 1997; Nishiumi 1998), presumably so as to have large enough sample sizes for analysis under conditions of high nest predation risk in the wild. In such situations, the buccal cell sampling would obviously reduce the stress for the nestlings male female M BL BS BL BS 400bp & 300bp & Fig. 1. Band patterns of male and female Ryukyu Robins, using the template DNA from two types of samples, blood (BL) and buccal swabs (BS). The lanes marked ‘M’ contain 100 bp DNA ladder. Primers 3007F and 3112R give one short fragment in males (CHD-Z) and two fragments in females (CHD-Z and CHD-W),. 136 and their parents compared to blood drawing. The only negative aspect of buccal cell sampling is that DNA yields (1.3+0.7mg, Mean+SD, rough esti- mates using a spectrophotometer) are low compared to blood sampling, nevertheless it is sufficient for sexing, and also for other molecular studies using PER? Urine and feathers are other less- or non-invasive sources of DNA used in avian molecular ecological studies (Wingfield 1999). Urine samples, however, contain PCR inhibiting substances and food-origin compounds (Reed et al. 1997; Yamauchi et al. 2000; Nota & Takenaka 1999; Robertson et al. 1999), and for the Ryukyu Robin sexing success using urine samples was just 15% (n=20) (Seki 2003), thus buc- cal cells were preferred. Feather samples, especially plucked feathers, are also a reliable source of DNA (Segelbacher 2002), but care must be taken when re- moving growing feathers from nestlings (Wingfield 1999) and feathers are, of course, not available from naked nestlings. Non-invasive sampling is becoming important for molecular studies of free-ranging animals (Segel- bacher 2002), thus buccal cell sampling provides a preferable, easily performed, safe technique suitable for most studies of adults that requires only small amounts of DNA. ACKNOWLEDGMENTS I thank Noritomo Kawaji, Akihiro Yamane, and Kiyoshi Yamauchi for their helpful advice on molecular sexing, and Tsuneaki Yabe and Akira Endo for their comments. I also thank Toshima Village Government, Takeshi Ogura, Kazuyoshi Nagata, Yoshito Kobayashi, Mutsuhiro Koizumi, Koji Kaji, Mizuho Orita, Hiroyuki Tanaka, Naho Mitani, Yuri Inoue, and Masamichi Same- jima for their support during the course of the fieldwork. Two anonymous reviewers provided valuable comments on the manuscripts. The subject species was captured and sampled with the permission of the Ministry of En- vironment and the Agency for Cultural Affairs of Japan. This study was partly supported by Grants-in Aid (No. 14760108) from the Ministry of Education, Culture, Sports, Science, and Technology. REFERENCES Ellegren H & Fridolfsson A-K (1997) Male-driven evo- lution of DNA sequences in birds. Nat Genet 17: 182-184. Sexing Ryukyu Robins from buccal cells Feigelson HS, Rodriguez C, Robertson AS, Jacobs EJ, Calle EE, Reid YA & Thun MJ (2001) Determinants of DNA yield and quality from buccal cell samples collected with mouthwash. Cancer Epidemiol Bio- mark & Prev 10: 1005-1008. Fridolfsson A-K & Ellegren H (1999) A simple and uni- versal method for molecular sexing of non-ratite birds. J Avian Biol 30: 116-121. Hashimoto C, Fruichi T & Takenaka O (1996) Matrilin- eal kin relationships and social behavior of wild bonobos (Pan paniscus): Sequencing the D-loop re- gion of mitochondrial DNA. Primates 37: 305-318. Kawaji N & Higuchi H (1989) Distribution and status of the Ryukyu Robin Erithacus komadori. J Yamashina Inst Ornithol 21: 224—233. Komdeur J, Daan S, Tinbergen J & Mateman C (1996) Extreme adaptive modification in sex ratio of Sey- chelles warbler’s eggs. Nature 385: 522-525. Kwon HM, Jackwood MW, Brown TP & Hilt DA (1993) Polymerase chain reaction and biotin-labeled DNA probe for detection of infectious bronchitis virus in chickens. Avian Dis 37: 149-156. Nota Y & Takenaka O (1999) DNA extraction from urine and sex identification of birds. Mol Ecol 8: 1235-1238. Nishiumi I (1998) Brood sex ratio is dependent on fe- male mating status in polygynous great reed warblers. Behav Ecol & Sociobiol 44: 9-14. Reed JZ, Tollit DJ, Thompson PM & Amos W (1997) Molecular scatology: the use of molecular genetic analysis to assign species, sex and individual identity to seal faeces. Mol Ecol 6: 225-234. Robertson BC, Minot EO & Lambert DM (1999) Mole- cular sexing of individual kakapo, Strigops habrop- tilus Aves, from faeces. Mol Ecol 8: 1347-1350. Segelbacher G (2002) Noninvasive genetic analysis in 137 birds: testing reliability of feather samples. Mol Ecol Notes 2: 367-369. Seki S-I (2002) Estimating the survival rate and disper- sal distance of Ryukyu Robin, Erithacus komadori (Preliminary report). Kyushu J For Res 55: 171-172 (in Japanese). Seki S-I (2003) Molecular sexing of individual Ryukyu Robin, Erithacus komadori. Kyushu J For Res 56: 228-229 (in Japanese). Wingfield JC (1999) Minor manipulative procedure. In: Gaunt AS, Oring LW (eds) Guidelines for the Use of Wild Birds in Research, 2nd ed. pp 29-34. The Or- nithological Council, Washington, D.C. Yamauchi K, Hamasaki S-I, Miyazaki K, Kikusui T, Takeuchi Y & Mori Y (2000) Sex determination based on fecal DNA analysis of the amelogenin gene in Sika deer (Cervus nippon). J Vertebr Med Sci 62: 669-671. Zimmerman K, Schwarz HP & Turecek, PL (2000) De- oxyribonucleic acid preparation in polymerase chain reaction genotyping of transgenic mice. Comp Med 50: 314-316. APPENDIX 1 Buccal cells can also be kept in 99% ethanol when long-term preservation at room temperature is neces- sary, by rotating the sampling swab in the microtube filled with 1 ml of ethanol for about 10 seconds. Be- fore the extraction, I centrifuged the microtube at 6,000g (8000rpm) for 10 minutes, discarded the ethanol without disturbing the pellet of cells, and al- lowed the pellet to dry. Although the DNA yield was lower by this method than when keeping cotton swabs in the buffer, it was still enough for sexing. aliso trvourd ofer 2hiabR wy dey y in ooh y acs eat ae op of B60-Tetiflemieb’s (Pegthtitiqnamich (S000 pd ee acai p ugeH Una sy doChGyiiivMine) Aide whinge wiatiintleg te (ozsnaqel ai) i 4a a soe tehemalohi4EOOL) lrg dae L a hy ehtendibn wry baka iicrenocs Retest APES Set, i, av lon ine Jock (PO? b) i, UiadgelW i bad A to}. Wd an eb Aa i oa WO ba, NN RO 5) to) LC norqidag Ve lee muro’) | jcngohorny, A ase “uM 1 f a a x” vee reyial a ee { row & Fa try Mee i aE aan f aed tian Bait” oak fins T Kos y "pak “+o y wi weer 4 Mesa oT} ae wt Witirioatnit\ Loin enrtetoty Pah ee th ay Bion’ Sided Ftc) rr her) a aire: li oat artlay nid ra Hit sos Caplan Rive Tv OTe Or ‘hed } \ ii , sine . a CL ZIGAATSA \ L ~ , } , ‘ pe | ’ Pee ROO nT HEIN od nels age alien lace tf Onin rrr? bed helenrhiaan | 4 if Vi! Pe iw adit rind SK) sethaien Yo Vine LOT Tet 1A Wile to DP IMS : rt a Lyra hate of De Aes weirs heciclp id See OF FE GOORT HOGS dy PoP ebey! Sats eichinss w alae Rit rte \y AVG at Waghothte Wal of Safe ate ews ; 4 norlw Fort} Dorion ahi yd wwol 7 (Qiu "ee. a Si tisdew fl BS or a Hed iss ie es HA Pridolfssen AK p}997) s a a | dose pstimncneneme eile Cennerena 62000) imu i OblAmge tite eden riety yi Diving vmone yews: earravlieny Roshemniicceed in re AUER OO OL «che hei KORE Nenad i As a a to yikes winosian a), bed Linc and foalher bf) the OE dang ‘ 2 AAPA Q cloned & Tid Bh bin Jowohuntlsk dee, beg: aetdenpiand H re a ais) a eS ‘ se Wan Sheen AC atmos neigh A HO it na AY rat % ae 40) Mud pa he hig nf z aa a fo “te sles Am P} ite gy j 4 ad ae Bilal poe ie ni ma? y Hep Na SP AS eae Py) i cH + be Ait fet toMy Haul paris pais aa : lath avi Had sha ‘anita wa vee 0m), (TE aR nuntvAC Sam rile} 4 smeniit SANE PERE? PG Rit: fan ere oe ‘nie sofeenasmioan rei afl), dtu Yeece coll eri 28 Punting peck bolas maidens tht msl adertigy Wap ind ¢ vars of DNA: Pt~0 > dnidoioee (TOG]) W aomA & Mq noagearad Tr US, iff Sissy i lug yi he Tay. Pc KARINE yilinshs laut fore at bas. rer RSII9ge ae 3 3 MELEE: etree ¢ phon ( €2 12, AAG eoneinascl ld PR apn PT ren agorrad woquatyks vuldy dik dol aga Ot 1b WA Ble Siiodl, ao) mat mh, pK ee PROF nuseny amyl Sh niry ‘ 4 RAINE PS: { t} i "eal Kojt'Kas. 2 Misi ri ean , Veieri’ - yin hi < ad Mut 7 we this sip ier litt ‘ thet CORT af fie ? ie ie encwiymouk eviews = mele “\ (Spe Ps pied With ie Perinrsa in | af heh 1s “ey fin Colt wad Alt re ry inscrimis, Tie subicc it’ fief th Ag ¢ ipa . dy Was cord) erpported by Orang Ab ‘ ° aa : re > ioe PACHA) frtrm ite Minetes of Edie rn ~~ ¢ 28 ll. byewtt, Screho:, anlTechndlogy, — a RES ERE NC iS iuiikne of DNA sequences. if bigds,= ae — - 1n2-154 ; 7 eae — a ISSN 1347-0558 ORNITHOLOGICAL SCIENCE Vol.2 2003 =A The Ornithological Society of Japan ORNITHOLOGICAL SCIENCE Official journal of the Ornithological Society of Japan Editor-in-Chief Keisuke Ueda, Rikkyo University, Tokyo Associate Editors Teruaki Hino, Forestry and Forest Products Research Institute, Kyoto Hidetsugu Sakai, Nihon University, Tokyo Editorial Board Masahiko Nakamura, Joetsu University of Education, Joetsu Isao Nishiumi, National Science Museum, Tokyo Kazuo Okanoya, Chiba University, Chiba Eiichiro Urano, Yamashina Institute for Ornithology, Abiko Advisory Board Alexander V. Andreev, Institute of Biological Problems, Magadan Walter J. Bock, Columbia University, New York Jiro Kikkawa, The University of Queensland, Brisbane Woo-Shin Lee, Seoul National University, Suwon Bernd Leisler, Max-Planck-Gesellschaft, Radolfzell Anders P. M@ller, Universite Pierre et Marie Curie, Paris Richard Noske, Northern Territory University, Casuarina Pilai Poonswad, Mahidol University, Bangkok Lucia Liu Severinghaus, Academia Sinica, Taipei Navjot S. Sodhi, National University of Singapore, Singapore Jeffrey R. Walters, Virginia Polytechnic Institute and State University, Blacksburg John C. Wingfield, University of Washington, Seattle Jeong-Chil Yoo, Kyung-Hee University, Seoul CONTENTS OF VOLUME 2 Number 1 SPECIAL FEATURE Ecology of seed dispersal Ueda K Introduction. i! Kominami Y, Sato T, Takeshita K, Manabe T, Endo A & Noma N Classification of bird-dispersed plants by fruiting phenology, fruit size, and growth form in a primary lucidophyllous forest: an analysis, with implications for the conservation of fruit-bird interactions. 3 Takanose Y & Kamitani T Fruiting of fleshy-fruited plants and abundance of frugivorous birds: Phenological correspondence in a temperate forest in central Japan. 25) February 2003 Hayashida M Seed dispersal of Japanese stone pine by the Eurasian Nutcracker. Bhs) Fukui A Relationship between seed retention time in bird’s gut and fruit characteristics. 41 Rakotomanana H, Hino T, Kanzaki M & Morioka H The role of the Velvet Asity Philepitta castanea in regeneration of understory shrubs in Madagascan rainforest. 49 Printer: Kokusai Bunken Insatsusha Co., Ltd., Takada-no-baba 3-8-8, Shinjuku-ku, Tokyo 169-0075, Japan. Tel: +81-3-3362- 9741, Fax: +81-3-3368-2822. Cover design: Eiichiro Urano ORIGINAL ARTICLES Fujita M & Kawakami K Head-bobbing patterns, while walking, of Black-winged Stilts Himantopus himantopus and various herons. 59 Tajima K & Nakamura M Response to manipulation of partner contribution: A handicapping experiment in the Barn Swallow. 65 SHORT COMMUNICATION Nagata H & Sodhi NS Low prevalence of blood parasites in five Sylviidae species in Japan. is} A List of Referees 2002 US) Abstracts of the Japanese Journal of Ornithology, Volume 51 iS Number 2 September 2003 INVITED ARTICLES Komdeur J Adaptations and maladaptations to island living in the Seychelles Warbler. 79 Pierre JP Translocations in avian conservation: reintroduction biology of the South Island Saddleback (Philesturnus carunculatus carunculatus). 89 ORIGINAL ARTICLES Mizuta T, Yamada H, Lin R-S, Yodogawa Y & Okanoya K Sexing White-rumped Munias in Taiwan, using morphology, DNA and distance calls. 97 Matsubara H Comparative study of territoriality and habitat use in syntopic Jungle Crow (Corvus macro- rhynchos) and Carrion Crow (C. corone). 103 Hamao S Reduction of cost of polygyny by nest predation in the Black-browed Reed Warbler. 1T3 Peh KS-H & Chong FL Seed dispersal agents of two Ficus species in a disturbed tropical forest. JS SHORT COMMUNICATIONS Yoshimura M, Hirata T, Nakajima A & Onoyama K Ants found in scats and pellets taken from the nests of the Japanese Wryneck Jynx torquilla Japonica. 127, Kawakami K & Higuchi H Interspecific learning by the Ogasawara Islands Honeyeater Apalopteron familiare from the Japanese White-eye Zosterops japonicus on Hahajima, the Bonin Islands, southern Japan. JU. TECHNICAL NOTE Seki S-I Molecular sexing of individual Ryukyu Robins Erithacus komadori using buccal cells as a non-invasive source of DNA. LSS) ORNITHOLOGICAL SCIENCE HS AIY He Bentholomteal Society oF Japan rrr nn a. 2v iboe Dttataga | A imedtevea 3 Mal do sonslavorg wot nidviocirs Matilde slinky enratier, uit Q -bsgk neupak esiooge aabiivivd us ~ arene tessa eaqotraikl alia bey ie é. = ; ubiren’ noord uaOneY Welter J Need. Chena Diivereiiy- Now oe " au Research |i devi Keowee Tie |) rer i ie aeroeta ap COS easreafl Yo hd A Wiese it ae~Apal SN See get kg )\ a rt th a Myatt , i a 1 aie aap he aft re ta lea rvol 2eotRgeh e430 Hoerieds - ~~ thio Ter dtiley ee she 12 tibia doloryiaw Pigiy Pe a¢ Malido) Univeraty, Tahibot” wo. Tar tse Se Bt Be ps8 das, Acadeontd Sirs,’ ; vyvpa) DS Seethi, Nay vel Ligierrny af Singage ~~. ” Al ridite fee i) “ji Aealt { vit) . i* 7 o, tier Vin vith si yiee fiz teed oa ee rity ky iw‘ ksbuty Wragg teh, Univeriity of | Washodeton ne BOOS wedmatqee, aS scm. ip Hee Lm cruly, Souy 7 Jd gnod’) BH-2A 04 age mol Gwi ty siege lozvoqeit hea oes ; toe legecpertl badunzls £ tuetei of enoitéiqubalart bre ano CONTENTS OR VOLUME Dew calls ee aig 2VOrTAAUMIMOD TAHOHe Bi » & emilee’ Jf aed .\Meieiiay February 2003 Pa : SovmvIsedo naive ai A amayonO a bnalel dived peti Yo ygolad 6 St qiut apkhatliog bas eteoe ot Dadt cis H aveshh eee ”s oe 4 bly Ww aumuitestlW) ao sy ead fi ut roy oeoundal sill to atenn 7) i dtewrol wT apanesé ons “ni At WyiPoany ‘ . ‘Te ha, yee : 4° H idsuyiH 2 A imadeved 23.I2ITA oe iM abe sie wit Call atarioavahitp pas setanert Nietmng ie iti. Tyo) pega survsiqu lak twins rorret : rts out and fruit ober eyumat 2 t r - - ‘a Vat OG, PALER CA ml’ 13 gortatyul apis H Soe TT ee cdat Rgronagitalh 9 tik Rati ouce, .eapuntel ie wl atin JACS. , se | socrnreth b brig LAG 390 Ca one Wore ‘i, (Oree De = es 7 } ule of the Velvet Asityal hide | Ary }\ ines 9O 08 , 7 T! _ 7 pay 4 ~ STOKLA AOA ESRD Y all eee aid 9178 (4) SMEASRUHRERS AHES 9AH18A (+) ASRHR AY — FER Bass 9A19H (8H) WAP-HR MB YYRVVSA ABS 9208 (AtiAa) BSR BAS 4. SMA LIAA KAICBMEHNSAAIL, BCUIUFZOKSH LUSAKEDO RMS DPaAL-, bU KEOA) GRKS - HRCI) ie Db AN Gy ARS ORBLE OD AIt LEOBMEANRERBICE LOL AHESHHET SS EE BDAPA VOB CHER LT BARS SILT RA CORM ERO ED, THA LBACKAS EBA ECHEYD CSV, MHL AbBaL, AM MEIC LOD pd§7F7 714 VICELOTHEH D< HEV, KBVZFEDOFUFIAICCAMVE LED, BLEACH ANSAD, Mile FMS E—- LER SB ESE ENTICMAMOET, oR Se Lp BN ESN eBEORAEL) 7-H: Pe, BAKE ARF: 9A19AR (A) FSSA Bam : RRR K Same BA (9 bABA3A) OFFIC LSRBBER LC AAYT—FICLAMESESET, HA Hie MPMRABRLKAD =7 EICBRLET, 10. @#A 1) KSSnm# 7A 16 AfAEC —AR3500 “PAE 2500 7H 17 ALB —fix 4000 FJ “PE 3500 A HAF -ERO BRASH CH FBAICILBM CE EL) 2) BALE 7A 16 AfeAKC —e 4000 A “FE 3000 FI 7517 BLAB —fix 4500 FY =7E 3500 3) PAB SEOR —ik- SLE HI 30008 TABICIL, Re CIAAD REAR CAA < FES, — Bi Sete BTR OV DA ICD oPReVE LEA. LDL, BUBEMALERAKSICBMCARKS RoKAICI, mins eee DiAO LET, RK, RUUASOREBAMRISFRSAMAICIT CAA CE ERA, P(E RIATE : OBB 00960 - 0 - 112026 OHEAR AABAS 2004 HXSERESS 11. ROT Y Sal LIA” (BMH Li”, RES BHRSER ORM, BRAGA) DRO Y IX, 7A 168 (@) (GHEIA%) CT, CUA, SMB BRSRITAVIBLERVET, EK, ABB Lae i OMG C, FAT LIAD REDO) OURO LIABE BEIT CE RU BAM H Y pom Dba CMo TKK SV, 12. @2EAK ABFA PCRS OS PE SI, FROM RRA SNE, MAICSMSNRVAIL, K SH LIAAAMOKEDOMABERICHARAIDOL, KABBRE CHIEN KES. 13. KERB MAICBUYC, PRBICEDT ORB URS ORM Se ABaNADAI ROX< B<. BX EBRS O1lrAR (8A 198A) ECC. AABFARBEAS - BINFRBSRICOMROL, WE ARE CFR ESV, HOPUMAROAR YRICIIMG LPREFOCITR waters=river, lake, pond and estuary; forest=indigenous forests, woodlot and bamboo thicket; riparian=reed bed, riparian grassland and grassland; cropland=paddy field, cropland, and orchard; human habitations=park, town area, and human human habitations. K. EGUCHI and H. E. AMANO and has established populations successfully and panded its range throughout mainly a southern half of the country. Most exotic species are established in disturbed habitats near humans such as reed beds, riparian grasslands, croplands, and towns (Table 1), as are cases seen in other countries (Diamond & Veitch 1981; Moulton & Pimm 1983; Green 1984; Case 1996). Such a tendency in habitat selection may be a result of the competition with native species (Case 1996). However, it appears that it is also a reflection of the environments of their original habitats. For ex- ample, the original habitats of major introduced species from taxa such as Passeridae (Estrildinae and Ploceinae) and Sturnidae are savanna, cropland or dry woodland. Such habitats are similar to riparian grassland, cropland and parks in human habitations in Japan. Exceptions are species of the Timaliidae fam- ily, such as the Red-billed Leiothrix and the Melodi- ous Laughing Thrush, which have invaded indige- nous forests (Eguchi & Amano 2000; Sato 2000; Amano & Eguchi 2002a, b). EFFECTS OF INTRODUCTION Increasing population size of exotic species can exert a negative influence on the local ecosystem and native biota. In general, biotic invasion causes, 1) crop damage or other economic damage, 2) distur- bance or destruction of habitats and ecosystem, and 3) decline or extinction of native biota (Long 1981; Lever 1987). 1) Crop damage and other economic damage In America, 56% of 75 introduced species caused damages on agriculture and other economic resources (Temple 1992). Because many introduced birds in- habit disturbed habitats (e.g., cropland, human habi- tations), it is likely that the increase of such species may have an influence on agriculture and other human activity. In Japan, several introduced species cause the crop damage or other economic damage. The Feral Pigeon is one of important pest birds damaging soybeans or other vegetables, and flocks roosting in storehouses, buildings or apartment houses cause the pollution with droppings (Yamashina Institute for Ornithology 1979). The Light-vented Bulbul invaded the Okinawa Island in mid 1970s and badly damaged fruits and vegetables (Kinjo et al. 1987; Kinjo 1998). The Black-billed Magpie builds a nest on an electricity 6 pole and this practice has been spreading among indi- vidual birds in recent years (Eguchi 1996), causing an increase in electrical accidents. The Ring-necked Parakeets are important pest bird of cereals and fruits both in its original habitats and introduced areas in North America and Europe (Feare 1996). This species has been increasing around Tokyo and ad- verse effects on agriculture in the suburbs are possi- ble. 2) Influence on ecosystem and habitat No cases of habitat destruction by introduced birds have been reported in Japan. However, if large birds are introduced and increase in numbers, adverse ef- fects on the environment may be possible. 3) Influence on native avifauna Predation The Common Myna, Red-wiskered Bulbul Pyc- nonotus Jocosus and Red-vented Bulbul P. cafer can be aggressive towards native bird species; e.g., egg breaking in Mascarenes and Tahiti (Jones 1996; Thibault et al. 2002; Blanvillain et al. 2003). If brood parasite species are introduced, new host species in- capable of egg rejection would suffer serious losses in reproduction, e.g., the Brown-headed Cowbird Molothrus ater (Rothstein et al. 1980). So far, no cases have been reported of predation or brood para- sitism by introduced species in Japan. However, Vidua species are a brood parasite in its original habi- tats. In Japan, two species of Vidua are known to breed (Table 1). Although host species have not been known yet, these species may influence on the breed- ing of native species. The Black-billed Magpie is a notorious predator of songbirds’ nests in Britain (Gooch et al. 1991), but no such predation has been reported in Japan. Disease In the Hawaiian Islands, avian malaria has accom- panied the invasion of introduced birds and reduced the number of native species (van Riper et al. 1986; Dobson & May 1991). In the Mascarenes Islands, avian pox and parrot diseases derived from intro- duced birds have prevailed (Jones 1996). Species on small and remote islands are particularly susceptible to new pathogen and parasites brought by introduced birds. In Japan, prevalence of blood parasites is low in grassland living warblers and buntings (Sodhi et al. 1996, 1999; Nagata & Sodhi 2003). The mechanism Exotic birds in Japan maintaining low parasite prevalence in grassland bird communities is unknown and it is possible that these parasite-free communities are vulnerable to invasion of new parasites. Introduced birds could act as a large reservoir of various parasites and pass them to native parasite-free birds. Hybridization and Introgression Introduced species can hybridize with native closely related species. This may cause the extinction of endangered native species. Sterility in hybrids re- sults in wasted copulation, which causes a population decline. If it is fertile, introgressive hybridization re- sults in a loss of original morphological traits in ei- ther species (Rhymer & Simberloff 1996; Allendorf et al. 2003). If introduced species outnumbers a na- tive species, original traits of native species would disappear. In Anatidae, interspecific hybridization occurs often and, in many cases, hybrids are fertile (Parkin 1996). In particular, introduced Mallards or domestic races have hybridized with native subspecies or na- tive species, causing a decline in native species (Hunter 1996). The Hawaiian Duck Anas wyvilliana in Hawaii (Browne et al. 1993) and Grey Duck A. su- perciliosa in New Zealand (Gilllespie 1985; Rhymer et al. 1994) have been decreasing due to introgressive hybridization with introduced Mallards’ A. plathyrhynchos. In Japan, hybridization between do- mestic races of the Mallard and the Spot-billed Duck A. poecilorhyncha has been reported (Nakamura 1994), but effects on the latter’s populations are not SO severe yet. In a polytypic species, introduction beyond a range of a particular subspecies causes a loss of original traits of each subspecies. In Japan, the Green Pheas- ant Phasianus colchicus, which comprises four sub- species lost originality in each subspecies due to in- troduction for hunting (Omithological Society of Japan 2000). Reintroduction of endangered species should be carefully planned to reduce the likelihood of hybridizations. Interspecific competition The introduced Ruddy Duck Oxyura jamaicensis threatens the native White-headed Duck O. /euco- cephala with extinction through hybridization and in- terspecific competition in Britain (Hughes 1996). Hole-nesting introduced birds such as the European Starling Sturnus vulgaris, Common Myna and Ring- necked Parakeet can compete with native species for nesting holes (Long 1981; Lever 1987; Feare 1996). In the Mascarenes Islands, there are some cases of in- terspecific competition, which caused a decline of na- tive species, e.g., the Common Myna and Red- wiskered Bulbul affecting on the Mauritius Bulbul Hypsipetes olivaceus and Mauritius Cuckoo-shrike Coracina typica, the Madagascar Fody Foudia mada- gascariensis on the Mauritius Fody F. rubra, and the Ring-necked Parakeet on Mauritius Parakeet Psittac- ula eques (Jones 1996). In Hawaiian native forest, the Japanese White-eye Zosterops japonica, Melodious Laughing Thrush and Red-billed Leiothrix caused a decline of native species probably due to competition for food (Mountainspring & Scott 1985). In general, however, it is difficult to demonstrate the existence of competition. Some authors think that extinctions of native species due to the competition with introduced species have been few (Diamond & Veitch 1981; Case 1991,1996). Introduced species tend to inhabit environments disturbed by humans (Table 1), whereas native species inhabit undisturbed environments such as indigenous forests. Such a sep- aration of habitat may be due to a difference in habi- tat selection, not interaction between these species (Diamond & Veitch 1981). Because of adapting to local habitats, native species would have a competi- tive advantage over introduced species. Therefore, in- troduced species seldom establish in indigenous forests due to competition from native species. Case (1996) attributes the success of introduced birds on islands to environmental changes by humans. How- ever, between particular species, introduced species can outnumber native species even in indigenous forests due to demographic or environmental stochas- ticity or environmental heterogeneity (Sax & Brown 2000). In the Hawaiian Islands, generalist species such as the Japanese White-eye and Timaliidae outnumbered native species specializing in a particular food re- source (Mountainspring & Scott 1985). In oceanic is- lands, both habitat size and abundance of individual food resources are small. Usually, successful invasive species are a generalist (Jones 1996). Even in oceanic islands, introduced generalist species can become es- tablished and maintain a population by virtue of its wide niche. Such a introduced generalist may con- stantly interact with native species specializing in a particular resource and, because of its small niche width and a limited available resource, a native spe- cialist may eventually decline its population size through interspecific competition. K. EGUCHI and H. E. AMANO e Red-billed Leiothrix has also invaded the de- ciduous broadleaved forests in Japan. This species dominates over the avifauna of such habitats in Kyushu (Eguchi & Masuda 1994; Eguchi & Amano 2000). In the avifauna of deciduous broadleaved forests, major species occupying a similar niche to this species are the Japanese Bush-warbler Cettia di- phone, Great Tit Parus major, Varied Tit P. varius, Coal Tit P. ater, Willow Tit P. montanus and Long- tailed Tit Aegithalos caudatus. Although there is no evidence to confirm the presence of interspecific competition between introduced and native species for food resources or nest sites, predation on Japanese Bush-warbler nesting in the same nesting habitat as the Red-billed Leiothrix has been found to be higher than in other regions where the latter species has not yet invaded (Amano & Eguchi 2002a, b). Even if there is no direct competition, a high density of Red- billed Leiothrixes may cause the decline of the repro- ductive success in sympatric species through appar- ent competition (Martin & Martin 2001) mediated by predator attraction. PROBLEMS CAUSED BY INTRODUCED BIRDS IN JAPAN Problems caused by avian introductions are differ- ent from those in other organisms in some areas. First, the introduction of birds does not cause as much public concern as in the cases of other organ- isms. As mentioned, the effects of introduced birds on the ecosystem, biota and human society are not obvious. Most introduced birds inhabit disturbed habitats. Therefore, introduced birds replace native species that have disappeared due to human distur- bance and hence play a role in maintaining ecosys- tems and mitigating environments for human being (Case 1996). For example, introduced game birds oc- cupy a similar niche held by now-extinct or rare na- tive species and facilitate seed dispersal and germina- tion of native plants in disturbed habitats in Hawaii (Cole et al. 1995). Therefore, people believe that be- cause introduced birds are harmless compared to other animals such as mongooses in the Ryukyu Is- lands and feral goats in the Bonin Islands, they should be allowed to exist in disturbed habitats until they start to exhibit negative impacts on the ecosys- tem and humans. Second, it is difficult to clarify the process and scale of introduction of birds in Japan. In recent years, there have been few large-scale introductions like the ones done in New Zealand and Hawaii from the late nineteenth and early twentieth centuries. In- stead, accidental and small-scale deliberate releases of caged birds are major sources of introductions into Japan. This makes it difficult to clarify the frequency and scale of introduction. Establishment and popula- tion growth of these escapees progress without at- tracting much attention, and such populations sud- denly show up in large numbers that make regulation and eradication practically impossible. Many birds have been imported to Japan (Nash 1993; Melville 1994); more than two million every year by some es- timates (Agency of the Environment 1986). Most of them are non-native species. Because many exotic birds have been imported into Japan, many escapees would establish self-maintaining populations. Finally, it is nearly impossible to eradicate estab- lished exotic birds, both technically and psychologi- cally. Generally, people cannot understand why a population already established in a habitat should be eradicated. For example, the Black-billed Magpie is an introduced species in Japan. Its population in northern Kyushu has increased in size over the past 20 to 30 years (Eguchi & Takeishi 1997). However, few people believe that this species should be eradi- cated just because it is an introduced species. The same situation is seen in the case of the Red-billed Leiothrix increasing in deciduous broadleaved forests throughout Japan. Exotic species can establish populations in an undisturbed habitat and exert a negative effect on na- tive species (Sax & Brown 2000). For example, timaliid birds like the Red-billed Leiothrix or Melodi- ous Laughing Thrush have invaded into native forests in Japan and Hawaii, where they interact directly or indirectly with native species (Mountainspring & Scott 1985; Eguchi & Amano 2000; Amano & Eguchi 2002a, b). Increasing introductions of exotic birds increase the likelihood of escapees establishing populations in undisturbed habitats. This may in- crease the impact of such species on native species. So far, native species have declined in islands, such as the Hawaiian Islands, Tahiti, and Mascarenes Is- lands. In Japan, it is likely that the impact by intro- duced species will be serious in small islands such as the Ryukyu Islands. LEGAL CONTROL OF AVIAN INTRODUCTION IN JAPAN One basic measure against biotic introduction is Exotic birds in Japan the legal control of introductions. The Ministry of the Environment has made a new Biodiversity Strategy in 2002 (http://www/biodic.go.jp/nbsap.html). It in- cludes three main actions for biotic introduction, pre- vention, assessment and management (regulation and eradication). However, current legislation cannot con- trol introduction effectively. Because most of the in- troduced birds and mammals in Japan are escaped captive animals, a legal barrier against the import of animals is necessary. Basically the import of exotic species should be strictly regulated. The permission of import should be issued based on the risk assessment involving the abundance of the species in original countries, proba- bility of establishment after escaping into non-origi- nal habitats, estimation of influences on native biota and difficulty of eradication of naturalized popula- tion. After issuing the permit, each individual should be registered and identified with a metal ring. An owner of the bird keeps the registration number and form. When a bird escapes or dies, the owner should notify its registration number and the escape or death to the permit issuing organization. Campaigns for preventing introductions of exotic species, including the threat of biotic invasion on the biodiversity, should be conducted for agents of import and bird fanciers. Exotic species already naturalized should be eradi- cated as soon as they are recognized. Eradication in an early stage of introduction is essential. Eradication and regulation project of introduced birds should be conducted with a well-laid plan that has a feedback system that allows the project to be adjusted accord- ing to the result of each trial. Nation-wide monitoring of establishment of introduced birds is necessary to make plans for eradication and regulation. So far, in- vestigations of introduced birds have been scarce and regional. Campaigns based on accurate and adequate information make it easy to obtain public approval for executing eradication and measurement project. ACKNOWLEDGMENTS This study is partly supported by the Pro Natura Fund from the Nature Conservation Society of Japan, the Grant-in-Aid for the Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan (No. 08454252), and the Grant for the Global Environment Re- search Program (No. F-3 (2001-2003)) by the Ministry of the Environment of Japan. The following organizations helped us in the inquiry of introduced birds; Ornithological Society of Japan, Japanese Society for Preservation of Birds, Wild Bird Society of Japan (WBSJ), Sapporo Branch of WBSJ, Ki- takyushu Branch of WBSJ, mailing list “jeconet”, and mailing list “banding”. We thank all the people who have provided in- valuable information to our inquiry, especially Dr. M. Izawa, Mr. Y. Nakamura, Mr. M. Kamogawa, Mr. S. Nakamura, and Mr. H. Kuroda. REFERENCES Agency of the Environment (1986) A report of the proj- ect on the registration system of caged birds. Agency of the Environment, Tokyo, (in Japanese). Allendorf FW, Leary RF, Spruell P & Wenburg JK (2003) The problems with hybrids: setting conserva- tion guidelines. Trends Ecol Evol 16: 613-622. Amano HE & Eguchi K (2002a) Nest-site selection of the Red-billed Leiothrix and Japanese Bush Warbler in Japan. Ornithol Sci 1: 101-110. Amano HE & Eguchi K (2002b) Foraging niches of in- troduced Red-billed Leiothrix and native species in Japan. Ornithol Sci 1: 123-132. Blackburn TM & Duncan RP (2001) Determinants of establishment success in introduced birds. Nature 414: 195-197. Blanvillain C, Salducci JM, Tutorial G & Maura M (2003) Impact of introduced birds on the recovery of the Tahiti Flycatcher (Pomarea nigra), a critically en- dangered forest bird of Tahiti. Biol Conserv 109: 197-205. Browne RA, Griffin CR, Chang PR, Hubley M & Mar- tin AE (1993) Genetic divergence among populations of the Hawaiian Duck, Laysan Duck, and Mallard. Auk 110: 49-56. Case TJ (1991) Invasion resistance, species build-up and community collapse in metapopulation models with interspecies competition. Biol J Linn Soc 42: 239-266. Case TJ (1996) Global patterns in the establishment and distribution of exotic birds. Biol Conserv 78: 69-96. Cole FR, Loope LI, Medeiros AC & Raikes JA (1995) Conservation implications of introduced game birds in high-elevation Hawaiian shrubland. Conserv Biol 9: 306-313. Diamond JM & Veitch GR (1981) Extinctions and intro- ductions in the New Zealand avifauna. Science 211: 499-501. Dobson AP & May RM (1991) Parasites, Cuckoos, and avian population dynamics. In: Perrins CM, Lebreton J-P & Hirons GJM (eds) Bird population Studies-Rel- evance to Conservation and Management, pp 391- 412, Oxford University Press, Oxford. Duncan RP (1997) The role of competition and intro- duction effort in the success of passeriform birds in- K. EGUCHI and H. E. AMANO duced to New Zealand. Am Nat 149: 903-915. Duncan RP, Blackburn TM & Veltman CJ (1999) Deter- minants of geographical range sizes: a test using in- troduced New Zealand birds. J Anim Ecol 68: 963- 975. Eguchi K (1996) Recent increase of nesting on utility poles by the Black-billed Magpie Pica pica sericea. Jpn J Ornithol 45: 101-107. Eguchi K & Amano HE (1999) Naturalization of birds in Japan. Jpn J Ornithol 47: 97-114 (in Japanese with English summary). Eguchi K & Amano HE (2000) Problems of introduc- tions and naturalizations in birds. Jpn J Conserv Ecol 5: 131-148 (in Japanese with English summary). Eguchi K & Kubo H (1992) Origin of the Black-billed Magpie Pica pica sericea: by investigation of histori- cal records. J Yamashina Inst Ornithol 24: 32—39 (in Japanese with English summary). Eguchi K & Masuda T (1994) Habitats of the Red-billed Leiothrix Leiothrix lutea in Kyushu. Jpn J Ornithol 43: 91-100 (in Japanese with English summary). Eguchi K & Takeishi M (1997) The ecology of the Black-billed Magpie in Japan. Acta Ornithol 32: 33— 38. Feare CJ (1996) Rose-ringed parakeet Psittacula krameri: a love-hate relationship in the making? In: Holmes JS & Simons JR (eds) The Introduction & Naturalisation of Birds, pp 107-112, Stationery Of- fice Publ Centre, London. Gilllespie GD (1985) Hybridization, introgression, and morphometric differentiation between Mallard (Anas platyrhynchos) and Grey Duck (Anas superciliosa) in Otago, New Zealand. Auk 102: 459-469. Gooch S, Baillie SR & Birkhead TR (1991) Magpie Pica pica and songbird populations. Retrospective in- vestigation of trends in population density and breed- ing success. J Appl Ecol 28: 1068-1086. Green RE (1997) The influence of numbers released on the outcome of attempts to introduce exotic bird species to New Zealand. J Anim Ecol 66: 25-35. Green RJ (1984) Native and exotic birds in a suburban habitat. Aust Wildl Res 11: 181-190. Hughes B (1996) The ruddy duck Oxyura jamaicensis in the Western Palearctic and the threat to the white- headed duck Oxyura leucocephala. In: Holmes JS & Simons JR (eds) The Introduction & Naturalisation of Birds, pp 79-86, Stationery Office Publ Centre, Lon- don. Hunter ML Jr. (1996) Fundamentals of Conservation Biology. Blackwell, Oxford. lozawa H (2000) 550 Wild Birds of Japan: Land Birds. Bun-ichi Sogo Shuppan Press, Tokyo (in Japanese). Jones CG (1996) Bird introductions to Mauritius: status 10 and relationship with native birds. In: Holmes JS & Simons JR (eds) The Introduction & Naturalisation of Birds, pp 113-123, Stationery Office Publ. Centre, London. Kinjo T (1998) The biology of the Chinese bulbul, Pyc- nonotus sinensis, and the methods to prevent the damage to crops in Okinawa Island. Plant Prot 52: 397-402 (in Japanese). Kinjo T, Nishimura M & Nakamura K (1987) Invasion of the Chinese bulbul, Pycnonotus sinensis, to Oki- nawa Island and the status of damage on vegetables by it. Plant Prot 41: 428-432 (in Japanese). Kiyosu Y (1978) The Birds of Japan. Kodansha, Tokyo (in Japanese). Kuroda N (1913) Flocks of the Java Sparrow in Haneda village. Zool Mag 25: 563-566 (in Japanese). Kuroda N & Komiya T (1987) Taxonomy of Phasiani- nae. In: Kuroda N & Morioka H (eds) Animals of the World 10(1), pp 66-126, Dobutsusha, Tokyo (in Japanese). Lever C (1987) Naturalized Birds of the World. Long- man, London. Long JL (1981) Introduced Birds of the World. Reed, Wellington. Martin PR & Martin TE (2001) Ecological and fitness consequences of species coexistence; a removal ex- periment with Wood Warblers. Ecology 82: 189-206. McLain DK, Moulton MP & Redfearn TP (1995) Sex- ual selection and the risk of extinction of introduced birds on oceanic islands. Oikos 74: 27-34. Melville DS (1994) International trade in birds with spe- cial reference to trade from China through Hong Kong to Japan. In: Status Report on Wild Bird Poach- ing, pp 1-27, National Wild Bird Poaching Counter- measure Committee, Tokyo. Moulton MP (1985) Morphological similarity and the coexistence of congeners: an experimental test with introduced Hawaiian birds. Oikos 44: 301-305. Moulton MP (1993) The all-or-none pattern in intro- duced Hawaiian passeriforms: the role of competition sustained. Am Nat 141: 105-119. Moulton MP & Pimm SL (1983) The introduced Hawai- ian avifauna: biogeographic evidence for competition. Am Nat 121: 669-690. Moulton MP & Pimm SL (1986) The extent of competi- tion in shaping an introduced avifauna. In: Diamond JM & Case TJ (eds) Community Ecology, pp 80-97, Harper & Row, New York. Mountainspring S & Scott JM (1985) Interspecific com- petition among Hawaiian forest birds. Ecol Monogr 55: 219-239. Nagata H & Sodhi NS (2003) Low prevalence of blood parasites in five Sylvidae species in Japan. Ornithol Exotic birds in Japan Sci 2: 73-74. Nakamura K (1990) Tree Sparrow and Cabbage Butter- fly are introduced species. PHP Institute Press, Tokyo (in Japanese). Nakamura K (1994) On Naturalized Animals. Giho-do Press, Tokyo (in Japanese). Narusue M (1981) Exotic birds seen in a metropolitan area of Tokyo. Yurikamome 309: 6—7 (in Japanese). Nash S (1993) Sold for a song, TRAFFIC International, Cambridge.(Japanese version translated by WING Volunteer Asia Club. Wild Bird Society of Japan). Ornithological Society of Japan (2000) A checklist of birds of Japan, 6th edition. Ornithological Society of Japan, Obihiro. Parkin D (1996) Colonisation and hybridisation in birds. In: Holmes JS & Simons JR (eds) The Introduction & Naturalisation of Birds, pp 25-35, Stationery Office Publ. Centre, London. Rhymer JM, Williams MJ & Braun MJ (1994) Mito- chondrial analysis of gene flow between New Zealand Mallards (Anas platyrhynchos) and Grey Ducks (A. superciliosa). Auk 111: 970-978. Rhymer JM & Simberloff D (1996) Extinction by hy- bridization and introgression. Ann Rev Ecol Syst 27: 83-109. Rothstein SI, Verner J & Stevens E (1980) Range ex- pansion and diurnal changes in dispersion of the Brown-headed Cowbird in the Sierra Nevada. Auk 97: 253-267. Sato S (2000) Naturalization of exotic Hwamei Garru- lax canorus in northern Kyushu, Japan. Jpn J Ornithol 48: 233-235 (in Japanese with English summary). Sax DF & Brown JH (2000) The paradox of invasion. Global Ecol & Biogeogr 9: 363-371. Simberloff D (1992) Extinction, survival, and effects of birds introduced to the Mascarenes. Acta Oecol 13: 663-678. Simberloff D & Boecklen W (1991) Patterns of extinc- tion in the introduced Hawaiian avifauna: a reexami- nation of the role of competition. Am Nat 138: 300- B27. Simberloff D & Stiling P (1996) Risks of species intro- duced for biological control. Biol Conserv 185-192. Sodhi NS, Bennett GF & Nagata H (1996) Absence of blood parasites in the Japanese Reed Bunting Ember- iza yessoensis. Jpn J Ornithol 45: 115-117. Sodhi NS, Adlard RD, Nagata H & Kara AU (1999) Low prevalence of blood parasites in six Emberiza species in Japan. Jpn J Ornithol 47: 65-67. Sol D & Lefebvre L (2000) Behavioural flexibility pre- dicts invasion success in birds introduced to New Zealand. Oikos 90: 599-605. Sol D, Timmermans S & Lefebvre L (2002) Behav- ioural flexibility and invasion success in birds. Anim Behav 63: 495-502. Sorci G, Moller AP & Clobert J (1998) Plumage dichro- matism of birds predicts introduction success in New Zealand. J Anim Ecol 67: 263-269. Tanaka S & Takehara K (2003) Distribution and present status of feral Common Peafowls Pavo cristatus in the Sakishima Islands, the Ryukyu Islands. Bull Oki- nawa Pref Mus 29: 19-24 (in Japanese). Temple SA (1992) Exotic birds: a growing problem with no easy solution. Auk 109: 395-397. Thibault J-C, Martin J-L, Penloup A & Meyer J-Y (2002) Understanding the decline and extinction of monarchs (Aves) in Polynesian Islands. Biol Conserv 108: 161-174. van Riper C III, van Riper SG, Goff ML & Laird M (1986) The epizootiology and ecological significance of malaria in Hawaiian land birds. Ecol Monogr 56: 327-344. Veltman CJ, Nee S & Crawley MJ (1996) Correlates of introduction success in exotic New Zealand birds. Am Nat 147: 542-557. Yamashina Institute for Ornithology (1979) A report of a study on pest control of feral pigeons. Yamashina Inst Ornithol, Tokyo (in Japanese). reted opb bid otha? jo or He bien 0) mriatiwiele’ wilh sitive finds div Sider 11 190% Deee Crean 5 ZO VON Thee erent Many RAR ASA WO) gor wi Henaimi2. offal ueerd bhatibAl Auerinomraritenlienlt ftond oigelGicbtl bah —~ ‘Lomeien ev #A AD vom 2 ibe? Gita edie Pechernlt inc’ tiene’ bead ee top lead yieel ieeiaepead yetwtiee comes, Tod heyapeaUnitiay me A hort MnQolend |. 2 eceaaret mipingen Len at red Uietwirdsl (2 2B ye 1 GH trwlbA 24 iilbo® Haeatatph Ay Taree snd oe st eres id Ase) Padi tty odreese reed esr tachyinest: NT SFE Senha ea, Wad, au oFatc fe lortitinO 10 .neqhlyaresinapl OMI 9h Gene! arte iocfie vn /evermanties tewni vaio, (OL). svivesia tT BR OD lok tomas yrolsde brit |W linens pam oitre MA CCWPGebsobd anpaivniieodth ereaiiary PrqOhOl > dihah! i fit ue POOWR. Cea TOs! Qpyinidiz ‘taStyol dat) tin neiTitre dh ph b (COTS) Deo deta ud PA chine 0 ia? 4 ‘ Dm! iy eo boul nl ive bey Ree! = bai adi Fle a aol ) : iy ln Oiet@belal at 2 when VRE Poy at Havre Np adler hy LAO Vb itviedal Y QA OM OF for} =Sysitt hevhritbersn Fee et ligt / Titian = a a ortboum ashen “bud 1 wien ' ir yt! NS Moria hob 4 OOS LEG | Fates Taree A eects BtROLT EME fine Dh , OU citi before adi Oe) eliotn ie b Hh A dire — ats ond wifi omy 10 rhea: : vio Lipwanies? @niani (eeeitowestese | won Thad lav de tecetitey? tind) titel /f wretch wy AVR MP Sel egiliads? 5m ARM : FET date (eenttere we ' dengncpa ind dae Oh cOC Gah el wel ont (AL AOR AAR oh seed ry 6 calnid oitoxd (8001) AQ Signet © :0wieeiel vollenA naieesigoria be : Wh Oh 2 CR 4) oud neislapwe a Kaien oe : F imtyy FE TUN Pena neal a Své iene s coleh wit MM) Dv gigseidt eo van UK e) Wensvertk t usa g 60% jonrtoel » Soitarerrvstys) $9 SL) oh WD” meryema aad wig tints ! aid init orbodt' tl alae i ne Land perenapin Sovu oplleta® werd arth’ nb bro Lt Lata. ata vi he ebol = () Riedl ov INORS OH (ei tage aenglt wore TW eQineit ine Voit Hdhude nol ean liu ggotolinggiy ea RAR panier Pry Lyte AO Veneta ie cea) lin: riotepwol Wheelers JL cieineroe sity Ades desaigaing .i2 Bk ~! (Panam tip Mituimp om! (Oy ‘Mb CORNEUM Rate O aan eD Rey ee) DNA NOORD AVE daa be Weed) See hy mop aie hi oases hoe Vereen exntinnad (SORE) Tit Chih baaPivets ~ Eh Pre dA dantlaa otie ab c iy goto thr O) tt. aie) erie yY Nyc. an Apert “n ta yo ans etiverserieay 9) vee y wees bo unis (Pin Riot 4 meryt ton) coeds beta (ab nn oa) 2 anton Akjinwal bistitibe rhe ; , yraihe role on $e ion 11 10=-T19 : k - PA Pin SE eo oy Teeter Ser ow /Finuacums frie raphic ov dence Fors St1 | UB) The extent * re = P > AM ict) ail aura: Sie fa rniuntlf\' Be WORF, ; Ll erie e Rai Ne tor d a as a7 ¢ > ipesiiennng & Scorn JM (TORS) ine feito woo Mewaran twirest binds b ¢ o f qe iy 4) 4 = 4 aed 1 ’ $ At) ) Liaw shew ne rare ; Svivicdag pe ty lay ] SPECIAL FEATURE Ornithol Sci 3: 13—21 (2004) Invasive bird species The spread of the introduced Melodious Laughing Thrush Garrulax canorus in Japan Kazuto KAWAKAMI'* and Yoshimori YAMAGUCHI ‘Tama Forest Science Garden, Forestry and Forest Products Research Institute, Todori 1833, Hachioji, Tokyo 193-0843, Japan ? Lake Tanzawa Visitor Center, Kurokura 515, Yamakita, Kanagawa 258-0202, Japan ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2004 Abstract The Melodious Laughing Thrush Garrulax canorus is an introduced species in Japan, and it has been recorded in the wild since the 1980s. The distribu- tion of this species was estimated based on questionnaires. Four populations of this species have been confirmed in the western Kanto, northern Kyushu, southern To- hoku and Nagano Prefecture. There is a possibility that the expansion of this species is limited by elevational conditions and snowfall. Nevertheless, the distribution is still expanding. Therefore, it is necessary, as soon as possible, to assess the effect of the Melodious Laughing Thrush on the native birds. Key words Conservation, Expansion of distribution, Garrulax canorus, Introduced birds, Melodious Laughing Thrush The Melodious Laughing Thrush Garrulax canorus 1s an introduced species in Japan, and it has been observed in the wild since the 1980s (Sato 2000). The reason why this species naturalized in Japan is not clear. It is naturally distributed in south- ern China, Taiwan, Laos and Vietnam (Long 1981; Zheng 1982). It is a popular cage bird in China and Southeast Asia and a large number of individuals have been exported from Hong Kong (Nash 1999). This species has been imported into Japan since the seventeenth century (Isono & Uchida 1992). This species was introduced to the Hawaiian Islands in the early 1900s and to California in 1941 accidentally and purposely (Long 1981). While it failed to estab- lish a population in California, it succeeded in natu- ralizing in the Hawaiian Islands, where it became the most common species below 1200 meters in eleva- tion (Long 1981; Mountainspring & Scott 1985). The cause of disappearance in California is not known. The Melodious Laughing Thrush prefers dense un- derstory as breeding and foraging habitat and is sedentary in China and the Hawaiian Islands (Long 1981; Mountainspring & Scott 1985; Cheng et al. 1987). It is omnivorous and forages mainly on the ground for arthropods, earthworms, fruits and other (Received 11 September 2003; Accepted 10 November 2003) * Corresponding author, E-mail: kazzto@ffpri.affrc.go.jp food (Long 1981; Cheng et al. 1987; Yamaguchi 2000; Kawakami 2002). This species was shown to have out competed native species on the Hawaiian Is- lands (Mountainspring & Scott 1985). Thus, there is concern that this species might give an impact the na- tive species such as thrushes and bush-warblers that utilize the same prey and microhabitats (Kawakami 2002). Thus far, the Melodious Laughing Thrush has been observed in northern Kyushu, in the Kanto, and in Fukushima Prefecture in Japan (Sato 2000; Yama- guchi 2000; Kawakami 2002). While it appears to be expanding its range in Japan (Sato 2000; Kawakami 2002), there is no study on its nationwide distribution to provide essential information for a management strategy. The aim of the present study is to show the past and present distribution of the species and to dis- cuss factors limiting its population. With question- naire, the distributions were clarified. With the data of recent distributions, we discussed on two limiting factors of range expansion: elevation and deepest snow depth of normal years. As the Melodious Laughing Thrush has been observed in lowland mon- tane areas (Sato 2000; Yamaguchi 2000), there is a possibility that elevation limits it’s distribution. Since it frequently forages on the ground and is resident, deep snow might prevent it from foraging. We there- fore compared its distribution with maps showing the K. KAWAKAMI and Y. YAMAGUCHI pest snow depth of normal years. METHODS 1) Questionnaires Questionnaires were used to collect information on its distribution. As the Melodious Laughing Thrush spends much of its time in bushes, it is not easy to observe. However, its song is loud and conspicuous (Kawakami & Yamaguchi unpublished). Therefore, it is easy to detect its presence. Questionnaires were distributed through some mailing lists (Jeconet ML, Wildlife ML, Jbnest ML), newspapers (Sankei Shim- bun 18 January 2000, Tokyo Simbun 27 January 2000, Asahi Shimbun 2 February 2000) and maga- zines (“Yacho” of Wild Bird Society of Japan, “Watashitachi no Shizen” of Japanese Society for Preservation of Birds, “Yacho Fukuoka” of WBSJ Fukuoka branch, “Kusunoki” of The Nature Inter- preter Society of Fukuoka). The questions were simple: 1) Have you seen or heard the Melodious Laughing Thrush? 2) If yes, when and where did you see or hear it? The morpho- logical character was described with the questions as follows. “The body size is similar to the Dusky Thrush Turdus naumanni. Body color is bright brown, and there are white teardrop patterns around the eyes. The color of bills and legs is yellow”. Ques- tionnaires in newspapers and magazines also in- cluded photographs of the bird. The Grey Starling Sturnus cineraceus is a very common bird species in Japanese lowlands, and it has similar morphological characters to the Melodious Laughing Thrush (i.e. similar body size, brown bod- ies, white pattern on their cheeks and yellow bills and legs). The two species can be easily confused with one another and some responses were misidentifica- tion. However, their ecological habits are very differ- ent. Especially, the Grey Starling prefers open habi- tats, while the Melodious Laughing Thrush prefers forests or bushes. Therefore, we excluded question- able answers from the analysis that involved ecologi- cal habits of Grey Starlings, such as “perching on wires side by side”, “nesting under a roof” and “for- aging at open habitats in residential areas”. Re- sponses were collected from January 2000 to August 2001. 2) Literature citation Though the questionnaire method is useful to cover a wide area, its credibility is not high (Fujimaki & Konishi 1996). In order to compensate for question- able information, we referred to the past reports of observations, specifically the Kanagawa chapter of WBSJ (1998), Kunihiro (1999), Sonoda (1999), Sato (2000), Yamaguchi (2000) and the records of in- quiries on Melodious Laughing Thrushes at the Ya- mashina Institute for Ornithology. 3) Field detection Based on the information of the questionnaire method, we investigated the existence of Melodious Laughing Thrushes at the edges of their distributions in western Kanto and southern Tohoku and Nagano Prefecture. The playback method was used to identify the species. This species readily responded to songs played on a tape recorder both in the breeding and non-breeding seasons (Kawakami & Yamaguchi un- published). We played the songs in forests that had grown bushes near the edge of the distributions and recorded the existence of the species. This survey was conducted from January 2000 to August 2001. 4) tion We made distribution maps with grid maps consist- ing by square grids for each year based on the above information. Each grid was about 25km? that was consisted of 5X5 grids based on standard grid data (Ministry of Environment 1997a-l). We calculated areas of 100% minimum convex polygon connecting the midpoint of each grid for each population for each year. Based on these values, we got the regres- sion formula between the year and the square root of the area. The distribution and expansion of each popula- 5) Snow depth and elevational conditions In order to assess the influence of snowfall, the dis- tribution map of the species was superimporsed on the map showing the deepest snow depth of normal years (Japan Meteorological Agency 1993). And, in order to clarify the elevational conditions of the habi- tats, the elevation of each plot in which the species was recorded was classified into 6 categories, 0-200 m, 200-400 m, 400-600 m, 600-800 m, 800- 1000 m and more than 1000m. As there were plots that can not be recognized their reliable elevation, we excluded such data from analysis. The frequency of elevation data belonging to each category was shown for each population. The spread of Garrulax canorus RESULTS I) tion Four populations were identified at northern Kyushu, western Kanto, southern Tohoku and Nagano Prefecture (Fig. 1). We obtained 173, 557, 37 and 39 records by questionnaire for the four popula- tions, respectively, and 184, 32 and 5 records by field detection for latter three populations, respectively. Figs. 2—5 show the distribution and expansion of each population. The Kyushu population originated in Fukuoka Prefecture in 1985 and expanded to south- ward. Melodious Laughing Thrushes were found in Fukuoka, Oita, Saga and Kumamoto prefectures in 2000. No map could be drawn for 2001 because no additional information was obtained for that year. The first record of the Kanto population was re- ported at Fujino Town, Kanagawa Prefecture in 1987. The distribution extended in all directions along mon- The distribution and expansion of each popula- tane habitats. The eastward expansion stopped at the urban area. The species occurred in Kanagawa, Tokyo, Yamanashi, Saitama, Gunma and Shizuoka prefectures in 2001. The Tohoku population origi- Fig. 1. The distribution of introduced Melodious Laughing Thrushes in Japan. a) The Kyushu population, b) The Kanto population, c) The Tohoku population, d) The Nagano popula- tion. e) 2000 Fig. 2. The distribution of introduced Melodious Laughing Thrushes by year in Northern Kyushu. The gray area shows the range where the deepest snow depth of normal years is more than 20 cm. 15 K. KAWAKAML and Y. YAMAGUCHI a) 1990 e) 2001 The distribution of introduced Melodious Laughing Thrushes by year in Western Kanto. The symbol is the same as Fig. 2. Fig. 3. nated in Fukushima City, northern Fukushima Prefec- ture in 1997, and expanded toward the east and south. By 2001, this population had spread to southern Miyagi and northern Ibaraki prefectures. The Nagano population appeared at Saku City in 1995 and the dis- tribution was restricted to near the city. Although the area was expanded once, it has subsequently de- creased. We excluded the following records from the analy- sis, because no record was obtained in subsequent years near the site; at Hachioji, Tokyo in 1981; at Sakai City, Osaka in 1994; and at Mt. Tsukuba, Ibaraki in 1998. We also excluded the record from Yasato City, Ibaraki in 2000, because it was far from main populations and could not be classified. Fig. 6 shows the relationship between the square root of the area of 100% minimum convex polygon occupied by Melodious Laughing Thrushes (y) and the year (x) for each population. The relationships be- tween them are expressed as y=2.30x—4560 (a, 1985— 1994, r=0.947) and y=3.03x-5990 (b, 1995-2000, 16 t=0.979) for the Kyushu population, y=1.29x—2570 (c, 1987-1993, r=0.896) and y=8.90x—17700 (d, 1994-2001, r=0.990) for the Kanto population and y= 13.4x—26800 (e, r=0.959) for the Tohoku popula- tion. 2) Snow depth and elevational conditions The areas that have more than 20cm deepest snow depth of normal years (heavy snow area) were shown in Fig. 2-5. There was no heavy snow area near the center of the distribution for the Kyushu population (Fig. 2). At the Kanto population, Melodious Laugh- ing Thrushes extended its range mainly without the heavy snow area. Though the distribution extended across the heavy snow area, the route was along the low elevation area where is considered to have rela- tive light snow (Fig. 3). At the Fukushima popula- tion, the distribution seemed to spread avoiding the heavy snow area (Fig. 4). There was a light snow area about 20 km west of the Nagano population, though the species occurred in a heavy snow area The spread of Garrulax canorus c) 2000 d) 2001 Fig. 4. The distribution of introduced Melodious Laughing Thrushes by year in Southern Tohoku. The symbol is the same as Fig. 2. (Fig. 5). There was no record for elevations higher than 1400 m. The ranges of elevations of the four popula- tions were 0—1000m (Kyushu), 0-1400m (Kanto), 0-600 m (Tohoku) and 400-800 m (Nagano; Fig. 7). DISCUSSION Nationwide distribution There were four populations of Melodious Laugh- ing Thrushes: in northern Kyushu, western Kanto, southern Tohoku and Nagano Prefecture. The oldest information was from 1985 in northern Kyushu, while Sato (2000) noted that the species had already been observed in the early 1980s at Rikimaru Dam in northern Fukuoka Prefecture. These four populations had discrete origins and oc- curred discontinuously since the 1980s. The Red- billed Leiothrix Leiothrix lutea was also naturalized at various region of Japan since the 1980s, most likely due to releases by individuals and shops (Eguchi 2002). The patterns of distributions were similar between the two species. The same likely oc- 1) curred with Melodious Laughing Thrushes, though the cause of the naturalization is unclear. One pre- sumable cause of naturalization is the increase of dense understory. This species prefers dense under- story as breeding and foraging habitats (Long 1981; Cheng et al. 1987; Sato 2000; Kawakami 2002). In lowland forests, shrub cover has increased and be- come thicker because of the depression of the forestry industry and insufficient management of rural forests for past few decades (Takeuchi et al. 2001; Forestry Agency 2003). The increase in shrub cover might accelerate the naturalization of this species. Since the 1980s, four introduced Timalliinae species, i.e. the above two species plus the Masked Laughing Thrush G. perspicillatus and the White- browed Laughing Thrush G. sannio, have increased in Japan (Nakamura 1993; Eguchi & Amano 1999; Kawakami 2002). These four species and Chinese Bamboo Partridge Bambusicola thoracica, which be- came feral in Japan in the early twentieth century, have similar original distributions, mainly in southern China (China Wildlife Conservation Association K. KAWAKAMI and Y. YAMAGUCHI c) 1999 The distribution of introduced Melodious Laughing Thrushes by year in Nagano Prefecture. The symbol Fig. 5. is the same as Fig. 2. © Kyushu @ Kanto A Tohoku (L Nagano 100 = — © 2 © 80 S 2 = 2 60 — Smal 2 Ze) ® 2 40 _ — (2) S 20 oO a ® o = 0 oC (op) 1980 1985 1990 1995 2000 2005 Fig. 6. The square root of the area of 100% minimum con- vex polygon occupied by Melodious Laughing Thrushes in the four population zones of Kyushu, Kanto, Tohoku and Nagano Prefecture. d) 2001 100 WB Kanto Kyushu — 80 [] Tohoku = — Nagano — a 60 ¢ ) =) oD 40 iadle < H. TOJO and S. NAKAMURA Number of territories in 8 ha plot Family Common name Scientific name 1994 1995 Phasianidae Chinese Bamboo Partridge Bambusicola thoracica* ae + Copper Pheasant Syrmaticus soemmerringii + 0.5 Columbidae Rufous Turtle Dove Streptopelia orientalis 1.0 0.5 Japanese Green Pigeon Sphenurus sieboldii + Cuculidae Fugitive Hawk Cuckoo Cuculus fugax 0.5 Oriental Cuckoo Cuculus saturatus 0.5 1.0 Little Cuckoo Cuculus poliocephalus 1.0 1.5 Picidae Japanese Green Woodpecker Picus awokera 0.5 1.0 Japanese Pygmy Woodpecker Dendrocopos kizuki 5) Pie) Pycnonotidae Brown-eared Bulbul Hypsipetes amaurotis DS) ile) Laniidae Bull-headed Shrike Lanius bucephalus 3 Troglodytidae Winter Wren Troglodytes troglodytes 0.5 1.0 Turdidae Japanese Robin Erithacus akahige af 0.5 Siberian Blue Robin Luscinia cyane ] 1.5 White’s Thrush Zoothera dauma + + Siberian Thrush Turdus sibiricus oF Grey Thrush Turdus cardis 1.0 0.5 Sylviidae Short-tailed Bush Warbler Urosphena squameiceps 4.5 3h) Bush Warbler Cettia diphone 1355) 10.5 Arctic Warbler Phylloscopus borealis + Eastern Pale-legged Leaf Warbler Phylloscopus borealoides 3p Crowned Willow Warbler Phylloscopus coronatus 1.0 1.0 Muscicapidae Narcissus Flycatcher Ficedula narcissina 1.0 15, Blue and White Flycatcher Cyanoptila cvanomelana 0.5 1.0 Brown Flycatcher Muscicapa dauurica oF Aegithalidae Long-tailed Tit Aegithalos caudatus 2.0 3.0 Paridae Willow Tit Parus montanus + 1.5 Coal Tit Parus ater 1.0 1S Varied Tit Parus varius 1.5 DES Great Tit Parus major 8.5 WS Zosteropidae Japanese White-eye Zosterops japonicus 1.5 3.0 Emberizidae Siberian Meadow Bunting Emberiza cioides 0.5 1.0 Fringillidae Japanese Grosbeak Eophona personata 1.0 0.5 Ploceidae Tree Sparrow Passer montanus Ste Corvidae Jay Garrulus glandarius 1.0 1.0 Jungle Crow Corvus macrorhynchos + ar Timaliidae Red-billed Leiothrix Leiothrix lutea* 28.5 33.5 Total number of territories WS 85.0 Number of species that occupied territories 23 27 +: presence in the plot without occupying territories. * introduced species. kizuki, Brown-eared Bulbul Hypsipetes amaurotis, Short-tailed Bush Warbler Urosphena squameiceps, Long-tailed Tit Aegithalos caudatus, Varied Tit Parus varius and Japanese White-eye. The Siberian Meadow Bunting, an edge species, had territories near Miyukigahara. Tree Sparrows Passer montanus, which live almost exclusively in residential areas in Japan and are seen in Miyukigahara year round, were also recorded in 1995. 2) Density and fates of nests Six complete nests were found in a 1 ha area of the plot in 1995 (Fig. 2). Of those, four nests contained clutches and one nest contained a broken eggshell. No egg was observed in the remaining one nest, pos- sibly predated soon after laying. Other than the 6 Breeding density of exotic Red-billed Leiothrix and native bird species Fig. 2. Distribution of leiothrix territories and nests in 1995. The square shows the | ha intensive area for nest survey. Closed circles show complete nests and an open circle shows an incomplete nest. Arrows indicate re-nest- ing by same males after nest predation. A Incomplete nest 1 Complete nest NS Egg laying Wi Incubating {i Nestling Fledged Predated Predated (broken eggshell remained) No egg was observed “Uy Mf Z 1|>?>y Yy Nest was not completed Yfrmsw0@( YYHI_ 5 10 15 20 25 30 4 5 10 15 = MAY JUNE Fig. 3. Stages and fates of leiothrix nests in the | ha area in the plot. Numbers of each bar indicate nests in Fig. 2. Numbers in laying periods in bars show clutch size of each nests. Both two replacement nests were outside the | ha area. complete nests, an incomplete nest was found in the | ha area. We felt that it were difficult for us to locate all the nests in the dwarf bamboo bushes even within 1 ha area. All males associated these territories were color banded and two of them re-nested after predation of their nests, though females were not identified in both cases. Two out of the 6 complete nests in the | ha area are supposed to have fledged young, and two ringed young from one nest were recaptured within the breeding season (Fig. 3). Of the four remaining nests, at least 3 nests had predated eggs. Although leiothrix nest from April to September on Mt. Tsukuba (pers. obs.), the six complete nests H. TOJO and S. NAKAMURA und here were most likely to be first nesting at- tempts. Therefore, their density would represent min- imum breeding density in early breeding season, at least 600 pairs per 100 ha in this case, though sample area is small. 3) Comparison with Haga’s list Our list of birds that were found in the plot lacks Common Cuckoo Cuculus canorus and Great-spotted Woodpecker Dendrocopos major, which were listed in Haga’s list. The Great-spotted Woodpecker is rarely seen on Mt. Tsukuba, but in much smaller numbers than Japanese Green Woodpecker Picus awokera, another medium-sized woodpecker. The Common Cuckoo is basically not a bird of forest, al- though it is rarely heard around Miyukigahara. We recorded three tropical migrant species, Hors- field’s hawk Cuckoo Cuculus fugax, Siberian Thrush Turdus sibiricus, Pale-legged Warbler Phylloscopus borealoides that Haga’s list lacks. While latter two species were transients, the Horsfield’s hawk Cuckoo, which parasitizes on flycatchers and robins, occupied a territory in 1995. 4) Comparison with other Japanese deciduous forests The average number of breeding native bird species in the six forests compared was 26.8 (20- 32.5, Fig 4 (1)), similar to that on Mt. Tsukuba (24). Mt. Tsukuba lacked Great-spotted Woodpecker and Nuthatch Sitta europaea, which bred in all six forests. From Haga’s list, it is clear that the Nuthatch was ab- sent on Mt. Tsukuba even before leiothrix invasion. The Japanese Green Pigeon Sphenurus sieboldii, White’s Thrush Zoothera dauma and Brown Fly- catcher Muscicapa dauurica, which bred in four out of the six forests, did not occupy territories on Mt. Tsukuba, although were present in the plot at least in one study years. In contrast, Japanese White-eye, a common breeder on Mt. Tsukuba, did not occupy ter- ritories in the six forests compared. The breeding density of native birds of Mt. Tsukuba (616 pairs per 100 ha) is in the range of the six previous studies (370-620 pairs, average 505 pairs, Fig. 4 (2)). Including leiothrix, total breeding bird density on Mt. Tsukuba reached around 1000 pairs per 100 ha. The presence of leiothrix, thus, does not seem to depress breeding density of native birds. Total biomass is markedly affected by presence of some large species such as Copper Pheasant Syrmati- which weigh about 1000g. cus soemmerringil, (1) wW >) Number of species N is) (2) i Red-billed Leiothrix L) Native Species _ S S i=) Density (pairs/100ha) (3) Biomass (kg/100ha) Mt. Tsukuba a Fig. 4. 1) Number of species 2) total density 3) total bio- mass of breeding birds on Mt. Tsukuba and in six Japanese natural deciduous forests. See Figure | for abbreviations. Bars show SD for data collected in two years. Again, the biomass of leiothrix (17.5kg per 100 ha) does not seem to depress that of native birds of Mt. Tsukuba (42.3kg), which is in the range of the six previous studies (21.5-43.5kg, average 34.1 kg, Fig. 4 (3)). DISCUSSION The territory mapping showed that leiothrix breed in Mt. Tsukuba at a high density, 350-400 pairs per 100 ha, which is comparable to the total breeding bird density reported in some Japanese deciduous forests (e.g. Suzuki et al. 1983; Kobayashi & Fujimaki 1985; Inoue & Kondo 1983). Such a high breeding density is uncommon for Japanese forest bird species, and the sole species comparable to this within the six studies Breeding density of exotic Red-billed Leiothrix and native bird species compared is Coal Tit Parus ater in Kayanodaira (Nakamura 1986), which breed at 211 pairs per 100 ha, or at an even much higher density (Nakamura et al. 1987). The nest survey ascertained the high density of leiothrix and even suggested that it may be underestimated. For leiothrix that nest on dwarf bam- boo bushes, nest site is less likely to be a limiting fac- tor of breeding density as in many hole nesters (Amano & Eguchi 2002). It has no doubt that leiothrix have undergone a population explosion peculiar to introduced species since the first record in 1980 on Mt. Tsukuba, al- though details of the increase are unknown. Haga (1988) wrote that this species was seen in rather large flocks in winter. When we mist-netted in July and September 1990, leiothrix was already the most dom- inant species. As there were no nearby large leiothrix breeding areas from which mass immigration can occur to Mt. Tsukuba, this rapid increase was caused by successful breeding of the present population. In spite of the high breeding density of leiothrix, the native avifauna does not seem to be altered. Species that have disappeared in the decade are mini- mal, if any. The absence, or very low density, of two trunk foragers, Great-spotted Woodpecker and Nuthatch characterizes the avifauna in beech forest on Mt. Tsukuba, but there is little possibility that they have been excluded by leiothrix. Supposed native competitors on Mt. Tsukuba, such as Bush Warbler and Great Tit bred in higher densities than in any of leiothrix-free forests compared in this study. Both total breeding density and total biomass of native species on Mt. Tsukuba are in the range of those in the six compared forests and, therefore, do not seem to be depressed by those of leiothrix. All-in-all, we did not find any sign of leiothrix im- pact on native bird species. The total avifauna of Mt. Tsukuba seems to be a simple sum of native species and a huge amount of leiothrix. Therefore, we sug- gest that leiothrix have invaded Mt. Tsukuba without severe competition with native bird species. Eguchi and Masuda (1994) found similar results in Kyushu and suggest that leiothrix may have invaded into a vacant niche, a lower layer in the deciduous forests. However, it would be premature to conclude that leiothrix is harmless to native birds. Interspecific competition may have an effect only during infre- quent lean periods, as Mountainspring and Scott (1985) suggest. On the other hand, even if leiothrix have lowered breeding success of certain native bird species, breeding density of those species may change little, because enough recruit may come from surrounding “leiothrix-free” breeding areas for the species. In such cases, however, the effects on breed- ing density may become apparent when expanding leiothrix breeding area reached a threshold that re- maining leiothrix-free breeding areas of the native species no longer produce enough recruits. ACKNOWLEDGMENTS We thank Kazuhiro Eguchi for inviting us to write this man- uscript. Our research was partially supported by the Ministry of Environment, Japan (as Global Environmental Research Fund, project leader: K. Goka). REFERENCES Ali S (1977) Field Guide to the birds of the eastern Hi- malayas. Oxford University Press, London. Amano HE & Eguchi K (2002) Nest-site selection of the Red-billed Leiothrix and Japanese Bush Warbler in Japan. Ornithol Sci 1: 101-110. Berger AJ (1981) Hawaiian Birdlife. 2nd ed. University Press of Hawaii, Honolulu. Bibby CJ, Burgess ND & Hill DA (1992) Bird census techniques. Academic Press, London. Case TD (1996) Global patterns in the establishment and distribution of exotic birds. Biol Conserv 78: 69— 96. Diamond JM & Veitch CR (1981) Extinctions and intro- ductions in the New Zealand avifauna: cause and ef- fect? Science 211: 499-501. Dunning Jr. JB (1993) CRC handbook of avian body masses. CRC Press, Boca Raton. Ecological Society of Japan (2002) Handbook of alien species in Japan. Chijinshokan, Tokyo. Eguchi K & Masuda T (1994) A report on the habitats of Peking Robin Leiothrix lutea in Kyushu. Jpn J Or- nithol 43: 91-100 (in Japanese with English sum- mary). Eguchi K & Amano HE (2000) Problems of introduc- tion and naturalizations in birds. Jpn J Conserv Ecol 5: 131-148 (in Japanese). Fujimaki Y (1988) Breeding bird community of a Quer- cus mongolica forest in eastern Hokkaido, Japan. Jpn J Ornithol 37: 69-75 (in Japanese with English sum- mary). Fujimaki Y (1986) Breeding bird community in a decid- uous broad-leaved forest in southern Hokkaido, Japan. Jpn J Ornithol 35: 15-23 (in Japanese with English summary). Fukamachi K, lida S & Nakashizuka T (1996) Land- scape patterns and plant species diversity of forest re- H. TOJO and S. erves in the Kanto region Japan. Vegetatio 124: 107— 114. Haga K (1988) Faunistic notes on Mt. Tsukuba. Nihon no seibutsu 2(7): 35-41 (in Japanese). Hino T (1990) Palaearctic deciduous forests and their bird communities: Comparisons between East Asia and West-Central Europe. In: Keast A (ed) Biogeog- raphy and ecology of forest bird communities. pp 87-94, SPB Academic Pub. Hague. Ibaraki Prefectural Forest Experiment Station (1980) Map of Actual Vegetation of Ibaraki Prefecture (in Japanese). Inoue M & Kondo Y (1983) A comparison of breeding bird communities in coniferous and broad-leaved woods. Hard-Wood Research 2: 55—72 (in Japanese with English summary). Ishii S (1992) Tsukubasan no sousichou (Red-billed Leiothrix on Mt. Tsukuba). In: Tsukuba Science City Nature Club (ed) 7sukubasan: p 101. STEP, Tsukuba (in Japanese). Kiyosu Y (1966) Encyclopedia of wild birds. Tokyo-do Shuppan, Tokyo (in Japanese). Kiyosu Y (1978) The birds of Japan, revised and en- larged edition. Kodansha, Tokyo (in Japanese). Kobayashi S & Fujimaki Y (1985) Breeding bird com- munity in a deciduous broad-leaved wood and a larch plantation. Tori 34: 57-63 (in Japanese with English summary). Long JL (1981) Introduced Birds of the World. David & Charles, London. MacKinnon J & Phillipps K (2000) A field guide to the Birds of China. Oxford University Press, Oxford. Mitsuishi H (1970) Ecological study of the bird commu- nity in Okususobana Natural Garden, Nagano Prefec- ture. Bull Inst Nat Educ Shiga Heights Shinshu Univ 30 NAKAMURA 9: 1229-1238 (in Japanese with English summary). Mountainspring S & Scott M (1985) Interspecific com- petition among Hawaiian forest birds. Ecol Monogr 55: 219-239. Nakamura H (1986) Ecological studies of the beech for- est bird community on Kayanodaira Heights. Bull Inst Nat Educ Shiga Heights Shinshu Univ 23: 9-20 (in Japanese with English summary). Nakamura H, Murayama K, Kubokawa A, Suzuki R, Takizawa T & Shigemori K (1987) Bird community in the Kayanodaira scientific reserved beech forest in breeding season. Bull Inst Nat Educ Shiga Heights Shinshu Univ 24: 33-41 (in Japanese with English summary). Nakamura T (1983) On the breeding bird community structure in the mature beech forest, Kayanodaira. In: Taiga N (ed) Gendai seitaigaku no danmen, pp 301— 307. Kyouritu Shuppan, Tokyo (in Japanese). Ralph CJ, Fancy SG & Male TD (1998) Demography of an introduced Red-billed Leiothrix population in Hawaii. Condor 100: 468-473. Suzuki T, Saito S & Saito M (1983) Bird populations during the breeding season in a deciduous broad- leaved forest at Iwamizawa, central Hokkaido. Bull Hokkaido For Exp Stn 21: 95-103 (in Japanese with English summary). Tojo H (1994) Population increase of the Red-billed Leiothrix Leiothrix lutea in the Massif Tsukuba. Jpn J Ornithol 43: 39-42 (in Japanese with English sum- mary). Uramoto M (1961) Ecological study of the bird commu- nity of the broad-leaved deciduous forest of central Japan. J Yamashina Inst Ornithol 3: 1-32. Williamson K (1964) Bird census work in woodland. Bird Study 11: 1-22. Breeding density of exotic Red-billed Leiothrix and native bird species Appendix. Avifaunal list on Mt. Tsukuba prepared by Haga (1988) Compared with present list Excluded from comparison Family Common name Scientific name * Common name Scientific name Ardeidae c Japanese Night Heron Gorsakius goisagi Accipitridae a Black Kite Milvus migrans a Sparrowhawk Accipiter nisus a Grey-faced Buzzard-eagle Butastur indicus Falconidae a Kestrel Falco tinnunculus Phasianidae Chinese Bamboo Partridge Bambusicola thoracica b Pheasant Phasianus colchicus Copper Pheasant Syrmaticus soemmerringii Columbidae Rufous Turtle Dove Streptopelia orientalis Japanese Green Pigeon Sphenurus sieboldii Cuculidae Common Cuckoo Cuculus canorus Oriental Cuckoo Cuculus saturatus Little Cuckoo Cuculus poliocephalus Strigidae ce Collared Scops Owl Otus lempiji c¢ Brown Hawk Owl Ninox scutulata c Ural Owl Strix urarensis Caprimulgidae c Jungle Nightjar Caprimulgus indicus Apodidae d White-rumped Swift Apus pacificus Alucedinidae e Kingfisher Alcedo atthis Picidae Japanese Green Woodpecker Picus awokera Great Spotted Woodpecker Dendrocopos major Japanese Pygmy Woodpecker Dendrocopos kizuki Alaudidae b Skylark Alauda arvensis Hirundidae d House Swallow Hirundo rustica d Red-rumped Swallow Hirundo daurica Motacillidae e Grey Wagtail Motacilla cinerea e White Wagtail Motacilla alba e Japanese Wagtail Motacilla grandis g Olive-backed Pipit Anthus hodgsoni f Water Pipit Anthus spinoletta Pycnonotidae Brown-eared Bulbul Hypsipetes amaurotis Laniidae Bull-headed Shrike Lanius bucephalus Troglodytidae Winter Wren Troglodytes troglodytes Prunellidae g Alpine Accentor Prunella collaris g Japanese Accentor Prunella rubida Turdidae Japanese Robin Erithacus akahige g Red-flanked Bushrobin Tarsiger cyanurus Siberian Blue Robin Luscinia cyane f Daurian Redstart Phoenicurus auroreus White’s Thrush Zoothera dauma g Brown Thrush Turdus chrysolaus Grey Thrush Turdus cardis f Pale Thrush Turdus pallidus f Dusky Thrush Turdus naumanni Sylviidae Short-tailed Bush Warbler Urosphena squameiceps _ g Goldcrest Regulus regulus Bush Warbler Cettia diphone Arctic Warbler Phylloscopus borealis Crowned Willow Warbler — Phylloscopus coronatus Muscicapidae Narcissus Flycatcher Ficedula narcissina Blue and White Flycatcher Cyanoptila cyanomelana Brown Flycatcher Muscicapa dauurica Aegithalidae Long-tailed Tit Aegithalos caudatus Paridae Willow Tit Parus montanus Coal Tit Parus ater Varied Tit Parus varius Great Tit Parus major 31 idix. (Continued) H. TOJO and S. NAKAMURA Compared with present list Family Common name Scientific name Zosteropidae Emberizidae Japanese White-eye Siberian Meadow Bunting Fringillidae Japanese Grosbeak Ploceidae Tree Sparrow Sturnidae Corvidae Jay Jungle Crow Zosterops japonicus Emberiza cioides Eophona personata Passer montanus Garrulus glandarius Corvus macrorhynchos Total number of species Excluded from comparison * Common name f Rustic Bunting g Black-faced Bunting g Grey Bunting f Brambling b Oriental Greenfinch f Siskin f Rosy Finch f Long-tailed Rosefinch g Bullfinch f Hawfinch b Grey Starling b Azure-winged Magpie b Carrion Crow 39 Scientific name Emberiza rustica Emberiza spodocephala Emberiza variabilis Fringilla montifringilla Carduelis sinica Carduelis spinus Leucosticte arctoa Uragus sibiricus Pyrrhula pyrrhula Coccothraustes coccothraustes Sturnus cineraceus Cyanopica cyana Corvus corone 4] * reasons for exclusions: a: raptors, b: farmland species, c: nocturnal species, d: aerial species, e: aquatic species, f: not breed in central Japan, g: breed in higher altituidal areas. 32 SPECIAL FEATURE Ornithol Sci 3: 33-42 (2004) Invasive bird species Australasian bird invasions: accidents of history? Barry W. BROOK* Key Centre for Tropical Wildlife Management, Charles Darwin University, Darwin 0909, Northern Territory, Australia ORNITHOLOGICAL SCIENCE © The Omnithological Society of Japan 2004 Abstract Exotic bird introductions to Australia, New Zealand and surrounding is- lands, have been aggregated into one of the best documented and most completely analysed datasets available on biological invasions. Of the >242 species introduced by Europeans to Australasia during the 18th—20th centuries, at least 32% established long-term viable populations. A review of the literature reveals the most robust pre- dictors of introduction success to be total number of individuals liberated, and the number of separate attempts at introduction. Using generalized linear modelling on a combined regional dataset, I confirm this result, and demonstrate that together these two characteristics of historical introductions correctly explains the observed outcome in 89.3% of cases in Australasia. Further, I show that a simple stochastic population dynamics model, derived for a sub-set of 44 species from entirely independent long- term studies, is also able to achieve a high degree of predictive success (83%). Fi- nally, a suite of meta-analyses have shown the strongest life history and environmen- tal correlates of introduction success to be large body size, low propensity to migrate, climatically matched habitats across the native and invasive geographical range, sexu- ally monochromatic plumage, dietary generalism, and greater behavioural flexibility. The collective results of these analyses on Australasian introductions provide a poten- tially powerful framework for predicting the probable outcomes of future bird inva- sions worldwide. Key words Zealand The historical introduction of exotic birds to Aus- tralasia (Australia, New Zealand, New Guinea, and surrounding islands) represents what is almost cer- tainly the most completely documented (Thompson 1922; Long 1981; Newsome & Noble 1986; Lever 1987; Blackburn & Duncan 2001a) and thoroughly analysed (e.g. Diamond & Veitch 1981; Veltman et al. 1996; Duncan 1997; Green 1997; Sorci et al. 1998; Duncan et al. 1999; Legendre et al. 1999; Sol & Lefebvre 2000; Cassey 2001b; Duncan et al. 2001; Forsyth & Duncan 2001; Moulton et al. 2001; Dun- can & Blackburn 2002) source of empirical informa- tion and natural experiments on invasibility available anywhere in the ecological literature (Kolar & Lodge 2001; Duncan et al. 2003). Bird introductions to the Australasian realm were associated closely with the expansion of European settlement in the 19th century and the formation of ‘acclimatization societies’ (dat- (Received 29 August 2003; Accepted 1 November 2003) * Email: barry.brook@cdu.edu.au 33 Australia, Biological invasions, Bird introductions, Life history, New ing from the 1860s), whose primary goal was to es- tablish a variety of “beneficial or desirable species” in the newly settled lands (Thompson 1922). The main reasons for introducing birds into Australasia are summarized by Long (1981): (i) aesthetics (e.g. European birds were considered superior songsters to native species), (11) for food, hunting and sport (usu- ally game birds such as pheasants and ducks), (ili) as biological pest control agents (especially sparrows, starlings and mynas), and (iv) accidentally, as es- capees from cages or transport vessels. Invasions by exotic species present a major risk to biodiversity at a global scale (Vitousek et al. 1997; Duncan et al. 2003), yet, paradoxically, commonly increase the biodiversity of ecological communities at local scales (Lodge 1993; Sax et al. 2002). For in- stance, New Zealand’s avifauna has suffered substan- tial extinctions in the past, but thanks to the second highest number of attempted bird introductions on any landmass (after Hawaii: Long 1981), now sup- ports a more greater diversity of species (albeit, less B. W. BROOK distinctive) than at any other time in its past (Wilson 1997; Cassey 2001a; Sax et al. 2002). \bout 60% of the bird species introduced to Aus- tralia and New Zealand originated from the Palaearc- tic and Australasian biogeographic realms (Duncan et al. 2003). Globally, introductions of birds have been highly non-random with respect to both taxonomic composition and the locations of sources and destina- tions (Blackburn & Duncan 2001a; Cassey 2002). Of the approximately 9672 extant bird species, about 213 (2%) have been transported by humans and es- tablished successfully outside their natural geo- graphic ranges (Cassey 2002), although the number of failed attempts is considerably higher (perhaps as many as 90% for all invasive species: Williamson 1996, and at least 50% for birds; Blackburn & Dun- can 2001b). Fewer still of the established species have become widespread and abundant in their natu- ralized ranges (Duncan et al. 2003). The taxonomic bias in introductions is obvious: over two thirds of worldwide avian invasives come from just 6 of the 145 bird families (i.e. Anatidae, Columbidae, Fringillidae, Passeridae, Phasianidae, and Psittacidae: Blackburn & Duncan 2001a). Australasia ranks 8 out of 11 geographical regions in terms of ease of invasi- bility (Blackburn & Duncan 2001b), but nevertheless represents approximately 20% of the total docu- mented count of global bird introductions (Blackburn & Duncan 2001a). Due to the often highly negative impacts of intro- duced species on the environment, we need to have better predictability of invasion success. Recognised stages of invasion are transport, introduction, estab- lishment, a lag period during which abundances re- main low and distribution restricted, spread, and sub- sequent ecological and human impacts (Kolar & Lodge 2001; Sakai et al. 2001). A key requirement for developing a robust predictive framework from which to understand these processes and potential to intervene to inhibit invasion lies in the identification and ranking of factors which, under a range of cir- cumstances, predispose particular species or groups to being successful (or failed) invaders of new lands (Williamson 1996). These factors can be broadly cat- egorized as acting at the level of the species (e.g. life history traits), location (e.g. environment at the intro- duction site) and event (e.g. number of individuals re- leased) (Blackburn & Duncan 2001b; Duncan et al. 2003). Detailed empirical data are available for these factors on the bird introductions of Australasia, thereby providing an ideal case-study of natural ex- 34 periments from which to develop and test the theory of invasion ecology (Kolar & Lodge 2001). MAJOR DETERMINANT OF INVASION SUCCESS—INTRODUCTION EFFORT The most consistent and powerful predictor of in- vasion success for birds in Australasia has been shown repeatedly to be introduction effort by hu- mans—both in terms of the number of individual birds released (often termed ‘propagule pressure’), and the number of separate attempts at release (see Newsome & Noble 1986; Veltman et al. 1996; Dun- can 1997; Green 1997; Duncan et al. 1999; Black- burn & Duncan 2001a; Duncan et al. 2001; Duncan et al. 2003). Similarly strong relationships have been found when examining large mammal and insect in- vasions (see Daehler & Strong 1993; Forsyth & Dun- can 2001 and references cited therein). There are a number of obvious explanations for this pattern. Stochastic hazards associated with very small population size, such as demographic stochas- ticity, genetic deterioration, and other “Allee effects” which act to reduce reproductive success at low den- sities (Shaffer 1981), are offset as founder population size increases (Pimm et al. 1993; Ryan & Siegfried 1994; Green 1997; Legendre et al. 1999; Forsyth & Duncan 2001). Multiple introductions of a given species may provide a demographic ‘rescue effect’, and act to supplement genetic diversity, thereby cir- cumventing problems like inbreeding depression and loss of adaptive fitness associated with population size bottlenecks (Sakai et al. 2001). In addition, some species (e.g. many game birds), had high failure rates because of human hunting pressure (Long 1981; Duncan et al. 2001), poisoning, non-human preda- tion, and lack of suitable habitat (Duncan & Black- burn 2002), which never permitted them to establish large enough resident populations to remain viable over the long-term. The minimum adequate statistical model devel- oped by Green (1997) to explain establishment suc- cess for a sub-set of New Zealand land birds included only propagule size, whilst Cassey’s (2001b) results supported the additional inclusion of number of re- lease events, and geographical range size of a species in their native realm (the latter factors being associ- ated with a higher likelihood of finding matching habitats in newly invaded regions). Precisely this same, three-factor model, was arrived at by Duncan et al. (2001) using independent data for Australian Australasian bird invasions bird introductions, after having controlled for phy- logeny. A lack of migratory tendency was substituted for range size for New Zealand passeriforms (Duncan 1997), and likewise for all New Zealand birds by Veltman et al. (1996). The latter’s model explained 97.4% of the variance in introduction success for the 79 species they examined. Supporting these similari- ties, Sol (2000) found no difference in establishment success for birds invading continental Australia ver- sus New Zealand. To better understand the degree to which individ- ual components of introduction effort could explain establishment success of exotic birds across the entire scope of Australasia, I used generalized linear model- ling (logistic regression), supported by information- theoretic model selection, to re-analyse the combined datasets of Newsome & Noble (1986), Veltman et al. (1996), and Duncan et al. (2001), yielding N=131. My modelling framework considered the total num- ber of individuals released [NV], number of release events [E], and an interaction term [/], calculated as NXE. Individual species were classed as successful (code=1) or unsuccessful (code=0). For this simple analysis, I deemed it unnecessary to control for phy- logenetically-related statistical non-independence (Harvey & Pagel 1991), because compared to life his- tory/ecological correlates (see next section), introduc- tion effort by humans is likely to be only weakly bi- ased taxonomically (Blackburn & Duncan 2001a; Cassey 2002). The Akaike Information Criterion, corrected for small sample size (AIC..) was used as an objective means of regression model selection, based on considerations of both predictive power and parsi- mony (detailed in Burnham & Anderson 2002). The most parsimonious model for predicting inva- sion success included only number of release events [E], G=65.6, df=1, P<0.0001, Concordance= 86.9% (see Table | for complete model-selection sta- tistics). However, the fully-specified model also re- ceived considerable support from the data (ranked 3/7 overall), G=67.7, df=3, P<0.0001, and explained more of the variance (C=89.3%), the equation being: Invasion success (P)=1/[1+ l/exp(1.27 Xlog,)[N] +6.73 Xlog )[E ]—1.36Xlog,)[NXE]—5.34)] (1) A goodness of fit test on equation (1) revealed no significant deviations from the assumptions of a lo- gistic generalized linear model (y7=102.1, df=97, P=0.342), and no _ over-dispersion of variance 35 Table 1. Model selection results, relating likelihood of suc- cessful invasion by exotic birds in Australasia (i.e. percentage of introduced species which establish viable populations in Australia and/or New Zealand) to the log-transformed total number of individuals released [N], the log-transformed num- ber of release events [£], and an interaction term [/], calcu- lated as NXE. Model selection criteria are the maximized log- likelihood [log(Z)], number of parameters (K; includes regres- sion intercept and coefficients), information criterion (AIC.,), difference from best model (A,), and Akaike weight (w,). Model log(L) K AIC, A, w, E 5) sil3) 2 108.3 0.0 0.444 N+E = le 79) 3 109.8 1.4 0.219 NDE iSO 4 110.4 2.1 0.159 E+I yal 3 110.4 2.1 0.156 if —55.48 wy) 115.1 6.7 0.016 N+I —55.44 3 117.1 8.7 0.006 N —60.25 2 124.6 16.3 0.000 x me a b. x ie ae) 2 50 5 z 25 0 0.75 1.50 2.25 3.00 3.75 03 06 O09 12 £415 log i) (Propagule size) log 9 (No. release events ) Fig. 1. Relationship between likelihood of successful inva- sion by exotic birds in Australasia (i.e. percentage of intro- duced species which establish viable populations in Australia [n=52] and New Zealand [n=79]) to a. total number of indi- viduals (“propagules’) released and b. number of release (in- troduction) events. Fitted lines were derived using logistic re- gression (models ‘N’ and ‘E”’ of Table 1, respectively). (¢=1.05). Thus (unsurprisingly), the strong explana- tory power of introduction success holds when taken across the region, but interestingly, the number of re- lease events, not the propagule pressure, was identi- fied by information-theoretic model selection as being the single most important component of inva- sion success. EVOLUTIONARY, LIFE HISTORY AND ENVIRONMENTAL CORRELATES After factoring out the very strong influence of in- troduction effort on establishment success (1.e. the ‘accidents of history’), an obvious question arises: B. W. BROOK ny worthwhile explanatory power remain for ‘fe history and environmental attributes of invasive bird species, and if so, which ones? Whilst compar- isons among taxa cannot in themselves determine causation, such correlations can at least provide in- sight into causes, and highlight potential evolutionary influences that are related to variability in introduc- tion success (Cassey 2002). Kolar and Lodge (2001) found a remarkably high number of potential corre- lates of invasion success (68) had been evaluated across all species in the literature they reviewed. Those traits associated with increased introduction success tended to be related (directly or indirectly) to increases in adult survival and high rates of reproduc- tion, the ability to tolerate or thrive in human modi- fied landscapes, and previous success at invasion (see also Diamond & Veitch 1981; Case 1996; Williamson 1996). Below I present a brief overview of the meta- analyses on ecological correlates which are of most relevance to bird introductions in Australasia. The major correlates examined were geographical range size, migratory tendencies, habitat and dietary gener- alism, competitive ability, behavioural flexibility, and surrogates for population dynamics, such as body mass and clutch sizes. Blackburn & Duncan (2001a) showed that species chosen for introduction tended to be abundant in their native range, were hence presumably easy to obtain, and were therefore collected and released in greater numbers. Certain groups in particular, such as wild- fowl and other game birds, were chosen preferen- tially for introduction. Blackburn and Duncan (2001a) and Duncan et al. (2001) were concerned that because species common in Europe also tend to pos- sess life history traits that distinguish them from less abundant species, species-level traits may be con- founded by event-level effects such as introduction effort. However, recent work by Cassey (2002) showed no correlation between habitat generalism and geographic range, implying there is little support for the idea that generalist species enjoy high success rates simply because they are more abundant. Australia’s exotic birds show a high proportion of ground nesters which prefer grazing, cultivated, and urban landscapes, and feed predominantly upon seeds and fruit (Newsome & Noble 1986). Conversely, Duncan et al. (2001) found no evidence to support the hypothesis that species with faster reproductive traits, non-migratory, or gregarious behaviour, in- creased introduction success in Australia. The latter results may be rationalized when one considers that 36 those traits associated with increased introduction success may be quite different from those which pro- mote successful transportation (Kolar & Lodge 2001), or long-term persistence subsequent to estab- lishment (Duncan et al. 2003). For example, r-se- lected traits (e.g. rapid development, high reproduc- tive potential, short generation intervals, and often weak intra-specific competition) may favour rapid expansion from small founder size (Sakai et al. 2001), but their often concomitantly higher inherent natural population variability also leave such species more prone to local extinction (Pimm et al. 1993). These confounding attributes can result in often con- fusing or apparently contradictory results when eval- uating correlates of introduction success, especially if one relies on studies of fully naturalized birds (Cassey 2001b; Duncan et al. 2001). Introduction success in New Zealand was lower for species with dichromatic plumage and more general- ist diets. The former result may be due to a reduction in ‘invasion fitness’ caused by higher levels of intra- specific sexual selection pressures (McLain et al. 1999), making such taxa more vulnerable to environ- mental fluctuations and inter-specific competition, more visible to predators, and reducing their geneti- cally effective population size (Sorci et al. 1998). Be- havioural flexibility, such as the ability to rapidly ad- just cognitive processes to suit new environments (measured by relative brain size, after controlling for allometry and phylogeny) is another possibility. Species with larger brains and a greater number of documented feeding innovations were shown by Sol and Lefebvre (2000) to be superior invaders. Ecologists commonly regard extrinsic forces, such as inter-specific competition, as an important deter- minant of community construction and stability. This idea motivated studies such as Moulton et al. (2001), who provided evidence that competition between alike species, measured by morphological over-dis- persion, might contribute to introduction success. This embraces the long-held view that species-rich communities are harder to invade than species-poor ones, because less realized niche-space is presumably available for exploitation (Elton 1958; Holway & Suarez 1999). Analyses examining the success of in- troduced passeriform birds to New Zealand (Duncan 1997), revealed a pattern which could be explained by either introduction effort or by competition, but Duncan also suggested that morphological over-dis- persion might more parsimoniously be explained by the propensity of acclimatization societies to put Australasian bird invasions greater effort into introducing bird species with a wide diversity of characteristics rather than concen- trating on only a few, similar taxa. In support of this conclusion, Duncan and Blackburn’s (2002) critique of Moulton et al. (2001) showed that most morpho- logically similar game bird species would, in most in- stances, never have had the opportunity to interact or compete with alike species during the early phase of establishment. Similarly, Case (1996) failed to iden- tify a global link between invasion success and species richness of the native avifauna. THE GLOBAL CONTEXT OF BIRD INVASIONS Australasian trends tend to be mirrored at the global level. In a comprehensive meta-analysis of worldwide bird introductions, Blackburn and Duncan (2001b) examined 1,466 introduction events for 398 bird species, and, after accounting for introduction ef- fort, showed the probability of invasion success was higher if only a small difference existed between the latitude of origin and latitude of introduction, when introductions were matched with suitable environ- ments, and when range and population size in native regions is comparatively large (see also Blackburn & Duncan 2001a; Duncan et al. 2001), but equally, un- covered no consistent relationship with biotic resist- ance (whereby species-rich areas such as the tropics were equally easy to invade as more species-poor, temperate and island areas), body mass, generation time, or population growth rate. Cassey (2002) found that increased habitat gener- alism, lack of migratory tendency, and sexual mono- chromatism together explain significant variation in the successful establishment of land birds worldwide. Further, increasing female body weight, and lack of migratory tendency, could together explain 76% of the variation in introduction success among families (see also Cassey 2001a). Parasites are often filtered out as a result of the population size bottlenecks asso- ciated with the invasion process (Sakai et al. 2001), and can lead to a demographic release, as apparently occurred for Common Starlings (Sturnus vulgaris) in- troduced into North America (Torchin et al. 2003). In Southeast Asia (the biogeographical division in clos- est proximity to Australasia), correlates of invasibil- ity included commensalism with humans, facultative colonial nesting, communal roosting, possession of a crop, and ability to congregate rapidly at ephemeral food sources (Lim et al. 2003; Yap & Sodhi 2004). 37 Phylogenetic effects accounted for an average of 2.1% of the variation in trait values within families for global bird introductions, and 4.7% across fami- lies (maximum of 24%), and are therefore surpris- ingly weak (Cassey 2002). INVASION SUCCESS AND POPULATION DYNAMICS Clearly, the probability of successful invasion of Australasia and elsewhere by exotic birds correlates with both introduction effort and certain life history attributes, and this knowledge can be combined to provide a robust phenomenological framework for predicting the probably outcome of future invasion attempts (Daehler & Strong 1993; Kolar & Lodge 2001; Cassey 2002). But on a more fundamental, mechanistic level, what is the population dynamical basis for such predictions? In an attempt to provide a preliminary answer to this question, I undertook the following analysis. I first extracted long-term population time-series data from the Global Population Dynamics Database (hereafter GPDD, see Inchausti & Halley 2001), from which sufficient information was available to charac- terize the population dynamics of 44 of the bird species introduced to Australia or New Zealand. For each species, I then fitted a simple stochastic model of exponential population growth via maximum like- lihood estimation: (2) where AN is the population size at time f, r is the in- trinsic rate of population increase, o, is the standard deviation of r, and z is a standard normal distributed random deviate (Barker & Sauer 1992; McCallum 2000; Morris & Doak 2002). A list of the species and associated parameter estimates are given in Table 2. Finally, a stochastic simulation representation of equation (2) was implemented in Microsoft® Excel, and used to estimate, for each species separately, the initial population size at which the likelihood of es- tablishment of a given species was >50%. This process is analogous to estimating a minimum viable population size in conservation biology (sensu Shaf- fer 1981; Ryan & Siegfried 1994; Reed et al. 2003), except that the aim in this case is to determine the number of founders required for a species to be more likely invade successfully than to fail. This population dynamics modelling approach pre- dicted successfully the observed outcome in 54 of 65 in iN...) 1-520, B. W. BROOK la 2. Predictions of invasion success for exotic birds in Australia and New Zealand, using a simple population dynamics nodel of stochastic exponential growth. A sub-set of 44 species was used, for which long-term and independent population time- series data was available to parameterize the model. The actual outcome was predicted correctly by the model (indicated by a * in the columns labeled ‘I’) in 83% (54/65) of individual cases. BCS Australia New Zealand Scientific name qd r O,. N (model) | N E | N E Aix sponsa 14 —0.022 0.149 529 =~ _ — 0* 10 5 Alauda arvensis 47 0.045 0.331 23 1 526 13 les 39] 1] Alectoris rufa 14. —0.037 0.177 2,478 0* Oly secre 0* 20 2 Anas acuta 48 —0.036 0.236 2,693 — — — 0* 102 3 Anas penelope 18 —0.070 0.798 846,754 — — — 0* 32 3) Anas platyrhynchos 48 —0.003 0.141 85 ils 86 4 1* 1,539 17 Anser anser 9 0.042 0.129 13 — — — 0* 7 2, Anser caerulescens 28 0.069 0.117 12 — —_ —_ 0* 10 ] Aythya fuligula 36 0.032 0.280 29 — —_ — 0* 5 2 Branta canadensis 21 0.000 0.168 51 0* 4 2 iIFS 60 10 Callipepla californica 23 0.024 0.600 146 = — — 1* 1,420 15 Carduelis cannabina 20 0.046 0.222 47 0* 32, 4 0 209 12 Carduelis carduelis 14 0.066 0.193 10 1* 223 5 1* 626 14 Carduelis chloris 16 —0.061 0.432 74,139 l 133 7 1 65 6 Carduelis flammea 20 0.106 1.374 503 = — — i 607 10 Carduelis spinus 20 0.012 1.081 6,297 0* 80 3 0* 54 3 Colinus virginianus 46 —0.013 0.400 507 — — — ib 1,156 17 Corvus frugilegus 13. —0.074 0.192 183,187 — — _ 1 182 10 Corvus monedula 15 0.214 0.218 4 —_ — = On 3 B Cygnus olor 116 0.006 0.077 50 l 12 5) 1 29 6 Emberiza citrinella 20 —0.043 0.490 12,941 0* 34 3 656 14 Emberiza schoeniclus 20 0.027 0.415 46 — — — 0* 9 2 Erithacus rubecula 29 0.001 0.188 170 Oe 47 3 0* 123 11 Fringilla coelebs 26 —0.029 0.509 3,500 0* 498 4 1 449 17 Fringilla montifringilla 20 —0.020 0.409 1,000 0* 78 0* 121 7 Lagopus lagopus 32 0.044 0.919 387 —= aa — 0* 4 2 Luscinia megahrynchos 34 0.010 0.246 49 0* 4 l 0* 7 4 Passer domesticus 26 0.029 0.726 215 It 414 20 Li 416 12 Perdix perdix 140 0.003 0.767 1,444 0* 4 2 0* 676 24 Phasianus colchicus 30 0.062 1.265 1,531 ] 750 12 1 244 Di Pluvialis squatarola 21 0.090 0.179 8 = = — 0* 3 2 Prunella modularis 30 —0.045 0.386 10,066 — — — l 245 14 Pyrrhula pyrrhula 26 0.029 0.540 62 0* 14 ] 0* 12 2 Streptopelia turtur 12. —0.058 0.674 146,902 0* 8 ] — — =e Struthio camelus 13 0.081 0.741 42 Lis 500 4 — _— a Sturnus vulgaris 4] 0.070 0.452 20 1M 292 9 IF 653 14 Sylvia atricapilla 26 —0.024 0.521 M335) —_— — — 0* 5 l Sylvia communis 20 0.001 0.147 66 -- = == 0* 2 Tetrao tetrix 40 0.029 0.453 49 -- — os 0* 13 js Turdus merula 26 0.018 0.327 46 Ne 102 8 ies 596 16 Turdus philomelos 26 0.067 0.532 28 INF? 129 8 1* 343 12 Tympanuchus phasianellus 20 0.137 0.732 29 — — — 0* 22 I Tyto alba 14. —0.031 0.410 DIAL — a 0* 7 l Vanellus vanellus 47 —0.040 0.544 IPE I)5 — — — 0* 124 8 Definitions: g=years of time-series data, r=intrinsic rate of increase, o,=standard deviation of r, [=invasion outcome (O=failure, | =success), N=total number of individuals introduced, E=number of release (introduction) events. Data sources: q, r, and o, are derived from the GPDD (http://cpbnts1.bio.ic.ac.uk/gpdd/); I, NV and E are given in from Long (1981), Veltman et al. (1996), Green (1997), and Duncan et al. (2001). 38 Australasian bird invasions cases involving these species (83%; see Table 2) - a considerably better result than the 50% success rate expected by chance (G=30.7, df=1, P<0.0001), de- spite there being a number of simplistic postulates as- sociated with equation (2), e.g. the lack of demo- graphic stochasticity or negative feedbacks, and the assumption of a globally consistent value for r and o,, across all environments of a species. The reason this semi-mechanistic model predicts invasion success so well is probably because: (i) a suite of relevant life history variables is effectively collapsed into its two basic parameters (Dennis et al. 1991; Morris & Doak 2002), and (ii) it represents such a fundamental and robust framework for population ecology, from which most other more complex models are derived (Turchin 2001). A good example in point is the case of the European Partridge (Perdix perdix), which failed to establish in New Zealand, despite enjoying at least 24 attempts at introduction with a total of 676 individuals being liberated. This failure vexed Velt- man et al. (1996), but is in fact predicted correctly by the population dynamics model, being very likely a product of the species’ high level of inherent popula- tion variability (Pimm et al. 1993). IMPACT AND MANAGEMENT OF EXOTIC BIRDS IN AUSTRALASIA Exotic birds now comprise a substantial element of Australasia’s avian biota, especially in urban, agricul- tural and other human-dominated landscapes (Cayley 1973; Kentish et al. 1995; Martin 1996; Pell & Tide- mann 1997a). For example, whilst a survey of resi- dent bird species in suburban Melbourne recorded only 9 of 43 species as being invasives, these few species were extremely abundant, representing over two-thirds (69%) of 2,856 sightings (Green 1984). Similarly, introduced Common Mynas (Acridotheres tristis) reach densities exceeding 120 birds km? in suburban Canberra, and are continuing to increase (Pell & Tidemann 1997a). The environmental, agricultural and concomitant economic impact of invasive birds is a substantial global problem—Pigeons (Columba livia) and Star- lings alone have been estimated to cost the United States economy over $US1,900 million annually in damage and control costs (Pimentel et al. 2000). Aus- tralasia faces comparable problems from naturalized exotic birds (Bomford & Sinclair 2002), including: damage to fruit crops (cherries, blueberries, grapes, olives) by Starlings, Blackbirds (Turdus merula), 3 House Sparrows (Passer domesticus), and Mynas; spreading of noxious weeds (Starlings and Black- birds); as vectors of human and livestock disease (Starlings and Sparrows); impacts on intensive cattle, pig and poultry production (which rely on high grain rations) by Starlings, Sparrows and Pigeons; defacing of buildings and as hazards to public amenities (by all of the abovementioned species); and perhaps of most concern, their often detrimental interactions with na- tive bird species (including direct and indirect com- petition for feeding and breeding resources, and ge- netic introgression) by Starlings, Mynas, Blackbirds, House Sparrows, Javan Sparrows (Lonchura oryzivora), Nutmeg Manikins (Lonchura punctulata) and Mallards (Anas platyrhynchos), and interference with, or predation upon, other taxa, such as arthro- pods, fish and mammals (for details, see Long 1981; Green 1984; Kentish et al. 1995; Martin 1996; Pell & Tidemann 1997b; Bomford & Sinclair 2002). Even given these realized and potential impacts of invasive bird species, it is usually argued that the eco- nomic and logistical resources devoted to their man- agement and control fall far short of being propor- tional to the risks involved (Pimentel et al. 2000; Bomford & Sinclair 2002). Successful, scientifically- based management, requires an inherently interdisci- plinary endeavour, involving ecology, economics, and mathematics. If combined effectively, an intellec- tually diverse approach can help to both determine realistic environmental goals, and evaluate the effec- tiveness of methods to arrive at these targets (Leung et al. 2002). For instance, the successful and on-going control of invasive House Crows (Corvus splendens) and Mynas (Acridotheres spp.) in Singapore owes its success to an integrated management programme in- volving governmental support, scientific monitoring, population and statistical modelling, direct on-ground actions such as shooting, and indirect activities such as habitat manipulation, and public education cam- paigns (see Yap et al. 2002; Brook et al. 2003; Lim et al. 2003). NEW FRONTIERS Without adequate management and control efforts, the problems associated with exotic birds are likely to increase in the future, because many invasive bird populations in continental Australia have probably not yet had sufficient time since colonization to reach an ecological equilibrium. For instance, the Common Myna, House Sparrow and Starling continue to ex- B. W. BROOK their distributional ranges within Australia Long 1981; Newsome & Noble 1986; Martin 1996), and are already beginning to invade ever further into the continent’s tropical regions (McCrie 2000). Eurasian Tree Sparrows (Passer montanus) have also entered the continent as far north as Darwin, but have to date been quickly identified and eradicated. This leaves the feral Pigeon as the only exotic bird cur- rently resident in northern Australia, although efforts are underway (e.g. shooting, restriction of homing pi- geon licences) to attempt to eliminate this species too (Chapman 2000). In New Zealand, however, the situation may be less worrisome. Duncan et al. (1999) have shown that the contemporary geographic range size of intro- duced birds does not depend on the length of time since they were introduced, but instead reflected pri- marily the extent and availability of preferred habitat (see also Diamond & Veitch 1981), suggesting that range expansion after initial founding events was rather rapid (within 50 years), and has reached equi- librium (carrying capacity) for most species. In other words, in this region at least, things have probably al- ready got as bad as they are ever likely to get, unless authorities and the populace relax their vigilance to permit deliberate or accidental introductions. CONCLUSION—ACCIDENTS OF HISTORY? A sythesis of historical, ecological, genetic, behav- ioural and evolutionary perspectives is essential for predicting the introduction success and subsequent spread of invasive species (Sakai et al. 2001)—a point exemplified by the wealth of analyses published on the introduction of exotic birds to Australasia. Al- though has been claimed that invasion success is largely unpredictable and case-specific (Lodge 1993; Williamson 1996), recent results (e.g. Veltman et al. 1996; Duncan et al. 2001) show that patterns of suc- cess and failure, during and after introduction, are in- deed highly predictable, provided key species- loca- tion- and event-level attributes are known. The fre- quent failure of life history traits to explain ade- quately establishment success in non-Australasian birds examples may simply reflect the universal and overriding importance—but generally severely con- strained knowledge of—introduction effort and envi- ronmental matching (Duncan et al. 2003). The post- release spread and proliferation of introduced species also appears to include a component of historical cir- cumstance (Duncan et al. 1999), whereby species 40 with larger contemporary range sizes were also those that benefited from the greatest introduction efforts, and were therefore able to exclude competitively later arrivals by sheer predominance in the landscape (Duncan et al. 2003)—truly ‘accidents of history’. The collective results of these quantitative, empiri- cally-grounded analyses on Australasian bird intro- ductions, provide a potentially powerful framework for predicting the likelihood of success of future avian invasions throughout the world, as well as mak- ing valuable contributions to development of envi- ronmental management policy. For example, the ex- isting insights regarding propagule size and fre- quency of release events are sufficient to inform guidelines for maximum individual holdings or total populations of exotic birds held in captivity. Research into establishment patterns of invasive birds also has the potential to make valuable contributions to pre- dictive ecology in general; for instance, quantifying the importance (and magnitude) of propagule size and number of events is likely to be informative for reintroduction programmes of threatened species (Green 1997), and may help advance our understand- ing of the ecological rules of community assembly (Kolar & Lodge 2001). ACKNOWLEDGMENTS I thank Lochran Traill and Guy Pardon for help with inter- rogating the GPPD database for relevant population dynamics data on the invasive bird species, Don Franklin for discus- sions, and Navjot Sodhi and Peter Whitehead and two anony- mous referees for comments on the manuscript. This work was undertaken by the Key Centre for Tropical Wildlife Manage- ment, Charles Darwin University, under funding from the Australian Research Council. REFERENCES Barker RJ & Sauer JR (1992) Modelling population change from time series data. In: McCullough DR & Barret RH (eds) Wildlife 2001: Populations. pp 182- 194. Elsevier Applied Science, London. Blackburn TM & Duncan RP (2001a) Establishment patterns of exotic birds are constrained by non-ran- dom patterns in introduction. J Biogeogr 28: 927-— 939: Blackburn TM & Duncan RP (2001b) Determinants of establishment success in introduced birds. Nature 414: 195-197. Bomford M & Sinclair R (2002) Australian research on bird pests: impact, management and future directions. Emu 102: 29-45. Australasian bird invasions Brook BW, Sodhi NS, Soh MCK & Lim HC (2003) Abundance and projected control of invasive house crows in Singapore. J Wild] Manage 67: 808-817. Burnham KP & Anderson DR (2002) Model selection and multimodel inference: a practical information- theoretic approach. Springer, New York. Case TJ (1996) Global patterns in the establishment and distribution of exotic birds. Biol Conserv 78: 69-96. Cassey P (2001a) Are there body size implications for the success of globally introduced land birds? Ecog- raphy 24: 413-420. Cassey P (2001b) Determining variation in the success of New Zealand land birds. Glob Ecol Biogeogr 10: 161-172. Cassey P (2002) Life history and ecology influences es- tablishment success of introduced land birds. Biol J Linn Soc 76: 465-480. Cayley NW (1973) What bird is that?: a guide to the birds of Australia. Angus & Robertson, Sydney. Chapman S (2000) Occurrence and eradication of a small population of the Eurasian tree sparrow Passer montanus in Darwin. North Territ Nat 16: 32-34. Daehler CC & Strong DR (1993) Prediction and biolog- ical invasions. Trends Ecol Evol 8: 380. Dennis B, Munholland PL & Scott JM (1991) Estima- tion of growth and extinction parameters for endan- gered species. Ecol Monogr 61: 115-143. Diamond JM & Veitch CR (1981) Extinctions and intro- ductions in the New Zealand avifauna: cause and ef- fect? Science 211: 499-501. Duncan RP (1997) The role of competition and intro- duction effort in the success of passeriform birds in- troduced to New Zealand. Am Nat 149: 903-915. Duncan RP, Blackburn TM & Veltman CJ (1999) Deter- minants of geographical range sizes: a test using in- troduced New Zealand birds. J Anim Ecol 68: 963- 975. Duncan RP, Bomford M, Forsyth DM & Conibear L (2001) High predictability in introduction outcomes and the geographical range size of introduced Aus- tralian birds: a role for climate. J Anim Ecol 70: 621-— 632. Duncan RP & Blackburn TM (2002) Morphological over-dispersion in game birds (Aves: Galliformes) successfully introduced to New Zealand was not caused by interspecific competition. Evol Ecol Res 4: 551-561. Duncan RP, Blackburn TM & Sol D (2003) The ecology of bird introductions. Annu Rev Ecol Syst 34: 71-98. Elton CS (1958) The ecology of invasions by animals and plants. Chapman & Hall, London. Forsyth DM & Duncan RP (2001) Propagule size and the relative success of exotic ungulate and bird intro- 4] ductions to New Zealand. Am Nat 157: 583-595. Green RE (1997) The influence of numbers released on the outcome of attempts to introduce exotic bird species to New Zealand. J Anim Ecol 66: 25-35. Green RJ (1984) Native and exotic birds in a suburban habitat. Aust Wildlife Res 11: 181-190. Harvey PH & Pagel MD (1991) The comparative method in evolutionary biology. Oxford University Press, London. Holway DA & Suarez AV (1999) Animal behaviour: an essential component of invasion biology. Trends Ecol Evol 14: 328-330. Inchausti P & Halley J (2001) Investigating long-term ecological variability using the global population dy- namics database. Science 293: 655-657. Kentish BJ, Dann P & Lowe KW (1995) Breeding biol- ogy of the common blackbird Turdus merula in Aus- tralia. Emu 95: 233-244. Kolar CS & Lodge DM (2001) Progress in invasion bi- ology: predicting invaders. Trends Ecol Evol 16: 199-204. Legendre S, Clobert J, Maller AP & Sorci G (1999) De- mographic stochasticity and social mating system in the process of extinction of small populations: the case of passerines introduced to New Zealand. Am Nat 153: 449-463. Leung B, Lodge DM, Finnoff D, Shogren JF, Lewis MA & Lamberti G (2002) An ounce of prevention or a pound of cure: bioeconomic risk analysis of invasive species. Proc R Soc Lond, B: 269: 2407-2413. Lever C (1987) Naturalized birds of the world. Long- man Scientific, New York. Lim HC, Sodhi NS, Brook BW & Soh MCK (2003) Un- desirable aliens: factors determining the distribution of three invasive bird species in Singapore. J Trop Ecol 19: 685-695. Lodge DM (1993) Biological invasions: lessons for ecology. Trends Ecol Evol 8: 133-137. Long JL (1981) Introduced birds of the world: the worldwide history, distribution, and influence of birds introduced to new environments. Universe Books, New York. Martin WK (1996) The current and potential distribu- tion of the common myna Acridotheres tristis in Aus- tralia. Emu 96: 166-173. McCallum H (2000) Population parameters: estimation for ecological models. Blackwell Science, Oxford. McCrie N (2000) A record of the common starling Stur- nus vulgaris in Darwin, Northern Territory. North Territ Nat 16: 26—27. McLain DK, Moulton MP & Sanderson JG (1999) Sex- ual selection and extinction: the fate of plumage-di- morphic and plumage-monomorphic birds introduced B. W. BROOK \to islands. Evol Ecol Res 1: 549-565. \Miorris WEF & Doak DF (2002) Quantitative conserva- tion biology: theory and practice of population via- bility analysis. Sinauer Associates, Sunderland. Moulton MP, Sanderson JG & Labisky RF (2001) Pat- terns of success in game bird (Aves: Galliformes) in- troductions to the Hawaiian islands and New Zealand. Evol Ecol Res 3: 507-519. Newsome AE & Noble IR (1986) Ecological and physi- ological characters of invading species. In: Groves RH & Burdon JJ (eds) Biological invasions: an Aus- tralian perspective. pp 1-20. Cambridge University Press, Cambridge. Pell AS & Tidemann CR (1997a) The ecology of the common myna in urban nature reserves in the Aus- tralian Capital Territory. Emu 97: 141-149. Pell AS & Tidemann CR (1997b) The impact of two ex- otic hollow-nesting birds on two native parrots in sa- vannah and woodland in eastern Australia. Biol Con- serv 79: 145-153. Pimentel D, Lach L, Zuniga R & Morrison D (2000) Environmental and economic costs of nonindigenous species in the United States. BioScience 50: 53-65. Pimm SL, Diamond J, Reed TM, Russell GJ & Verner J (1993) Times to extinction for small populations of large birds. Proc Natl Acad Sci USA 90: 10871-— 10875. Reed DH, O’Grady JJ, Brook BW, Ballou JD & Frankham R (2003) Estimates of minimum viable population sizes for vertebrates and factors influenc- ing those estimates. Biol Conserv (in press). Ryan PG & Siegfried WR (1994) The viability of small populations of birds: an empirical investigation of vulnerability. In: Remmert H (ed) Minimum animal populations. pp 3-22. Springer, Berlin. Sakai AK, Allendorf FW, Holt JS, Lodge DM, Molof- sky J, With KA, Baughman S, Cabin RJ, Cohen JE, Ellstrand NC, McCauley DE, O’Neil P, Parker IM, Thompson JN & Weller SG (2001) The population biology of invasive species. Annu Rev Ecol Syst 32: 305-332. Sax DF, Gaines SD & Brown JH (2002) Species inva- sions exceed extinctions on islands worldwide: a comparative study of plants and birds. Am Nat 160: 766-783. Shaffer ML (1981) Minimum population sizes for species conservation. BioScience 31: 131-134. Sol D (2000) Are islands more susceptible to be invaded than continents? Birds say no. Ecography 23: 687- 692. Sol D & Lefebvre L (2000) Behavioural flexibility pre- dicts invasion success in birds introduced to New Zealand. Oikos 90: 599-605. Sorci G, Moller AP & Clobert J (1998) Plumage dichro- matism of birds predicts introduction success in New Zealand. J Anim Ecol 67: 263-269. Thompson GM (1922) The naturalization of plants and animals in New Zealand. Cambridge University Press, Cambridge. Torchin ME, Lafferty KD, Dobson AP, McKenzie VJ & Kuris AM (2003) Introduced species and their miss- ing parasites. Nature 421: 628-630. Turchin P (2001) Does population ecology have general laws? Oikos 94: 17-26. Veltman CJ, Nee S & Crawley MJ (1996) Correlates of introduction success in exotic New Zealand birds. Am Nat 147: 542-557. Vitousek PM, D’Antonio CM, Loope LL, Rejmanek M & Westbrooks R (1997) Introduced species: a signifi- cant component of human-caused global change. NZ J Ecol 21: 1-16. Williamson M (1996) Biological invasions. Chapman & Hall, New York. Wilson KJ (1997) Extinct and introduced vertebrate species in New Zealand: a loss of biodistinctiveness and gain in biodiversity. Pacific Conserv Biol 3: 301— 305. Yap CAM, Sodhi NS & Brook BW (2002) Roost char- acteristics of invasive mynas in Singapore. J Wildl Manage 66: 1118-1127. Yap CAM & Sodhi NS (2003) Southeast Asian invasive birds: ecology, impact and management. Ornithol Sci 3: 57-67. Ornithol Sci 3: 43-55 (2004) SPECIAL FEATURE Invasive bird species Invasive birds in Hong Kong, China Michael R. LEVEN! and Richard T. CORLETT** ' Asia Ecological Consultants Ltd., 127 Commercial Centre, Palm Springs, Yuen Long, Hong Kong, China ? Department of Ecology & Biodiversity, University of Hong Kong, Pokfulam Road, Hong Kong, China Abstract The natural vegetation of Hong Kong is tropical evergreen forest, but this ORNITHOLOGICAL was almost entirely cleared by people before the eighteenth century. This clearance SCIENCE must have had a major impact on the bird community and undoubtedly caused the © The Omithological Society disappearance of all forest-dependant species, but these changes are undocumented. of Japan 2004 The documented history of Hong Kong’s avifauna begins following British coloniza- tion in 1860 and parallels a process of progressive restoration of forests, at least in the uplands. At least nine bird species that were present in Hong Kong in 1860 are depen- dant upon anthropogenic habitats and are therefore considered to have invaded before colonial times. Subsequently 41 species (30% of the breeding avifauna) have colo- nized. Of these, 22 species are believed to have spread unaided from southern China, while the remaining 19 species are considered to have been introduced by people from sources both within and beyond the region. Unlike the pattern of documented bird invasions elsewhere in Southeast Asia, most of these recent invaders are forest species, reflecting the recent pattern of habitat change. The possible ecological im- pacts of these invaders (both natural and human-assisted) are reviewed, but they are largely unknown. Hong Kong may provide a model for the evolution of bird commu- nities elsewhere in the region if current patterns of deforestation are permitted to con- tinue. Key words Birds, Exotic species, Invasions, Tropics The Hong Kong Special Administrative Region of the People’s Republic of China (hereafter, Hong Kong) is a biogeographically arbitrary 1100-km? sec- tion of coastal southern China, along with the adja- cent islands (Fig. 1) (Dudgeon & Corlett 1994). The landscape of southern China has been so greatly modified by human impacts that any attempt to re- construct its natural state is inevitably speculative to some degree. The whole region has a forest climate, with hot wet summers and cool dry winters. More- over, it forms part of a continuous belt of forest cli- mates stretching from the tip of the Malay Peninsula in the south to the arctic tree-line in the north, with no significant barriers to dispersal apart from the south-north climatic gradient. One important point in this gradient—the line between frost-free and frost- prone environments—runs through Hong Kong, but this is much more significant for plants than birds (Dudgeon & Corlett 1994). Hong Kong’s natural veg- (Received 26 August 2003; Accepted 9 October 2003) * Corresponding author, E-mail: corlett@hkucc.hku.hk etation would have been tall, species-rich, tropical evergreen forest, but less tall, less species-rich and less evergreen than forests nearer the equator. Natural open habitats in the Hong Kong region must have been rare, with coastal cliffs and beaches, and per- * Mai Po New Territories e Tai Po Kau Tai Mo Shan Fig. 1. the text. Map of Hong Kong showing the places mentioned in M. R. LEVEN and R. T. CORLETT s seasonally flooded riverine grasslands, the only yes extensive enough to support non-forest birds. \lthough early human settlements in the region must have had some impact on the biota, the major period of deforestation—leading eventually to the total transformation of the landscape of southern China—seems to have started following increased Chinese immigration in the eleventh and twelfth cen- turies (Marks 1998). The process accelerated with population increases in the sixteenth and seventeenth centuries, and was essentially completed by the end of the eighteenth century. There are no descriptions of Hong Kong itself before the mid eighteenth cen- tury, but earlier European visitors to the Pearl River delta region, west of Hong Kong, make no mention of forest (Corlett 1997). Historical settlement patterns in Hong Kong suggest that deforestation proceeded from the coast inland and from the lowlands upwards. Even the highest slopes were terraced for tea by the late seventeenth century—probably earlier—and it is hard to see where any substantial area of forest could have survived in Hong Kong past this date. The words “bleak”, “barren” and “bare” appear in all nineteenth century descriptions of the Hong Kong (Dudgeon & Corlett 1994). The lowlands were al- most entirely under cultivation, while the uplands were mostly grassland, maintained by fires and the cutting of hillside biomass for fuel. Forest cover sur- vived during this period of maximum human impact as feng shui woods near villages, as linear patches along upland streams, and as small stands in topo- graphically protected sites (Zhuang & Corlett 1997). None of these forest patches is likely to have been big enough to support populations of forest-depend- ent birds, even without the additional impacts from hunting and trapping (Corlett 2000). The absence from Hong Kong of forest pheasants, resident wood- peckers, most of the expected babblers, and a range of other forest species presumably dates from this pe- riod. The human population of Hong Kong is now approaching 7 million. Agriculture has been largely abandoned and the lowlands are being rapidly urban- ized. By contrast, the intensity of human impact in upland areas has declined. Shrubland and secondary forest are replacing the fire-maintained grasslands (Zhuang & Corlett 1997) and more than half the non- urban area is under some form of legal protection (Dudgeon & Corlett 1994). The situation in Hong Kong is thus very different from that in most of Southeast Asia (Wells 1999), The ecological transformation that is occurring cur- 44 rently in Southeast Asia, as forest is being replaced by a variety of non-forest habitats, occurred at least 300 years ago in Hong Kong (Corlett 2000). The loss of bird species from isolated and degraded forest fragments that is just starting in Southeast Asia (Wells 1999; Brook et al. 2003) has gone almost to completion in Hong Kong, where the expanding sec- ondary forests are dominated by habitat generalists and new arrivals of various origins (Kwok & Corlett 1999, Corlett 2000). Hong Kong may be the future of Southeast Asia— we hope not!—but it is also a place without an eco- logical history. Bird records go back only to 1860, when Robert Swinhoe visited Hong Kong (Swinhoe 1861), and anything before that is partly speculation. This complete absence of a pre-impact baseline is a major problem for many kinds of ecological research in Hong Kong. For invasive birds, the problem is somewhat less, as most of the obvious introductions have occurred in the last few decades, but it is still difficult to determine which of the bird species pres- ent in 1860 would have been present prior to defor- estation and which are actually early invaders. We also face a further difficulty in that, alongside the re- cent documented introductions of obviously exotic bird species, Hong Kong is experiencing colonization by a suite of species that are native to southeast China. All of these may have been present in Hong Kong’s primeval forests, though we cannot be certain of this. Some of these species have reached Hong Kong naturally, but for others there is strong evidence that the populations have been introduced. Accord- ingly, in assessing the impact of invasive bird species, we have divided our review into two periods: the pe- riod before 1860 for which there are no historical records (Table 1) and the period since. In this second period, we have distinguished between those species which appear to have colonized (or recolonized) Hong Kong naturally (Table 2) and those that appear to have been introduced by human agency (Table 3). SPECIES PRESENT IN 1860 The nine species listed in Table | are all wholly or largely associated with anthropogenic habitats in Hong Kong. All are widespread in southern China (Cheng 1987; Lewthwaite 1996) and Hong Kong (Carey et al. 2001). Two species (Eurasian Tree Spar- row and Common Magpie; bird scientific names are in the Tables) have a wide distribution in the Palearc- tic Region (Clements 2000) and it seems probable Invasive birds in Hong Kong Table 1. tats. Species known to be present in Hong Kong prior to 1860 which are solely or largely restricted to anthropogenic habi- Species in this and the following Tables have been ascribed to feeding guilds as follows: C=carnivore, F=frugivore, G=grani- vore (including seed predators and herbivores), I=insectivore, N=nectarivore and O=omnivore. For some species, distributions within Hong Kong are distinguished as HKI (Hong Kong Island) or NT (New Territories). Boece Enrich Name peeeee Seenige Name Feeding Anthropogenic habitats utilised Natural habitats guild (in approximate order of importance) utilised Spotted Dove Streptopelia chinensis G Farmland, urban, urban fringe, fishponds, = grassland with shrubs. Oriental Magpie Robin Copsychus saularis I Farmland, urban fringe, urban, fishponds. Mangroves. Masked Laughingthrush Garrulax perspicillatus I/F Urban fringe, farmland, fishponds, grassland == with shrubs. Scaly-breasted Munia Lonchura punctulata G Farmland, fishponds, grassland, shrubland. — Eurasian Tree Sparrow Passer montanus G Urban, urban fringe, farmland, fishponds. = Black-collared Starling Sturnus nigricollis O Farmland, fishponds, urban fringe, urban. — White-shouldered Starling Stwrnus sinensis I/F Farmland, fishponds, coastal shrubland, — urban fringe. Crested Myna Acridotheres cristatellus O Farmland, fishponds, urban, urban fringe, — coastal shrubland. Common Magpie Pica pica O Farmland, fishponds, urban, urban fringe, = shrubland, grassland. that they have reached Hong Kong and southern China from the north, as has previously been sug- gested for the Eurasian Tree Sparrow (Summers- Smith 1988). A further three species (Spotted Dove, Oriental Magpie Robin and Scaly-breasted Munia) have widespread distributions in the Oriental Region (Clements 2000). These three species are widespread in Hong Kong, although Scaly-breasted Munias are absent from urban areas (Carey et al. 2001). Spotted Doves and Scaly-breasted Munias are restricted to anthropogenic habitats, but Oriental Magpie Robins are common in mangroves and it has been suggested that this is the original habitat of the species in this region (Carey et al. 2001). The distributions of the re- maining four species (Masked Laughingthrush, Black-collared Starling, White-shouldered Starling and Crested Myna) are largely restricted to southern China (Clements 2000), so they are presumably na- tive to this region, but in Hong Kong they are re- stricted to habitats that are wholly anthropogenic in origin (Leven 2000; Carey et al. 2001). Although we have no information on the pre-1860 bird trade or bird release practices, the continuity of suitable an- thropogenic habitats between Hong Kong and the probable source areas of the ten species means that all these species could have reached Hong Kong as natural colonists, without direct human assistance. 45 NATURAL COLONISTS SINCE 1860 Forty-one species have become established as breeding birds in Hong Kong since 1860 (Tables 2 & 3) (Kershaw 1904; Herklots 1967; Chalmers 1986; Carey et al. 2001). Twenty-two species are consid- ered to have reached Hong Kong without direct human agency (Table 2), though some are dependent on anthropogenic habitats. The other 19 species (Table 3) are believed to have populations founded by individuals that were released by humans. Separation into the two categories has been based on review of the pattern of establishment, together with an assess- ment of the natural vagility of the species and the proximity of potential source populations (Carey et al. 2001, pp 108-113). Characteristics considered likely to be indicative of natural colonists include: species with potential source populations in South- east China; species known to be migratory or irrup- tive with a pattern of a progressive increase in records of transient individuals prior to the establish- ment of a breeding population; species where the first Hong Kong populations became established in the northern New Territories (i.e. near the border with the rest of China); and species that are relatively infre- quent in the cage bird trade. Conversely, the patterns of colonization of species considered likely to have been introduced by humans show one or more of the following features: species not occurring naturally in Table 2. M. R. LEVEN and R. T. CORLETT Species that have colonized Hong Kong since 1860 without direct human agency. Species English Reasons for Black Baza Eagle Besra Cuckoo Cuckoo Little Swift Great Barbet Grey-chinned Minivet Scientific Name Diet Breeding habitats Pattern of invasion : ; Name Invasion Aviceda leuphotes V/C Forest Increasing summer visitor Forest maturation, from 1972; now stable in low _ perhaps natural numbers. range expansion in China. Black-winged Kite Elanus caeruleus (Ce Famland, wetlands Irregular throughout the year Range expansion from 1984; has bred, still rare in China. and probably not established. Crested Serpent Spilornis cheela C Forest First recorded 1940, gradual _—_- Forest maturation. increase from 1950s, further increase in 1980s & 90s. Now widespread in NT. Crested Goshawk = Accipiter trivirgatus C Forest Pre 1980s status confused, but Forest maturation. undoubted increase since. Accipiter virgatus C Forest Pre 1990s status confused, Forest maturation. probable increase since 1980s. Emerald Dove Chalcophaps indica G Forest, shrubland First recorded 1960, Forest maturation. widespread (but low density) from 1980s. Chestnut-winged = Clamator I Forest, shrubland First recorded 1957 but rare Spread of Greater coromandus until c. 1980, subsequent rapid Necklaced increase. Laughingthrush (host). Hodgson’s Hawk _— Hierococcyx fugax I Forest One record 1971, annual since Forest 1994. maturation? Spread of host? Apus affinis It Widespread but only First recorded 1941, colonized Nest site breeds in urban areas during 1950s, now in all urban availability? areas. Megalaima virens I/F Forest, NT only 1920s, widespread by 1950s __ Forest maturation. (HKI colonized 1930s—1960s then died out). White Wagtail Motacilla alba I Wetland, farmland, Sporadic breeding species un- Unknown. urban til 1990s, rapid increase, now common and widespread. Pericrocotus solaris I Forest First record 1957, first in Forest maturation. summer 1981, first proven breeding 1984, continued increase and spread in 1990s (NT only). Scarlet Minivet Pericrocotus I Forest First record 1950, first in Forest maturation. flammeus summer 1975, subsequent gradual increase and spread central NT. Chestnut Bulbul Hypsipetes I/F Forest First recorded 1936, increased Forest maturation. castanonotus in 1960s but largely as winter visitor, regular breeding during 1980s, continued spread since. Orange-bellied Chloropsis I/F/N _ Forest One record 1934, otherwise Forest maturation, Leafbird hardwickii 46 since 1984, first breeding 1989, gradual spread in 1990s. probable natural spread but some evidence of human introduction. Invasive birds in Hong Kong Table 2. (Continued). Species English Scientific Name Diet Breeding habitats ; 2 Reasons for Pattern of invasion Name invasion Orange-headed Zoothera citrina I/F Forest First recorded 1956, sporadic Forest maturation. Thrush until 1980s, then gradual increase, still rare. White-bellied Yuhina zantholeuca I/F Forest First recorded 1980, Forest maturation, Yuhina subsequent slow increase probable natural and spread. spread but could be derived from human introduction. Hainan Blue Cyornis hainanus ] Forest First recorded 1958, gradual _—- Forest maturation. Flycatcher subsequent spread (perhaps more rapidly in 1990s). Buff-bellied Dicaeum ignipectus I/F Forest First recorded 1954 but Forest maturation. Flowerpecker perhaps previously overlooked, undoubted increase from 1960s, breeding from 1975, continuing gradual increase. Fork-tailed Aethopyga christinae 1/N Forest, parks, urban _ First recorded 1959, slow Forest maturation, Sunbird fringe increase during 1960s, rapid ornamental since 1970s, now common tree planting? and still spreading. White-rumped Lonchura striata G Forest, shrubland, First recorded 1913 but scarce Unknown, widely Munia urban and irregular until 1970, then traded and rapid increase and spread. breeding population may have, at least, been have, at least, been bolstered by released birds. Large-billed Crow Corvus O Farmland, wetlands, First recorded 1934, Unknown. macrorhynchus fringe open country, urban progressive spread through HK (from North) reaching HKI by 1950s, numerical increase since. Southeast China; species not known to be naturally migratory or irruptive; a relatively sudden appearance of a breeding population with few prior observations of transients; initial population foci in or near urban areas or on Hong Kong Island; birds showing physi- cal or behavioral evidence of former captivity; and species occurring frequently in trade (Leven 2000; Carey et al. 2001). For most species, we can be confi- dent that their assignment to either Table 2 or Table 3 is correct, though for some, the evidence is not over- whelming either way. In addition, for at least one species not included in Tables 2 and 3, White-rumped Munia, the pattern of records suggests that both natu- ral colonization and large-scale Buddhist releases 47 have occurred (Vaughan & Jones 1913; Webster 1976; Carey et al. 2001). The overwhelming majority (18 out of 22) of the species that are considered to be natural colonists occur in forest or shrubland and their establishment in Hong Kong since 1860 presumably reflects the ex- pansion of these habitats over this period (Zhuang & Corlett 1997; Corlett 1999; Corlett 2000). Most are insectivores or insectivore/frugivores. Most insecti- vores and many insectivore/frugivores which breed in Southeast China are migratory or irruptive (Cheng 1987; Lewthwaite 1996), doubtless because of the marked seasonality of food availability (Kwok & Corlett 1999; Leven 2000; Kwok & Corlett 2002) M. R. LEVEN and R. T. CORLETT 3. Species that have become established in Hong Kong since 1860 through human agency. Species English Name Rock Dove Yellow-crested Cockatoo Rose-ringed Parakeet Red-whiskered Bulbul Streak-breasted Scimitar Babbler Rufous-capped Babbler Chinese Babax Greater Necklaced Laughingthrush Black-throated Laughingthrush White-browed Laughingthrush Silver-eared Mesia Scientific Name tae Breeding habitats Columba livia G Urban, farmland Cacatua sulphurea G Urban parks, urban fringe forest HKI only Psittacula krameri G Urban parks HKI (formerly also farmland NT) Pycnonotus jocosus I/F Urban, farmland, grassland, forest fringe Pomatorhinus I/F Forest, shrubland ruficollis Stachyris ruficeps I Forest, shrubland Babax lanceolatus I/F Montane forest, shrubland Garrulax pectoralis I/F Forest, shrubland Garrulax chinensis I/F Forest, shrubland Garrulax sannio I/F Shrubland, farmland Leiothrix argentauris I/F Forest, shrubland 48 Pattern of invasion Possible reasons for invasion Poorly documented, present by 1953, widespread in urban and rural areas, no known change in status since. First in 1960s. Population c. 150—slow increase. First in 1903-1913; population now <20 birds, formerly 100+. Apparently not present 1860, common in 1904, no subsequent change. Single records 1949, 1978. Rapid spread since 1986 with foci on HKI and central NT. First recorded 1985, rapid spread in 1990s. First recorded 1959 then small population established on TMS (above c 650 m). No recent change. First recorded 1969. Slow spread in 1970s and 1980s, rapid and accelerating in 1990s. First recorded 1913, but not recorded again until 1950s when found to be widespread on HK], scattered reports from NT in 1980s, increase and spread in 1990s. First record from Cheung Chau (1941), released birds HKI 1961, widespread records from mid-1970s, decline during 1990s. One record 1970, many from 1987, widespread and rapid increase in 1990s. Released cagebirds and/or farmed birds. Released cagebirds (extralimital species). Released cagebirds (extralimital species). Probably human agency (not necessarily in HK). Released cagebirds but facilitated by availability of suitable habitat. Released cagebirds but facilitated by availability of suitable habitat. Released cagebirds. Released cagebirds but facilitated by availability of suitable habitat. Released cagebirds. Recent spread facilitated by availability of suitable habitat? Released cagebirds. Released cagebirds (extralimital species). Invasive birds in Hong Kong Table 3. (Continued). Species English Scientific Name beedine Breeding habitats Pattern of invasion Possible eens Name guild for invasion Red-billed Leiothrix lutea I/F Forest, shrubland Inrequent records from V&J = Released Leiothrix (1913) on but marked upsurge c. agebirds. during late 1980s and early 1990s, then marked subsequent decline. Blue-winged Minla cyanouroptera \/F Forest First recorded 1992, rapid Released Minla subsequent increase. cagebirds (extralimital species). Vinous-throated Paradoxornis /G Montane shrubland, First recorded 1971, breeding Released Parrotbill webbianus dwarf bamboo on TMS from 1980 where cagebirds. above 650m small population persists. Yellow-cheeked Parus spilonotus I Forest First recorded 1988, Released Tit widespread records from cagebirds. 1989, numbers peaked early 1990s then declined. Velvet-fronted Sitta frontalis I Forest First recorded 1989, limited Released Nuthatch spread in 1990s, now stable? cagebirds Baya Weaver Ploceus philippinus G Wetlands First recorded 1993, breeding Released confirmed in 1995. Population cagebirds of c. 30 birds still present. (extralimital species). Common Myna Acridotheres tristis O Farmland, fishponds, First recorded 1952. Regularly Released villages recorded since mid-1950s, cagebirds population fluctuating but (extralimital small (c. 100 birds). species). House Crow Corvus splendens O Urban, container First recorded 1974, 12 Probably port records by 1998, recent increase to c. 100 birds. ship-assisted (extralimital species). and are thus pre-adapted to find and utilize newly available habitat. The remaining four species (Black-shouldered Kite, Little Swift, White Wagtail and Large-billed Crow) are all associated with open country habitats. As such, their colonization of Hong Kong in the last 50 years or so perhaps represents a continuation of the pattern of spread in southern China of species that are dependent on anthropogenic habitats detailed in Table 1. We have direct evidence of such spread in southern China in the case of Black-winged Kite, which now occurs throughout Southeast China (MR Leven pers. obs.) despite its being considered by Cheng (1987) to be rare and largely restricted to the southwestern Provinces of Yunnan and Guangxi. 49 HUMAN INTRODUCTIONS SINCE 1860 Nineteen species are considered to have become established in Hong Kong through human agency (Table 3). Apart from the House Crow, which most probably arrived on a ship as it has done in many other parts of the world (Madge & Burn 1994; Ryall 2002), all these species are thought to have been brought to Hong Kong as cagebirds. Some wild pop- ulations may have been established by escaped birds, but deliberate releases, usually by Buddhists, are more likely to include multiple individuals of the same species. Other possible reasons for releasing birds include bird dealers getting rid of diseased or surplus stock or, in the case of the parrots, pet owners tiring of the responsibility of looking after long-lived and noisy pets. These 19 species fall into two sub- groups: those which do not occur naturally in south- M. R. LEVEN and R. T. CORLETT China: and those which are native to southeast China, but for which there is evidence that the Hong Kong populations are not of natural origin. The first sub-group includes nine species: Rock Dove, two parrots (Rose-ringed Parakeet and Yellow- crested Cockatoo), two babblers (Silver-eared Mesia and Blue-winged Minla), Velvet-fronted Nuthatch, Baya Weaver, Common Myna and House Crow. Rock Doves are now widespread in Hong Kong and occur in both urban and rural areas. Presumably, they were introduced to Hong Kong as domesticated birds at some time after the establishment of the colony by the British in 1842. However, the history of this species is obscure and the earliest reference to its presence appears to be that of Webster and Phillipps (1967) by which time it was already widespread. Though Rose-ringed Parakeets were present in Hong Kong from at least 1913 (Vaughan & Jones 1913) and flocks of up to 87 birds were recorded throughout the territory during the 1970s and 1980s, the population declined substantially subsequently and is now re- stricted to a limited urban area on Hong Kong Island (Carey et al. 2001). Yellow-crested Cockatoos were first noted in Hong Kong in 1961 (Viney 1973) and numbers increased to an estimated at 60-100 birds during the late 1990s and at least 150 birds in 2003 though the population remains largely restricted to the urbanised north side of Hong Kong Island (Carey et al. 2001; MR Leven pers. obs.). Both Silver-eared Mesias and Blue-winged Minlas have become established in forest and shrubland, the former since 1987 and the latter since 1992 and both species appear to be spreading rapidly in these habi- tats (Leven 2000; Carey et al. 2001). Velvet-fronted Nuthatches were first recorded in Hong Kong in 1989 and quickly became established at one forest site (Leven 1993; Kwok & Corlett 1999) but, despite iso- lated records elsewhere (Carey et al. 2001) have shown little further spread (Carey et al. 2002). Baya Weavers have been recorded in Hong Kong since 1970 and breeding has occurred regularly at Mai Po Nature Reserve, in the northwest New Terri- tories, since 1995 (Carey et al. 2001). The population appears to have remained stable at about 30 birds since then (Carey et al. 2002) and there is no evi- dence of any further spread (MR Leven pers. obs.). Common Mynas were first noted in Hong Kong in 1952 and a small population appears to have become established in farmland areas in the New Territories by the end of the 1950s, at which time flocks of up to 30 birds were noted occasionally (Herklots 1967). 50 There appear to have been no subsequent changes in numbers and this species remains localized (Carey et al. 2001). House Crows were first recorded in Hong Kong in 1974 (Chalmers 1986) but by 1998 there were only a total of 12 records of one or two individ- uals and this species was not considered to be estab- lished in Hong Kong by Carey et al. (2001). Numbers have increased since and a population of approxi- mately 100 birds is now present in the urban area of Kowloon with a distribution centred around the con- tainer port and associated cargo handling facilities (PJ Leader pers. com.). The second sub-group, of ten species, includes Red-whiskered Bulbul, seven babblers and laugh- ingthrushes (Streak-breasted Scimitar-babbler, Ru- fous-capped Babbler, Chinese Babax, Greater Neck- laced, Black-throated and White-browed Laughingth- rushes and Red-billed Leiothrix), Vinous-throated Parrotbill and Yellow-cheeked Tit. Red-whiskered Bulbul is only tentatively included in this group as it appears to have reached Hong Kong in colonial times, but the point of origin and vector is uncertain. Swinhoe (1861) recorded this species in Guangzhou in adjacent Guangdong Province but not in Hong Kong, though it had reached Hong Kong by 1903 (Kershaw 1904). It is unclear whether Hong Kong was colonized from Guangdong Province and it is also unclear whether the Guangzhou population had itself been introduced by people. However, this species is almost entirely absent from forest in both Hong Kong and Guangdong Province (Lewthwaite 1996; Kwok & Corlett 1999, 2000) and, on this basis, it appears likely to be alien to the region. The seven species of babblers and laughingth- rushes in this category were first recorded in Hong Kong as follows: Black-throated Laughingthrush and Red-billed Leiothrix (pre 1913), Streak-breasted Scimitar-babbler (1949), Chinese Babax (1959), White-browed Laughingthrush (1961), Greater Neck- laced Laughingthrush (1969) and Rufous-capped Babbler (1985) (Vaughan & Jones 1913; Herklots 1967; Chalmers 1986; Carey et al. 2001). These species, with the exceptions of Streak-breasted Scim- itar-babbler and Rufous-capped Babbler, were treated as natural colonists by Chalmers (1986). Subse- quently, Streak-breasted Scimitar-babblers were also considered to be of natural origin by Leader (1993). However, a review of the patterns of occurrence and spread of all seven species resulted in their being treated as derived from captive stock by Carey et al. (2001). Invasive birds in Hong Kong Streak-breasted Scimitar-babblers, Rufous-capped Babblers, and Greater Necklaced and Black-throated Laughingthrushes now have well-established and ex- panding populations in forest and shrubland (Leven 2000; Carey et al. 2001; Carey et al. 2002). Chinese Babax has only become established above 650 m ele- vation on Tai Mo Shan in the central New Territories, Hong Kong’s highest mountain, where a small popu- lation appears to have persisted since at least the date of the first observation in 1959 (Carey et al. 2001). Red-billed Leiothrixes were first noted in Hong Kong prior to 1913 (Vaughan & Jones 1913) but it was not until the 1980s that they became widespread (again in shrubland and forest) (Chalmers 1986) and, after peaking in the mid-1990s (Carey et al. 2001), the population has since decreased (Carey et al. 2002). Unlike the other species in this group, White-browed Laughingthrushes do not occur in forest but are found in shrubland and farmland (Leven 2000). Numbers of this species apparently peaked in the late 1970s and have since declined (Carey et al. 2001). Vinous-throated Parrotbills were first recorded in Hong Kong in 1960 and a population has become es- tablished at above 600 m elevation on Tai Mo Shan since 1980 (Carey et al. 2001). Whilst irregular records from elsewhere in Hong Kong were consid- ered to relate to birds of captive origin (Chalmers 1986), the Tai Mo Shan population was thought to be derived from natural colonists by Leader (1993). However, a review of the pattern of records in Hong Kong and Guangdong Province led to this population being treated, once again, as derived from captive stock by Carey et al. (2001). The first record of a Yellow-cheeked Tit in Hong Kong occurred in 1988 and breeding was first recorded in Tai Po Kau forest—Hong Kong’s largest contiguous forest area—in 1989 (Chalmers 1990). A small population has persisted at this site but, despite occasional records from other forest localities as well as in shrubland and urban parks, no other populations have become established (Carey et al. 2001). This species was originally considered to have colonized Hong Kong naturally (Chalmers 1990), however Carey et al. (2001) considered that all records of this species in the territory were derived from birds of captive origin. ECOLOGICAL IMPACTS OF INVASIVE BIRD SPECIES INHONG KONG The species listed in Table 1 are long-established 51 components of Hong Kong’s avifauna. Seven of the nine species (all except for Black-collared and White- shouldered Starlings and Scaly-breasted Munia) are among Hong Kong’s 20 most widespread breeding bird species (Carey et al. 2001). In contrast, of the 19 species in Table 2, only one, the relatively long-estab- lished Red-whiskered Bulbul, falls within this cate- gory. This bulbul is Hong Kong’s second most wide- spread breeding bird, occurring in 72% of |km squares in the territory (Carey et al. 2001). Taken to- gether, however, exotic species are now a substantial element of the breeding avifauna, comprising over 20% of the 141 species that were recorded breeding in Hong Kong during the mid 1990s (Carey et al. 2001). The diversity and abundance of exotic bird species in Hong Kong suggests that they must be of some ecological significance, but there is little direct evi- dence of their ecological impact. The long-estab- lished species, which comprise all but one of the most widespread exotics, would be expected now to be in equilibrium with the native flora and fauna, so any impacts of their arrival would be impossible to detect. Furthermore, all of these species largely or en- tirely utilize anthropogenic habitats, so they interact only with other exotics and the most tolerant of na- tive species. Of the species that have invaded Hong Kong in historical times, both Yellow-crested Cockatoos and Rose-ringed Parakeets cause temporary damage to trees in city parks by feeding on growing shoots (Herklots 1967). The Yellow-crested Cockatoo also causes similar damage to native trees in secondary forest on Hong Kong Island and occasionally de- stroys whole crops of unripe fruits (T Corlett pers. obs.). These birds are not abundant enough for these impacts to be serious, but any future increase in par- rot populations should be viewed with some concern. Perhaps more significantly, it has been suggested that these species may have been implicated in the disap- pearance of the native Great Barbet Megalaima virens from Hong Kong Island, presumably by com- petition for nesting holes (Carey et al. 2001). This can be excluded as an explanation in the case of Yel- low-crested Cockatoo, as Great Barbets vanished from Hong Kong Island prior to the occurrence of the cockatoos, but the disappearance of the barbets did coincide with the period of peak abundance of the parakeets. As noted above, the Red-whiskered Bulbul is the most widespread of the species that have invaded M. R. LEVEN and R. T. CORLETT Hong Kong in historical times. This bird has been considered a potential pest species elsewhere in the world where it has been introduced (Long 1981). In Hong Kong, it is one of the commonest species in urban parks (Lock 2000) and it is also abundant in farmland and shrubland, where it co-exists with and is outnumbered by the closely related Chinese Bulbul Pycnonotus sinensis (Leven 2000). However, it is rare in forest (Kwok & Corlett 1999, 2000) and there is no reason to suppose that it has had any impact on the numbers and distribution of any native species in Hong Kong in natural habitats. The Red-whiskered Bulbul is highly frugivorous and probably the third most important seed dispersal agent in Hong Kong (after P. sinensis and Z. japonica) in terms of num- bers of seeds dispersed (Corlett 1998, 2002; MR Leven unpubl, S So unpubl.). Its impact on woody succession in open habitats is therefore likely to have been positive. Both the Common Magpie, which is considered to have invaded during colonial times, and the Large- billed Crow, which arrived in the 1930s, are known as nest predators (Ali & Ripley 1986). However, both species are largely absent from wooded habitats (Kwok & Corlett 2000; Leven 2000), unlike the na- tive Red-billed Magpie which appears to be at least as effective as a nest predator in Hong Kong (Herk- lots 1967; Viney 1995). Although most of the species that have become es- tablished in forest habitats without human assistance are apparently generalist insectivore-frugivores, they also include: four predators on vertebrates, one a spe- cialist on snakes (Crested Serpent Eagle); the only species occurring regularly in Hong Kong which is capable of making its own nest holes in trees (Great Barbet); and the most nectarivorous bird species now occurring in Hong Kong (Orange-bellied Leafbird and Fork-tailed Sunbird). These species must have had some impact on the recovery of Hong Kong’s forest ecosystem and, while it is tempting to assume that as natural colonists from Southeast China their impacts have been benign, there is no firm evidence of this. Potentially of greater concern are the suite of species that have been introduced into forest habitats in Hong Kong as a result of human actions (five species of babbler, three laughingthrushes, Yellow- cheeked Tit and Velvet-fronted Nuthatch). In all these species, populations have become established since the 1950s and the numbers and distribution of six species (Streak-breasted Scimitar-babbler, Rufous- capped Babbler, Greater Necklaced Laughingthrush, Black-throated Laughingthrush, Silver-eared Mesia and Blue-winged Minla) are currently increasing (Carey et al. 2001; Carey et al. 2002). As far as is known, none of these species was released with the deliberate intention of establishing a wild population. More likely, their establishment in Hong Kong in re- cent years is a consequence of an increase in the area of suitable forest habitat co-incident with an upsurge in the bird trade from southern China and the in- creased popularity of releases of birds from temples (Melville & Lau 1994; Severinghaus 1999; Carey et al. 2001). Most of these introduced species have native popu- lations in Guangdong Province, and it seems likely that they were present in Hong Kong prior to defor- estation. As a result, their establishment has generally been welcomed as contributing to the restoration of Hong Kong’s avifauna and of bird-dependent ecolog- ical processes. The Greater Necklaced Laughingth- rush, in particular, is now the largest-gaped avian fru- givore in many areas of shrubland and secondary for- est, capable of swallowing fruits that are too big for any other common bird species and thus restoring seed dispersal services for species that were un- or under-dispersed previously (Corlett 2002). This species also appears to have permitted natural colo- nization by the Red-winged Cuckoo, a brood parasite for which this species appears to be a primary host (Becking 1981; Carey et al. 2001). However, relying on informal releases of traded birds as a means of reintroduction is risky at best. The source or sources of the released birds is usually unknown, but will rarely, if ever, be the nearest wild population to Hong Kong. At least one established sub-population of Greater Necklaced Laughingth- rushes shows characters of a race from western China (rather than the race G. p. picticollis, which is present in Guangdong Province) (Carey et al. 2001) and there is some evidence that more than one race of Streak- breasted Scimitar-babbler may have been introduced (MR Leven pers. obs.). Even where there are no visi- ble racial differences, birds captured in areas remote from Hong Kong are likely to be genetically distinct from the nearest populations and thus less suitable for reintroduction. It is also probable that some intro- duced populations have been established from very few released birds and will therefore be genetically impoverished. While these “informal reintroductions” have demonstrated the potential for ecological restoration in Hong Kong, they have also, presum- Invasive birds in Hong Kong ably, occupied the available niches for these species, making planned reintroductions of the appropriate races from the nearest available source population difficult or impossible. The two species introduced to montane shrubland, Chinese Babax and Vinous-throated Parrotbill, are also considered by Carey et al. (2001) to have poten- tially been present in Hong Kong prior to deforesta- tion. Populations of both species have persisted on one mountain in Hong Kong but appear to be re- stricted to habitat above 600m, which is similar to their lower altitudinal limit in Guangdong Province (Lewthwaite 1996). As such, the ecological impact of their introduction is probably limited. Three of the introduced forest species, Silver-eared Mesia, Blue-winged Minla and _ Velvet-fronted Nuthatch, do not occur in the wild in southeast China (Lewthwaite 1996; Carey et al. 2001). The ecological consequences of their introduction are unknown, though the Silver-eared Mesia may be competing di- rectly with the congeneric Red-billed Leiothrix. The ability of these exotic species to establish in semi- natural forest communities presumably reflects the very impoverished nature of Hong Kong’s forest avi- fauna, from which most of the expected babblers are missing. Their establishment may have pre-empted the reintroduction of species that were present before deforestation, although the Velvet-fronted Nuthatch occupies a niche with no obvious competitors in the present or past avifauna. The Common Myna is a highly invasive species elsewhere in the world and has had significant im- pacts on native species, especially on islands (Storer 1931; Feare & Craig 1998; Yap & Sodhi 2004). In Hong Kong, a population has persisted for 50 years but there is no evidence that it has increased beyond the size that it reached within a few years of colo- nization. The reasons for the failure of Common Myna to prosper in Hong Kong are unclear. However, the largest group in recent years has occurred in the only area where (feral) water buffalo remain in Hong Kong (Carey et al. 2001) and it may be that else- where in Hong Kong it is largely excluded by Crested Myna, a species with which it overlaps naturally only in a limited area in southwest China (Cheng 1987). Potentially the greatest ecological concern attaches to the recent increase in numbers of House Crows. Currently, the population is restricted to a limited ge- ographical area and is wholly urban. At present, it is unlikely to be interacting with native wildlife to a sig- nificant extent and the known roosts are in large trees 53 in small city parks (P.J. Leader pers. com.), which are of limited importance for wildlife, though these parks are of considerable amenity value. However, the pop- ulation is geographically close to the only egretry in the Hong Kong harbour area (Wong 2002). This is of concern in view of the House Crow’s known adverse effects on ardeid colonies elsewhere in the world (Ryall 1992). Based on experience elsewhere, the population may already have reached a point where eradication would be very difficult (Ryall 2003). Numerous other exotic species have become estab- lished in Hong Kong over recent decades but, in con- trast to the birds, these are largely in lowland anthro- pogenic habitats (Dudgeon & Corlett 1994). The best-studied exotic group is the plants, which so far have not invaded upland shrubland and forest habitats (Ng & Corlett 2002). The same pattern seems to be true for other, less studied, groups, with a very few exceptions. The clearest analogy with the birds con- sidered in this paper is Pallas’s Squirrel (Callosciurus erythraeus), which became established in Hong Kong around 1970, from escaped or released pets (Dud- geon & Corlett 1994). Previously, there were no na- tive squirrels in Hong Kong, presumably as a result of deforestation. Squirrels are now widespread in for- est and shrubland on Hong Kong Island and in parts of the New Territories. Two different subspecies have become established as result of at least two separate introductions of squirrels from different parts of the species’ range. The squirrels have increased pre-dis- persal seed predation, as well as being responsible for bark stripping and bud damage. However, Hong Kong is within the natural range of C. erythraeus, so these impacts could be viewed as a restoration of nat- ural ecological processes. CONCLUSIONS As a consequence of early deforestation of the Hong Kong region, the pattern of invasion by exotic bird species is very different from that which is cur- rently being recorded elsewhere in tropical East Asia (Wells 1999; Yap & Sodhi 2004). With the exception of Red-whiskered Bulbuls, which appear to have col- onized Hong Kong in the latter part of the nineteenth century, and perhaps House Crows which may be on the brink of a rapid population increase, Hong Kong has not been colonized in historical times by open country generalists or granivores that have rapidly become abundant in anthropogenic habitats. Rather, recent invaders of Hong Kong are forest species that M. R. LEVEN and R. T. CORLETT have been brought to Hong Kong for sale as cage- birds or for release from temples. The majority of species are babblers and laughingthrushes, which are insectivore-frugivores (Corlett 1998; Leven 2000). Rather than being opportunists with widespread dis- tributions, these species are Indo-Himalayan in origin (Clements 2000) and the apparent source populations of most species are forests in southern China (Carey et al. 2001). Thus, entirely fortuitously, many of the species that have recently invaded Hong Kong are likely to have been present in Hong Kong’s primeval forests. This apparent paradox is explained when the prob- able origins of Hong Kong’s widespread open coun- try species are examined. Many of these species are wholly or largely restricted to anthropogenic habitats in Hong Kong that almost certainly had no equivalent in the territory before deforestation. The inescapable conclusion is that Hong Kong experienced a wave of colonization by adaptable open country bird species prior to the beginning of our historical records of bird distribution. The pattern of invasion of bird species in Hong Kong thus provides a possible model for the fu- ture evolution of bird communities elsewhere in the region. ACKNOWLEDGMENTS The authors wish to thank Paul Leader and Geoff Carey for various contributions to this paper and the Hong Kong Bird Watching Society for access to unpublished records. REFERENCES Ali S & Ripley SD (1986) Handbook of the birds of India and Pakistan. Vol 5. 2nd ed. Oxford University Press, New Delhi. Becking JH (1981) Notes on the breeding of Indian cuckoos. J Bombay Nat Hist Soc 78: 201-231. Brook BW, Sodhi NS & Ng PKL (2003) Catastrophic extinctions follow deforestation in Singapore. Nature 424: 420-423. Carey GJ, Chalmers ML, Diskin DA, Kennerley PR, Leader PJ, Leven MR, Lewthwaite RW, Melville DS, Turnbull, M & Young L (2001) The avifauna of Hong Kong. Hong Kong Bird Watching Society, Hong Kong. Carey GJ, Diskin DA, Lewthwaite RW & Turnbull M (2002) Systematic list. Hong Kong Bird Rep 1988: 18-95. Chalmers ML (1986) Annotated checklist of the birds of Hong Kong. Hong Kong Bird Watching Society, 54 Hong Kong. Chalmers ML (1990) Records Committee report 1989. Hong Kong Bird Rep 1989: 16-31. Cheng T-h (1987) A synopsis of the avifauna of China. Science Press, Beijing. Clements JF (2000) Birds of the world: a checklist. Pica Press, East Sussex. Corlett RT (1997) Human impact on the flora of Hong Kong Island. In: Jablonski NG (ed) The changing face of East Asia during the Tertiary and Quaternary. pp 400-412. Centre of Asian Studies, University of Hong Kong, Hong Kong. Corlett RT (1998) Frugivory and seed dispersal by birds in Hong Kong shrubland. Forktail 13: 23-37. Corlett RT (1999) Environmental forestry in Hong Kong: 1871-1997. Forest Ecol Manag 116: 93-105. Corlett RT (2000) Environmental heterogeneity and species survival in degraded tropical landscapes. In: Hutchings MJ, John EA & Stewart A (eds) The eco- logical consequences of environmental heterogeneity. pp 333-355. Blackwell Science, Oxford. Corlett RT (2002) Frugivory and seed dispersal in de- graded tropical East Asian landscapes. In: Levey DJ, Silva, WR & Galetti, M (eds) Seed dispersal and fru- givory: ecology, evolution and conservation. pp 451-465. CAB International, Wallingford. Dudgeon D & Corlett R (1994) Hills and streams. An ecology of Hong Kong. Hong Kong University Press, Hong Kong. Feare C & Craig A (1998) Starlings and Mynas. Christopher Helm, London. Herklots GAC (1967) Hong Kong Birds. 2nd ed. South China Morning Post, Hong Kong. Kershaw JC (1904) Birds of the Quangtung coast, China. Ibis 4: 235-248. Kwok HK & Corlett RT (1999) Seasonality of a forest bird community in Hong Kong, South China. Ibis 141: 70-79. Kwok HK & Corlett RT (2000) The bird communities of a natural secondary forest and a Lophostemon con- fertus plantation in Hong Kong, South China. For Ecol Manag 130: 227-334. Kwok HK & Corlett RT (2002) Seasonality of forest in- vertebrates in Hong Kong, China. J Trop Ecol 18: 637-644 Leader PJ (1993) Records Committee report. Hong Kong Bird Rep 1992: 4—S. Leven MR (1993) Velvet-fronted Nuthatches in Hong Kong. Hong Kong Bird Report 1992: 188-191. Leven MR (2000) Shrubland birds in Hong Kong: com- munity structure, seasonality and diet. Ph. D. thesis. The University of Hong Kong, Hong Kong. Lewthwaite RW (1996) Forest birds in Southeast China: Invasive birds in Hong Kong observations during 1984-1996. Hong Kong Bird Re- port 1995: 150-203. Lock, NY (2000) The ecology of urban birds in Hong Kong. Ph. D. thesis. The University of Hong Kong, Kong Kong. Long, JL (1981) Introduced birds of the world. Reed, Sydney. Madge S & Burn H (1994) Crows and Jays. Christopher Helm, London. Marks R (1998) Tigers, rice, silk, and silt; environment and economy in late imperial south China. Cam- bridge University Press, Cambridge. Melville DS & Lau A (1994) Hong Kong (with special reference to China). In: Nash SV (ed) Sold for a song. The trade in Southeast Asian non-CITES birds. TRAFFIC, Cambridge. Ng S-C & Corlett RT (2002) The bad biodiversity: alien plant species in Hong Kong. Biodiversity Sci 10: 109-118. Ryall C (1992) Predation and harassment of native birds by the Indian house crow Corvus splendens in Mom- basa, Kenya. Scopus 16: 1-8. Ryall, C (2002) Further records of range extension in the House Crow Corvus splendens. Bull Br Ornithol Club 122: 231-240. Ryall C (2003) Notes on ecology and behaviour of House Crows at Hoek van Holland. Dutch Bird 25: 167-171. Severinghaus LL & Chi L (1999) Prayer animal release in Taiwan. Biol Conserv 89: 301-304. Summers-Smith JD (1988) The Sparrows. T & AD Poyser, Calton. 55 Swinhoe R (1861) Notes on the ornithology of Hong Kong, Macao and Canton, made during the latter end of February, March, April, and the beginning of May 1860. Ibis 1861: 23-57. Vaughan RE & Jones KH (1913) The birds of Hong Kong, Macao and the West River or Si Kiang in Southeast China, with special reference to their nidifi- cation and seasonal movements. Ibis 1913: 17-76, 163-201, 351-384. Viney CA (1973) Systematic list 1972. Hong Kong Bird Report 1972: 5-42. Viney CA (1995) Mount Nicholson Hong Kong 1978— 1995. A natural history. Privately published, Hong Kong. Webster MA (1975) An annotated checklist of the birds of Hong Kong. 3rd ed. Hong Kong Bird Watching So- ciety, Hong Kong. Webster M & Phillipps K (1976) A new guide to the birds of Hong Kong. Sino-American Publishing Com- pany, Hong Kong. Wells, DR (1999) The birds of the Thai-Malay Penin- sula, Vol. 1. Academic Press, London. Wong CLC (2002) Summer 2002 report on egretry counts in Hong Kong, with particular reference to the Mai Po Inner Deep Bay Ramsar Site. Hong Kong Bird Watching Society, Hong Kong. Yap CAM & Sodhi NS (2004) Southeast Asian invasive birds: ecology, impact and management. Ornithol Sci 3: 57-67. Zhuang X & Corlett RT (1997) Forest and forest succes- sion in Hong Kong, China. J Trop Ecol 14: 857-866. 2M greoth eeeleid wae yale (r @MOll HhOEM Aypodgiwe “Atta bap panel arpa on, hag) twig adud 14. ped gist] ORI CLA Avert, es i gedtaael aye att Bon Deis alo. nis ) ence aan pit cae ONT if BN Dehoe jit f “hah 12h Votstan ad TH sant ne (fF at 7 >} . 4.40 ii Be cae! echt AUT E ENA | aT Th ebbr) Re i (hit, Yieovel evel. lose h nae fovut od Wi CBR FARE int RP GE ea ak 7 aE (rE ern afhhea b4' bri ised yw y ask yea gin PRAT wrigitné Al tw @plitec Vie ALTO Oe magi jon rid eR fis Qn Nt aE el a gine AY ee "i “QnOM Gantt eae ane’ ton Waa eke emid of@rGd ip Sdh. eis] aamiboowl iausiht senshi bows (nt Pennine (CANS 9 sented BOGS 90/2 qui’ rotted) ow bit Oty mocyye VWAmuipeesy tyke OAT Geedeaai iid Yuoghl. > ehpartine hl, cise) ieoiaibe oh ae gif anh vine yriiiosn bn eval s ortirod (HO0C) 24 iho? BAAD aay ’ jeoripeay yt } io oF AREF! ey at ne more tad Td-1< x tee een here nar POCO Ls 29) pala Kognstt’. yr “?) 44° meek bopaidt> decker fh nee -ehdy Ragan sn0H ol] 280) ynnuh oe 0 Chelan ML 1990) Reconhs Copthiiiele hepa Nites aby aL obi matin: 49 vagulens EME (Ov0S) A td Cygne yoo th 8o vata deld rad yalesstyih ate obras fr. Perfo Aepafha Aitrowe: «MAY, \er7Ratn ly Wnodbopribe rf ivr Pac SQtacr OretiqotrntD oot, bine nein eh ky E hoot Lyles. ini. dablraski NG. fedje ype tee paie Alby. Anis ahha se oF — = watts lpn. Rie Hong ae vo Luke WOoH Om wi obt y sO al A bev MNeReC DO? bob ve Ard Robin ft a olny ee Woredoiganaamofise ty Shy Var; 187 | ? Fipeat Fe oi a Talley: carccon aan AAG ay iad, (deck ibon layne ‘aolicl huni: MJ, Juby EAS Swear a str wate. to paneer hoe twi* ai Vashuniak me Sdrety ¢ m WP ott) Tithe as “is anti wipliagitn, de eiwoty chap: Leal tery om) Ale, needa Megat W GY ah ) un, widatie--tK io ~aneranton, Rey eg okeion. RO VAM oD teil a Daal ied | pt doable f Hose Aw Hong Kony stack Hemiqut e679 (QUOT) 4 Gre & CoobPtal (G0Riaeh tA. . Peaewternnys, aati KCL) qt x loth eer GAC (1967) dfong! aig Bad M Pom, Mong Range * (S04) Binie- ot ie oak ewes ; 7 ee : 4 & Corton RT 11999) Seen Hoke Kong Sone hs se (2000 PTs iad 20,000 birds; Siew et al. 1980). In Singapore, the House Crow population is estimated to be 130,000 birds (Brook et al. 2003b). Urban managers have faced great public concern over the nocturnal roosts of these birds (Peh & Sodhi 2002). Since 1973, the sole management measure carried out by the government of Singapore to control urban House Crow roosts has been periodic shooting at known roost sites to disperse roosting birds (Peh & Sodhi 2002). However, this management technique has proven ineffective in achieving long-term control. In a recent study by Peh & Sodhi (2002), it was found that the House Crow preferred roosting in tall, old trees such as Pfterocarpus indicus with large dense crowns surrounded by tall buildings and lo- cated in areas of much human activity. Their study recommended the following habitat modification measures to discourage the crows from roosting in af- fected areas: (a) avoid planting well-spaced tall trees (18m tall) in urban area; (b) making existing roost sites less attractive (e.g. by tree pruning); and (c) es- tablishing alternative roost sites. A study of the nest site selection of the House Crow in Singapore by Soh et al. (2002) found that it preferred nesting in Pel- tophorum pterocarpum, in trees with greater crown volume and diameter at breast height, in urban open habitats, with higher disturbance, and nearer to bin centres and food centres. They recommended the fol- C. A. M. YAP and N. S. SODHI habitat modification measures to discourage crows from nesting in affected areas: (a) minor changes to the design of existing bin centres (i.e. food source) to restrict access by crows; (b) planting alter- native, less suitable tree species; and (c) regular prun- ing of trees with large and dense crowns. Brook et al. (2003b) argued that if applied simultaneously, both population control and habitat management might work effectively for the long-term control of House Crows in Singapore. Control of Common and White-vented Mynas: These two species of mynas are well established and common in Singapore, where their populations are estimated to exceed 100,000 birds and 25,000 birds, respectively (Lim et al. 2003). They are considered pests in Singapore primarily because of their commu- nal nocturnal roosts, which frequently occur in resi- dential areas, where their noise and faecal droppings disturb neighbouring residents. This has led to calls for management, and to attempts by authorities in Singapore to remove the birds by scaring, poisoning, and thinning or removing the trees (Hails 1985; Kang et al. 1990). These measures were partially effective short-term solutions to the problem, probably because they were not integrated with serious efforts by the authorities to provide alternative roosting sites for the mynas in less humanly populated areas (Kang et al. 1990). Schmidt & Johnson (1984) proposed that the difficulties of roost dispersal might be reduced if stress was imposed on the birds by removing sources of food, water or shelter from the vicinity of the roost. Kang et al. (1990) argued that this strategy would be unsuccessful with the mynas because they had a highly diverse diet and choice of nesting sites, and they could range over a large area (e.g. 308.0 ha [Kang 1989] and 14.0ha [Yap 2003]). Ecological studies have been carried out on mynas in Singapore with the aim of formulating habitat modification pro- grammes to discourage mynas from roosting in af- fected areas. These studies found that mynas pre- ferred to roost in tall, old trees with dense canopies, such as Pterocarpus indicus and Eugenia grandis, in urban areas sheltered from winds by nearby buildings or embankments, situated closer to food centres (i.e. open-air eateries) and surrounded by more vegetation than random non-roost trees (Hails 1985; Kang & Yeo 1993; Yap et al. 2002). Their studies recom- mended these management strategies: (a) roosts should be discouraged from forming in undesirable areas through a combination of bioacoustic and habi- tat modification control measures such as thinning of OWlnsg 64 canopies; (b) attractive alternative sites should be cre- ated in other areas (such as roadside verges or round- abouts situated away from residential areas) in num- bers that kept pace with the myna population; (c) re- fuse should be stringently controlled at food centres; (d) planting of mono-specific rows of tall trees with dense canopies such as Pterocarpus indicus and Eu- genia grandis in urban areas, especially near to food centres should be avoided, and trees with flattened, less dense canopies, or with leaves that close at night, such as Samanea samanea should be planted instead; and e) all these measures should be adopted on a long-term basis for effective control of myna roosts. Hails (1985) also recommended that roosts should not be disturbed unless they posed a serious nuisance, because of the costs and logistics involved in roost dispersal, and that public education be carried out at the same time. Control of pigeon populations: A possible method of control of Feral Pigeon populations using modifi- cation of habitats would involve the elimination of food sources (including deliberate feeding of pigeons by people), and the cleanup of food waste and spillage. The control of food supply was the basis for a successful control programme of a Feral Pigeon population in Basel, Switzerland (Haag-Wackernagel 1993a). Prior experience in Basel showed that the population of Feral Pigeons was about 20,000 indi- viduals, which numbers had been maintained in the face of the inadequate trapping and shooting of 100,014 birds in the period 1961 to 1985 (Haag-Wak- ernagel 1993b). The new programme envisioned pop- ulation reduction, not elimination, by means of highly restricted feeding of just a few birds. A loft was cre- ated and the birds were fed nearby in a public place. A bird-keeper maintained the loft and took eggs and young when the population exceeded a certain prede- termined number. The programme included public education on the advantages of population control, and the public advantages of having healthy, un- stressed birds. CONCLUSION There are 16 established invasive bird species in Southeast Asia. These species may have some nega- tive impacts on the native biodiversity and human economy. However, we need more research to assess the impacts of these species precisely. Control of in- vasive bird species in Southeast Asia is possible, but it requires a multi-pronged approach including both Southeast Asian invasive birds population and habitat management. ACKNOWLEDGMENTS Our research on invasive bird species in Singapore was sup- ported by the Ministry of National Development and Agri- food & Veterinary Authority of Singapore. We thank Malcolm Soh, Lian Pin Koh, Tien Ming Lee and two anonymous re- viewers for their comments. REFERENCES Ali S & Ripley SD (1968-74) Handbook of the birds of India and Pakistan. Oxford University Press, Bom- bay. Baker ECS (1922) The fauna of British India, including Ceylon and Burma. Taylor & Francis, London. Bernstein HA (1861) Uber nester and eier javanischer vogel. Journal of Field Ornithology 1861. Boosey EJ (1958) Foreign bird keeping. Cage Birds, London. Brook BW (2004) Australasian bird invasions: accidents of history? Ornithol Sci 3: 33-42. Brook BW, Sodhi NS & Ng PKL (2003a) Catastrophic extinctions follow deforestation in Singapore. Nature 424: 420-423. Brook BW, Sodhi NS, Soh MCK & Lim HC (2003b) Abundance and projected control of invasive house crows in Singapore. J Wild] Manage 67: 808-817. Cassey P (2002) Life history and ecology influences es- tablishment success of introduced land birds. Biol J Linn Soc 76: 465-480. Cheng T (1963) Chung-kuo ching-chi tung-wu chich- niao lei (China’s Economic Fauna: Birds). Science Publication, Pieping. Cooper JE (1996) Health studies on the Indian house crows. Avian Pathol 25: 381-386. Deignan HG (1945) The birds of northern Thailand. U. S. Natural History Museum, Smithsonian Institute Bulletin 186. Washington. Dickinson EC, Kennedy RS & Parkes KC (1991) The birds of the Philippines. An annotated check-list. British Ornithologists Union, Tring. Dorzhiev TZ (1978) K voprosu o polovozrelosti sizogo golubya (The question of sexual maturity of Rock Pi- geons). Gnezdovaya zhizn’ptits: 27-32. Duncan RP, Blackburn TM & Sol D (2003) The ecology of bird introduction. Annual Review of Ecology, Evol Syst 34: 71-98. Dyer MI & Ward P (1977) Management of pest situa- tions. In: Pinowski J & Kendeigh SC (eds) Granivo- rous birds in ecosystems. pp 267-300. Cambridge University Press, Cambridge. Dyer MI, Pinowski J & Pinowska B (1977) Population dynamics. In Pinowski J & Kendeigh SC (eds) Granivorous birds in ecosystems. pp 53-105. Cam- bridge University Press, Cambridge. Ehrlich PR & Ehrlich AH (1981) Extinction. Random House, New York. Elton CS (1958) The ecology of invasion by animals and plants. Methuen, London. Feare C & Craig A (1999) Starlings and mynas. Prince- ton University Press, New Jersey. Fitzwater WD (1967) Hindu PCO’s have problems American PCO’s haven’t thought of. Pest Control 10: 70-78. Fitwater WD (1994) House sparrows. In: Hyngstrom SE, Timm A & Larson GE (eds). Handbook of pre- vention and control of animal damage. pp E101- E108. University of Nebraska, Lincoln. Flint EP (1994) Changes in land use in South and Southeast Asia from 1880 to 1980: a data base pre- pared as part of a coordinated research program on carbon fluxes in the tropics. Chemosphere 29: 1015— 1062. Gibbs D, Barnes E & Cox J (2001) Pigeons and doves: a guide to the pigeons and doves of the world. Yale University Press, Connecticut. Haag-Wackernagel D (1993a) Street pigeons in Basel. Nature 361: 200. Haag-Wackernagel D (1993b) Zur Biologie der Strassentaube. Verlag Medizinische Biologie, Basel. Hails CJ (1985) Studies of problem bird species in Sin- gapore: |. Sturnidae (Mynas and Starlings). A report submitted to the commissioner for Parks and Recre- ation, Ministry of National Development, Republic of Singapore, Singapore. Huong SL & Sodhi NS (1997) Status of the Oriental Magpie Robin Copsychus saularis in Singapore. Malay Nat J 50: 347-354. Johnston R & Johnson S (1990) Reproductive ecology of feral pigeons. In: Pinowski J & Summers-Smith D (eds) Granivorous birds in the ecological landscape. pp 237-252. Polish Ecological Publishing, Warsaw. Johnston RF & Janiga M (1995) Feral Pigeons. Oxford University Press, New York. Jones C (1996) Bird introductions to Mauritius: status and relationship with native birds. In: Homes J S & Simons J R (eds) The introduction and naturalisation of birds. pp 113-123. Station Office Publication Cen- tre, London. Kang N (1989) Comparative behavioural ecology of the mynas, Acridotheres tristis (Linnaeus) and A. javani- cus (Cabanis) in Singapore. Ph. D. thesis. National University of Singapore, Republic of Singapore. Kang N (1992) Radiotelemetry in an urban environ- C. A. M. YAP and N. S. SODHI ment: a study of mynas Acridotheres spp. in Singa- pore. In: Priede IG & Swift SM (eds) Remote moni- toring and tracking of animals. pp 633-641. Ellis Horwood, Chichester. Kang N, Sigurdsson JB, Hails CJ, and Counsilman JJ (1990) Some implications of resource removal in the control of Mynas Acridotheres spp. in Singapore. Malay Nat J 44: 103-108. Kang N & Yeo VYY (1993) Roost site selection and the waking and roosting behaviour of mynas in relation to light intensity. Malay Nat J 46: 255-263. Kolar CS & Lodge DM (2001) Progress in invasion bi- ology: predicting invaders. Trends Ecol Evol 16: 199-204. Komdeur J (1996) Breeding of the Seychelles Magpie Robin Copsychus sechellarum and its implications for its conservation. Ibis 138: 485-498. Leopold A (1939) A biotic view of land. Journal of Forestry 37: 727-730. Leven MR & Corlett RT (2004) Invasive birds in Hong Kong, China. Ornith Sci 3: 43-55. Lever C (1987) Naturalized birds of the world. Wiley, New York. Lim HC, Sodhi NS, Brook BW & Soh MCK (2003) Un- desirable aliens: factors determining the distribution of three invasive bird species in Singapore. J Trop Ecol 19: 685-695. Lodge DM (1993a) Biological invasions: lessons for ecology. Trends Ecol Evol 8: 133-137. Lodge DM (1993b) Species invasions and deletions: community effects and responses to climate and habi- tat change. In: Karieva PM, Kingsolver JG & Huey RB (eds) Biotic interactions and global change. pp 367-387. Sinauer, Sunderland. Long JL (1981) Introduced birds of the world: the worldwide history, distribution and influence of birds introduced to new environments. Universe Books, New York. Madge S & Burns H (1994) Crows and jays: a guide to the crows, jays and magpies of the world. Houghton Mifflin, Boston. Martin CM & Martin LR (1982) Pigeon control: an inte- grated approach. Proc Vertebr Pest Conf 10: 190-192. Mayr E (1965) The nature of colonizations in birds. In: Baker HG & Stebbins GL (eds) The Genetics of Col- onizing Species. Proceedings of the first international union of biological sciences symposia on general bi- ology. pp 29-47. Academic Press, New York. McNeely JA (ed) (2001) The great reshuffling: human dimensions of invasive alien species. \UCN, Gland & Cambridge, Cambridge. Morris JG (1969) The control of feral pigeons and spar- rows associated with intensive animal production. 66 Aust J Sci 32: 9-14. Murton R, Thearle R & Thompson J (1972) Ecological studies of the feral pigeon Columba livia var. I. Popu- lation, breeding biology and methods of control. J Appl Ecol 9: 835-874. Myers N (1979) The sinking ark. Pergamon Press, Ox- ford. Myers N (1992) The primary source: tropical forests and our future. W. W. Norton, New York. Peh KSH & Sodhi NS (2002) Characteristics of noctur- nal roosts of house crows in Singapore. J Wildl Manag 66: 1128-1133. Pell AS & Tidemann CR (1997) The ecology of the Common Myna in urban nature reserves in the Aus- tralian Capital Territory. Emu 97: 141-149. Pimental D, Lach L, Zuniga R & Morrison D (2000) Environmental and economic costs of nonindigenous species in the United States. BioScience 50: 53-65. Pinowski J & Kendeigh SC (eds) (1977) Granivorous birds in ecosystems. Cambridge University Press, Cambridge. Restall R (1997) Munias and mannikins. Yale Univer- sity Press, New Haven. Ricklefs RE (1969a) An analysis of nesting mortality in birds. Smithson Contrib Zool 9: 148. Ricklefs RE (1969b) Natural selection and the develop- ment of mortality rates of young birds. Nature 223: 922-925. Robson C (2000) A Guide to the birds of Southeast Asia: Thailand, Peninsular Malaysia, Singapore, Myanmar, Laos, Vietnam and Cambodia. Princeton University Press, New Jersey. Ryall C (1992) Predation and harassment of native bird species by the Indian House Crow Corvus splendens, in Mombassa, Kenya. Scopus 16: 1-8. Ryall C (1994) Recent extensions of range in the house crow Corvus splendens. Bull Br Ornithol Club 114: 90-100. Sakai AK, Allendorf FW, Holt JS, Lodge DM, Molof- sky J, With KA, Baughman S, Cabin RJ, Cohen JE, Ellstrand NC, McCauley DE, O’Neil P, Parker IM, Thompson JN & Weller SG (2001) The population biology of invasive species. Annu Rev Ecol Syst 32: 305-332. Schmidt RH & Johnson RJ (1984) Bird dispersal recordings: an overview. Proc Verteb Pest Control 4: 43-65. Sengupta S (1968) Studies of the Common Myna Acridotheres tristis tristis (Linnaeus). Proc Zool Soc Calcutta 21: 1-27. Shuyler HR (1963) Bird control in Kansas City. Pest Control 31: 9-17. Siew YC, Fong FW, Ng SM & Poon SK (1980) Nest Southeast Asian invasive birds dispersion of the House Crow, Corvus splendens vieillot. Trop Ecol Dev 1980: 677-682. Simberloff D (1996) Impacts of introduced species in the United States. Consequences: Nat Implic Environ Change 2: 13-22. Soh MCK, Sodhi NS, Seoh RKH & Brook BW (2002) Nest site selection of the house crow Corvus splen- dens, an urban invasive bird species in Singapore and implications for its management. Landscape Urban Plann 59: 217-226. Weber WJ (1979) Health hazards from pigeons, star- lings and English sparrows. Thomson Publications, California. Wiens JA & Johnston RF (1977) Adaptive correlates of 67 granivory in birds. In: Pinowski J & Kendeigh SC (eds). Granivorous birds in Ecosystems. pp 301-340. Cambridge University Press, Cambridge. Wilson EO (1988) The current state of biological diver- sity. In: Wilson EO (ed). Biodiversity. pp 3-18. Na- tional Academy Press, Washington. Yap CAM (2003) A study of the changes in the ranges of white-vented mynas in the presence of an increas- ing house crow population in Singapore. Raffles Bull Zool 51: 159-163. Yap CAM, Sodhi NS & Brooks BW (2002) Roost char- acteristics of invasive mynas in Singapore. J Wildl Manage 66: 1118-1127. oly pe dy CMY i anna eal RA) Dero wiitE dub carey Tee) 4s a hog lee to yao Uh By beetle } nvTTuy area oh abor OA adel i ahnetie 1! patter aetna yee dene, uy orth Mr lene AC) Mafeacpar oy 200 angi Reinse-aiiw bo Teor a ee rede earre Thagrt ety AAA 12 las : vita eidts » eres MAD git Loyd TPC Pages deqy ea scciend, Cie Aa sgoaulA: 6 ! anal «54 (R801) OF aeiiw — i eee lent Zuntgh ¥ & Vit De: cydoeiye. ante. or) avo). spt, | na SAd4t W ‘DBO wethid brs! Nt sagt Finye LaoelronitntosatnaalyDWeb Lhd Hower oikpok del DoleyupeeduBcdae bérin Apr Lon OBAS-874 COA < 8 (0G) WY B vlvon a ab HAS dose, bebe aya tn) wow wasdd sii 16 pole bie TWN Binoge baie Ovieavniny mudd sap sadunn]\\nteenduymany eet tae | is SH. © Soathi WS (2012) Chute ave Awe cease | «THA bes oprah hen A at a —rfiblulid: EpetoalE erty Neila Met AS © Tideminn CK df! The 1/1 atabeeaD PWiracbé CMM ait x ior. Em 0% als ww alialt ¢ akf mai Exvironiy wet nid etondénind (eet Cinited Staree BjoSeienen wi 1& Kemicigh SC (ede) (78 / MP PAN feed OTM TPT. Cambridge ve iy t “oe truiiee ~ ae “< 19) i. YY Mtunixe zedt ‘Thrit iy Pree Now tive i (ickiefs (UE (f909a PAR augers orn bids, Smithton Contih Zook er eiels Bi (10006) Neva sclectiia ae ye feabalh i Tay wi of aveuiity vite. of me in a an lhetyere, | Aes) ON lagoons én pea" iutivad, ~Pentaaitlur Whee Wier tac Vidlhom and Cun Livny seruly. wa Me icc \ 992} Pralaticwy ahd hicasaul ‘ Indiad House CARY ui wth a flys Konya, aS Oyhue it. ‘ Riek (A) Recent e7esionk OL Torn: splendéas. Hall ea es Dy tie ni (yW) % eh 9 ; iL’ rE PW Hi sie Ll, Wal Baugh wgunad S. Ces Vit sey oe icy eP4 i“: € Welles a Pinel!) ' of inva) porte A Rey's si ate : t lohreod #4] (@R4) fi io) Major axis of the larger testis. > Combined mass of both testes as a percentage of body mass. spermatogonia, and by well-developed interstitial tis- sue, while mature testes show active spermatogene- sis, the seminiferous tubules have expanded enor- mously, and the interstitial tissue between adjacent tubules is greatly compressed (Lofts & Murton 1973). Based on the condition of the seminiferous tubules and interstitial tissues, we distinguished im- mature (Fig. la) and mature (Fig. 1b) testes. In fe- males, the ovary was removed and weighed to the nearest 0.01 g on an electronic balance. Breeding fe- males, particularly females that have laid eggs, have long, enlarged oviducts, but the oviducts become thinner during the non-breeding season (Gilbert 1979). We regarded long, and enlarged oviducts (Fig. 1c) as developed oviducts. RESULTS 1) Males Roosting males had significantly lighter bodies, shorter wings and tails than breeding males; while tarsus length was marginally, but not significantly shorter in roosting males (Table 1). Roosting males had significantly lighter and shorter testes, and a lower testes/body percentage than breeding males. However, overlap in these values was evident (Table 1). The frequency distribution of testes mass also showed considerable overlap (Fig. 2). All 15 breed- ing males and 17 of 70 roosting males (24.3%) had mature testes, while the remaining 53 roosting males (75.7%) had immature testes (Fig. 2). Based on age determination using plumage characteristics, all 15 breeding males and 23 of 70 roosting males (32.9%) were adults. The testes of 47 juveniles were imma- ture. Of 23 adult males, 17 males had mature testes, Roosting individuals (N=70) SY SSSSXSS Ven WW yw — Breeding individuals (N=15) Number of birds =} 10 15 20 25 3.0 3.5 Testes mass (g) Fig. 2. Frequency distribution of testes mass (combined mass of both testes). Hatched bars are males with mature testes. but six had immature testes. 2) Females Roosting females had significantly lighter bodies, shorter wings and tails than breeding females, while tarsus length was marginally, but not significantly shorter in roosting females (Table 2). Roosting fe- males had significantly lighter ovaries, and a lower ovary/body percentage than breeding females, al- though the values overlapped (Table 2). The fre- quency distribution of ovary mass showed consider- able overlap between breeding and roosting females (Fig. 3). All 20 breeding females and 19 of 80 roost- ing females (23.8%) had long, enlarged oviducts, while 61 roosting females (76.2%) had undeveloped oviducts (Fig. 3). All 20 breeding females and 15 of 80 roosting females (18.8%) were identified as adults using plumage characteristics. All 15 adults had developed oviducts. Of 65 juvenile females, 61 had Tih M. NAKAMURA and S. MURAYAMAR Table 2. Comparison of body and ovary size of roosting and breeding females. Figures are means+SD. Roosting individual Breeding individual t P N 80 20 Body mass (g) 537.9+43.0 602.6+39.7 6.10 <0.0001 Wing length (mm) sise14 32 18087) 4.46 <0.0001 Tail length (mm) 19.1+1.9 20.3+0.6 2.82 <0.01 Tarsus length (mm) 6.0+0.2 6.1+0.2 1.81 =0.073 Ovary mass (g) 017=0:12 0.36+0.21 3). <0.0001 Ovary/body! (%) 0.03+0.02 0.06+0.03 4.61 <0.0001 ' Mass of ovary as a percentage of body mass. Roosting individuals (N=80) \ ¥ Number of birds 20 Breeding individuals (N=20) 02 04 O06 O08 10. 1.2 Ovary mass (g) Fig. 3. Frequency distribution of ovary mass. Hatched bars are females with developed oviducts. undeveloped oviducts, while four had developed oviducts. DISCUSSION It is clear that 114 of 150 roosting crows (76.0%) were non-breeders because they had immature repro- ductive organs. Based on plumage characteristics, we considered 112 of the 150 roosting crows (74.7%) to be juveniles. The reproductive organs of 108 of the 112 juveniles (96.4%) were immature. These results suggest that the spring roost was composed mainly of juveniles with immature reproductive organs. How- ever, Six roosting males estimated as adults had im- mature testes. Four roosting females had developed oviducts but were considered juveniles. Thus, plumage characteristics are not always consistent with sexual organ development. Breeding adults likely roost near their nests to pro- tect eggs and nestlings from predators. However, 17 males and 19 females with mature reproductive or- gans assembled in the spring roost. There are three possible explanations for this. First, individuals that bred near the spring roost assembled in the roost. Kurata and Higuchi (1972) observed that some breed- ing adults assembled in a spring roost that was 4km from their nests. Second, individuals that failed in breeding assembled in the spring roost because they did not need to protect eggs and nestlings. The Car- rion Crow has two distinct social types: territorial pairs, and flocks of non-territorial individuals (Yoshida 2003). Pairs breed in territories (Nakamura 1998; Yoshida 2003), thus, sexually mature individu- als that bred near the spring roost and failed in breed- ing must have had territories. Third, non-territorial adults assembled in the spring roost. Yoshida (2003) found that non-territorial individuals live in flocks until they can acquire a breeding territory, and the flock is composed of juveniles and individuals over two years old. Only rarely are females with devel- oped oviducts non-territorial. Therefore, the individu- als congregating in the spring roost would be sexu- ally mature non-territorial males. We concluded that spring roosts were composed mainly of non-territorial juveniles with immature re- productive organs and some sexually mature territo- rial adults that had attempted to breed near the roost but failed, or sexually mature non-territorial males. ACKNOWLEDGMENTS We are indebted to Honorary Professor T. Nakamura of Joetsu University of Education for his encouragement and ad- vice on our work. We thank Professor S. Daigobo of the uni- versity for his technical advice in making paraffin sections. Professor A. Chiba of Nippon Dental University School of Dentistry at Niigata kindly gave histological information on reproductive organs. We are grateful to members of the Labo- ratory of Animal Ecology of the university for helping with the field work. We also express our thanks to two anonymous reviewers for their many constructive comments and sugges- tions. We could not have carried out this work without the sup- Carrion Crows in spring roosts port of the members of the Joetsu branch of the Japanese Hunting Association (Dainippon Ryoyukai), K. Nozaki, Y. Hayashi, and staff members of the Environmental Planning Section in the Joetsu City Office. REFERENCES Busse P (1984) Key to sexing and ageing of European Passerines. Beitr Naturkunde Niedersachs. 37 (Sonderheft): 71. Gilbert AB (1979) Female genital organs. In: King AS & McLelland J (eds) Form and Function in Birds. Vol. 1, pp 237-360. Academic Press, London. Goodwin D (1986) Crows of the World. 2nd ed. British Museum, London. Haneda K & lida Y (1966) Life history of Eastern Car- rion-crow (Corvus corone orientalis) I. Breeding sea- son (1). Jpn J Ecol 16: 97-105 (in Japanese with Eng- lish summary). Koshio C, Kawakami M, Mori Y, Morikami T, Tojo N, Komemoto K, Yoshida T, Nakata S, Simizu K, Sato K & Takahata Y (1996) Investigation of the natural environment in Sukunoumi Cove (V): The roosting behavior of crows around Naruto University of Edu- cation. Bull Naruto Univ Educ 11: 15—20 (in Japanese with English summary). 73 Kurata A & Higuchi Y (1972) Roosting behaviour of two species of crows in Mie prefecture. J Yamashina Inst Ornithol 6: 89-106 (in Japanese with English summary). Lofts B & Murton RK (1973) Reproduction in birds. In: Farner DS, King JR & Parkers KC (eds) Avian Biol- ogy. Vol. III, pp 1-108. Academic Press, London. Nakamura S (1998) The territorial behaviour of the Car- rion Crows Corvus corone corone in Japan. Jpn J Or- nithol 46: 213-223 (in Japanese with English sum- mary). Nakamura S (2003) The seasonal and annual cycle of Crows’ roost in northeastern Osaka prefecture. Strix 21: 177-185 (in Japanese with English summary). Tamada K & Fujimaki Y (1993) Breeding biology of Corvus corone and C. macrorhynchos in central Hokkaido. Jpn J Ornithol 42: 9—20 (in Japanese with English summary). Tanaka E (1989) Comparison of roosting and feeding behavior between Carrion and Jungle Crows. Gradua- tion thesis, Joetsu University of Education, Niigata, Japan (in Japanese). Yoshida Y (2003) Foraging areas and roost utilization of non-territorial individuals of the Carrion Crow Corvus corone. J Yamashina Inst Ornithol 34: 257— 269 (in Japanese with English summary). “meio ye PLOW pepit te especial mn? J tittoreroem wild ¢ ? x7? : ; Srwigcdel meh erect’? Pio eng WAT ! tpl &) Gel a4 Ay ul { jloJ 13! if, (7s 7 j $y 7 r Svea t jest sitst Lan “u¥ son. bo a) le opyaiod igi | GX ‘ muineagh Lett decsl n 7 ” yr erg fwalgped oti ol al . OP lodiin i Cutt ele, Jew &ne danurese cuirnnday' rv WHIDSINTO axar fri Fyn rue ihestgersd ; Ti -t¢% Hod waibenttl (Ft init @& ft alee] j > hi Asi tyre” ; a weer PANG.) Bemegnlinl) US-0 ote | ) L tat .obtei6r : ; rand¢YPEeaa ried oet Ft vey oP tip eae nw om npn J TUSOi. Of edaagl aT bare rye dahl mn asew Tact oivariod canal A Gerptne eS ane rTTr a Poet owt) nod’ si) nadal ruts wen hxwhiati: Ba Aig ‘wit yb, , , ; tect teen ar ir STi Luin pititn tone teal wi fleirTige.t dvi i ceevtlyyt STL 17aret ‘ . sid seary gal TTROTE \ ‘ tir) * re 4 nG J rif = TES tye peattonw brea. eg, fabs) Lt ai at 1 st aa i romans lati) be acon aera DAO aggro - way ¥ daw Ab cde hooey Holl ~ aftr) \anaguinived off Wy nahiyoe ‘Wase (tate whee: _— auRO 1Ooe 7 ey Mm 1h ”, . a. ‘ aq VARTA “0 Beaqoud WW yrioge fs uate on yur We Af) alaprester . ohapiwisl “a O45 9 71 ve git il-ngie fasion tne vey, . Robaod 2207 sinrebagA me fend by pet Anan $i\i Wes Nt rod, ace Wing es assr) By + stb} Bat ant BRS ‘ltd NIG asi Aas fda a Be ana Iti alba ¥DY ex Wiassentbied. ay the elie arg. i At Rh ie) Pre ROPE a OP BE dl zx Wie 4 ae Ft al litiinn ay id LaHeginins PQOGsly Sautide apy Paige Deady Lath eit AD inuiaor tocniiall de simp wi} 0022 thd ve wert ca i ral CLG Te) | sia: iolas ermiiead nm Pas ny ihe roiynte ; (eosial Loge fom lerytingiy! indivik 0 10 or nive n breeding i ‘ soot al juveniles HY OW, Waly faysly ure ies & oped. deiducts nom terion i The ’ us ihe spring 20pnf ve iMWreTaline Me poate none rihutial frit, - q Ve Sonciuded that iping Walaise mun "vents ih noe hyve“ ary scone vos thaill ym L wehaleeinas Mile atic rextyes firesd oi i ? sehuslly in el ie ? ACKNCIWT FG MENTE * ah ™ , ~ a ! HAnhkear ry Keane we Pia hig tes Oe Whe cher wed ¥ yor). We thank Prifessor S) Daa a ri “ivi ! oh Vig is ital Lai ern is 7 eer Livlinttaudl in bs, iw i ‘ coy Udals 16 el uctve Gunners ~. 4 ice! Oottpit ork wit ORIGINAL ARTICLE Ornithol Sci 3: 75—84 (2004) Habitat use and foraging behavior of male Black-and-white Warblers (Mniotilta varia) in forest fragments and in a contiguous boreal forest Cynthia A. PASZKOWSKI’, Navjot S. SODHI*, Scott JAMIESON and Sandra A. ZOHAR** Department of Biological Sciences, University of Alberta, Edmonton, Alberta T6G 2E9, Canada ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2004 Abstract We examined habitat use and foraging behavior of male Black-and-white Warblers (Mniotilta varia) in fragments (2-140ha) of mature forest dominated by trembling aspen (Populus tremuloides) and a similar contiguous forest (> 1000 ha) in central Alberta, Canada. Vegetation structure and composition differed significantly between occupied and unoccupied fragments, and between plots within and outside territories in occupied fragments. Territories in fragments were characterized by a high abundance of willow (Salix spp.), which was the primary foraging site for males in this habitat. Territories in contiguous forest differed significantly in vegetation structure from territories in fragments, and were characterized by a high abundance of trembling aspen trees, which served as the primary forest foraging site. The use of foraging substrates and foraging methods did not differ between individuals in frag- ments versus contiguous forest. In the boreal mixedwood ecoregion, the Black-and- white Warbler appears flexible in its habitat use. Key words In the last two decades, the proposition that frag- mentation of forest habitat negatively affects song- bird communities in temperate and tropical regions has become widely accepted by avian ecologists and conservationists. Some bird species will not occupy fragments (Schmiegelow et al. 1997), whereas in oth- ers, males establishing territories in fragments may not attract mates (Villard et al. 1993), or pairs nesting in fragments experience low reproductive success (Hinsley et al. 1999). The inferiority of fragments as nesting sites has often been linked to “negative edge effects” that create habitats that differ from contigu- ous forest in vegetative structure, production of food organisms, or risk from predation or brood parasitism (Hoover et al. 1995; Bayne & Hobson 1997; Burke & Nol 1998; Smith et al. 2000; Zanette et al. 2000). (Received 18 August 2003; Accepted 28 November 2003) * Corresponding author, E-mail: cindy.paszkowski @ualberta.ca Present address: Department of Biological Sciences, National University of Singapore, Blk S2, 14 Science Drive 4, Singapore 117543, Republic of Singapore ** Present address: National Science & Engineering Research Council of Canada, 350 Albert Street, Ottawa, Ontario KIA 1H5, Canada * 75 Canada, Disturbance, Neotropical migrants, Vegetation, Wood warbler Surprisingly few studies (e.g. Nour et al. 1997; Flem- ming et al. 1999; Miller & Cale 2000) have actually compared the behavior of birds in fragments versus contiguous forest, either in terms of habitat selection or foraging, to understand the possible reasons be- hind the apparent failure of fragments to offer suit- able habitat or compensatory responses of birds in- habiting fragments. The Black-and-white Warbler (Mniotilta varia) is a monogamous, ground-nesting species (Peck & James 1983: Ehrlich et al. 1988; Kricher 1995) that breeds in Nearctic forests but winters mostly in the Neotrop- ics (O'Connor 1992). It is the only wood warbler species (Embrizidae: Parulinae) that frequently feeds on arthropods found on bark (Morse 1989). The Black-and-white Warbler is considered to be “area- sensitive” because in parts of its breeding range it does not occur in small forest patches (e.g. <10 ha in area; Villard et al. 1995) and tends to occupy forest- interior (>100m inward from forest edge; Askins et al. 1987; Freemark & Collins 1992; Morneau et al. 1999). It has become locally extinct (Litwin & Smith 1992; Terborgh 1992) or shown long-term population declines (Sauer et al. 1996) in some areas, suggesting C. A. PASZKOWSKI et al. that it may be vulnerable to changing patterns of land use. However, the Black-and-white Warbler does dis- play some degree of flexibility in habitat use as the species breeds in forests of different vegetation com- position and ages across its range (Kricher 1995; Schieck et al. 1995; Dettmers et al. 2002). We determined habitat use and foraging behavior of male Black-and-white Warblers in forest fragments and in a contiguous forest tract in the aspen-domi- nated boreal mixedwood forest of central Alberta near the northwestern edge of the species’ range. Knowledge is slowly growing concerning the ecol- ogy of this species and other passerines in the west- ern boreal forest, but populations certainly face rap- idly changing conditions due to expansion of forestry, agriculture, and the energy sector (Schneider et al. 1999; Hobson & Schieck 1999; Hobson & Bayne 2000; Sykes & Hannon 2001). The objective of our study was to examine to what degree songbirds occupying forest fragments display behavioral flexibility while documenting the specific habitat requirements and responses of male Black- and-white Warblers in the threatened boreal mixed- wood forest. To do so, we addressed the following questions: (1) Within forest fragments, do Black-and- white Warblers establish territories in areas with par- ticular habitat characteristics? (2) Do territories of the Black-and-white Warbler in forest fragments differ in their habitat characteristics from territories in con- tiguous forest? and (3) Do foraging patterns of male Black-and-white Warblers differ between forest frag- ments and contiguous forest? Based on continental and regional patterns for the species, we did not ex- pect the Black-and-white Warbler to be a true habitat generalist, however, we did expect some flexibility in the species’ behavior. METHODS The study was conducted around Meanook Biolog- ical Research Station (54°37'N, 113°20'W) near Athabasca, Alberta, Canada. Today, approximately 30% of the landscape is covered by forest dominated by trembling aspen, balsam poplar, willow, alder, white birch, and white spruce (scientific names of trees and shrubs in Table 1). In the past 50 years, the remaining area has been gradually cleared and con- verted to pasture or crop land. We selected eight for- est fragments, dominated by mature (>80yr old) trembling aspen, that ranged from 2 to 140 ha in area (2, 4, 6, 9, 32, 50, 113, and 140 ha; areas were calcu- 76 lated from 1 : 30,000 aerial photographs using a Pla- con® digital planimeter). We defined a forest frag- ment as a wooded area separated from other wooded areas by =30m and/or connected to another wooded areas by a hedge/fencerow <10m wide (cf., Villard et al. 1995). All forest fragments were located within a 9km?* area and excluded patches open to cattle grazing. For comparison with forest fragments, a 30- ha plot was established in a contiguous forest at Nar- row Lake, about 13km from the nearest fragment. This area was >1000 ha, dominated by mature trem- bling aspen trees, and known to contain Black-and- white Warblers. It should be noted that throughout this study sample sizes are relatively small, simply because the Black-and-white Warbler is generally un- common. The species does not occur at high densities anywhere within its breeding range (mean=0.27 ter- ritories/ha [range=0.03-0.61]; data from 1993 Breeding Bird Census published in the Journal of Field Ornithology 1994 [supplement]). The study population, at the periphery of the species’ range, dis- played low densities (0.02—0.18 territories/ha; based on occurrences of individuals in occupied fragments and the contiguous forest plot). Bird locations All areas (fragments and contiguous forest) were flagged into 100 100m grids. Grids were surveyed between early May and the end of June three times in 1993 (fragments only) and twice in 1994 (fragments and contiguous forest) to locate male Black-and- white Warblers. Locations of male Black-and-white Warblers were plotted on gridded maps and later transferred to 1:30000 aerial photographs. Outer- most sightings for individual warblers were joined to obtain territorial boundaries. Although at least eight visits are recommended for spot-mapping (Interna- tional Bird Census Committee 1970), we believe our effort was sufficient because our primary interest was the presence/absence of warblers at a locality. In late June of both years, forest fragments unoccupied by male warblers were surveyed again to verify the ab- sence of birds. At 50m intervals, Black-and-white Warbler songs were played for 2 min, followed by 5 min pauses. No new warbler was detected during sur- veys, and because access to occupied areas within fragments took us regularly through unoccupied areas it is unlikely that we missed any territorial birds. 1) 2) Vegetation sampling Vegetation composition was measured in unoccu- Black-and-white Warbler habitat use pied fragments (UNF), in occupied fragments both within (WTF) and outside territories (OTF), and within territories in contiguous forest. Vegetation was sampled between early and mid-July using a modifi- cation of the circular sampling plot method (James & Shugart 1970). At randomly selected plots within each site (e.g. an unoccupied fragment), we recorded species/genus and number, height (using a clinome- ter) and diameter at breast height (using a dbh meas- uring tape) of all trees (>1.75 m tall), shrubs (woody plants 0.5—1.75 m tall) species/genus and number, and maximum shrub height. At a vegetation plot, trees were sampled within an 11.2m radius, but shrubs were sampled within a 5m radius. Further, at 20 ran- domly-selected locations within an 11.2 m radius of a vegetation sampling plot, we recorded presence or absence of canopy (>5m tall) and ground cover by looking through an ocular tube (4 mm in diameter) to estimate relative canopy closure and herb layer cover, respectively. Vegetation data were collected from 50 plots within 10 territories in four occupied fragments, 15 plots outside of territories in three occupied frag- ments, 20 plots within four unoccupied fragments, and 20 plots within four territories in contiguous for- est. We assumed that vegetation changed little be- tween years, therefore forest fragments were sampled only in 1993, except for one newly located territory in 1994. We used multivariate analysis to compare vegeta- tion composition within and among sites in the three forest fragments categories (UNF, WTF, OTF), and between territories in fragments and contiguous for- est. Principal components analysis (PCA) was per- formed using CANOCO (ter Braak & Smilauer 1998) Table 1. based on the 27 variables listed in Table 1, all of which were square-root transformed to approximate a normal distribution. Preliminary inspection of the data using Detrended Correspondence Analysis pro- duced gradient lengths <2, indicating that PCA was the appropriate ordination technique (ter Braak 1987). In order to determine whether habitat differed among the three occupancy categories in fragments and between territories located in fragments versus contiguous forest, multiple-response permutation pro- cedure (MRPP), a non-parametric equivalent of dis- criminant analysis (PCORD; McCune & Mefford 1997), was performed on vegetation data. 3) Foraging behavior and habitat use During May—June 1994, foraging data were col- lected for nine males in forest fragments and four males in contiguous forest. Individual warblers were observed weekly between 05:00h and 12:00h, on days without heavy rains or strong winds, for a total of 156h. Each male was visited on six to eight dates. Males were located using vocalizations and then tracked for up to 90min. Foraging data were col- lected by composing a timed, continuous record of an individual’s movements until it was lost from view. When an individual was relocated, a new timed se- quence began. While making observations, we also recorded foraging locations on gridded maps to vali- date territorial boundaries generated by spot-map- ping. We recorded: (1) species or genus of tree on which a foraging sequence (a series of movements that began when a bird landed on a plant and ended when it left) occurred (hereafter foraging tree), (2) sub- Vegetation variables measured from plots in forest fragments and a contiguous forest in central Alberta. Category Vegetation type No. trees (woody plants >1.75 m tall) Willow (Salix spp.), Trembling aspen (Populus tremuloides), Balsam poplar (P. balsamifera), White birch (Betula papyrifera), Alder (Alnus spp.), White spruce (Picea glauca) Tree sizes No. shrubs (woody plants 05—1.75 m tall) Mean tree height (m), Mean dbh (mm) Red-osier dogwood (Cornus stolonifera), Wild gooseberry (Ribes oxyacan- thoides), Wild red raspberry (Rubus idaeus), Honeysuckle (Lonicera spp.), Prickly wild rose (Rosa acicularis), Snowberry (Symphoricarpos spp.), Caragana sp., Viburnum spp., Prunus spp., Saskatoon berry (Amelanchier alnifolia), Maple sapling (Acer sp.), Balsam poplar sapling, Willow sapling, Trembling aspen sapling, Birch sapling Shrub height (mm) Cover* Mean maximum shrub height Canopy, Ground * See Methods for estimation of these variables. 77 C. A. PASZKOWSKTI et al. strates where foraging attempts were made (see Table 2 tor substrate categories), and (3) foraging method. Foraging methods observed were: glean (prey picked from substrate while perched), hawk (prey captured in mid-air while in flight), hover (prey picked from substrate while in flight without landing), and poke (bill inserted into a substrate to extract prey). Because Black-and-white Warblers usually feed on small prey (N. S. Sodhi, pers. obs.), we could not reliably deter- mine the success of capture attempts, therefore, we treat all attempts equally as indicators of foraging ef- fort. We used Redundancy Analysis (RDA; CANOCO; ter Braak & Smilauer 1998) to assess patterns of for- aging behavior (use of trees, substrates, and methods) of individual males within forest fragments and con- tiguous forest as constrained by habitat features of their territories. RDA is the appropriate form of direct gradient analysis to complement the indirect gradient analysis, PCA, that we previously chose for assessing vegetation patterns alone (ter Braak 1987). Frequency of occurrence of the behavioral categories described above were used to compare the total foraging records of nine territorial males from fragments with those of four males from contiguous forest. Vegeta- tion data used in RDA were based on samples from five plots within each territory, and consisted of mean abundance of the five tree species where foraging was observed: alder, balsam poplar, trembling aspen, white birch, and willow. All values were square-root transformed. MRPP analyses of foraging data matri- ces were used to determine whether behavioral pat- terns differed significantly between warblers occupy- ing territories in the two habitat types. RESULTS 1) Habitat characteristics of forest fragments Each year nine male Black-and-white Warblers oc- cupied four (2, 50, 113, and 140ha in area) of eight fragments. Occupied and unoccupied fragments re- mained the same in both years. Eight (89%) of nine territories found in 1993 were also occupied in 1994. In a given year, each transect encountered between one and four territories. Two of nine males in 1993 and four of nine males in 1994 fledged young (see also Sodhi & Paszkowski 1997). In our analysis of vegetation composition, the first two principal component axes accounted for 48% of the variation among plots (30% and 18% for axis | and 2, respectively). On the ordination plot for PC 78 axes | and 2, the three categories of vegetation plots from forest fragments (WTF, OTF, UNF) were clearly aggregated (Fig. 1A). Vegetation composition differed significantly in pairwise comparisons be- tween the three plot types (MRPP analyses: WTF vs. UNF, P=0.000; WTF vs. OTF, P=0.000; and OTF vs. UNF, P=0.007). Strongly positive “species” scores on the first PC axis for willow and alder reflected the dominance of these plants within the overstory of territories in oc- cupied fragments (Fig. 1A). The shrub layer on terri- tories was characterized by raspberry and gooseberry. Negative scores on the first axis for trembling aspen trees and saplings coincided with high densities of this species in areas within occupied fragments but outside territories (Fig. 1A). Vegetation on unoccu- pied fragments was also characterized by a relatively high abundance of aspen and absence of willow. Shrub species composition, however, was most influ- ential in distinguishing plots from unoccupied frag- ments, where the understory was dominated by dog- wood, rose, and viburnum. 2) Characteristics of territories in contiguous forest versus fragments Five males were located in the contiguous forest in 1994; three of these males successfully produced fledglings. In our comparison of vegetation character- istics of the territories of warblers located in contigu- ous forest versus territories in forest fragments, the first two PC axes accounted for 56% of the variation among plots (36% and 20% for axis | and 2, respec- tively). Plots differed distinctly between the two for- est types (Fig. 1B), with the first axis separating sam- ples from territories in contiguous forest versus frag- ments, and the second axis further separating territo- ries within fragments. MRPP confirmed that vegeta- tion profiles of territories in the two habitats differed significantly (P<0.001). As noted in the general analysis of vegetation patterns for fragments (Fig. 1A), territories were distinguished by willow in the overstory and by raspberry and gooseberry in the un- derstory, reflected in negative scores for these species on the first PC axis. In contrast, high positive scores on axis | for trembling aspen trees reflected domi- nance of this species within territories in contiguous forest with viburnum and rose contributing to a tall shrub layer (Fig. 1B). Black-and-white Warbler habitat use | Poplar Dogwood Honeysuckle Gooseberry Aspen Aspen sapling Ground cover Raspberry | + Aspen sapling Rose Viburnum | Willow Poplar Willow sapling Snowberry Raspberry Viburnum Tree height Tree DBH ® Occupied fragment within territory Occupied fragment outside territory A Unoccupied fragment Willow Alder Raspberry Gooseberry ——> ® Territory - fragment © Territory — contiguous forest Aspen Viburnum ————> Rose Shrub height + Raspberry Dogwood Gooseber Fig. 1. Principal component analysis ordinations based on 27 vegetation variables (see Table 1), comparing habitat features of plots in aspen-dominated forest fragments (2-140 ha) that were not occupied by Black-and- white Warblers, plots outside territories in occupied fragments, and plots inside territories (A), and comparing habitat features between plots within territories in fragments versus within territories in an aspen-dominated con- tiguous forest (>1000ha; B). Vegetation variables listed had the greatest positive or negative “species” scores on PC axis | or axis 2. 3) Foraging behavior in fragments versus contigu- ous forest Of the three aspects of foraging behavior that we examined via multivariate analysis (use of tree species, substrate, and method), only use of foraging trees differed significantly between males on forest fragments and those in contiguous forest (MRPP; P=0.001). Thus only results from RDA examining 19) tree use are presented here (Fig. 2). The first and sec- ond axes explained 38% and 11%, respectively of the variation in foraging behavior among individuals. The constrained ordination separated birds occurring in contiguous forest versus fragments. Males in con- tiguous forest most commonly foraged on trembling aspen, whereas males in fragments favored willow and balsam poplar as foraging trees (see also Table C. A. PASZKOWSKT et al. WILLOW Fig. 2. +Foraging tree e Male - fragment © Male — contiguous forest Redundancy analysis triplot summarizing the of use of tree species as foraging sites (crosses), behavior of individual birds (circles), and occurrence of these same tree species on territories (vectors) for 13 male Black- and-white Warblers in aspen-dominated forest fragments versus an aspen-dominated contiguous forest. 2). As indicated by the triplot, use of foraging trees paralleled the abundance of the five species in con- tiguous forest versus fragments. Foraging method and foraging substrates did not differ between male warblers in forest fragments and contiguous forest (MRPP: substrate, P=0.618; method, P=0.611). In both habitats, birds captured prey mostly by gleaning, which accounted for 90% and 78% of all feeding attempts observed in frag- ments and contiguous forest, respectively (Table 2). The most frequently used foraging substrate was branches, the site of 47% and 37% of all attempts in fragments and contiguous forest, respectively (Table 2). DISCUSSION Black-and-white Warblers at the southern edge of the boreal forest in Alberta displayed distinct patterns of habitat use at several spatial scales and behavioral patterns that, in turn, reflected the composition of the defended forest patch. The species was not a special- ist on forest-interior habitat as some authors have proposed (Freemark & Collins 1992; Morneau et al. 1999) and occupied half of the forest fragments sur- veyed. Occupancy was not dictated by patch size (the 80 smallest 2 ha fragment supported a territory), but was clearly related to finer scale vegetation patterns within individual patches. Several recent studies in the eastern United States have also shown the Black- and-white Warbler to be more variable in its habitat associations than previously described. Golet et al. (2001) and Lichstein et al. (2002) both reported that landscape features were important in determining the occurrence and abundance of the species at individual sites. The former study, for example, found that Black-and-white Warblers occurred in red maple (Acer rubrum) swamps as small as | ha in area if suf- ficient swamp forest was available in the larger land- scape. All fragments in our study were chosen based on a qualitative dominance of mature trembling aspen, yet the density of other woody species proved more im- portant in determining if a particular fragment sup- ported a warbler territory and where that territory was located. High densities of willow stems in the over- story characterized Black-and-white Warbler territo- ries in these forest fragments. All 50 of the plots sam- pled within territories contained willow trees, with an average of 9 stems per plot. Areas in occupied frag- ments that were outside of territories actually sup- ported higher densities of aspen trees and saplings Black-and-white Warbler habitat use Table 2. Foraging behavior of male Black-and-white War- blers in forest fragments (2-140 ha; n=9 individuals) and con- tiguous forest (>1000ha in area; n=4 individuals) in central Alberta. Values represent the mean and range of percentage of occurrence of foraging sequences (tree species) or prey at- tempts (substrate and method) associated with each variable from males from the two habitat types. (See Table 1 for scien- tific names of plants.) Contiguous forest Mean (range) Forest fragments Mean (range) Foraging tree Willow 50.7 (0-100) 15.6 (0-50) Alder 5.4 (0-33.3) 0 Trembling aspen 10.0 (0-71.4) 63.7 (40-88.4) Balsam poplar 20.5 (0-71.4) 20.6 (10-33.3) White birch 0 13.4 (0-100) No. of foraging 100 45 sequences® Foraging substrate? Trunk 17.0 (0-34.1) 24.3 (0-50) Limb 6.1 (0-16) 5.7 (0-S1.7) Branch 56.9 (32-100) 35.5 (28.5—42.5) Twig 12.6 (0-21.3) 19.4 (3.3-42.8) Leaf 3.0 (0-9.1) 7.1 (0-13.3) Air 4.0 (0-13.1) 13.2 (3.8-28.5) Foraging method Glean 90.1 (77.7-100) 79.6 (71.4—92.3) Hawk 4.1 (0-14.0) 11.5 (0-28.5) Hover 1.4 (0-8.3) 4.6 (0-12.3) Poke 3.3 (0-14.8) 7.1 (O-7.6) No. of captures* 410 139 “ Foraging sequence defined as series of movements that began when a bird landed on a plant and ended when it left that plant. > Foraging substrates (after Sodhi & Paszkowski 1995): trunk =wood>12cm diameter, limb=wood 2.6—12cm diameter, branch=wood 1.0—2.5cm diameter, and twig=wood<1 cm diameter. “ Number of prey captures was used to calculate the percent- age use of both foraging substrates and methods. than areas inside, but willow was virtually absent, with only one tree encountered in 15 plots. Domi- nance by trembling aspen, but with a decline in wil- low and a change in the composition of the shrub layer, typified the four unoccupied fragments. Ordi- nation indicated that the two categories of “unoccu- pied” plots were more similar to each other in vegeta- tion composition and structure than either was to plots within territories. It is of interest that plant species composition appeared to affect occupancy patterns of forest fragments more than did broader structural features of the canopy, shrub layer or 81 ground cover. Although our results cannot elucidate the specific causes behind this pattern for Black-and- white Warbler, Whelan (2001) demonstrated experi- mentally for three other species of wood warbler that the foliage structure of different deciduous tree species created distinctive foraging microhabitats that influenced warbler behavior. Based only on results from forest fragments, we might conclude that willow thickets offer preferred conditions for breeding Black-and-white Warblers in a boreal mixedwood setting. Yet when plots from ter- ritories in fragments are compared with vegetation samples from within territories in contiguous forest, the species’ patterns of habitat use become more complex (Fig. 1B). In the contiguous forest tract, re- sults more closely matched expectations for habitat preferred by Black-and-white Warblers in north-cen- tral Alberta derived from other studies (e.g. Schieck et al. 1995; Hobson & Schieck 1999), i.e., the most distinguishing feature within territories was a domi- nance of large and small trembling aspen. Aspen trees were present in 95% of the plots within territo- ries, and willows in only 60%; plots that did contain willow averaged only 3.4 stems per plot. In turn, in contiguous forest and fragments, males used the dominant tree species as their primary forag- ing site, aspen and willow respectively (Fig. 2). Dif- ferent tree species typically harbor different inverte- brate prey, or differ in the abundance or accessibility of shared prey types (Robinson & Holmes 1982; Holmes & Schultz 1988; Adams & Morrison 1993; Whelan 2001). Other studies have reported that bird species may or may not respond to shifts in plant species composition by using different foraging tac- tics in different habitat settings (Franzreb 1983; Block 1990; Petit et al. 1990; Miller & Cale 2000). We found that the use of aspen versus willow was the only behavior of Black-and-white Warblers that dif- fered significantly between the two habitats, and basi- cally mirrored tree availability (Table 2). Foraging followed the pattern typical of wood warblers, with a predominance of simple gleaning and some use of flight, a tactic that was somewhat more common in contiguous forest. Males displayed the species’ pecu- liar nuthatch-like behavior on both aspen and willow, with about one third of all prey captures occurring on larger diameter limbs and trunks where birds occa- sionally probed into bark for prey. The foraging methods of the Black-and-white Warbler may be ap- propriate for a variety of habitats and lend the species considerable versatility in the types of forest that it C. A. PASZKOWSKT et al. occupy even within a single geographic region Lichstein et al. 2002). Strong (2000) reported a simi- lar situation for over-wintering Ovenbirds (Seiurus aurocapillus) in the West Indies. This species’ ground foraging behavior, which focuses on capturing ants and is unusual for a wood warbler, allowed it to use a surprisingly diverse number of habitat types. The key distinction that we found between Black- and-white Warblers inhabiting human-created frag- ments and birds inhabiting a larger, contiguous forest tract was the strong link between warblers and wil- lows in fragments. One explanation for this pattern is that warblers were selecting for particular features other than dominant tree species when choosing frag- ments and establishing territories therein, and those other features were incidentally associated with wil- lows. The warblers’ foraging techniques allowed them to feed on willows as the most commonly en- countered component of the overstory vegetation. We documented, for example, strong differences in the abundance of a variety of shrub species between habitat types and in areas inside versus outside terri- tories. Black-and-white Warblers are ground nesters. The structure and composition of the shrub layer might affect habitat selection by shaping nest cover and microhabitat conditions, and affecting the likeli- hood of predation or parasitism (Martin & Roper 1988; Steele 1992; Ricketts & Ritchison 2000). War- blers might be establishing territories to encompass suitable nest sites. A second explanation for the observed linkage is that Black-and-white Warblers actively select frag- ments with willows because these trees are excep- tionally good foraging sites (based on arthropod abundance or delectability) and occur more densely embedded within open aspen stands. An abundance of willows can perhaps compensate for other short- comings of forest fragments as breeding sites (e.g. higher risk of nest predation). In a previous study (Sodhi et al. 1999), we reported a similar strong asso- ciation between American Redstarts (Setophaga ruti- cilla) and willows in some of the same mixedwood fragments. Bisson and Stutchbury (2000) reported that Hooded Warblers (Wilsonia citrina), a species often characterized as requiring large areas of mature forest, inhabited and had good breeding success in woodlots that contained a dense understory. They suggested that, at a finer spatial level, Hooded War- blers in mature stands might similarly be attracted to scattered canopy gaps associated with well-developed shrub layers. We found that all Black-and-white War- 82 bler territories in the contiguous forest contained some willow as well. Nonetheless, even in fragments, dependence on willow as a foraging site was not ab- solute as one of nine males was never seen foraging in willow and three additional males used other plant species more frequently. In summary, our study suggests that the Black-and- white Warbler will nest in relatively small patches of forest in the boreal mixedwood ecoregion. Warblers do not, however, behave as if these fragments were scaled-down versions of undisturbed stands. Different features of the vegetation determine where males es- tablish territories, and to some extent, males adopt different foraging tactics. In short, birds in fragments “adjust” to local conditions. Like the Black-and- white Warbler, many species of North American for- est passerines have wide ranges, but nowhere are they particularly common. Their biology undoubtedly varies across the continent, but is seldom documented because studies (such as ours) of these species 1n spe- cific ecological settings will be based on data from relatively few individuals. If we wish to learn about these species, and not be content with limiting re- search to very common or very rare species, or mak- ing faulty generalizations about the ecology of un- common species, we must acknowledge and accept the logistical constraints inherent in such studies. ACKNOWLEDGMENTS Thanks to E. Nash and S. Boss for assistance in the field and at the computer, and to R. Moses, C. Machtans, J. Kricher, and two anonymous reviewers for making comments on an earlier draft. This study was supported by the Recreation, Parks and Wildlife Foundation of Alberta, Canadian Circum- polar Institute, Natural Sciences and Engineering Research Council of Canada (grant to CAP), and James Anderson McAfee Postdoctoral Fellowship (to NSS). REFERENCES Adams EM & Morrison ML (1993) Effects of forest stand structure and composition on Red-breasted Nuthatches and Brown Creepers. J Wild! Manage 57: 616-629. Askins RA, Philbrick MJ & Sugeno DS (1987) Rela- tionship between the regional abundance of forest and the composition of forest bird communities. Biol Conserv 39: 129-152. Bayne EM & Hobson KA (1997) Comparing the effects of landscape fragmentation by forestry and agricul- ture on predation of artificial nests. Conserv Biol 11: 1418-1429. Black-and-white Warbler habitat use Bisson IA & Stutchbury BJM (2000) Nesting success and nest-site selection by a neotropical migrant in a fragmented landscape. Can J Zool 78: 858-863. Block WM (1990) Geographic variation in foraging ecologies of breeding and nonbreeding birds in oak woodlands. Stud Avian Biol 13: 264-269. Burke DM & Nol E (1998) Influence of food abun- dance, nest-site habitat and forest fragmentation on breeding Ovenbirds. Auk 115: 96-104. Dettmers R, Buehler DA & Franzeb KE (2002) Testing habitat-relationship models for forest birds of the southeastern United States. J Wildl Manage 66: 417-424. Ehrlich PR, Dobkin DS & Wheye D (1988) The birder’s handbook. Simon & Schuster Inc., New York. Flemming SP, Holloway GL, Watts EJ & Lawrance PS (1999) Characteristics of foraging trees selected by Pileated Woodpeckers in New Brunswick. J Wildl Manage 63: 461-469. Franzreb KE (1983) A comparison of avian foraging be- havior in unlogged and logged mixed-coniferous for- est. Wilson Bull 95: 60-76. Freemark K & Collins B (1992) Landscape ecology of birds breeding in temperate forest fragments. In: Hagan JM & Johnston DW (eds) Ecology and conser- vation of Neotropical migrant landbirds. pp 443-454. Smithsonian Institution Press, Washington, D.C. Golet, FC, Wang Y, Merrow, JS & DeRagon WR (2001) Relationship between habitat and landscape features and the avian community of red maple swamps in southern Rhode Island. Wilson Bull 113:217—227. Hinsley SA, Rothery P & Bellamy PE (1999) Influence of woodland area on breeding success in Great Tits Parus major and Blue Tits Parus caeruleus. J Avian Biol 30: 271-281. Hobson KA & Bayne E (2000) Effects of forest frag- mentation by agriculture on avian communities in the southern boreal mixed woods of western Canada. Wilson Bull 112: 373-387. Hobson KA & Schieck J (1999) Changes in bird com- munities in boreal mixedwood forest: harvest and wildfire effects over 30 years. Ecol Appl 9: 849-863. Holmes RT & Schultz JC (1988) Food availability for forest birds: effects of prey distribution and abun- dance on bird foraging. Can J Zool 66: 720-728. Hoover JP, Brittingham MC & Goodrich LJ (1995) Ef- fects of forest patch size on nesting success of Wood Thrushes. Auk 112: 146-155. International Bird Census Committee (1970) An inter- national standard for a mapping method in bird cen- sus work. Audubon Field Notes 24: 722-726. James FC & Shugart HH (1970) A quantitative method of habitat description. Audubon Field Notes 24: 83 727-736. Kricher JC (1995) Black-and-white Warbler (Mniotilta varia). In: Poole A & Gill F (eds) The Birds of North America, No. 158. pp 1-20. The Academy of Natural Sciences and The American Ornithologists’ Union, Washington, D.C. Lichstein JW, Simons TR & Franzreb KE (2002) Land- scape effects on breeding songbird abundance in managed forests. Ecol Appl 12: 836-857. Litwin TS & Smith CR (1992) Factors influencing the decline of neotropical migrants in a northeastern for- est fragment: isolation, fragmentation, or mosaic ef- fects. In: Hagan JM & Johnston DW (eds) Ecology and conservation of Neotropical migrant landbirds. pp 483-496. Smithsonian Institution Press, Washing- ton, D.C. Martin TE & Roper JJ (1988) Nest predation and nest- site selection of a western population of the Hermit Thrush. Condor 90: 51-57. McCune B & Mefford MJ (1997) PC-ORD. Multivari- ate analysis of ecological data, Version 3.18. MjM Software Design, Gleneden Beach. Miller J & Cale P (2000) Behavioral mechanisms and habitat use by birds in a fragmented agricultural land- scape. Ecol App! 10: 1732-1748. Morneau F, Doucet GJ, Giguere M & Laperle M (1999) Breeding bird species richness associated with a pow- erline right-of-way in a northern mixed forest land- scape. Can Field-Nat 113: 598-604. Morse DH (1989) American warblers. Harvard Univer- sity Press, Cambridge. Nour N, Matthysen E & Dhondt AA (1997) Effects of habitat fragmentation on foraging behaviour of tits and related species: does niche space vary in relation to size and degree of isolation of forest fragmenta- tion. Ecography 20: 281-286. O’Connor RJ (1992) Population variation in relation to migrancy status in some North American birds. In: Hagan JM & Johnston DW (eds) Ecology and conser- vation of Neotropical migrant landbirds. pp 64-74. Smithsonian Institution Press, Washington, D.C. Peck GK & James RD (1983) Breeding birds of On- tario: nidiology and distribution. Vol 2, Passerines. The Royal Ontario Museum, Toronto. Petit DR, Petit KE & Petit LJ (1990) Geographic varia- tion in foraging ecology of North American insectiv- orous birds. Stud Avian Biol 13: 254-263. Ricketts MS & Ritchison G (2000) Nesting success of Yellow-breasted Chats: effects of nest site and terri- tory vegetation structure. Wilson Bull 112: 510-516. Robinson SK & Holmes RT (1982) Foraging behavior of forest birds: the relationship among search tactics, diet, and habitat structure. Ecology 63: 1918-1931. C. A. PASZKOWSKI et al. Sauer JR, Schwartz S, Peterjohn BG & Hines JE (1996) The North American breeding bird survey. Home Page Version 95.1. Patuxent Wildlife Research Cen- ter, Laurel. Schieck J, Nietfeld M & Stelfox JB (1995) Differences in bird species richness and abundance among three successional stages of aspen-dominated — boreal forests. Can J Zool 73: 1417-1431. Schmiegelow FKA, Machtans CS & Hannon SJ (1997) Are boreal birds resilient to forest fragmentation? An experimental study of short-term community re- sponses. Ecology 78: 1914-1932. Schneider R, Lee P, Shank C, Farr D, & Stelfox B (1999) Conceptual framework and rationale for moni- toring forest biodiversity in Alberta. In: Farr DR, Franklin SE, Dixon EE, Scrimgeour G, Kendall S, Lee P, Hanus S, Winchester NN & Shank CC (eds) Monitoring forest biodiversity in Alberta: Program framework. Alberta Forest Biodiversity Monitoring Program Technical Report 3. pp 1-45. Edmonton. Smith JNM, Cook TL, Rothstein SI, Robinson SK & Sealy SG (Eds) (2000) Ecology and management of cowbirds and their hosts. University of Texas Press, Austin. Sodhi NS & Paszkowski CA (1995) Habitat use and for- aging behavior of four parulid warblers in a second- growth forest. J Field Ornith 66: 277-288. Sodhi NS & Paszkowski CA (1997) The pairing success of male Black-and-white Warblers Mniotilta varia in forest fragments and a continuous forest. Can Field- Nat 111: 457-458. Sodhi NS, Paszkowski CA & Keehn S (1999) Scale-de- pendent habitat selection by American Redstarts in 84 aspen-dominated forest fragments. Wilson Bull 111: 70-75. Steele BB (1992) Habitat selection by breeding Black- throated Blue Warblers at two spatial scales. Ornis Scand 23: 33-42. Strong AM (2000) Divergent foraging strategies of two neotropical migrant warblers: implication for winter habitat use. Auk 117: 381-392. Sykes AK & Hannon SJ (2001) Fencerows as habitat for birds in an agricultural lanscape in central Al- berta, Canada. Ecoscience 8: 441-449. Terborgh J (1992) Why American songbirds are vanish- ing. Sci Am 267: 98-104. ter Braak CJF & Smilauer P (1998) Canoco for Win- dows Version 4.0. Centre for Biometry Wageningen, CPRO-DLO, Wageningen. ter Braak CJF (1987) Ordination. In: Jongman RHG, ter Braak CJF & van Tongeren OFR (eds) Data analysis in community and landscape ecology. pp 91-169. Pudoc, Wagenningen. Villard M-A, Martin PR & Drummond CG (1993) Habi- tat fragmentation and pairing success in the Ovenbird (Seiurus aurocapillus). Auk 110: 759-768. Villard M-A, Merriam G & Maurer BA (1995) Dynam- ics in subdivided populations of neotropical migra- tory birds in a fragmented temperate forest. Ecology 76: 27-40. Whelan CJ (2001) Foliage structure influences foraging of insectivorous forest birds: an experimental study. Ecology 82: 219-231. Zanette L, Doyle P & Tremont S (2000) Food shortage in small fragments: evidence from an area sensitive passerine. Ecology 81: 1654-1666. Ornithol Sci 3: 85-92 (2004) ORIGINAL ARTICLE Estimation of hearing range in raptors using unconditioned responses Yumiko YAMAZAKI'*, Hiroko YAMADA7?, Mikio MUROFUSHP, Hiroshi MOMOSE* and Kazuo OKANOYA? ' Biopsychologie, Institut fiir Kognitive Neurowissenschaft, Fakultat fiir Psychologie, Ruhr-Universitéit Bochum, 44780 Bochum, Germany ? Department of Cognitive and Information Science, Faculty of Letters, Chiba University, 1-33 Yayoi-cho, Inage-ku, Chiba 263-8522, Japan 4 Ueno Zoo, 9-83 Ueno Kouen, Taito-Ku, Tokyo 110-8711, Japan * Landscape and Ecology Division, Environment Department, National Institute for Land and Infrastructure Management, Min- istry of Land, Infrastructure and Transport, I Asahi, Tsukuba, Ibaraki 305-0804, Japan Abstract We developed a new method to estimate the auditory abilities of animals ORNITHOLOGICAL using responses elicited by the presentation of auditory stimuli, without restraining or SCIENCE training the subjects. Using this method, we examined the hearing ranges of four rap- © The Omithological Society tors (a Mountain Hawk-eagle Spizaetus nipalensis, Northern Goshawk Accipiter gen- of Japan 2004 tilis, Common Buzzard Buteo buteo, and Grey-faced Buzzard-eagle Butastur indicus) kept in Ueno Zoo, in Japan, by presenting pure tones and white noise at two sound- pressure levels. Unconditioned responses, such as pupillary dilation and physical movements, were observed in all subjects. We then presented paired video clips of the raptors, with and without auditory stimuli, to human assayers, who were asked which clip contained the auditory stimulus. The accuracy of the human perceptual assay (HPA) suggested that the Mountain Hawk-eagle and Northern Goshawk hear frequen- cies from | to 5.7 kHz best, which is comparable to the results of an experiment with an American Kestrel Falco sparverius and European Sparrowhawks Accipiter nisus. The assayers reported that they used movements of the neck, head, and eyes, and changes in the pupils of the raptors as critical cues. Our method reliably reflected the hearing ranges of the raptors, and should be helpful for estimating the auditory capa- bilities of rare animals, such as the Mountain Hawk-eagle studied here. Key words Hearing range, HPA, Raptors, Unconditioned response Hawks and eagles, which belong to the order Fal- alarmed against them (Klump et al. 1986), that it is coniformes, are diurnal birds that hunt insects, birds, | dangerous advance, or that hatchlings/fledglings are and small mammals, such as rabbits, voles, and mice. hungry, and so on. Few studies, however, have exam- They usually sit on high perches, and dive at high _ ined the auditory abilities of Falconiformes, which do speed when they attack prey. Therefore, they are not appear to be exceptional among avian species strongly dependent on their excellent visual capacity (Fig. 1). Trainer (1946) measured hearing in the when hunting. Their excellent visual ability is sup- | American Kestrel Falco sparverius using conditioned ported by the evidence that some raptors have much __ responses elicited by, or emitted in response to, elec- bigger eyes than humans, relative to their body size, _ tric shocks that were preceded by pure tones. He and much greater visual acuity than humans (e.g., showed that the kestrel responded to sounds at fre- Wedge-tailed Eagle Aquila audax, Reymond 1985). quencies of 14kHz. Klump et al. (1986) measured Auditory stimulus is also important for raptors, be- the audible range of European Sparrowhawks Accip- cause it indicates that prey is approaching or that itis iter nisus using discrimination training through an operant conditioning technique, and found that the (Received 30 July 2003; Accepted 25 November 2003) hawks were most sensitive to sounds from | to 4 kHz. # Corresponding author, E-mail: yamyam@rondo.plala.or.jp Their best frequency was 2 kHz, while they could not 85 Y. YAMAZAKT et al. r sounds at frequencies higher than 8 kHz. This hearing range is average in relation to birds in gen- eral, but is much narrower than that of owls and some other raptors (Dyson et al. 1998). The differences in the hearing abilities of these birds likely evolved be- cause hawks are diurnal and they catch their prey from great distances using visual cues, whereas owls are nocturnal and they catch their prey from short dis- tances using auditory cues. For conservation purposes, it is necessary to know the auditory capabilities of raptors, especially those living in forested or mountain habitats, because such birds are often affected by human habitat-alternating activities such as logging, dam construction, and resi- dential developments. Noise from these projects may influence the birds’ reproductive or foraging success. Therefore, we need to be able to identify noise levels and understand how such noise might affect raptors. Unfortunately, we do not have sufficient data on the auditory ability of raptors, mainly because many of their populations are endangered, which makes it impossible to undertake experiments that require sur- gery or long periods of time for the collection of data, such as electrophysiological studies and operant dis- crimination training. Therefore, in order to determine the auditory capabilities of raptors, we must develop non-invasive methods that can be completed over short periods. One solution is to use unconditioned responses to sound. For example, Megela-Simmons et al. (1985) used a “reflex modification” technique to obtain au- diograms of the Bullfrog Rana catesbeiana and Green Tree Frog Hyla cinerea. In their technique, the unconditioned reflex elicited by the electric shock was inhibited when the pre-stimulus, say, pure tone, preceded the presentations of the shock if the animal had detected the pre-stimulus. The audiograms meas- ured by reflex modification technique agreed well with neural sensitivity data (Ehret & Capranica 1980; Shofner & Feng 1981). Recently, Bala and Takahashi (2000) measured the hearing curve in American Barn Owls Tyto alba using the pupillary dilation response elicited by the presentation of sound, and found that it was consistent with the data that Quine and Konishi (1974) obtained using operant conditioning. Al- though methods using unconditioned responses to sound have marked advantages over the operant con- ditioning method in that they do not require prior training, the subjects must be restrained during these experiments, and aversive stimulus or surgical opera- tion is necessary to obtain small behavioral changes. 86 Here, we establish a method of estimating the hear- ing ability of unrestrained, captive raptors using re- sponses to sound presentation, such as pupillary dila- tion, blinking, orienting responses, and small move- ments of the body. The subjects were four Japanese raptors (a Mountain WHawk-eagle Spizaetus nipalensis, Northern Goshawk Accipiter gentilis, Common Buzzard Buteo buteo, and Grey-faced Buz- zard-eagle Butastur indicus) kept in a zoo. We as- sumed that unconditioned responses would be ob- served only when sound was presented, and that such responses could be used as an index of hearing. We videotaped the behavior of the raptors with and with- out sound presentation, and then asked human assay- ers who were unaware of the experimental conditions to judge whether a given video clip contained sound presentation. We called this the Human Perceptual Assay (HPA). We assumed that if the assayers could correctly categorize the clips using responses to sound presentation as discriminative cues, the HPA would then reflect the hearing capability of the rap- tors. MATERIALS AND METHODS 1) Study Area Two experimenters recorded the raptors’ behavior on days when Ueno Zoo, Tokyo, was closed to the public, from 16 March 2001 to 9 July 2001. One ex- perimenter controlled the stimulus presentation, and the other videotaped the behavior of the subjects. Two loudspeakers were placed just in front of the birds’ metal cage, and the video camera was set at least 3m behind the loudspeakers. The distance be- tween the loudspeakers and the birds was 3 to 6m. The sound pressure levels were adjusted for distance so that they were equal where the birds were. After the recording, we conducted the HPA in a laboratory at Chiba University. 2) Subjects The subjects were four raptors kept in Ueno Zoo: a male Mountain Hawk-eagle, a male Northern Goshawk, a male Common Buzzard, and a Grey- faced Buzzard-eagle of unknown sex. The Mountain Hawk-eagle had been shot in the wild and taken to the zoo; it had recovered and was flying freely in its cage at the time of the experiment. It was kept in a cage by itself. The wings of the Northern Goshawk and Common Buzzard had been injured, and neither bird could fly; these two hawks were kept in a cage Hearing range raptors with five Ural Owls Strix uralensis. The Grey-faced Buzzard-eagle could fly, and lived with two Common Kestrels Falco tinninculus. Three male volunteers (22—30 years-old) took part in the HPA as assayers. 3) Materials The auditory stimuli were controlled by a CD player (SL-SW404, Panasonic) and broadcast from two loudspeakers (YST-M100, Yamaha). The behav- ior of the raptors was recorded with a digital video camera (VL-MRI1 PRO, Sharp). We used a personal computer (PCV-J15, Sony) and video presentation software (Adobe Premiere 5.1 for Windows, Adobe) for the HPA. 4) Stimuli The sound stimuli were created using Avisoft-SAS Lab Pro software (Avisoft) with a sampling fre- quency of 44.1 kHz. The stimulus duration was 0.8 sec with rise-fall times of 50ms. We prepared pure tones at eight different frequencies (0.25, 0.5, 1, 2, 4, 5.7, 8, and 11.3kHz) and white noise. The frequen- cies of the pure tones were determined using the oc- tave scale, except for 5.7 kHz, which was the geomet- ric mean between 4 and 8 kHz. Each sound had two sound pressure levels, 50+4dB and 74+4dB, and the difference between the two levels was 24 dB. 5) Procedure Sound presentation and video recording of the hawks The test sounds were presented from in front of the cage when the birds were not moving and there was relatively little external noise. The sounds were presented using the constant stimuli method (Klump et al. 1995); 1.e., the stimuli were presented in ran- dom sequences within and among sessions. The inter- stimulus interval was at least 30sec, and it was usu- ally much longer, either because the birds were mov- ing or there was external noise. The sound and no- sound conditions were alternately presented after the inter-stimulus interval. In the sound condition, we presented an auditory stimulus 3 seconds after push- ing the button on the CD player. The procedure in the no-sound condition was the same as in the sound con- dition, except that no sound was broadcast. We video- taped mainly the upper part of each bird, including the head, shoulders, and abdomen, in these sessions. A session consisted of 18 sound conditions, which consisted of one sound at each of the nine frequencies and two sound-pressure levels, and 18 no-sound con- 87 ditions. Experiments were conducted on three sepa- rate days for each subject, with one session per day. Therefore, 54 video clips with sound and 54 clips without sound were obtained for each subject. HPA To determine whether the birds responded to the sound stimuli and not to the no-sound condi- tion, we conducted the HPA. In the sound condition, the original video records were edited so that in a 4.8- sec video clip, the sound presentation period was in the middle 0.8 sec of the clip. A 4.8-sec clip of the no-sound condition was edited similarly. Two video clips, one from the sound condition and the other from the no-sound condition, were randomly paired and presented to the assayers of the HPA in windows on each side of a PC monitor, without any auditory information. Before the assay, we told the assayers five things: 1. Look carefully at the video clips presented on the left and right sides of the screen, one at a time; 2. Each clip contains a raptor, and is about 5 sec- onds long; 3. In one clip, a sound was played to the hawk for about half the length of the clip; no sound was played in the other clip; 4. Determine which clip contains the sound; 5. After the presentation, check the blank space with the corresponding trial number on your response sheet using a pencil. Note that because the assayers were not told about critical cues that might suggest the presentation of the sound, they were free to use any cues in the clips to make their decision. Before presenting the test clips to each assayer, eight training trials were run using video clips that were not used in the subsequent test session: four each from the sound and no-sound conditions. In the training session, the clips in the sound condition con- tained obvious behavioral changes to ensure that the assayers understood their task. During training, the assayers were shown one clip per trial, and were asked to answer yes or no to whether the clip con- tained a sound presentation. The researcher gave feedback (correct or incorrect) after each answer. The sequence of the trials was such that the same condi- tion (sound or no-sound) was not repeated more than three times in a row. The assayers had more than 50% correct responses to each bird condition. The training session was only given once regardless of how the as- sayers performed. In the test session, clips were presented on the left Y. YAMAZAKT et al. »ht sides of the monitor in each trial. In one clip, sound was presented to the bird, and in the other it was not. The conditions (sound or no-sound), posi- tions of the clips (left or right), and the starting clip of a given trial (left or right) were determined using quasi-random sequences, in which the same condi- tions were not repeated more than three times in a row. One block consisted of 18 trials, which included nine different sound conditions at two sound-pressure levels, and 18 clips from the no-sound condition. These clips were selected from the three separate ex- periments for each raptor. A session consisted of six blocks, and the 54 different clips were presented twice; therefore, 108 trials per session were run. The assayers were given all three test sessions, one for each raptor with a different sequence of trial blocks, and the orders for the four raptor species were coun- terbalanced among the assayers. After the HPA trials, the assayers were asked the following questions: 1. What were the critical cues or rules they used in making their decisions? Were there distinctive characteristics or con- sistent changes in the behavior of the different raptors? How difficult was it to make a decision, and what was your general impression of the tasks? Calibration Calibration was performed by plac- ing a sound-level meter with a 1/3 octave band filter at 3, 5, and 6m from the loudspeaker. These were the distances at which the birds perched during the ex- periment. Ambient noise level was also measured at the same distances for each 1/3 octave. For the 250- Hz test frequency, some of the harmonic distortion products were as intense as 10dB below the funda- mental. Therefore, the data for the 250-Hz tone was excluded from further analyses. For the other test fre- quencies, harmonic distortion within the audible range (assuming 15 kHz at most) was at most 30dB below the fundamental at each test frequency for each harmonic. The 1/3 octave noise level was con- verted into a spectrum level by subtracting the loga- rithm of the bandwidth. The signal level was at least 50 dB above the ambient noise. This suggests that no masking occurred at the test frequencies, and that the thresholds obtained here reflect true, unmasked thresholds. i 6) Statistics We calculated the response accuracy of three as- 88 sayers in each bird condition for each sound condi- tion (Fig. 2). The average accuracy of the HPA was calculated for the three assayers in the six trials for each sound condition, in which each raptor condition was presented once per block. The accuracy at fre- quencies of | to 4kHz was calculated separately (upper panel of Fig. 3), because it is assumed that the raptors can hear sounds at these frequencies (Fig. 1; Trainer 1946; Klump et al. 1986). Consistency of the decisions made by the assayers was also calculated at these frequencies to determine whether they used the same cues for their decisions in each trial. If all the assayers made the same decision (sound presence or absence), then it was counted as |. If there was a dif- ference among the decisions, then it was counted as 0. We performed a binomial test to examine the dif- ference between the accuracy or consistency and ex- pected value. The chance level was set at 0.5 for ac- curacy, and 0.25 for consistency. To examine the dif- ference of the accuracy and the consistency among birds, one-way analysis of variance (one-way ANOVA) was conducted, and then Tukey’s HSD test was used for multiple comparisons. In addition, we calculated the kappa kK to measure the agreement of the judgment by the assayers (Siegel & Castellan 1988). 80, -—@~ American Kestrel 70+ -{- European Sparrowhawk = OW Y) ea) ce) 0 0:25:05 ay 2 4 8 10 Frequency (KHz) Fig. 1. Audiograms of the American Kestrel (Trainer 1946) and European Sparrowhawk (replotted from Klump et al. 1986). Hearing range raptors 0 @ SsP-low Q Mountain pee Northern ae Hawk-Eagle te Tota 20 Goshawk Common Buzzard 0.5 1 2 457 811.3 WN 051 2 457 811.3 WN Grey-Faced Buzzard-Eagle 051 2 45.7 811.3 WN Frequency (kHz)/ White Noise Fig. 2. Response accuracy of the human perceptual assay (HPA) of raptors, with SD. The numbers and letters on the horizontal axis represent the sound frequency and white noise (WN) presented to the raptors. There are separate plots for the low- (SP-Low) and high- (SP-High) pressure levels, and total (TOTAL), which was calculated regardless of the sound-pressure level. Note that the vertical axis is reversed (the origin is at the top of the y-axis) to compare the shapes of the curves with audiograms obtained in other studies (Fig. 1). RESULTS The average accuracy was calculated for the total and the two sound-pressure levels (Fig. 2). The total accuracy for the Mountain Hawk-eagle was 77.1%, for the Northern Goshawk 69.1%, for the Common Buzzard 60.1%, and for the Grey-faced Buzzard- eagle 52.8%. Comparison between the accuracy of the HPA at frequencies of | to 4kHz and the chance level probability (upper panel of Fig. 3) revealed a significant difference for the Mountain Hawk-eagle (87.0%, N=108, z=7.60, P<0.01) and for the North- ern Goshawk (75.0%, N=108, z=5.10, P<0.01), but not for the Common Buzzard (57.4%, N=108, z=1.44, P>0.05) or Grey-faced Buzzard-eagle (40.7%, N=108, z=1.83, P>0.05). There was a sig- nificant difference among the birds (N=12, F=9.25, 89 df=3, P<0.01), and following multiple comparison revealed that there was significant difference between the Grey-faced Buzzard-eagle and the Mountain Hawk-eagle (N=3, P<0.01), and the Northern Goshawk (N=3, P<0.05). The assessment of the assayers (Fig. 2) was that they were 80% accurate for sounds at frequencies of 1 to 5.7kHz in the Mountain Hawk-eagle, in both sound-pressure conditions. There is a steep decrease in accuracy from 5.7 to 11.3kHz under low sound- pressure conditions, and from 8 to 11.3kHz under high sound-pressure conditions. For the Northern Goshawk, the accuracy level of HPA was more than 80% from 1 to 5.7kHz, but only under high sound- pressure conditions. The difference in the human per- formance between the sound-pressure conditions was clearer for this bird than for any of the other three Y. YAMAZAK et al. sy “e s Accuracy ® 80 S 60 x2 40 B 20 [> Ong * »<0.01 se 100 Consistency > 80 = ou i) a 40 S 20 O 0 MHE NGH CB GFBE Fig. 3. Response accuracy (upper) and consistency (lower) of the three human assayers in the HPA of each raptor (MHE: Mountain Hawk-eagle; NGH: Northern Goshawk; CB: Com- mon Buzzard; GFBE: Grey-faced Buzzard-eagle). The data were collected from the responses to the video clips with sound frequencies of 1, 2, and 4kHz. In the upper panel, the solid line indicates the chance level of performance (50%). Consistency means the percentage of trials in which all three assayers made the same decision (sound presence/absence). The solid line indicates the chance level (25%). Asterisks indi- cate statistically significant differences between the data and a given chance level. raptors. Under the low sound-pressure condition, the accuracy did not reach 80% at any frequency. The shapes of the curves obtained from the HPA of these two species were similar to those of the audiograms in Trainer (1946) and Klump et al. (1986). By contrast, the performance in the HPA for the Common Buzzard and Grey-faced Buzzard-eagle (Fig. 2) oscillated markedly, depending on the sound frequency. In the Grey-faced Buzzard-eagle, the greatest accuracy under the high sound-pressure con- dition was for 11.3-kHz, the frequency at which the humans were not so accurate for the other three rap- tors. For the Common Buzzard, it was the highest for 5.7 kHz, and the lowest for 8 kHz. In these two birds, no consistent change in accuracy was observed with sound condition, for either frequency or sound pres- sure level. Except for the Grey-faced Buzzard-eagle, the accu- racy in response to the white noise condition was 90, comparatively high, and for the higher sound-pres- sure level it increased to around 80%. The interviews of the assayers after the HPA ex- periment revealed that all of them used movements of the neck, head, and pupil, blinking, and behavioral changes seen in the video clip as cues for their deci- sions. Although the Mountain Hawk-eagle was not generally active, the assayers reported that its behay- ior changes were the most distinct of the four species, and one assayer noticed whether it blinked with one or both eyes. In the Northern Goshawk, two of the as- sayers reported that they divided the video clips into two categories based on the distinctiveness of the be- havior changes. In the Common Buzzard, all of the assayers had difficulty detecting pupillary changes, because the border between the pupil and the iris is not clear in this species. The consistency of the decision among the assay- ers (lower panel of Fig. 3) was above chance level (25%) in the Mountain Hawk-eagle (83.3%, N=36, z=7.89, P<0.01), the Northern Goshawk (50.0%, N=36, z=3.27, P<0.01), and the Common Buzzard (58.3%, N=36, z=4.43, P<0.01), but not in the Grey-faced Buzzard-eagle (40.7%, N=36, z=0.19, P>0.05). The difference among them was significant (one-way ANOVA, N=24, df=3, F=9.96, P<0.01). Further comparison between each species revealed that the consistency of the Mountain Hawk-eagle was higher than that of the Northern Goshawk (Tukey’s HSD test, HSD=1.73, N=6, P<0.05), and the Grey- faced Buzzard-eagle (N=6, P<0.01), and that of the Common Buzzard was higher than that of the Grey- faced Buzzard-eagle (N=6, P<0.05). Thus, it was clear that both the response accuracy and consistency were high for the Mountain Hawk-eagle, whereas they differed for the Northern Goshawk and the Com- mon Buzzard (higher response accuracy and lower consistency in the former, and they were reversed in the latter). By contrast, both of them were lower in the Grey-faced Buzzard-eagle. Kappa statistic also showed the same trend, that the agreement of the judgment was significantly above chance for the Mountain Hawk-eagle (k=0.51, z=2.18, P<0.05) and Common Buzzard (K=0.43, z=4.30, P<0.01), but not for the Northern Goshawk (k=0.11, z=0.76, P>0.05) or the Grey-faced Buzzard-eagle (k=0.01, z=0.09, P>0.05). DISCUSSION This study showed that the responses of the raptors Hearing range raptors to sound presentation and the accuracy of our HPA could be used to measure their hearing ranges. The validity of our method was suggested by the similar- ity between the response accuracy to the Mountain Hawk-eagle and Northern Goshawk and the audio- grams obtained in the prior experiments. The Ameri- can Kestrel and European Sparrowhawk heard fre- quencies from | to 4kHz best (Fig. 1), and the re- sponse accuracy of the HPA in the Mountain Hawk- eagle and Northern Goshawk was greater to sounds from | to 5.7 kHz (Fig. 2). In addition, the Mountain Hawk-eagle and Northern Goshawk showed clear and consistent responses to sounds within that range, as suggested by the performance that was signifi- cantly different from chance in those species. The high accuracy and consistency in HPA in the Moun- tain Hawk-eagle suggest that the assayers used the same behavioral cues that were well coincident with the sound presentation and had small variability. By contrast, in the Northern Goshawk, although all the assayers reported that it was easy to detect behavioral changes in this species, the high accuracy with lower consistency than that of the Mountain Hawk-eagle suggests that the assayers used the various behavioral cues that coincided with the sound presentation. It is possible that the Northern Goshawk was more sensi- tive to the sound stimuli than any of the other species. Assuming that there should not be a big difference in the auditory abilities of these species of raptor, the audibility range was from 0.5 to 8 kHz, with the best hearing range from | to 6 kHz. This study might be criticized on the basis that the hearing ability of the raptors may not be identical with the response accuracy in the HPA. We also admit that our methodology cannot be used in all cases. Our observations indicated that the behavioral changes in response to sound presentation in the Common Buzzard and Grey-faced Buzzard-eagle were inconsistent. These birds often moved actively after being presented with sounds at frequencies of 8 or 11.3kHz, which are reported to be beyond the hearing range of European Sparrowhawks (Klump et al. 1986), whereas they did not respond to sounds at | or 2 kHz, which are thought to be within their ranges. This inconsistency in their responses is why the per- formance of the HPA for these species was not statis- tically different from chance. By contrast, the Moun- tain Hawk-eagle and Northern Goshawk consistently showed behavioral changes to the sounds, and this was reflected in the higher HPA performance (upper panel of Fig. 3). The reasons for these differences 9] may be owing to the native habitats of these species. Mountain Hawk-eagles and Northern Goshawks usu- ally perch in trees while foraging and roosting, whereas Common Buzzards and Grey-faced Buzzard- eagles live and hunt in open fields in flatlands. The former two species would be more used to hearing sounds distorted by obstacles than the latter two species. In the present experiment, pure tones were broadcast from speakers set in front of each bird’s cage. Thus, it is possible that the tones were more strange to the Mountain Hawk-eagle and the North- ern Goshawk than to the Common Buzzard and the Grey-faced Buzzard-eagle, and that they would be more sensitized. In addition, such habitat conditions would affect their movements on hearing the sounds. Mountain Hawk-eagle and Northern Goshawk may move actively in order to detect environmental changes visually, whereas Common Buzzard and Grey-faced Buzzard-eagle may move inactively in order not to be seen by preys in flatlands. Since our method relies on unconditioned responses to sound, it is not useful for species with small or few behavioral changes, such as Common Buzzards and Grey-faced Buzzard eagles. It is necessary to consider the spe- cific ecology and behavior of the subject before ap- plying our method. To confirm the hypothesis out- lined above, we must increase the number of subjects to determine the generality of the behavioral charac- teristics to the sound observed in this study. One could argue that the birds kept in a zoo may have become de-sensitised or habituated to certain auditory stimuli because they are continuously ex- posed to the sounds made by visitors. However, the pure tones used in the present study do not exist in a natural environment, and it was presumed to be the first time for the birds to hear these sounds. Addition- ally, the coincidence of the behavioral change with the sound presentation frequently observed in the subjects, precluded the possibility of loss of sensitiv- ity to such kind of sounds. Because the raptors other than the Mountain Hawk-eagle were kept in their cages together with other birds such as owls which are known to have greater sensitivity to the sound than hawks, consider- ation must be given to the fact that the subject birds could have used their behavioral changes to the sounds. However, such an effect would have been small, because the Common Buzzard, which showed inconsistent behavioral changes to the sounds, was kept in the cage with the Northern Goshawk which proved to be more sensitive than the buzzard. Y. YAMAZAK et al. Ve cannot say that the function (Fig. 2) represents the absolute hearing threshold in these raptors. How- ever, the accuracy curve of the HPA measured using two different sound-pressure levels could be consid- ered an equal-loudness curve for the raptors, reflect- ing the audible range under given conditions. By run- ning additional tests using sounds at the best-heard frequencies and different sound-pressure levels, we should be able to estimate other auditory properties, such as absolute thresholds and frequency cutoffs. In conclusion, our method is useful for estimating the hearing ranges of some raptors, and does not re- quire training or restraints. The accuracy of the HPA reflected the hearing ranges of the birds. In addition, because we presented the sounds in front of the birds’ usual cages, the possible effects of interference should have been minimal. Our method can be ap- plied to animals in captivity and possibly even to those in the wild. ACKNOWLEDGMENTS We thank the staff of Ueno Zoo for their assistance over the course of this study. This study was supported by the Research on Conservation Techniques Program of the Ministry of Land, Infrastructure, and Transport of Japan. Portions of this paper were presented to the meeting of the Technical Committee of Psychological and Physiological Acoustics in 2001 in Chiba University, Japan, and at the 4th International Symposium on Physiology and Behaviour of Wild and Zoo Animals in 2002 in Berlin, Germany. We also thank the various referees for their critical comments on the manuscript. REFERENCES Bala ADS & Takahashi TT (2000) Pupillary dilation re- sponse as an indicator of auditory discrimination in the barn owl. J Comp Physiol A 186: 425-434. Dyson ML, Klump GM & Gauger B (1998) Absolute hearing thresholds and critical masking ratios in the European barn owl: a comparison with other owls. J Comp Physiol A 182: 695-702. Ehret G & Capranica RR (1980) Masking patterns and filter characteristics of auditory nerve fibers in the green tree frog (Hyla cinerea). J Comp Physiol 141: 1-12. Klump GM, Dooling RJ & Stebbins WC (1995) Meth- ods in comparative psychoacoustics. Birkhauser Ver- lag, Basel. Klump GM, Kretzschmar E & Curio E (1986) The hear- ing of an avian predator and its avian prey. Behav Ecol Sociobiol 18: 317-323. Megela-Simmons A, Moss CF & Daniel KM (1985) Be- havioral audiograms of the bullfrog (Rana cates- beiana) and the green tree frog (Hyla cinerea). J Acoust Soc Am 78: 1236-1244. Quine DB & Konishi M (1974) Absolute frequency dis- crimination in the barn owl. J Comp Physiol 93: 347-360. Reymond L (1985) Spatial visual acuity of the eagle Aquila audax: A behavioural, optical and anatomical investigation. Vision Res 25: 1477-1491. Shofner WP & Feng AS (1981) Post-metamorphic de- velopment of the frequency selectivities and sensitivi- ties of the peripheral auditory system of the bullfrog, Rana catesbeiana. J Exp Biol 93: 181-196. Siegel S & Castellan NJ (1988) Nonparametric statis- tics for the behavioral sciences. McGraw-Hill, New York. Trainer JE (1946) The auditory acuity of certain birds. Dissertation (Ph.D.), Cornell University, New York. A List of Referees 2003 The editors of Ornithological Science extend hearty gratitude to the referees, who kindly and voluntarily helped the advancement of ornithology. The figures in parentheses denote the number of papers they refereed. Hitoha AMANO, Shigeki ASAT, Richard CORLETT, Yasuo EZAKI, Shoji HAMAO, Yuko HAYASHI (2), Kazuto KAWAKAMI, Kazuhiro EGUCHI (2), Shigeru MATSUOKA, Taku MIZUTA, Masashi MURAKAMI, Hisashi NAGATA, Sumio NAKAMURA, Yoshito OHSAKO, Hidetsugu SAKAI, Shigeho SATO, Navjot SODHI (5), Hitoshi TOJO, Noriyuki YAMAGUCHI, Hoshiko YOSHIDA Rewriter: Mark BRAZIL Keisuke UEDA, Yoshiyuki BABA, Barry BROOK Go FUJITA (2), Hiroko FUJIWARA Aki HIGUCHI, Teruaki HINO (5) Taku MAEDA, Hajime MATSUBARA Mizuki MIZUTANI, Hiroshi MOMOSE Hisashi NAKAGAWA, — Masahiko NAKAMURA (2) Nariko OKA, Chan-Ryul PARK Yuji SAWARA, Tetsuo SHIMADA E1ichiro URANO (2) Abstracts of the Japanese Journal of Ornithology, Volume 52 Number 1 REVIEW Egg production and its physiological regulation in domestic fowl. Kiyoshi IMAI Domestic fowl, Gallus gallus domesticus, is the most important poultry (domesticated birds that serve as a source of egg or meat to human beings) in the world. The process of egg production in the female bird is composed of follicular rapid growth in the ovary, ovulation of the largest follicle, egg formation in the oviduct and oviposition. In the present review, the Ovipository pattern, the egg producing process and its physiological regulation, especially endocrine con- trol, in the hen are described. The period required to follicular rapid growth is 8 days with a highest fre- quency, ranging from 7 to 10 days in almost all the follicles examined. Ovulation of the largest follicle in the ovary is mainly controlled by pituitary LH and follicular progesterone. Egg formation, being consti- _ tuted by egg white, shell membrane and egg shell surrounding yolk, is conducted in the oviduct during about 24 hours. Vasotocin and prostaglandins play an important role participating in oviposition. Jpn J Ornithol 52: 1-12. 2003. ARTICLES Mate Guarding in the Black-faced Bunting Em- beriza spodocephala. Akira KOIWAI The overlap of home ranges, copulation and mate guarding behaviour were studied in socially monoga- mous Black-faced Buntings Emberiza spodocephala from April to July in 1992 at Sugadaira (36°32'N, 138°20’E, 1,250 m alt). Each male bunting had a sta- ble home range and a song area within that. Since males frequently intruded into neighbouring home ranges to seek extra-pair copulations, the home ranges overlapped greatly. Song areas also over- lapped with each other. Once males acquired mates, they did not sing frequently. These suggest that males did not maintain the song area as a territory after pair formation. Males guarded their mates by spending uch time near their mates and by following them closely. The shortness of intra-pair distance and the percentage of flights in which the female was fol- lowed by her mate peaked during the presumed fer- tile period. Males started mate guarding soon after pair formation. All 29 intra-pair copulations were ob- served during the fertile period. Aggressive behaviors between males due to attempts of extra-pair copula- tion were observed between pair formation and the fertile period. The half of 10 extra-pair copulations were observed during the fertile period. It means that males could not defend their mates completely. These results suggest that male Blackfaced Buntings guard their mates rather than exclusive areas, while they seek opportunities to attempt extra-pair copulations. Black-faced Buntings live in the brushy and visually occluded habitats. These habitats would prevent males from performing mate guarding and territorial defence simultaneously. Males responded aggres- sively to extra-pair male intrusions which started and continued after pair formation and took part in incu- bation. Male-biased sex ratio in this area also makes males to give priority to mate guarding over territo- rial defence to assure their paternity. Jpn J Ornithol 52: 13-23. 2003. COMMENT Management of the Great Cormorant-comments on Hayama (2002). Masahiro FUJIOKA Jpn J Ornithol 52: 24-28. 2003. SHORT NOTES Food item found in pellets of Long-eared Owls wintering in Ehime, Japan. Satoshi KAWAGUCHI and Takahito YAMAMOTO Sixty seven pellets of Long-eared Owls, Asio otus, were collected in a village, Ehime Prefecture, during 19 February 2001 to 6 April 2001. The bones of 164 Mus musculus, ten Micromys minutus, two Apodemus speciosus, two Rattus sp., seventeen Pipistrellus abramus, two Myotis macrodactylus, twelve Cro- cidura dsinezumi and a bird were found in the pellets. M. minutus, P. abramus, M. macrodactylus and C. dsinezumi were new records in diet of the Long-eared owls. Jpn J Ornithol 52: 29-31. 2003. 94 Distribution of feeding sites of wintering Greater White-fronted Geese in Lake Izunuma-Uchinuma. Tetsuo SHIMADA Distribution of feeding sites of wintering Greater White-fronted Geese Anser albifrons around Lake Izunuma-Uchinuma (roost) was determined by direct observations during winter of 1997, 1998 and 1999. The 68.6% of 110 feeding flocks were observed in rice fields within 6 km from the lake. Jpn J Ornithol 52: 32-34. 2003. Effects of body mass on daily mass increment in Rhinoceros Auklet chicks receiving different amount of foods. Tomohiro DEGUCHI and Tatsuhiko KAGAMI Daily measurement of chick mass could be a useful method to estimate daily food intake under the as- sumption that the effects of the body mass on daily mass increment do not differ between chicks receiv- ing different amount of food. To evaluate this as- sumption, 40 g and 60 g anchovy per day were fed to Rhinoceros Auklets chicks in semi-captivity. Nega- tive relationships between chick mass and daily mass increment were found in 40 g- and 60 g-fed chicks. There was no significant difference in the slope of re- gression equation between 40g-fed chicks (—0.037+0.018) and 60g-fed chicks (—0.041+ 0.016). The intercept in 60g fed chicks (17.578+ 3.491 g¢) was larger than that in 40g fed chicks (10.950+3.491 g). Therefore, the daily amounts of food intake in Rhinoceros Auklet chicks can be esti- mated from body mass and daily mass increment. Jpn J Ornithol 52: 35-38. 2003. A parasitological survey of Hwamei Garrulax canorus and Red-billed Leiothrix Leiothrix lutea (Passeriforms: Terimiidae). Tomoo YOSHINO, Kazuto KAWAKAMI, Hitoshi SASAKI, Kenji MIYAMOTO and Mitsuhiko ASAKAWA, As one of ecological surveys of alien avian species in Japan, internal and external parasites of 4 individuals of Hwamei Garrulax canorus and 4 individuals of Red-billed Leiothrix Leiothrix lutea collected in Kanagawa, Tokyo, and Fukuoka Prefs., Japan during from July 1999 to November 2001, were investi- gated. Ornithoica bistativa, Ornithomya avicularia cf. daobatoensis, Haemaphysalis flava and Cen- trorhynchus turdi were obtained from Hwamei, and Ornithonyssus sylvialum and Anonchotaenia sp, were obtained from the Leiothrixes, respectively. All of these external parasites and the acanthocephalan species are newly recorded from these two avian species in Japan. Jpn J Ornithol 52: 39-42. 2003. NUMBER 2 REVIEWS Roles of behavioral science in conservation at the population level. Go FUJITA Application of recent advances in behavioral science is a growing subject in conservation biology. One possible reason for the growth of this approach is that behavioral scientists have become aware of the im- portance of conservation through their experience during field studies. In addition, there are three more plausible factors that can explain the approach. |. As conservation biology has developed, both conserva- tionists and behavioral scientists have recognized the importance of behavioral science in many cases of conservation activity. 2. Behavioral ecology has ma- tured and many behavioral ecologists have shifted their interest to the application of their subject. 3. Studies of interactions between individual behavior and population dynamics have become active. Here, I briefly review model studies that have attempted to apply the behavioral sciences to conservation activi- ties. The transfer and captive breeding of endangered populations often requires intensive conservation ac- tivity, such as the mating of individuals under condi- tions that differ from their natural habitats. Such ac- tivities require knowledge of the natural behavior and habits of the target species. Predicting the population dynamics of target animals is usually required to evaluate the effects of habitat change resulting from human disturbance on these populations. There are new approaches that incorporate behavioral processes into models of population dynamics. Certain models that include behavioral processes, foraging patch choice under game theory parameters, have suc- ceeded in predicting population parameters such as winter mortality. Mating systems, sexual selection, regulation of the birth sex ratio and helping behavior are all suspected of affecting the effective population size of the target species. Helpers, in addition, can also serve to the function of absorbing the impacts of environmental changes on the population. Jpn J Ornithol 52: 71-78. 2004. How can behavioral studies on individual animals contribute to wildlife conservation? Masato MINAMI The roles of behavioral studies of individuals in the conservation of wildlife, which should be important for the conservation of birds also, are briefly intro- duced and discussed in the light of two cases studies of the Asiatic Black Bear Ursus thibetanus and Sika Deer Cervus nippon in Japan. Some local populations of the bear, one of Japan’s largest mammal species, are in danger of extinction. At Karuizawa town, Nagano Prefecture, central Japan, some bears have been visiting garbage disposal sites. Sixteen bears were captured, collared, and tracked using radio- telemetry, from July 1998 to September 2001. Three individuals were killed because they were staying within the town and judged as dangerous individuals. The other 16 bears were translocated. This individu- als based management with radio-telemetry data can minimize risks of local population extinction of bears. On Kinkazan Island, Miyagi Prefecture, north- east Japan, most members of a 150-strong Sika Deer population were individually recognizable and have been tracked since 1989. Changes in body conditions, such as morphological and nutritional condition of in- dividuals, and lifetime reproductive success were recorded. The data suggested that variances in life- time reproductive successes among individuals were high, and the population dynamics can be strongly re- lated to survival and reproductive success of the vari- ation among individuals. This long-term study of the deer showed that reproductive strategies of individu- als affected population dynamics. While, this study also showed that the population density, food avail- ability, and stochastic variation in environment might seriously affect on survival and reproductive success of each individual. Jpn J Ornithol 52: 79-87. 2004. Behavioral approaches to agricultural damage and goose conservation. Katsumi USHIYAMA, Tatsuya AMANO, Go FU- JITA and Hiroyoshi HIGUCHI Conflict between geese and agriculture has become a widely acknowledged problem in recent decades. Some of the goose species involved in such conflict are coincidentally considered to be of high conserva- interest. Behavioral ecology provides a useful ework for uniting agricultural damage manage- jent and conservation in a common scientific disci- pline. This paper summarizes the behavioral studies of habitat use by geese and their application in allevi- ating goose-agriculture conflicts. Behavioral ap- proaches to habitat use by geese are useful in deter- mining the underlying mechanism whereby damage occurs in a particular space and time, and enables ef- ficient targeting and setting of management measures. Such management measures include the provision of alternative feeding areas combined with scaring de- vices, and habitat manipulation through the manage- ment of farming methods and human disturbance. From a broader perspective, understanding site selec- tion by geese has important implications in guiding redistribution schemes, which is particularly impor- tant as the concentration of a population in a limited number of sites can have deleterious effects both on agriculture and on goose conservation. This behav- ioral knowledge can be integrated into a predictive model of goose distribution and population dynamics. Given evolutionary backgrounds, behavior-based models have the capability of predicting population level consequences of animals in response to habitat change. Such a model may serve as a cornerstone in predictive management and conservation. Jpn J Ornithol 52: 88-96. 2004. ARTICLES Formation of foraging flocks using recruitment calls in Jungle Crows Corvus macrorhynchos. Masayo SOMA and Toshikazu HASEGAWA Recruitment calls are reported as one mechanism used by some birds and mammals for attracting con- specifics once an individual has found a food source, hence leading to the formation of foraging flocks/ groups. Jungle Crows Corvus macrorhynchos some- times give “kakaka” calls while foraging in flocks. We conducted field observations and playback exper- iments to examine whether these calls serve as re- cruitment calls leading to flock formation. Crows gave “kakaka” calls while foraging, and the size of foraging flocks was positively correlated with the number of calls given. Broadcasting “kakaka” calls attracted more Jungle Crows. There was also a posi- tive correlation between pecking rates and the size of foraging flocks. The results indicate that “kakaka” calls are effective recruitment calls and that Jungle 96 Crows forage more efficiently in flocks. Jpn J Ornithol 52: 97-106. 2004. A Comparison of the feeding behavior of sym- patric Varied Tit Parus varius and Great Tit P. major. Akiko URATA and Keisuke UEDA To study the food size preferences of Great Parus major and Varied tits P. varius we provided three sizes of sunflower seeds at a feeder at the forest edge on Mt Aburayama, Fukuoka, southern Japan. The Varied Tits preferred larger seeds than the Great Tits, and selected the size of seed that gave them greater feeding efficiency. The Great Tits carried larger seeds away from the feeder, whereas they ate smaller seeds immediately at the feeder. The Varied Tits always se- lected larger seeds at the feeder. It is suggested that the larger bill of the Varied Tit may affect this species difference in food size selection. Jpn J Ornithol 52: 107-111. 2004. SHORT NOTES Partly albino fledglings in two Carrion Crow Corvus corone orientalis families. Masaki KURO-O and Reiko KATAKURA Four partly albino fledglings were observed in two Carrion Crow Corvus corone orientalis families in Hirosaki City, Aomori Prefecture, Japan, from 30 June to 9 August 2002. They all had white parts on the sub-terminal regions of the upper and under pri- mary coverts, upper and under greater coverts, pri- maries, secondaries, and tertiaries. Two fledglings in one family had white in the subterminal regions of the rectrices as well as in the remiges. The latter two birds stayed with their parents for about 20 days longer than their single normally plumaged sibling. Jpn J Ornithol 52: 112-115. 2004. The parasitic helminths of avian species in Niigata Prefecture, Japan. Shigeru NAKAMURA, Tomoo YOSHINO, Jun SATO, Akira CHIBA and Mitsuhiko ASAKAWA As part of ongoing research into the conservation of wild avian species in Niigata Prefecture, helmintho- logical examination was conducted between March and November 2002 in Niigata Prefectural Bird Pro- tection Center, Japan. A total of 50 individuals of 28 avian species was investigated, with parasitic helminthes collected from 26 individuals. The para- sitic helminths belonged to 13 nematode, two acan- thocephalan, and three trematode genera, and uniden- fied cestodes were also collected. Three genera, Epo- midostomum (host: Anser albifrons), Viktorocara (host: Fulmarus glacialis) and Diomedenema (host: Ardea cinerea) were new locality records in Japan. Jpn J Ornithol 52: 116-118. 2004. OBSERVATIONAL DATA The capture record of Taczanowski’s Grasshopper Warblers Locustella pleskei at the Hososhima, Miyazaki Prefecture. Yasuhiro YAMAGUCHI, Hiroshi KIKUCHI, Tetsuro KAMETANTI, Kyoko YAMAGUCHI and Shinichirou 97 UENO Jpn J Ornithol 52: 119-121. 2004. Record of Dovekie Alle alle in Japan. Yutaka NAKAMURA, Yutaka TAKANOHASHI, Akira ASO, Toshiko TAKANOHASHI, Naomi SYUTO, Sayuri SATO and Isako SHIOMOKAWA. Jpn J Ornithol 52: 122-123. 2004. A record of Ferruginous Duck Aythya nyroca in Yonago Warterbirds Sanctuary, Tottori Prefec- ture. Keisuke KIRIHARA Jpn J Ornithol 52: 124-125. 2004. Ornithological Society of Japa Membership Applicatio Office of the Ornithological Society of Japan, Laboratory of Biodiversity Science, School of Agriculture and Life Sciences University of Tokyo, Yayoi 1-1-1, Bunkyo-ku Tokyo 113-8657, JAPAN Phone: +81-3-5841-7541 Fax: +81-3-5841-8192 Email: osj@lagopus.com Contact information ([] Contact details have shaneed L] No change) Name: Address (include postal code): Phone: Fax: Email: Membership categories Type of membership (Please check one of the following boxes): C] Ordinary member Annual dues ¥5,000 (Regular member) CL] Student member Annual dues ¥3,000 (Members with confirmation of student status) C) Sustaining member Annual dues ¥10,000 (Member supporting the Society economically) CL] Institutional member Annual dues ¥10,000 (Universities, companies, non-governmental organizations, research groups) Certification of student status (to be completed by student’s supervisor): Er , certify that is a Student registered at the institution below. Name of university: Supervisor’s signature: )ther payments Be ations to the Society ve Purchase of “Check-list of Japanese Birds” (¥ 4,380/copy) ¥ Subtotal ¥ Total of membership dues and other payments: ¥ Methods of payment (Please check selected method of payment) a International Postal Money Order (Address to the Ornithological Society of Japan) o Direct deposit into the Society’s bank account Account information: Bank: Mizuho Bank, Nezuekimae Branch _ Account type and number: Regular account 1743740 Account administrator: President of the Ornithological Society of Japan, Dr. Hiroyoshi Higuchi OU Credit card (Please check either VISA or Mastercard. We are unable to accept any other cards.) C1 VISA CL] Mastercard Credit card number: Expiry (mo/yr): / Name on card (First-last): _ a The Ornithological Society of Japan President Hiroyoshi HIGUCHI (The University of Tokyo, Tokyo) Vice President Hiroshi NAKAMURA (Shinshu University, Nagano) Council Members Manabu ABE (Raptor Japan, Tokyo) Kazuhiro Eguchi (Kyushu University, Fukuoka) Yasuo EZAKI (Museum of Nature and Human Activities, Sanda) Yuzo FUJIMAKI (Bibai) Masahiro FUJIOKA (Tsukuba University, Shizuoka) Koichi KARASAWA (Ichikawa) Noritomo KAWAJI (Forestry and Forest Products Research Institute, Tsukuba) Ken ISHIDA (The University of Tokyo, Tokyo) Hisashi NAGATA (National Institute for Environmental Studies, Tsukuba) Masahiko NAKAMURA (Joetsu University of Education, Joetsu) Kiyoaki OZAKI (Yamashina Institute for Ornithology, Abiko) Keisuke UEDA (Rikkyo University, Tokyo) Satoshi YAMAGISHI (Yamashina Institute for Ornithology, Abiko) Instructions to Authors Ornithological Science is the English journal published by the Ornithological Society of Japan (OSJ) twice a year. Member- ship to OSJ is not required for contributors. The contribution can be classified into four categories: Reviews, Original Arti- cles, Short Communications, Technical Notes and Comments. Comments are only limited to the papers published on Ornithological Science. Accepted papers automatically be- come the copyright property of OSJ and may not be repro- duced, in whole or part, without a written permission from OSJ. Submission of Manuscripts Manuscripts have to be original and so far unpublished. Papers prepared by Microsoft WORD or PDF files are recommended to be submitted as an e-mail. Authors who do not have access to the Internet may send 3 complete photocopies (the original not required at the time of reception) of manuscript, tables and figures. Floppy disks (3.5-inch) containing a text file of the manuscript are also required after acceptance. Manuscripts are peer-reviewed by two anonymous referees. In case the contributors are requested to revise their manuscripts, they are supposed to resubmit then in consideration of the comments by reviewers within a month with a note describing how revision has been carried out. The final acceptance or re- jection for publication is decided by the Editor-in-Chief. The address for submission is: Dr. Teruaki Hino, Editor-in-Chief Kansai Research Center, Forestry and Forest Products Re- search Institute, Kyoto 612-0855, JAPAN. E-mail: tkpk @affre.go.jp TEL: +81-75-611-1201, FAX: +81-75-611-1207 Organization of Manuscripts Manuscripts should be printed, double-spaced, on one side only of A4 sheets (210X297 mm), with one-inch margins and 20 lines per sheet. Original Articles should be presented in the following order: (1) Title page (2) Abstract (3) Introduction (4) Materials and Methods (5) Results (6) Discussion (7) Ac- knowledgements (8) References (9) Tables (10) Figures and Legends. Short Communication is required not to exceed four printed pages and no abstract is needed. All manuscript sheets should be consecutively numbered. The use of footnotes is not recommended. A charge of 2,500 yen will be added to every page that ex- ceeds 20 printed pages. Each color page will be charged 50,000 yen. General Instructions 1) Cover page The title page should contain the title, author names, affilia- tions, addresses, contact name and address (phone and/or fac- simile number together with an e-mail address) for correspon- dence. Authors with different affiliations should be identified with superscripts. On the same page, a running title (less than 40 letters) and the number of tables and figures are also re- quired. To discriminate the family name(s) from the first and/or the middle owns, contributors are encouraged to spell their family names in all capitals. 2) Abstract and key words Abstracts should be less than 300 words, followed five or less key words in alphabetical order. 3) Names of animals and plants Scientific names of species should be given both in the ab- stract and in the article at the first mention. Scientific and Eng- lish names of birds should follow an authoritative source. Cap- itals should be used for the initial letters for each word of Eng- lish names (e.g., Black-billed Magpie) but not for a group name (€.g., Crows). 4) References References in the text: Tokugawa (1995) or (Tokugawa 1995) Oda and Smith (1996) or (Opel & Smith 1996) Mori et al. (1997) or (Mori et al. 1997; Richard et al.1999a) Journals: Yamaguchi N & Kawano KK (2001) Effect of body size on the re- source holding potential of male varied tits Parus varius. Jpn J Ornithol 50: 65-70. Books: Cambell RC (1974) Statistics for biologists. 2nd ed. Cambridge Univ Press, London. Chapters in a book: Dawson WH (1996) Energetic features of avian thermoregulatory responses. In: Carey C (ed) Avian energetics and nutritional ecol- ogy. pp 85-124. Capman & Hall, New York. Papers written in a language other than English: Yamagishi S (1981) Mozu no yomeiri—toshikouen no mozu no- seitai wo saguru (The bridal of shrikes — ecological research of bull-headed shrikes in an urban park). Dai-Nippon-Tosho, Tokyo (in Japanese). Electronic materials: Rahbek C & Graves GR (2001) Multiscale assessment of patterns of avian species richness. Proc Natl Acad Sci USA 98: 45344539 (online). Prast W & Shamoun J (1997) Bird Remains Identification System, Springer-Verlag, Berlin and Heidelberg (CD-ROM). Manuscripts accepted for publication but not yet published should be listed in the reference section as “in press”. All con- tributors are advised to prepare a list of references cited in the text arranged alphabetical by the author names. The sources of the reference should follow the commonly accepted abbrevia- tions for journal titles (refer to the ‘International List of Peri- odical Title Abbreviations’). The use of “in preparation,” “submitted” or “personal communication” is not allowed in the reference list. 5) Units and Abbreviations All measurements should be given in SI units. Time should be shown as 0930 and 1500 and the geographical location as 35°20'25”N, 136°10'20"E. Statistical symbols should be capi- talized except df (e.g., SD, SE, P, N). Italicized expressions are not preferred. 6) Tables and Figures The contributors are supposed to take the size of the printed page into consideration when they prepare their tables and fig- ures: the actual print is on 165X223 mm in double columns with 80 mm width. ORNITHOLOGICAL SCIENCE Volume 3. Number 1 SPECIAL FEATURE Invasive bird species Sodhi NS & Eguchi K Introduction Eguchi K & Amano HE Spread of exotic birds in Japan Kawakami K & Yamaguchi Y The spread of the introduced Melodious Laughing Thrush Garrulax canorus in Japan Tojo H & Nakamura S Breeding density of exotic Red-billed Leiothrix and native bird species on Mt. Tsukuba, central Japan Brook BW Australasian bird invasions: accidents of history? Leven MR & Corlett RT Invasive birds in Hong Kong, China Yap CAM & Sodhi NS Southeast Asian invasive birds: ecology, impact and management 23 35 43 aT January 2004 Contents ORIGINAL ARTICLES Nakamura M & Murayama S Are Carrion Crows that congregate in spring roosts juveniles or adults? 69 Paszkowski CA, Sodhi NS, Jamieson S & Zohar SA Habitat use and foraging behavior of male Black-and-white Warblers (Mniotilta varia) in forest fragments and in a contiguous boreal forest 75 Yamazaki Y, Yamada H, Murofushi M, Momose H & Okanoya K Estimation of hearing range in raptors using unconditioned responses 85 A List of Referees 2003 93 Abstracts of the Japanese Journal of Ornithology, Volume 52 93 Published by the Ornithological Society of Japan Printed by Kokusai Bunken Insatsusha Co., Ltd. ISSN 1347-0558 HISTORY MUSEUM 22 DEC 2004 PURCHASED TRING UBRARY ORNITHOLOGICAL SCIENCE Vol.3 No.2 October 2004 The Ornithological Society of Japan. ORNITHOLOGICAL SCIENCE Official journal of the Ornithological Society of Japan Editor-in-Chief Teruaki Hino, Forestry and Forest Products Research Institute, Kyoto Associate Editor Hidetsugu Sakai, Nihon University, Tokyo Editorial Board Kimiya Koga, Akan International Crane Center, Akan Masashi Murakami, Hokkaido University, Tomakomai Masahiko Nakamura, Joetsu University of Education, Joetsu Isao Nishiumi, National Science Museum, Tokyo Kazuo Okanoya, Chiba University, Chiba Navjot S. Sodhi, National University of Singapore, Singapore Eiichiro Urano, Yamashina Institute for Ornithology, Abiko Advisory Board Alexander V. Andreev, Institute of Biological Problems, Magadan Walter J. Bock, Columbia University, New York Jiro Kikkawa, The University of Queensland, Brisbane Woo-Shin Lee, Seoul National University, Suwon Bernd Leisler, Max-Planck-Gesellschaft, Radolfzell Richard Noske, Northern Territory University, Casuarina Pilai Poonswad, Mahidol University, Bangkok Lucia Liu Severinghaus, Academia Sinica, Taipei Jeffrey R. Walters, Virginia Polytechnic Institute and State University, Blacksburg John C. Wingfield, University of Washington, Seattle Jeong-Chil Yoo, Kyung-Hee University, Seoul Editorial Policy Ornithological Science publishes reviews, original articles, short communications and comments covering all aspects of ornithology. Manuscripts are judged on the basis of their contribution of original data and ideas or interpretation. All articles are peer-reviewed by at least two researchers expert in the field of the submitted paper. Manuscripts are edited where necessary for clarity and economy. Ornithological Science aims to publish as rapidly as is consistent with the requirements of peer-review and normal publishing constraints. Submission Manuscripts and editorial correspondence should be ad- dressed to: Dr. Teruaki Hino, Editor-in-Chief, Forestry and Forest Products Research Institute, Kyoto 612-0855, Japan. Tel: +81-75-611-1201, Fax: +81-75-611-1207, e-mail: tkpk @affre.go.jp For detailed instructions concerning the submission of man- uscripts, please refer to the Instructions to Authors at the in- side of the back cover of each issue of the journal or visit the OSJ web page: http://wwwsoc.nii.ac.jp/osj/ Subscriptions Membership of the Ornithological Society of Japan (OSJ) is open to anyone interested in ornithology and the aims of the OSJ. The annual fee for an ordinary member is Japanese Yen 5,000. Members are entitled to receive two journals. Ornithological Science (in English; two issues per year) and Japanese Journal of Ornithology (with English summary; two issues per year). To join the OSJ, please apply to: The OSJ Office, Laboratory of Biodiversity Science, School of Agricultural and Life Sciences, The University of Tokyo, Yayoi 1-1-1, Tokyo 113-8657, Japan. Tel: +81-3-5841-7541, Fax: +81-3-5841-8192, e-mail: osj@lagopus.com Copyright Submission of a manuscript implies: that the work de- scribed has not been published before; that is not under con- sideration for publication elsewhere; that its publication has been approved by all co- authors, if any, as well as by the re- sponsible authorities at the institute where the work has been carried out; that, if and when the manuscript is accepted for publication, the authors agree to automatic transfer of the copyright to the OSJ; and that the manuscript will not be pub- lished elsewhere in any language without the consent of the copyright holders. ISSN: 1347-0558. All articles published in the Ornithological Science are pro- tected by copyright, which covers the exclusive rights to re- produce and distribute the article (e.g. as offprints) as well as all translation rights. No material published in this journal may be reproduced photo-graphically or stored on microfilm, in electronic data bases, video disks, etc., without first obtaining written permission from the copyright holders. Printer: Kokusai Bunken Insatsusha Co., Ltd., Takada-no-baba 3-8-8, Shinjuku-ku, Tokyo 169-0075, Japan. Tel: +81-3-3362- 9741, Fax: +81-3-3368-2822. Cover design: Eiichiro Urano An announcement of “The Workshop for Novice and Amateur Ornithologists” BrFest—ls2e BOR] OMS CAB RS CS, BAUR OHEEIIDAKLH, KEVV REVO LORE ECORMRBMATI oO TWOET. ZO-KRELT, HRDERKILHASDAEP AH, MICLOLAOKAE DPSEKWEWISREMRIC, F HhGteitzx— (BOR REST SZCLICBROELK. B2M4AW FOLIC HVEF. 2G BOFR: mCeBCS |! BECAME PHERECAUSICBHSE TIE CULKID. Ek, FHHOELHADMOPSTFICEAY Tit TCLs dD. MAPR Ms HoH lc KAA NCC Za EUCHY, KOMAICRWU6EEF. SHOR DDFs He * POLED AGEN, mMEBX, SR MRRICMe SHB CLAAHBLET. Amid FEI A D Fe ts (Cee OC ° AHR ° BU ERT CARL, MTS CECH. MMEHBEKW LUI HROBM AEBESELTWOEF. COBMSHOREBITHA T+ —-—AlOLKMoT, HABICHSAZ< BHA -—IVT (Fax, MR ay) HLUATESW. BK, FHOBROFRISE IES BRE COMETH bDNET. Woes BTeSat, MMtEOKS, SSZUIEECACHBOKCEMKUT VF a THRE, EASE * EFA A CEO MBMULCEO EWA, BAHEE CILARSEBMUO LET. A Aimee <4 BH, MCHA KK CE* BUTIOU KOC EIT ARE MOCOKHNH TY « Hi, (AE TH NIE) BMA VN-OF—4F CRM ERHLTHESWEN LOW CORR HET -VeRMeEL, MARTY eH, MaRS OT, MMILERRMULHROULPKE RHA ms ff: ABR (ED SE ORNBAAABAOAB, mK VV cA hb CoRRSE FA PAREZ (NPOVEAZS— FY FRE, Strix ME, HHI HO & PR: BRR tis BAA el CIR A BRERA 5 GEAR 8 3h) http://www. ins.kahaku.go.jp/index.html Bf : 2005 4F2 9 220 CX) ~ 248 B A: treet QAO) ALIAS MWBpots : ee T 108-0071 FRR ABE Ae 5-21-5 FEEL 4 Be Ot BRAC Tel: 03-3441-7176, Fax: 03-3441-7012 E-mail: hamao@kahaku.go.jp FLIAAOBIZ FOC LABS PFA + BAG A (7S IL EHE TS) 9 ELEM “mes 777 ABS "BPL-NT RV S2IPHH Snisesi led) q CSHBEA) NP Rte OB EE CAH OP bee HISTORY MUSEUM 22 DEC 2004 PURCHASED TRING UBRARY a ee Ornithol Sci 3: 99-112 (2004) ORIGINAL ARTICLE Estimation of nutrients delivered to nest inmates by four sympatric species of hornbills in Khao Yai National Park, Thailand Pilai POONSWAD!*, Atsuo TSUJI’ and Narong JIRAWATKAVI' ' Department of Microbiology, Faculty of Science, Mahidol University, Bangkok, Thailand ? Department of Mathematics, Faculty of Science and Technology, Meijo University, Nagoya, Japan Abstract Hornbills are omnivorous and the breeding male delivers all food required ORNITHOLOGICAL by the nest-confined female and chicks. The contributions of different food types, in SCIENCE terms of breeding nutrition, have not previously been documented. In Khao Yai Na- © The Ornithological Society tional Park, Thailand, we sampled the identity and number of food items delivered of Japan 2004 daily to the nest, during each week of the nesting cycle, by two small and two large sympatric species of hornbills. We then recorded the mass and estimated the nutrient content of each food type from analyses of protein, fat, carbohydrate, calcium, and energy. The overall pattern of nutrient delivery during the nesting cycle was the same for each of the four hornbill species, and was related to sequential demands for egg, feather, and chick development. The two larger species delivered mainly carbohy- drates (Great Buceros bicornis 50%, Wreathed Aceros undulatus 57%) and less fat and protein. The smallest, Oriental Pied Hornbill Anthracoceros albirostris, also de- livered mostly carbohydrate (45%), but the small White-throated Brown Hornbill An- orrhinus austeni delivered equivalent proportions of protein (32%), fat (30%), and carbohydrate (37%). Comparison of the incubation and nestling phases showed that more protein was delivered during the nestling phase for all species, except for Great Hornbill where the compression of egg production, incubation, and molt had to be completed by midway through the nestling phase and so high levels of fat and protein were delivered during incubation. We confirmed that fruits are an important source of all nutrients, especially fat, for all four hornbill species, but suggest that delivery of animal protein may be linked, in some way, to breeding success. Oriental Pied Horn- bill broods, that received protein at about 1.05% of brood mass per day, had the high- est breeding success (96%) whereas Wreathed Hornbills received only 0.57% protein and had only 67% success, while the other two species delivered intermediate amounts of protein and had intermediate breeding success. Key words Breeding nutrition, Frugivore, Hornbill, Nesting diet, Omnivore, Thai- land Breeding hornbills, like many other birds, must nest cavity about a week before egg-laying and, once collect sufficient food to satisfy both their own nutri- laying commences, starts her annual molt of rectrices tional requirements and those of any offspring they and remiges, further increasing the nutritional de- raise. In hornbills, provision of these nutrients is the | mand imposed on the male (Poonswad 1993; Kemp responsibility of the male because the female seals 1995). herself into the nest and is fed by the male throughout There is some variation between hornbill species in the egg-laying, incubation and, with the chick(s), the the timing of the female’s emergence or whether as- nestling phase (Kemp 1995). In most hornbill sistance is available to the male from members of a species, the breeding female, after a phase of group. Species also differ in size, duration of the courtship feeding by the male, seals herself into the nesting cycle and number of eggs laid and chicks raised (Poonswad et al. 1987; Kemp 1995), all of (Received 13 May 2003; Accepted 25 July 2004) which influence the nutrient requirements of nest in- * Corresponding author, E-mail: scpps@mucc.mahidol.ac.th mates. Food delivered to the nest also varies between 99 P. POONSWAD, A. TSUJI and N. JIRAWATKAVI species, habitats, seasons and phases of the nesting cycle. It may depend in part on availability, but also on nutrient demands, such as calcium for egg or skeleton formation, or particular amino acids for feather growth. The diets of many bird species are predominately either frugivorous or carnivorous during breeding, but in omnivores, such as forest hornbills, both food types occur regularly in the diet (Poonswad et al. 1987; Kemp 1995). The proportion of nutrients con- tributed by different food types at different phases of the nesting cycle has not previously been recorded for hornbills. Four species of hornbills breed in Khao Yai National Park, central Thailand, two large species, Great Hornbill Buceros bicornis and Wreathed Hornbill Aceros undulatus, and two smaller species, White-throated Brown Hornbill Anorrhinus austeni and Oriental Pied Hornbill Anthracoceros al- birostris. Their dietary choices have been reported (Poonswad et al. 1987, 1998), as have differences in their size, nesting cycle and breeding biology (Poon- swad 1993: Table 1). Here, we concentrate on intra- specific patterns and sources of nutrients delivered to nest inmates during the incubation and _ nestling phases of the nesting cycle, based on the identity and analysis of food items delivered to the nest inmates; an attempt at inter-specific comparisons is also pre- sented. METHODS We conducted this study in the 70 km? Khao of Yai National Park (ca. 14°15-30'N, 101°20—24’E), Thai- land, an area of tall monsoon forest (62 km’) with patches of open grassland (8km*; Poonswad et al. 1998). We observed hornbill nests during the breed- ing seasons (January-June) of 1982-1985 and recorded all food items delivered (0700-1700) to the nesting female and/or chicks by the breeding male and any other helpers. We observed each hornbill species at 2-10 day intervals, to provide weekly sam- ples of full-day observations that covered the whole nesting cycle (Table 2). Breeding success, i.e. the number of chicks fledging per nest, was also recorded. We observed food items delivered to the nest using binoculars (8 X30), a spotting scope (X20 and X40), or took photographs (using 400-800 mm telephoto lenses), depending on the distance from the nest. We identified items as fruit or animal and classified them by direct observation, by collection of debris below 100 Attributes of size and breeding biology for four species of hornbills that breed in Khao Yai National Park, Thailand (from Poonswad 1993 unpublished data; Kemp 1995, 2001). Table 1. Exceptional Female + mean Mean brood size, Clutch size, range Nesting cycle, weeks Median female Hornbill species breeding biology brood mass, g ee (range, sample) (incubation/nestling) (ratios) (ratios) (mean) mass (g) Female emerges 6 weeks after chicks hatch 71 3,316 (2.1) 1(0,N 20 (7/13) (1.4) 14 PON Great 60 29 3,900 (2.5) =27) 1 (0, N 1-3 1-5 1,950 20 (7/13) (1.4) 755 Wreathed Cooperative group, 1—5 2,718 (1.7) 2.6 (1-4, N=15) 15 (5/10) (1.1) White-throated Brown helpers for male at nest 14 (5/9) (1.0) 1-3 1.5 (1-2, N=12) 1,560 (1.0) 29 624 Oriental Pied Nutrition of nesting hornbills in Thailand Table 2. Sample sizes for observations at nests of four sympatric species of hornbills in Khao Yai National Park, Thailand. In- cubation and nestling phases were determined from direct observation or published information, and the incubation phase was as- sumed to include the pre-laying interval of one week (Poonswad 1993; Kemp 1995). No. of full-day observations, mean days/week and phase duration No. of Total no. Hornbill species pair- of days Incubation phase Nestling phase years observation Full-days Days/week Phase duration Full-days Days/week Phase duration Great 16 437 78 11.1 Weeks 1-7 119 9.2 Weeks 8-20 Wreathed 9 227 35 5.0 Weeks 1-7 63 48 Weeks 8-20 White-throated 14 258 46 9.2 Weeks 1-5 92 Oo Weeks 6-15 Brown Oriental Pied 24 449 62 12.4 Weeks 1-5 157 17.4 Weeks 6—14 the nest, or by comparison with specimens collected nearby. Some plant samples were sent for further identification to the Forest Herbarium, Royal Forest Department, Bangkok, Thailand. Using an average weight for each food item, we estimated the total wet weight consumed for each item per day. Average weights of fruits were obtained from samples dropped below nests or collected later from fruiting trees. Average weights of animals were obtained from fresh specimens dropped below nests or from specimens of similar size caught elsewhere in the study area (of the same group but not always of the same species, due to the difficulty of collecting and identifying most small animals in tropical forest). We also determined nutritional values in g/wet weight (protein, fat, carbohydrate (CHO), calcium, and energy content) for each type of food. Fruits were analyzed by the Food Analysis Laboratory, Institute of Nutrition, Mahidol University at Salaya, Nakhon Pathom, Thailand. A few animal food items were also sent there for analysis, but most values for ani- mal foods were obtained from published sources (De- partment of Health 1978, 1984; Puwastien & Sung- puag 1983). Samples of specific animals were diffi- cult to obtain from the forest, so, where necessary, data for animals of the same genus or family as those identified as hornbill food items were substituted in the analyses. Food samples were weighed, to obtain the wet weight in grams, blended with an ordinary mixer; then frozen for 24 hours before being transferred into a freeze drier, if the sample was to be analyzed later. Nutritional values, including protein, fat, and calcium were determined by Kjeldahl AOAC 981.10, Soxhlet AOAC 945.16, and atomic absorption AOAC 975.03, respectively. Carbohydrate was obtained by subtract- 101 ing the sum of moisture (drying AOAC 925.45), pro- tein, fat, and ash (dry ashing) from 100. Energy was calculated using the general factors of 16.7, 37.5, and 16.7 kJ/g from protein, fat, and carbohydrate, respec- tively. A comparison of the quantities of food and nutri- ents delivered to nest inmates among these four horn- bill species, involves correcting for differences in fe- male body mass, clutch and brood sizes, and the length of nesting cycle and its phases (Table 1). Therefore, although mean daily delivery rates of food types and nutrients are compared within species as absolute amounts, useful for ecological comparisons, biological comparison requires some sort of correc- tion for inter-specific differences. In an attempt at correction, ratios were calculated between values for each species by taking the species with the lowest values, the smallest Oriental Pied Hornbill (Table 1), as one unit. The mean values for each species were then adjusted using either ratios of the mean mass of nest inmates or ratios of the duration of the total nest- ing cycle. The mean mass of nest inmates for each species was calculated as the sum of the median fe- male mass plus the mean brood size multiplied by the median female mass, except for Great Hornbill where the female leaves the nest early in the nestling phase and is scored for only half of her mass contribution (Table 1). Chick mass at different ages and growth curves were unavailable for any of the hornbill species, so the mass of the adult female, less than the adult male, was substituted for the mass of a chick at fledging. Any excess of the adult female mass over the real fledgling mass would offset any additional nutritional requirements of chicks during their devel- opment. Statistical analyses were performed with SigmaSta P. POONSWAD, A. TSUJI and N. JIRAWATKAVI Great Hornbill (a) Wreathed Hornbill (b) > 700 N=16 = 700 > N=9 = mo} S) 600 Sb 600 + 2 500 #500, = 400 ra 400 7 2 300 = 300 + 2 200 3 200 - ~ et > 100 il I g wo | | YT 3s. 5 OT Oe ee ree 12°33 5 FF Oo TPL 18 Sie Week of breeding cycle Week of breeding cycle Incubation i Ne tlin = + Incubation t Nestling © Animal ; ; me Fruit j i i White-throated Brown Hornbill (c) Oriental Pied Hornbill (d) > 700 2 700 > N=24 S 600- N=i4 J 600 - ob ~— > 500- © 5004 =| BY = 400 > 400 5 3 300- 2 300- 3 200- 3 2005 3 | 3 100 caliiiiiiniias SUL WiwH ae ot . | 3 5 7 9 il catiuiaats 1 3 5 Wl 9 11] Ei wis 7 Al) Week of breeding cycle +——— Week of breeding cycle + Incubation Neestling —— Fig. 1. tIncubation+————— Nestling Quantity of food (estimated g/day) and proportion as fruits and animals delivered by wet weight to the nest inmates of four hornbill species: Great Hornbill (a), Wreathed (b), White-throated (c), and Oriental Pied (d) in Khao Yai National Park, Thailand, at weekly intervals during the nesting cycle. * time of female emergence. 2.0 (Jardel Corporation 1995) We used Mann-Whit- ney Rank Sum Test (T, for medians) for intra-specific comparisons of each phase of the nesting cycle and the Kruskal-Wallis ANOVA (F, for means) or One Way ANOVA on Ranks (H, for medians) for All Pair- wise Multiple Comparison Procedures (Dunn’s Method, for significance at P<0.05) of inter-specific values. RESULTS 1) Intra-specific patterns of food delivery by mass and food type during the incubation and nestling phases The weekly pattern of food delivery over the whole nesting cycle was similar for each hornbill species (Fig. 1), taking into account the differences in the lengths of their nesting cycles (Table 1). There were two main peaks, the first during early incubation nd the second midway through the nestling phase, | though the quantities and proportions of fruits and animals in the diet varied between species. Hornbill food delivery rose to a peak eve Great 102 halfway through the incubation phase (Week 4) and nestling phase (Week 13, also the time of the female’s emergence) but then dropped off markedly until fledging (Fig. la). Only the delivery of animal food was significantly higher after hatching (T=5121.5, N,=78, N,=119, P<0.001), whereas fruit delivery showed no significant change between the incubation and nestling phases. The female Great Hornbill would have consumed all animal food delivered dur- ing incubation, but once she emerged from the nest, all animal food would have become available to the chick. Wreathed Hornbill food delivery also showed a low peak midway through incubation (Weeks 2-4), a drop before hatching, but then a high and sustained feeding rate throughout the nestling phase with a peak in Weeks 13-15, all of it supplied by the male alone to the female and a single chick (Fig. 1b). The incubating female received little animal food, but there was a significant increase in animals and fruits once the chick had hatched (T=913.0, N,=35, N,=63, P<0.001 and T=1322.0, N,=35, N,=63, P=0.002, respectively), the increase in animals suffi- Nutrition of nesting hornbills in Thailand ciently obvious to confirm hatching. White-throated Brown Hornbill food delivery by the male and nest helpers appeared to increase markedly during egg-laying (Week 2), especially for fruits (Fig. 1c), although there was no significant dif- ference in the rates of fruit delivery before and after hatching. Food delivery peaked again midway through the nestling phase (Week 9), probably through the significantly higher delivery rate to the Table 3. female and several chicks of animal foods after hatching (T=2126.0, N,=46, N,=92, P<0.001), and then declined towards fledging. Oriental Pied Hornbill food delivery also peaked during egg-laying (Week 2) and then, after hatching, rose steadily to an extended peak midway through the nestling phase (Weeks 8-11) before declining to- wards fledging (Fig. 1d). The Oriental Pied Hornbill delivered significantly more fruit and animal foods Statistical comparison of daily rates (range, mean, SD, median) of protein, fat, carbohydrate, calcium, and energy de- livered to nest inmates by wet weight during the incubation and nestling phases for four sympatric species of hornbills in Khao Yai National Park, Thailand. GH=Great Hornbill, WH=Wreathed Hornbill, BH=White-throated Brown Hornbill, PH=Oriental Pied Hornbill. Incubation phase Nestling phase GH WH BH PH GH WH BH PH N (weeks) 7 7 5 5) 13 13 10 9 Protein (g/day) Range 6.1-23.8 3.2-9.0 61-113 4.7-6.5 1.8-28.3 7.2-25.7 9.8-22.0 5.7-15.1 Mean 11.0 6.5 8.6 5D IP 15.6 14.5 10.9 SD 6.1 23 2) 0.7 7.8 5.6 4.1 3.1 Median 9.5 6.4 Ie 5.5 16.3 15.6 13.6 10.5 T=48.5 T=31.0 T=19.0 T=17.00 P=0.05 P<0.001 P=0.01 P=0.008 Fat (g/day) Range 6.5-56.1 4.5-35.2 7.2-11.6 6.0-9.2 1.7-14.9 8.9-61.6 8.6-17.9 7.4-14.0 Mean 19.7 19.7 9.6 TES ities) 21.6 2 10.8 SD 16.4 il? il 1.2 3.5 14.3 2h Pal Median 15.2 18.8 10.2 Hes 128 20.0 V2. 10.9 T=100.00 T=73.0 T=25.0 T=18.0 P=0.04 P=1.0 P=0.08 P=0.01 Carbohydrate (g/day) Range 20.7-61.0 15.5-43.0 11.1-19.3 7.7-19.2 5.9-39.2 37.3-79.3 7.1-19.0 4.4-22.9 Mean 32.6 30.1 13.8 13.3 26.3 5312 13.3 ISS SD 13.9 10.7 3.3 4.1 10.1 14.1 3.8 6.0 Median 26.3 30.6 IAs) 13.1 26.9 49.5 13.4 16.9 T=78.0 T=36.0 T=40.5 T=29.0 P=0.75 P=0.003 P=1.0 P=0.29 Calcium (g/day) Range 0.39-1.95 0.49-0.90 0.27-0.78 0.30-0.42 0.02-2.74 0.82—2.21 0.40-1.22 0.51-1.05 Mean 1.13 0.72 0.48 0.36 1.43 137 0.73 0.74 SD 0.56 0.13 0.25 0.06 0.77 0.45 0.26 0.18 Median 0.85 0.72 0.31 0.40 iS) 1.4 0.67 0.76 T=61.0 T=29.00 T=28.0 T=15.0 P=0.34 P<0.001 P=0.16 P=0.003 Energy (kJ/day) Range 706-3,520 481—2,187 572-894 485-754 192-1,652 1,186-3,945 605-1,334 620-1,067 Mean 1,469 1,353 735 596 1,150 1,959 917 845 SD 937 625 137 99 401 789 210 156 Median 1,131 1,323 711 593 1,111 1,764 889 851 T=77.0 T=55.0 T=26.0 T=17.0 P=0.81 P=0.15 P=0.10 P=0.008 103 P. POONSWAD, A. TSUJI and N. JIRAWATKAVI fable 4. Mean daily rates (from Table 2) and proportions of energy-producing protein, fat, and carbohydrate delivered to nest inmates by wet weight during the incubation and nestling phases for four sympatric species of hornbills in Khao Yai National Park, Thailand.* % of nutrients delivered, not including crude ash, present in the actual food remains. Protein Fat Carbohydrate Total Hornbill species g/day % g/day % g/day % g/day %* Incubation phase Great 11.0+6.1 17.4 19.7£164 31.1 32.6+13.9 51.5 63.3 100 Wreathed G:52=2:3 11.5 19.7+11.1 35.0 30.1£10.7 33}, 56.3 100 White-throated Brown 8.6+2.5 26.9 9.61.7 30.0 13.843.3 43.1 32.0 100 Oriental Pied 5.5+0.7 20.9 oes .2 28.5 13.3+4.1 50.6 26.3 100 Nestling phase Great WZ 728 31.4 SEESsS 20.6 26.3+10.0 48.0 54.8 100 Wreathed 15.6+5.6 17.3 2).6£14.3) 23.9 53.2+£14.1 58.8 90.4 100 White-throated Brown 14.5+4.] 36.3 Poe 7/ 30.3 13.3+3.8 33.3 S2).8) 100 Oriental Pied 10.9+3.1 29.3 10.842.1 29.0 15.5+6.0 41.7 SH 100 Total Great 28.2 DBO 31.0 26.2 58.9 49.9 118.1 100 Wreathed 22.1 15.1 41.3 28.2 83.3 56.7 146.7 100 White-throated Brown 23.1 324 Pale 30.2 Dire 36.7 719 100 Oriental Pied 16.4 25.8 18.3 28.8 28.8 45.4 63.5 100 during the nestling phase than the incubation phase (T=4394.5, N,=57, N,=142, P<0.001 and T= 4319.5, N,=62, N,=157, P<0.001, respectively), to the female and several chicks. 2) Intra-specific patterns of nutrient delivery during the incubation and nestling phases The estimated weekly pattern of nutrients deliv- ered to nest inmates over the whole nesting cycle var- ied considerably by nutrient type and hornbill species (Tables 3 & 4, Figs. 2—5). Great Hornbill nutrient delivery comprised similar proportions overall of protein (23.9%) and _ fat (26.2%) but more carbohydrate (49.9%) (Table 4). Significantly more fat was delivered during incuba- tion than the nestling phase (Table 3), possibly an anomaly due to a peak soon after egg-laying (Week 3, Fig. 2b). There were no significant differences for other nutrients, including energy (Table 3), even though all other nutrients also appeared to peak in Week 3 (except for calcium, Fig. 2d) and the delivery rates for protein, calcium and to a less extent carbo- hydrate rose again midway through the nestling phase (Fig. 2a, d, and c). The peak in Week 3 of fat derived mainly from lipid-rich fruits, the peak of protein de- rived from similar quantities of fruit and animal foods and the peak of carbohydrate from fruit (Fig. 2b, a, and c respectively). Wreathed Hornbill nutrient delivery was low over- 104 all in protein (15.1%), medium in fat (28.2%) and high in carbohydrate content (56.7%) (Table 4). Mean delivery rates of protein, carbohydrate, and cal- cium were significantly higher during the nestling phase (Table 3, Fig. 3a, c, and d), and although appar- ently also higher for fat and energy these were not significant. White-throated Brown Hornbill nutrient delivery yielded similar proportions overall of protein (32.1%), fat (30.2%), and carbohydrate (36.7%) (Table 4). Only protein increased significantly during the nestling phase (Table 3, Fig. 4a), although other nutrients also appeared to increase during the same phase (Fig. 4). Oriental Pied Hornbill nutrient delivery yielded similar proportions overall of protein (25.8%) and fat (28.8%) but higher carbohydrate (45.4%) (Table 4). There were significant increases during the nestling phase for all nutrients (protein, fat, energy, and cal- cium) except carbohydrate (Table 3, Fig. 5). 3) Intra-specific patterns of nutrient delivery by food type Fruit and animal foods contributed to all nutrient classes measured in this study. However, the propor- tions of these foods delivered to nest inmates and the quantities of their contributions to the main nutrient classes, varied among hornbill species (Figs. 2—5). Nutrition of nesting hornbills in Thailand a _ 3000) 2 7 2 ~ SS — tan a Q ‘ S ri\ 2 25 (a) ff Great Hornbill = & 2500 (d) | I 2 | | - E a ~ 50 W\ /Wi C4 Animal > 2000; 1 > 154 OMT Fruit > 1500] Hl] 2 104 Vi 3 1000) io} mo) E € 5 2 500 £0 ' So! : a, IS IS Ws 19 eed pide a reli lee Us gl Sarl a 9: Teubati pee cee cele Week of breeding cycle a Sapins '~ Incubation Nesstling 4000 ( f > e) 70 2 3500: _ 604 ~ 3000; 2 2 3 504 = 2500; oD 2 40; = 2000} FS 2 5, 304 3 1500 5 3 7 > 10004 = 20; rer? oi ey i. v Week of breeding cycle & 500 & 10; + Incubation-+———_Nestling 0! 0! a i SOS 7 OA AAS alg iO ion sole OFM 13 15817 19 Week of breeding cycle——1 (~~ Week of breeding cycle + Incubation—t——— Nestling ——— t~ Incubation Nest ling Fig. 2. Estimated mean daily delivery rates of nutrients derived from fruit and animal food by wet weight, in- cluding protein (a), fat (b), carbohydrate (c), calcium (d) and energy (e) to nest inmates of Great Hornbill and a total of nutrient delivery daily during each week of the incubation (solid circles) and nestling (open circles) phases. * time of female emergence. Protein Both fruits and animals were important protein sources for nesting Great Hornbill (Fig. 2a) and their delivery rates did not significantly differ. In contrast, fruits were the major source of protein in the diet of the Wreathed Hornbill, with a significantly higher de- livery rate than for animals (T=586.0, N,=20, N,=20, P<0.001; Fig. 3a), both before and after hatching (T=77.0, N,=7, N,=7, P=0.002 and T=250.5, N,;=13, N,=13, P<0.001, respectively). Animal food was the major source of protein in the diet of White-throated Brown and Oriental Pied Hornbills, since delivery rates from fruits were sig- nificantly lower (T=121.0, N,=15, N,=15, P<0.001 and T=152.0, N,=14, N,=14, P=0.02, respectively; Figs. 4a and 5a). Fat The main source of fat delivered by all four horn- bill species was lipid-rich fruits, as shown by the sig- nificantly higher delivery rate for fruits than animals (Great Hornbill T=602.5, N,=20, N,=20, P<0.001; 105 Wreathed Hornbill T=610.0, N,=20, N,=20, P<0.001; White-throated Brown Hornbill T=345.0, N,=15, N,=15, P<0.001; Oriental Pied Hornbill: T=301.0, N,=14, N,=14, P<0.001; Figs. 2b, 3b, 4b, and 5b). This was especially evident for the two larger species, Great and the Wreathed Hornbills, where the delivery rate of fat from fruits was 8.9 times and 25.2 times higher respectively than from animals (Figs. 2b and 3b). Delivery rates of fats from fruits and animals were both higher after than before hatching for Great Hornbill (T=77.0, N,=7, N,=7, P<0.001 and T=252.5, N,=13, N,=13, P<0.001, respectively). Carbohydrate Fruits were the main source of carbohydrate for all four species (Figs. 2c, 3c, 4c, and 5c), with a significantly higher delivery rate than from animals (Great Hornbill T=610.0, N,=20, N,=20, P< 0.001; Wreathed Hornbill T=610.0, N, =20, N,=20, P<0.001; White-throated Brown Hornbill T=345.0, N,=15, N,=15, P<0.001; Oriental Pied Hornbill P. POONSWAD, A. TSUJI and N. JIRAWATKAVI (a) Wreathed Hornbill > 3° 3000 ) S oL [> Animal E 2500 : mm Fruit = 2000) F 1500, illll il f | “ e E 5001 a) ~ ~ a. : 7.9 V1 Ion ioe ly alo i 3; 5s TOP Ty 1S Spy aS Week of breeding cycle — t——— Week of breeding cycle Incubation + Nestling a + Incubation Nestling AT Fat delivery rate (g/day) hag jp oS Fim ie) Lah Se) SMA Ses t+——— Week of breeding cycle Incubation +——— Nestling Fig. 3. 9 +—— Week of breeding cycle + Incubation +———— Nestling 7 ae INL) REN ike) Bilg/ Biles Energy delivery rate (kJ/day) pt: 3) gens net Week of breeding cycle + Incubation+———— Nesting Sedillo Sp, MSc (hZ TPs Estimated mean daily delivery rates of nutrients derived from fruit and animal food by wet weight, in- cluding protein (a), fat (b), carbohydrate (c), calctum (d) and energy (e) to nest inmates of Wreathed Hornbill and a total of nutrient delivery daily during each week of the incubation (solid circles) and nestling (open circles) phases. * time of female emergence. T=299.0, N,=14, N,=14, P<0.001, and this differ- ence occurred both before and after hatching for Wreathed Hornbill (T=77.0, N,=7, N,=7, P<0.001 and T=260.0, N,=13, N,=13, P<0.001, respec- tively). Calcium The main source of calcium depended on hornbill species (Figs. 2d, 3d, 4d and 5d). Wreathed Hornbill acquired significantly more calcium from fruits than from animals (T=607.0, N,=20, N,=20, P<0.001; Fig. 3d), whereas White-throated Brown Hornbill ac- quired most calcium from animals (T=167.5, N,=15, N,=15, P=0.008; Fig. 4d). Both Great and Oriental Pied Hornbills acquired calcium about equally from fruits and animals, with no significant differences (Figs. 2d and 5d). Delivery rates of cal- cium from fruits were higher both before and after hatching for Wreathed Hornbill (T=77.0, N,=7, N,=7, P<0.001 and T=260.0, N,=13, N,=13, P<0.001, respectively; Fig. 3d). 106 Energy Energy can be derived from proteins, fats, carbo- hydrates or a combination of these nutrients in fruits and animal foods. Fruits were the main overall source of energy and were delivered significantly more than animals by all four species (Great Hornbill, T=598.0, N,=20, N,=20, P<0.001; Wreathed Hornbill, T=610.0, N,=20, N,=20, P<0.001; White-throated Brown Hornbill, T=330.0, N,=15, N,=15, P<0.001; Oriental Pied Hornbill, T=300.0, N,=14, N,=14, P<0.001), particularly the Wreathed otal bill (Figs. 2e, 3e, 4e, and Se). However, energy deliy- ery from animals significantly increased for all species after hatching, Great (T=249.0, N,=13, N,=13, P<0.001), Wreathed (T=260.0, N,=13, N,=13, P<0.001), White-throated Brown (T= 144.0, N,=10, N,=10, P=0.004), and Oriental Pied Horn- bills (T=126.0, N,=9, N,=9, P<0.001) (Figs. 2e, 3e, 4e, and Se). For Great and Wreathed Hornbills, fat and carbo- hydrate were equally important sources of energy by mass, although not necessarily by calorific value per Protein delivery rate (g/day) Nw Ww ay t=) | aoe =) ie Ao Week of breeding cycle - Incubation+———— Nestling 3} Nutrition of nesting hornbills in Thailand 15 White-throated Brown Hornbill => Animal me Fruit (c) ee) > S i=) (d) Calcium delivery rate (mg/day) 1 STE: 7!) mak Week of breeding cycle - Incubation +———— Nestling I SJ >) (b) RY w FuNna Aaa 5 & | Fat delivery rate (g/day) oo 13 3 -Incubation+——Nestling Se dln oe Deel Mil Week of breeding cycle Siler 9) Week of breeding cycle Incubation Nestling Hit 3) 30 ST Be Oe il 15 Week of breeding cycle - Incubation-—— Nestling 133 Fig. 4. Estimated mean daily delivery rates of nutrients derived from fruit and animal food by wet weight, in- cluding protein (a), fat (b), carbohydrate (c), calcium (d) and energy (e) to nest inmates of White-throated Brown Hornbill and a total of nutrient delivery daily during each week of the incubation (solid circles) and nestling (open circles) phases. * time of female emergence. unit weight. Mean delivery rates of both nutrition classes were significantly higher than the delivery rate of protein (H=19.5, df=2, P<0.001 and H=33.7, df=2, P<0.001, respectively; Table 4). For the White-throated Brown and Oriental Pied Horn- bills, fat was the major source of energy (F=33.8, df=2, P<0.001 and F=23.7, df=2, P<0.001, respec- tively). 4) Inter-specific comparisons of patterns of nutrient delivery Comparisons between species of the quantities of food and nutrients delivered to nest inmates is com- plicated by inter-specific differences in body, clutch and brood sizes, and in the duration of the nesting cycle and its phases (Table 1). Results of the cali- brated means are shown for each of the nutrients (Table 5) and for the quantities of nutrients and total energy (Table 6) even though protein and energy requirements are proportional to body-mass”’ (Kleiber’s rule; Kleiber 1961). No statistical test of any differences between species was considered ap- propriate, but adjustment of the means did reduce 107 variance between the species (cf. Table 3), suggesting more equivalent comparisons. Protein delivery rates adjusted for brood mass for all species were similar to that of the Oriental Pied Hornbill during the incubation phase, except for being lower for the Wreathed Hornbill (Table 5). During the nestling phase, adjusted delivery rates were higher than during the incubation phase for all species, especially the Oriental Pied Hornbill. Ad- justed fat delivery rates were low during incubation for White-throated Brown Hornbill and high for Great Hornbill, but were highest for nestling Oriental Pied Hornbill and lowest for Great Hornbill (Table 5). Adjusted carbohydrate delivery rates were most varied among species, lowest for White-throated Brown Hornbill during both incubation and nestling phases, but highest for incubating Great Hornbill and nestling Wreathed Hornbill (Table 5). Adjusted cal- cium delivery rates were lower during the incubation than the nestling phase for all species, but with the greatest increase for nestlings of Oriental Pied Horn- bill (Table 5). Adjusted energy delivery rates to the female during incubation were lowest for White- P. POONSWAD, A. TSUJI and N. JIRAWATKAVI k = 30004 Z 30 Oriental Pied Hornbill §,,.,, | pp 2s 2” ‘ : 2 2000 d = 20 (a) [3 Animal = (d) . ° > S 15 : me Fruit 5 1500 3 10 | By 3 1000 = 2 e s 51 2h 1 4 [ 5 500 sere rt in 3 titi Week of breeding ace a Week oF breeding tr Incubation + —Nestling | + Incubation 4+ Nestling 0) cy 70/ = 60; (c) = 50 $4000 70 40) S 3500 B en 2 30 ~ 3000; (e) >| (b) 3 20 Mem a 2 2500: = 50 Ailin LT Fee S )- tet o- > z v O+ i =| i: i 5304 lim loys 3 1500; = 49| Week a Meanie cycle > 2 = - Incubation Nesting = 3 i= Merete iin = Wee of bane efole + Incubation -———— Nestling Fig. 5 St Ws A OR Wei Week of breeding cycle ———‘+ - Incubation +———— Nestling ———I Estimated mean daily delivery rates of nutrients derived from fruit and animal food by wet weight, in- cluding protein (a), fat (b), carbohydrate (c), calcium (d) and energy (e) to nest inmates of Oriental Pied Hornbill and a total of nutrient delivery daily during each week of the incubation (solid circles) and nestling (open circles) phases. * time of female emergence. throated Brown Hornbill, but rose during the nestling phase for all species except Great Hornbill where the incubating female received more energy than the nestling (Table 5). Delivery rates for nutrients and total energy, once adjusted for the duration of the nesting cycle, reduced considerably the overall differences between species (Table 5 cf. Table 4). Adjusted protein delivery rates were especially important to Great and White- throated Brown Hornbills, and more important in the nestling than the incubation phases for all species. Adjusted fat delivery rates were important during in- cubation for the larger species, Great and Wreathed Hornbills, but especially during the nestling phase for the smaller species-pair of White-throated Brown and Oriental Pied Hornbills, although about equal in both phases for Wreathed Hornbill. Adjusted carbohydrate delivery rates were especially important to both larger species during incubation, particularly for Great Hornbill, and markedly increased during the nestling phase for Wreathed Hornbill, but varied least in both phases for the smaller species. Adjusted de- livery rates for total energy were only highest during 108 the incubation phase for Great Hornbill, being similar and higher during the nestling phase for all the other species, particularly for Wreathed Hornbill. The percentages of nutrients delivered to the nestlings also needs to be adjusted by the brood mass (Table 6), especially for any comparisons with fledg- ing success. The adjusted percentages of nutrients and energy for the Oriental Pied Hornbill were 1.05 for protein, 1.17 for fat, 1.85 for carbohydrate, and 92.38 for energy (Table 6) while the other three species had relatively lower protein intakes, espe- cially the Wreathed Hornbill during the incubation phase (Table 3). The Oriental Pied Hornbill had the highest breeding success (95.8%) and the Wreathed Hornbill the lowest (66.7%, Table 6). DISCUSSION 1) Food delivery pattern The overall pattern of food delivery to nest inmates was similar for all four hornbill species, despite dif- ferences in size, duration of nesting cycle and breed- ing strategy. There was a rise in the mass of food de- Nutrition of nesting hornbills in Thailand TIippl TIbrl Tibrhl voOOsl 1696 LISS CCS Ce GIG L, WGOr Ge WON erOrcdl Be S197 [e}0L 6 St8 6 St8 6 St8 0'SE8 98S 1816 SLOCT LI8L 68S61 TET 78h 6 IST SUI[}SON €°S6S €S6S €S6S b'S99 S OEP 9 EEL L796 Ors LOSET 8601 7869 6991 Uuoneqnouy (Aep/py) Adsoug Ol Ol Ol Ol 1L'0 Tal 6b'I r3'0 60°7 €8'I Gol 957 [ejoL, pl'0 ¢L'0 pL'0 99'0 €r'0 €L'0 860 $s°0 Ze 1 Z0'l 89°0 eri SuIpSoN 9€0 9€°0 9€0 br'0 870 8r'°0 1S'0 67 0 ZL'0 130 ps0 ell uoneqnouy (Aep/3) eD 8°87 8°8z HG 9p 6SI WG S'6S 9 CE €€8 47 0°87 68S [e}OL SSI XSI] SSI na Bie cel 08 aie GES 881 ral £97 SUIT}SON eel cel €€l CCl 1°8 gl ce 6 CTI 10€ €€7 SSI 9°7E uoneqnouy (Aep/3) OHO eI €81 €8I L6l LGI LIZ $67 SOI Cir 17 S'r1 O1E [ejoL, 801 801 8°01 O11 Le ee v'SI 9°8 917 ae v's Cll BUTTISON Syl CL Sh L’8 9°¢ 96 Vv 6L Ll lvl b'6 L'6l uoneqnouy (Aep/3) yey p91 p91 p91 O17 9'€I 1'€Z BSI 88 ee 707 rel €87 [e}OL 601 601 601 Cel 58 Srl ler 79 9SI a (6a) cia SUI[ISON KE GAG GG oa: eis 9°8 9 97 9 6L cs O11 uoneqnouy (Avp/s) UIa}01g (0'1) (01) isn{pe (1) (LI) ysn[pe (v1) ($2) jsn{pe (pI) (12) jsn{pe ON Wa a10jog ON Wd a10jog ON Wad a10jaq ON Wad a10jog aseyd SuljsoN qUSINN Hd Hd HM HD ({ 21qe], Woy ones yUotUYsN{pe ‘Aep/3) soyes A1OAT[Op URoU paysn{pe ‘soroads |jIquioy ‘[[IQUIOH Pat [BIUSHO=Hd ‘[[!QUJOH UMoIg poyeory}-o A = Ha TIquIOH pouyVaM—HM ‘[[IquUIOH WwaAIH=HO ‘puerylieyL Aled [euoHeN leA ory ul “(yx0} 998 ‘] a[QeT Woy) satoads [[Iquioy yore Fo (ON) ajoAo Sunsou pue (Wg) sseul pooig pue Apoq J0F poysnipe “(Zz [qe], Wor) saseyd Surpsou pue UoHeqnoul sy} SULINP JYSIOM JOM AQ So}PUIUT jSOU SY} 0} S}USTNU JO So}e1 AIOAIOp uv, “S PGBL 109 P. POONSWAD, A. TSUJI and N. JIRAWATKAVI Table 6. rotal delivery rates of nutrients to the nest inmates by wet weight (from Table 5), percentages of nutrients by body and brood mass (BM) and breeding success of each species of hornbills (from Table 1) in Khao Yai National Park, Thailand. GH=Great Hornbill, WH=Wreathed Hornbill, BH=White-throated Brown Hornbill, PH=Oriental Pied Hornbill. GH (3,316 g) WH (3,900 g) BH (2,718 g) PH (1,560 g) Nutrient Total % by Total % by Total % by Tota % by nutrient BM nutrient BM nutrient BM nutrient BM Protein (g/day) 28.2 0.85 22.1 0.57 23 0.85 16.4 1.05 Fat (g/day) 31.0 0.93 41.3 1.06 PA I) 0.80 18.3 IL-1 CHO (g/day) 58.9 1.78 83.3 2.14 27 1.00 28.8 1.85 Ca (g/day) 2.56 0.08 2.09 0.05 1.21 0.04 1.10 0.07 Energy (kJ/day) 2,618.8 78.97 3,309.0 84.85 1,651.7 60.77 1,441.20 92.38 Breeding success (%) 93.8 66.7 92.9 95.8 livered during the first few weeks after enclosure, a drop towards hatching time, an even higher rise until the chicks attained full size and then a drop until fledging (Fig. 1). The form of diet, whether fruits or animals, and of nutrients, whether protein, fat, carbo- hydrate, calcium or energy, varied between phases of the nesting cycle. Diet also varied between species and was influenced by inter-specific differences in duration of the nesting cycle and in body, clutch and brood sizes (Table 1). Due to their peculiar breeding strategy as part of which the males feed imprisoned incubating and brooding females, it is difficult to compare hornbills with other birds. However, peaks in food delivery by all four hornbill species during the early weeks of the nesting phase may be associ- ated with the need to supply nutrients for egg produc- tion, incubation, and the commencement of molt in Week 2 for smaller species (Fig. lc & 1d), and the restoration of any reserves used in egg production and in the regeneration of new feathers in Weeks 24 for larger species (Figs. la & 1b) (Walsberg 1983; Bryant 1997), and/or the establishment of new re- serves prior to rearing chicks (Poonswad 1993). 5 “= ) Breeding cycle and nutrients We realize that our methods for estimating nutrient delivery to nest inmates can only be applied in broad terms. We are also conscious of errors that may arise from assuming that food delivered and the nutrients analyzed match those absorbed by the female and/or chicks, especially for carbohydrates that included in- digestible crude fiber (Bolton 1955) and for fruits (Levey & Matinez del Rio 2001; Pryor et al. 2001). In terms of nutrients and energy, breeding and molt are the most demanding processes within the annual cycle of any non-migratory bird species (Payne 1972; 110 Walsberg 1983). In hornbills, the incubation and nestling phases may be more demanding for male hornbills than for any other birds (Klaasen et al. 2003). Among the four hornbill species studied in Thailand, only the female Great Hornbill emerges midway through the nestling phase to feed herself and so reduces the workload of her mate and helps provide food for the large chick. In the White- throated Brown Hornbill, the male is assisted by non- breeding helpers, which reduced the male's workload throughout nesting by 40% (Table 1; Poonswad 1993). We noted that delivery rates for protein by all four hornbill species rose to a peak during Weeks 2-4 of incubation (Fig. 1), as might be expected to replace resources used in egg-laying and commencement of molt. We also recorded levels for delivery of energy during incubation that were as high as during chick rearing, which may relate to the costs of female molt. Formation of eggs requires extra quantities of pro- tein to form albumen, fat to form yolk, protein and fat to provide energy, and calcium to form eggshells (Meijer & Drent 1999). Due to the relatively high de- mands of clutch production versus female mainte- nance in various large bird species (Meijer & Drent 1999), deposition of body reserves prior to laying would be expected in hornbills rather than an in- crease in food delivery. The coincidence of nesting and complete molt of all flight feathers is uncommon in birds, but total de- pendence on the male hornbill for delivery suggests that food availability is high, at least for the breeding female since the breeding male only molts after nest- ing (Poonswad 1993). In the Great Hornbill, where the female emerges earlier in the nesting cycle than other species (Table 1), high delivery rates of protein Nutrition of nesting hornbills in Thailand may also contribute to the regeneration of new feath- ers within a more restricted time frame. Video record- ings from inside a Great Hornbill nest have shown that flight feather molt commences during Week 2 and is virtually complete by Week 10 (P. Poonswad, unpublished). It should be noted that the most important sources of energy were fat and/or carbohydrate depending on species (Table 4). Non-significant differences in fat or carbohydrate delivery rates indicated that both were the main source of energy during the incubation and nestling phases for each species (Table 3), but from significant differences between them it seems that the amount of protein delivered was important in deter- mining breeding success (Table 6), particularly dur- ing the seven-day pre-laying phase for production of quality eggs (Gill 1990). Differences in nutrient delivery during the nestling phase probably reflected changes in requirements of the growing chick(s), such as for skeleton formation and feather growth (Gill 1990). In all four species there was a general increase in delivery rates during the nesting cycle, rising to a peak until chicks were expected to have completed their body growth mid- way through the nestling phase, followed by a drop until fledging (Fig. 1). However, this pattern did not hold for all nutrients and the exact amounts varied between species. For example, lower fat delivery rates during the nestling phase in the Great Hornbill indicated the im- portance of fat for the incubating female, whereas fat was important to nestling Oriental Pied Hornbills as an energy source, possibly as a reserve to insure against poor food delivery by parents during the days just after fledging (Gill 1990). Calcium delivery rates to incubating female Great and White-throated Brown Hornbills were as high as during the nestling phase, even though these species have different parental care strategies and brood sizes (Poonswad 1993; Table 1). Both species lay up to four eggs, thus high calcium delivery may replace that lost during egg production. Increased calcium may also help the female Great Hornbill to prepare for early emergence in Week 13, and the female White-throated Brown Hornbill to share reserves with its large brood of up to four chicks. In contrast, female Wreathed and Oriental Pied Hornbills, with similar modes of parental care but different brood sizes (Poonswad 1993, Table 1), have the whole nest- ing cycle to replace lost calcium, yet enjoy a signifi- cant increase in calcium delivery when the chick(s) 111 are growing. Calcium was derived from both fruits and animal foods, the latter were especially common in the nestling diets of White-throated Brown and Oriental Pied Hornbills with their larger broods, but some fruits, especially figs, are also known to contain calcium (Poonswad 1993; O’Brien et al. 1998), and were important in the diet of the Wreathed Hornbill throughout breeding (Poonswad et al. 1987). 3) Source of nutrients The overall proportion of fruits and animals in the diet had been recorded for nesting hornbills (Poon- swad et al. 1987), but the respective nutritional con- tributions of these foods were not estimated from the food types delivered. Generally, fruit was the major source of nutrients, particularly for fat and carbohydrate. Hornbills in Khao Yai consumed as much lipid-rich fruit as those in Borneo (Leighton 1982), but whether hornbills se- lected lipid-rich fruits or the majority of large trees produce fruits rich in lipids remains to be studied. Differences in nutrient sources among the four Thai hornbill species were related also to feeding niche of each species. Wreathed Hornbill is a fruit specialist whereas White-throated Brown Hornbill is an animal specialist. In contrast, Great and Oriental Pied Horn- bills are generalists (Poonswad 1993). Wreathed Hornbill delivers the least protein from animal sources and Oriental Pied Hornbill the most (Table 4, 5, and 6), which may affect breeding success, particu- larly during egg production and incubation. Wreathed and Great Hornbills feed mainly in tree canopies, but Great Hornbill was frequently observed seeking ani- mal food. White-throated Brown Hornbill occupies the forest under-story and hunts more animal food while Oriental Pied Hornbill forages from the ground up to the canopy and consumes the greatest diversity of food among these four species (Poonswad 1993; Poonswad et al. 1998). ACKNOWLEDGMENTS We thank the National Park Division, Royal Forest Depart- ment of Thailand for permission to work at KYNP and for providing accommodation and helpers. We are grateful to the late Dr. Tem Smitinand, Dr. Veerachai Na-Nakhon, and Vicharn Aiedthong for identifying plants, Kosol Charernsom and Jarujin Nabhitabhat for identifying insects and other ani- mals, all superintendents of KYNP for their cooperation, and Pak Manchaona and family for excellent support. We thank Dr. Alan Kemp for comments on the manuscript. We also thank Sopha Sa-nguanchat, Siriwan Nakkuntod, and Porntip Poolswat for manuscript typing. The research that formed the P. POONSWAD, A. TSUJI and N. JIRAWATKAVI basis of this paper was funded by the International Foundation for the Conservation of Birds, North Hollywood, USA. REFERENCES Bolton W (1955) The digestibility of the carbohydrate complex by birds of different ages. J Agr Sci 46: 420-424. Bryant DM (1997) Energy expenditure in wild birds. Proceedings of the Nutrition Society 56: 1025-1039. Department of Health (1978) Nutrient composition table of Thailand foods. Nutrition Division, Department of Health, Ministry of Public Health, Bangkok. Department of Health (1984) Nutrient composition table of Thailand foods. Nutrition Division, Department of Health, Ministry of Public Health, Bangkok. Gill FB (1990) Ornithology. WH Freeman and Com- pany, New York. Jandel Corporation (1995) SigmaStat 2.0. Jandel Corpo- ration, San Rafael. Kemp AC (1995) The Hornbills (Bucerotiformes). Ox- ford University Press, Oxford. Kemp AC (2001) Family Bucerotidae (Hornbills). In: del Hoyo J, Elliott A & Sargatal J (eds) Handbook of birds of the world. Vol 6. pp 436-523. Lynx Editions, Barcelona. Klaasen M, Brenninkmeijer A, Boix-Hinzen C & Mendelsohn J (2003) Fathers with highly demanding partners and offspring in a semidesert environment: energetic aspects of the breeding system of Mon- teiro’s Hornbills (Jockus monteiri) in Nambia. Auk 120: 866-873. Kleiber M (1961) The fire of life. John Wiley, New York. Leighton M (1982) Fruit resources and patterns of feed- ing, spacing and grouping among sympatric Bornean hornbills (Bucerotidae). PhD thesis, University of California, Ann Arbor. Levey DJ & Martinez del Rio C (2001) It takes guts (and more) to eat fruit: lessons from avian nutritional ecology. Auk 118: 819-831. Meijer T & Drent R (1999) Re-examination of the capi- tal and income dichotomy in breeding birds. Ibis 141: 339-414. O’Brien TG, Kinnaird MF, Dierenfeld ES, Conklin NL, Wrangham RW & Silver SC (1998) What’s so special about figs? Nature 392: 668. Payne RB (1972) Mechanism and control of molt. In: Farner DS & King JR (eds) Avian biology Il. pp 103-155. Academic Press, New York. Poonswad P (1993) Comparative ecology of sympatric hornbills (Bucerotidae) in Thailand. D Sc Thesis, Osaka City University, Osaka. Poonswad P, Tsuji A, Jirawatkavi N & Chimchome V (1998) Some aspects of food and feeding ecology of sympatric hornbill species in Khao Yai National Park, Thailand. In: Poonswad P (ed) The Asian hornbills: ecology and conservation. pp 137-157. BIOTEC- NSTDA, Bangkok. Poonswad P, Tsuji A & Ngampongsai C (1987) A com- parative ecological study of four sympatric hornbills (Family Bucerotidae) in Thailand. In: Acta XIX Con- gressus Internationalis Ornithologici. pp 2781-2791. Pryor GS, Levey DS & Dierenfeld ES (2001) Protein re- quirements of a specialized frugivore, Pesquet’s Par- rot (Psittrichas fulgidus). Auk 118: 1080-1088. Puwastien P & Sungpuag P (1983) Nutritive value of unconventional protein sources, insects. J Nutr Assoc Thailand 17: 5-12. Walsberg GE (1983) Avian ecological energetics. In: Farner JR, King JR & Parkes KC (eds) Avian biology. Vol VII. pp 161-220. Academic Press, New York. ORIGINAL ARTICLE Ornithol Sci 3: 113-117 (2004) Breeding success of the tropical Spotted Munia Lonchura punctulata in urbanized and forest habitats Ramesh C. SHARMA, Dinesh BHATT* and Romesh K. SHARMA Avian Diversity and Bioacoustics Laboratory, Department of Zoology and Environmental Science, Gurukula Kangri University, Hardwar 249 404, India ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2004 Abstract Information on breeding success of birds of the Indian subcontinent is al- most negligible. The present study, carried out during the breeding seasons of 1997— 1999 inclusive at Hardwar in northern India (29°55’'N; 78°8’E), compared the breed- ing success of Spotted Munia Lonchura punctulata in urbanized and more natural forested habitats. A significant difference was found in total breeding success be- tween the two habitats, being 48.7% in urban areas and only 31.6% in forest. Al- though differences in eggs hatched (70.3% vs 60.3%) and nestlings reared (57.7% vs 51.9%) were less, significantly more fledglings survived in urban areas (89.0%) than in forested habitat (70.9%). This indicates that the greater breeding success of Spotted Munia in urbanized habitats is due primarily to higher rates of post fledgling survival there. A number of factors may affect reproductive success differently between habi- tats. In forest, nests were built in isolated thorny trees (e.g. Acasia nilotica) outside of forest canopy cover. In urban areas, however, the trees or shrubs selected for nesting were mostly of introduced species (e.g. Thuja orientalis, Polyanthea longifolia), all densely foliaged and rendering predation difficult. Although the typical habitat of Hardwar town is not natural, the Spotted Munia has evidently adapted quickly and successfully in its landscape. Such shifts in behaviour are not instantaneous and newly acquired behaviour takes time to spread. It would be interesting to determine whether these behavioural shifts in Spotted Munia are based on culturally transmitted learning or on genetic change. Key words Adaptation, Breeding success, Lonchura punctulata India constitutes a major part of the Oriental Realm, with wide latitudinal (08°04’ to 37°60'N) and longitudinal (68°07’ to 97°25’E) extents, and is situ- ated with the tropical monsoon belt. Its varied topog- raphy, considerable insolation, and monsoon climate, impart to it enormous complexity and habitat diver- sity. Therefore it is possible to study the breeding bi- ology of a wide range of birds in similar habitats and of similar/same species in diverse habitats providing an excellent opportunity for simultaneous analysis of various ecological factors. In the tropics, avian breed- ing seasons extend throughout the year in keeping with the extended favourable conditions; neverthe- less, breeding in individual species is essentially peri- odic (Baker 1938; Misra 1962; Miller 1965; Thap- liyal 1978; Chandola & Thapliyal 1978; Chandola et (Received 2 June 2003; Accepted 13 February 2004) * Correponding author, E-mail: dd_bhatt@ yahoo.com 113 al. 1983). Spotted Munia (Lonchura punctulata) is a season- ally breeding, non-migratory waxbill (family Estril- didae) occurring throughout the Indian sub continent (Ali & Ripley 1987). It feeds mainly upon grass seeds and crop grains, which are maximally available from autumn to late spring (Ali & Ripley 1987; Bhatt 1982). The nestlings and fledglings are fed mainly with half ripe crop grains Oriza sativa (Sharma 2002). Prior to the onset of reproduction, food intake declines (in summer) as food becomes scarce. The food intake cycle of Spotted Munia is thus almost in- versely related to its reproductive cycle (Bhatt 1982; Chandola et al. 1983). In this bird it seems that day length entrains the already existing endogenous cir- cannual rhythm with the calendar year, thus not di- rectly triggering the onset of breeding, as observed in many temperate birds (Bhatt 1982; Chandola et al. 1983). R. C. SHARMA, D. BHATT and R. K. SHARMA. Our earlier studies indicated that in Spotted Munia the average height of the nest from the ground was 2.25+0.21 meters and the shape of the nest was roughly globular or spherical with a lateral entrance hole. In this species the chicks hatch asynchronously. It has recently been observed that Spotted Munia se- lect various introduced plants for nesting in, in urban- ized habitat (Sharma 2002). With this observation in mind, we studied the breeding success in different habitats, in order to understand how successfully this species has adapted to urbanized landscapes for breeding. STUDY AREA AND METHODS The breeding success of the Spotted Munia L. punctulata was carried out from 1997 to 1999 in two different habitats (urban and forest area) in the Hi- malayan foothills at Hardwar (29°55'N and 78°08'E) India. The forest area was about 10 km away from the urban study site. The urbanized habitat is composed of residential areas, gardens, roads, ornamental trees and shrubs including Thuja orientalis, Polyanthea longifolia, Citrus, and Bougainvellia species with hedges of species such as Pithecolobium dulce. The forest habitat is of moist deciduous type with consid- erable vegetation. The main tree species in this habi- tat is Sal Shorea robusta, which comprises about 60- 65% of the total vegetational cover. During the breeding season sites were visited daily or as required. Nesting activities were observed/ recorded from 6—-10m distance with the help of 7X50 binoculars, and a Sony video camera (TRS90E). We recorded date of laying the first egg, date of clutch completion, clutch size, and the date of last hatching. In the urbanized area 54 eggs in 10 nests were marked with indelible Indian ink for incu- bational studies while 28 nests were observed for hatching success. Similarly in the forest area 60 eggs (in 10 nests) were marked and 24 nests were ob- served. To reduce disturbance early in the breeding cycle, incubation was recorded at fewer of the nests than the other breeding characteristics. During the egg laying and hatching periods, nests were observed daily between 06:00 and 12:30 and from 18:00 to 20:00. The incubation period was ob- served using Skutch (1957) and Nice’s (1954) method i.e. the period was recorded from laying of the last egg to the hatching of the same. To investi- gate pre- and post-fledging predation each nesting site and foraging area of the birds was inspected twice daily and mortality/predation, if any, was recorded. In the Spotted Munia, fledglings forage along with their parents for 16-20 days and returned to their nest daily for about a week, so it was easy to watch and count the number of fledglings from each nest in this study. In each habitat mean values of clutch size, number of hatchlings, number of nestlings that survived until fledging and number of fledglings that survived for more than one week after fledging were calculated on a per nest basis. Hatching success was calculated on the basis of the total number of eggs that were laid. Similarly, nestling fledgling and post-fledging suc- cess were calculated using the total numbers of eggs that hatched. A two way ANOVA (Baily 1995) statis- tical test was used to compare the data (all breeding characteristics) between forest and urban habitats and between years as well as habitat x years. RESULTS In the urban area a total of 54 eggs were laid in 10 nests (20 eggs in 1997, 18 in 1998, and 16 in 1999), out of which 11 eggs were destroyed by predators and 5 eggs remained unhatched due to some infection (See Table 1). The incubation period was recorded for 10 nests. The three-year average incubation pe- riod (N=10) was found to be 13.97+0.71 (mean+ SD) days in urban area. Table 1. Comparison of incubation period (mean+SD) in Spotted Munia in urban and forest area. Habitat ae No. of | No. of Average incubation nests eggs period (days) Urban 1997 4 20 14.66+0.34 1998 3 18 13.66+0.34 1999 3 16 13.75+0.25 3 Years 10 54 13.97+0.70 Forest 1997 3 16 14.20+0.70 1998 4 22 12.50+0.15 1999 4 22 14.15+0.08 3 Years 11 60 13.30+0.27 ANOVA Habitat df=1 F=1.090 P>0.05 Years df=2 F=0.181 P>0.05 Habitat X Years df=2 F=3.015 P>0.05 Breeding success of the tropical Spotted Munia In the forest area a total of 60 eggs (16 in 1997, 22 in 1998, and 22 in 1999) were laid in 11 nests, out of which 6 eggs were eaten by predators and 3 eggs were found thrown out of their nests, thus 51 eggs were incubated. The three-year average incubation period (recorded from 11 nest) in the forest area was 13.30+0.27 days. The two way ANOVA test indi- cated no variations in incubation period between habitats or years or habitat x years (Table 1). Clutches hatched over periods of 2 to 4 days. The average hatching period in the urban area was 3.37+0.85 days. Unhatched eggs were removed by parents within 2 to 3 days of the last egg hatching. In the forest area the average hatching period was 3.40+0.57 days. ANOVA test indicated no variation in hatching period between habitat or years or habitat x years (Table 2). The three-years mean clutch size was 5.66+0.93 in urban habitat and 5.40+0.98 in forest. There were no variations in clutch size between habitats or years or habitat x years (Table 3). In the urban area 158 eggs were counted from 28 nests out of which 3.8% eggs were rejected by par- ents due to unknown factors, 22.8% were destroyed before hatching and 3.2% remained unhatched. Thus only 70.2% of all eggs hatched. In the forest area out of 131 eggs, 79 hatched. Six nests containing 34 eggs were destroyed by predators. A further 18 eggs from 9 nests remained unhatched due to infection or un- known factors. Thus the total hatching success (of 3 years) was 69.3% in the forest area. The three-years mean number of hatchlings was 3.97+2.00 in the urban area and 3.26+2.32 in the forest (see Table 3). Our observations revealed 57.7% fledging success in Spotted Munia in urban habitat out of which 9.0% were eaten by predators thus survival of fledglings in urban habitat was 48.6%. In forest habitat, out of 79 eggs hatched only 41 nestlings flew indicating 51.9% nestling success, out of which 20.2% were eaten by predators. Thus only 31.6% survival of fledglings was recorded. The three-years mean of nestlings that survived until fledging was 2.28+1.39 in urban habi- tat and 1.70+1.31 in forest habitat, while the mean number of fledglings that survived for more than one week after fledging in urban areas was 1.94+1.20 and in forest areas 1.03+1.18 (Table 3). Two way ANOVA test indicated no significant differences in number of hatchlings between habitats or years. However, significant differences in number of fledg- lings and one-week surviving fledglings have been observed between habitats (Table 3). 115 Table 2. Comparison of hatching period (mean+SD) in Spotted Munia in urban and forest habitats. Habitat cae No. of | No. of Average hatching nests eggs period (days) Urban 1997 3 13 3.25+0.73 1998 3 15 3.21+0.82 1999 2 10 3.65+0.64 3 Years 8 38 3.37+£0.85 Forest 1997 3 15 3.26+0.68 1998 4 21 3.10+0.80 1999 3 15 3.50+0.72 3 Years 10 51 3.40+0.57 ANOVA Habitat df=1 F=0.016 P>0.05 Year df=2 F=0.122 P>0.05 Habitat x Year df=2 F=0.544 P>0.05 DISCUSSION It is clear from the results that the total breeding success of Spotted Munia in the urban area (48.7%) was significantly higher than in the forest area (31.6%) reflecting less predation in urban habitat than in forest. There was no difference in hatching period or clutch size between the habitats (Table 2). The maximum hatching period was 4 days. It has been suggested by some workers that the hatching of eggs over a period of several days helps in protecting the clutch and brood from the dangers of predation (Van Tyne & Berger 1959), but Lack (1947) suggested that it is an adaptation to bring brood size and available food supply into correspondence. In case of an avian species breeding success depends to a great extent on its ability to decide where and when to nest (Cody 1985; Robertson 1995). In forest habitat Spotted Munia builds nests generally on thorny plants clearly visible from the outside. Although such nests are con- spicuous, nest predation by tree climbing or flying predators is largely prevented by the thorns of the trees. In the urban study area, the plants selected for nesting, such as Ashoka pendula, Thuja orientalis, Feronia elephantum, were dense and nests were well covered by foliage from all sides. It was therefore difficult for the predators to find the nests. In the forest area since the predation pressure was R. C. SHARMA, D. BHATT and R. K. SHARMA. Table 3. Comprison of clutch size, numbers of hatchlings, fledglings and one-week surviving fledglings (mean+SD) in Spotted Munia in urban and forest habitat. en Mia f s a! No. of No. of No. of surviving Habitat/ Year No. of nests Clutch size Hatchling Bereiaee lesley Urban 1997 12 5.50+1.08 3.91+2.02 P23) a) \55)5) 1.91+1.37 1998 7 5.70+0.75 4.00+2.00 2.42+1.27 2.14+1.06 1999 9 5.80+0.97 4.00+2.00 2.1141.36 Le eS) 3 years 28 5.66+0.93 3.97+2.00 2.28+1.39 1.94+1.20 Forest 1997 8 5.50+1.06 BS 235) 1.87+1.35 1.00+1.19 1998 9 5.60+1.00 Sy 5es2.29) 1.66+1.32 1.11+1.36 1999 i 5.10+0.89 3.00+2.44 NeSy 7/3 he 27/ 1.00+1.00 3 years 24 5.40+0.98 3.26+2.32 1.70+1.31 1.03+1.18 ANOVA-F values Habitat df=1 1.104 2.629 4.549* 8.209** Year df=2 0.438 0.219 1.478 1.102 Habitat X Year df=2 0.272 0.284 0.166 0.105 *P=<(!05; =42P=<0:01 high this species selected urbanized habitat for breed- ing purpose. It may be suggested that factors such as higher ambient temperature, greater food availability (since agricultural fields were closer to urbanized habitat) and fewer predators were contributing to en- hance the breeding success of the Spotted Munia in the urban habitat. Apart from the Spotted Munia our observations on the Red-vented Bulbul Pycnonotus cafer also indicated that the breeding success of this species was more in urbanized habitat than forest be- cause of low predation pressure (reduced post fledg- ling and adult mortality) in the urban area (Bhatt & Kumar 2003). Similarly, other avian species such as the European Woodpigeon Columba palumbus, Her- ring Gull Larus argentatus, Eurasian Blackbird Tur- dus merula, Mallard Anas platyrhynchus and Magpie Pica pica have been found to experience a higher breeding success in the urban environment than in wild habitat (Snow 1958; Lack 1966; Cramp 1972; Monoghan 1979; Tomialoje 1979; Bentz 1985; Ka- vanagh et al. 1989). From the present study it is clear that Spotted Munia has adapted quickly and successfully to urban- ized habitat, in particular in selecting introduced plants for nesting in. It may be mentioned that after breeding in urbanized habitat both parents and juve- niles leave the area and spend the remainder of the year (approximately from mid November to mid 116 June) in agricultural and wild habitats. Which fac- tor(s) (ultimate or proximate) stimulates the Spotted Munia to shift habitat is not currently clear but this is an interesting finding and needs further investigation. Actually such shifts in behaviour do not occur in- stantly and newly acquired behaviour takes time to spread. It would be interesting to resolve whether such a shift in habitat in the Spotted Munia is based on cultural transmission, or genetic change. Here it 1s worth mentioning that not only the Spotted Munia but at least one other member of the same family, the White-throated Munia L. malabarica is also attempt- ing to adapt to urban habitat for breeding (Sharma 2002). ACKNOWLEDGEMENTS The authors are grateful to Prof. B. D. Joshi, Head of the department, for providing necessary facilities to conduct this work. We thank Mr. Y. K. Negi, Mr. L. Sanjit and Mr. Vinaya Kumar Sethi for their help in statistical analysis and/or typing of this manuscript. REFERENCES Ali S & Ripley SD (1987) Compact handbook of the birds of India and Pakistan. Oxford University Press, Bombay. Baily NTJ (1995) Statistical methods in biology 3rd ed. Breeding success of the tropical Spotted Munia Cambridge University Press, Cambridge. Baker ECS (1938) The evolution of breeding seasons in birds. In: Debeer GR (ed) Essay on aspect of evolu- tionary biology. pp 161-177. Oxford University Press, London. Bentz PG (1985) Studies on some urban mallard (Anas platyrhynchos) populations in Scandinavia. 1. cause of death, mortality and longevity among Malma mal- lards as shown by ringing recoveries. Fauna Narv Ser C, Cinclus 8: 44-56. Bhatt D (1982) Circannual Rhythms in spotted munia with special reference to seasonal reproduction, PhD Thesis. B.H.U. Varanasi. Bhatt D & Kumar A (2003) Breeding behaviour of Red- vented bulbul Pycnonotus cafer in a suburban habitat with special reference to its nest characteristic. Proc of the PASOC, Coimbatore. Chandola A & Thapliyal JP (1978) Regulation of repro- ductive cycles of tropical spotted munia & weaver bird, In: Assenamacher I & Farner DS (eds) Environ- mental endocrinology. pp 61-63. Springer Verlag, Berlin. Chandola A, Bhatt D & Pathak VK (1983) Environmen- tal manipulation of seasonal reproduction in spotted munia Lonchura punctulata. In: Mikami S (eds) Avian endocrinology, environment environment and ecological perspectives. pp 229-242. Japan Sci Soc Press, Tokyo/Springer Verlag, Berlin. Cody ML (1985) Habitat selection in Bird. Academic press, New York. Cramp S (1972) The breeding of urban woodpigeons. Ibis 114: 163-171. Kavanagh BP, Jerzak L & Gorsla W (1989) Factors af- fecting the breeding performance of the magpie (Pica pica) in three European cities. Proc Intl Symp 117 Grainivorous birds. pp 70-81. Lack D (1947) The significance of clutch size. Ibis 89: 302-352. Lack D (1966) Population studies of Birds. Clarendon Press, Oxford. Miller AH (1965) Capacity for photoperiod response and endogenous factors in the reproductive cycles of an equatorial sparrow. Proc Natl Acad Sci USA: 97-101. Misra AB (1962) The recurrent sexual cycles of birds. Presidential address. Second all India congress zool- ogy. pp 1-14. Monoghan P (1979) Aspects of breeding biology of Herring gull Larus argentatus in urban colonies. Ibis 121: 475-481. Nice MM (1954) Studies in the life history of the song sparrow. Trans Linn Soc NY 6: 1-329. Robertson GJ (1995) Factors affecting nest site selection and nesting success in the common eider Somateria mollissima. Ibis 137: 109-115. Sharma RC (2002): Breeding Strategies in spotted munia Lonchura punctulata in urbanized and forest habitats, PhD Thesis, Gurukula Kangri University, Hardwar, India. Skutch AF (1957) The incubation pattern of birds. Ibis 99: 69-93. Snow DW (1958) The breeding of blackbird Turdus merula at Oxford. Ibis 100: 1-30. Thapliyal JP (1978) Reproduction in Indian birds. Pavo 16: 151-161. Tomialojc L (1979) The impact of Predation on urban and rural Woodpigeon (Columba palumbus L.) popu- lations. Pol Ecol Stud 5 (4): 141-220. Van Tyne J and Berger AJ (1959) Fundamentals of or- nithology. John Willey & Sons, New York. Li) ag tne nnnina | vi T iv ee1) GO Ada C2é—LOF » taiulon) po00)) CO goend taotzO ast j vi in Teac rT) he tA iv oe! euonsaobos Dre 57 .wonnqe, Igrioips. as Inl-~Te In52 InaTI97 oe | Cadi tld, mail nas? zenbbha lalogbien'l sinh coy 136) eA (UURh 4! mitigonol nol toy galnwh PEL eTY, thf! f it strut nr eibure- (BP Oh i sol OOF —| ( 902 anid anei) org A ‘ 4:(¢ ate aoaiiwia! ' Tug atk Tr az ate anujeon bebb [<0O1 VET etch ater le: "tae JA anni 70 @ lest : Aiioriie: IVR ties | ; i in i clin ew? OA bithaadt aint . whi Fyre) Pear] svete ee ORRIN aati ‘fi T board ef Petthels 4/Chamond CNL ade dpgephec yu: lay ean be : a UAT eA Lag ratpasd | 4 LOir henpal « 1] mtn ort} GAUL) oJ wy igalisis ma] mule ieee Levey bod Och BT (By 8 tami lal andi : i) LA igual Jina | L wy ‘ me L valle ¥é AgoOL sv yt ~ - rl : (tet jive Eis us ame bh twee ody RCE . J TPH apie ra set a J ROA, gt: i = te “ww raps entl ae ah AUD ENE Day RCSD gE Ma: i anoniee gribosid Io coiulove sdT (BERL 208 -uloes- Wy vee ne*vorrs (bn) re] reodad- ee oa fet wedi cre in iii — J 7 a ORIGINAL ARTICLE Ornithol Sci 3: 119-124 (2004) Notes on feeding structures of the Black-faced Spoonbill Platalea minor Cornelis SWENNEN*™* and Yat-tung YU Department of Ecology and Biodiversity, University of Hong Kong, Pokfulam Road, Hong Kong ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2004 Abstract The bills of the spoonbills differ from the bills of most birds by being wider near the downward curved tip than in the middle, and having the mandibles dorso-ventrally extremely flattened. The mandibles have rounded lateral borders and lack cutting edges. The inward directed sides have a dense cover of thin parallel ridges on the distal parts and rows of teeth-like tubercles in the proximate parts. Their skeletons have numerous small pits in the distal parts especially along the edges and the insides. These pits are similar to the spaces for sensory corpuscles for touch in the bills of Scolopacidae (sandpipers, snipes) and are presumed to have the same function in spoonbills. The bill seems adapted for tactile feeding with lateral movements (sweeping) and for pecking, but not for probing into sediments. The wide gape with gular pouch allows the swallowing of rather large food items. The muscular layer of the gizzard is weakly developed and the gizzard is more a digestive pocket than a chewing organ such as occurs for grinding hard shells and grains in molluscivorous and granivorous birds. The long legs are laterally flattened, perhaps for minimising resistance and not-disturbing prey when walking in water during feeding; the partly webbed feet with long toes allow walking over soft mud bottoms. Key words minor Spoonbills (Ciconiiformes, Threskiornithidae) are large wading birds of 1-2 kg with long legs, necks, and bills. Their bills are remarkably widened at the tip and not pointed such as in most birds. The six spoonbill species are very similar in shape and behaviour, mainly differing in size, colour of legs, bills and other bare parts, and in distribution (del Hoyo et al. 1992). Spoonbills are considered to be tactile feeders that feed by walking in shallow water meanwhile sweeping their bills from side to side through the water (Kushlan 1978; Hancock et al. 1992). The bill is the morphological structure that is used for collecting food. In this respect, the atypical bill of a spoonbill looks clumsy compared to the bill of a chicken or egret. This has given reasons for spec- ulations that the shape has a specific function without studying the morphology of the whole bill. That the bill can be used for sieving small prey out of sedi- (Received 20 November 2003; Accepted 12 August 2004) * Corresponding author, E-mail: swennen@wxs.nl * Present address: Doolhof 7, 1792 CM Oudeschild, Texel, The Netherlands 119 Bill, Black-faced Spoonbill, Digestive tract, East Asia, Leg, Platalea ment or water as can be done by the bills of ducks is already rejected by Allen (1942) who showed that the bill lacks the necessary rows of lamellae for sieving. Vestjens (1975) has tried to relate some relative size differences between the bills of the Royal Spoonbill P. regia and Yellow-billed Spoonbill P. flavipes and suggested that the longer bill of the latter was more suited for probing. Different speculations about the function of knobs or teeth in the mandibles were made by Allen (1942) and Vestjens (1975). That the bill can act as suction apparatus to disturb and moves benthic prey items has been suggested by Weihs and Katzir (1994) for which they specifically assumed that the upper mandible is convex in cross section. All in all, however, the mandibles have never been described in some detail, the specific feeding method of spoonbills is not well understood and there is con- fusion about their food. Food lists in the usual hand- books (Cramp & Simmons 1977; del Hoyo et al. 1992) give the impression that spoonbills are omniv- orous, but the morphology of the alimentary tract in which the food items are processed and the legs, C. SWENNEN and Y. YU which are as well of importance for the way the food can be captured are neglected in studies about the food of spoonbills. An accurate description of the morphological structures used in collecting and pro- cessing food is an indispensable base for studying and understanding the feeding of any organism. In this paper, we study these aspects in some detail. We focus on the Black-faced Spoonbill Platalea minor, Which is the spoonbill species present in our study area. Its distribution is confined to East Asia where they mainly breed around the North and East Coast of the Yellow Sea and winter from South Korea to Vietnam (Kennerley 1990; del Hoyo et al. 1992; Hancock et al. 1992). We give a description of the morphology of the main feeding organs (bill, intes- tinal tract, and legs). The feeding method and the potential foods of the species and of spoonbills in general will be deduced from these morphological data. MATERIALS AND METHODS Live specimens handled were wintering Black- faced Spoonbills that had been caught in the Mai Po Marshes Nature Reserve (Hong Kong) in the winters of 1998-99 and 1999-2000, where they were marked for a migration study (Melville et al. 1999). We have discriminated first winter birds from older ones by having a pale coloured bill, totally brown iris, and the presence, distribution and size of black on the wing feathers. In adults the bill is black, the iris red, and black on the wing feathers is lacking. Intermediate specimens were considered in their second winter. Before releasing the birds, measurements were taken of morphological characteristics from adult birds of unknown sex. Length of legs were measured from the ankle (intertarsal joint) and the feathers on the belly straight to the floor in the natural stand. Bill length was measured as the shortest distance between the tip and the start of the feathering over the length axis of the upper mandible. Further details of the bill were studied from some skulls of preserved adult speci- mens in private collections in Hong Kong and Tai- wan, and of two fledglings in a private collection in South Korea. One preserved alimentary canal was given to us in Hong Kong, it originated from a win- tering specimen that was found freshly dead. 120 RESULTS Bill Bill length of Black-faced Spoonbills wintering in Hong Kong is 181.7+13.0mm (163-207 mm) (mean+SD and range; N=22), while the width of the spoon is 51.0+2.2mm (47-55 mm). In lateral view, the bill looks thin, is bent down near the tip and shows a 24mm wide gap in the central part (Fig. la). In dorsal view the mandibles are wide (Fig. 1b). The upper mandible of the bill is wider than the lower in the distal part, but the lower mandible is wider than the upper in the proximal part (Fig. 1b,c). The slit-like external nostrils are close to the head in a suture that runs more or less parallel to the border of the upper mandible. The nail on the tip of the upper mandible is small and not pronounced 1) Fig. 1. (a) Lateral view of the bill showing the gap and the flatness of the mandibles. Drawn from a photograph. (b) Bill in dorsal view. (c) Cross sections of the mandibles (at the same scale). The position of the sections is indicated in the left drawing. (d) Comb-like incisions at the tip of the bill. A: Border of upper mandible in frontal view. B: Border of lower mandible dorsal view. (e) Parallel ridges on the ventral side of the lower mandible. Im: lower mandible, na: nail, no: nostril, sp: spoon, su: suture, tu: tubercles, um: upper mandible. Feeding structures of the Black-faced Spoonbill (Fig. 1b). The mandibles are rigid to lateral forces, and only slightly flexible to dorso-ventral forces. Both mandibles are dorso-ventrally extremely flattened. The thickness of the spoons is 2mm, gradually increasing proximally and reaching 7.6+0.3mm (mean +SD, N=10) in the upper mandible and 5.5+0.4mm in the lower mandible at the distal part of the nostrils (Fig. la, c). The epidermal layer is thin over the whole length of the mandibles, slightly thicker along the distal borders of the mandibles where both show a series of comb-like incisions on both sides of the median line (Fig. 1d A, B). A few short and inconspicuous parallel ridges are found on the outside of the lower mandible (Fig. le). The insides of the mandibles are densely covered by minute (<0.5mm high) seemingly parallel, epi- dermal ridges in the distal parts (Fig. 2a). The ridges split and unite during their course, and the distance between their crests varies between about 0.3mm where they converge near the tip and 1mm where they diverge in the widest part of the spoon. Rows of minute depressions can be seen with the help of a magnifying glass on and between the dried ridges, between the comb-like incisions, and in the smooth outer sides of the spoons of dry preserved specimens. These probably indicate sites of sensory corpuscles in the dermal layer. The central parts of the insides of the mandibles show two longitudinal rows of widely spaced, less than 1mm _ high teeth-like cuticular tubercles (Fig. 2a). These tubercles are lacking in fledglings, and develop during the first winter. The lateral sides of the mandibles are rounded and lack any sign of tomia, i.e. sharp cutting edges (Fig. Ic). The two bony bars of the lower mandible are con- nected by soft, elastic tissue, which may expand to a gular pouch when swallowing large prey. This pouch is longitudinally folded; its colour is black in all age classes (Fig. 2a). The short tongue is triangular; the proximate side is thinly fimbriated (Fig. 2b). The surface of the bony skeleton of the bill shows fields with small pits in the distal parts. These pits are generally wider than the holes for the blood vessels. They are most dense along the sides of the mandibles from the tip up to over half of the bill length, on the insides and the outsides of the widened distal parts (Fig. 3a A, B, C, D). Most pits have an oval aperture varying in size between 0.4X0.5mm and 0.7X 1.1 mm, and a few are round with a diameter varying between 0.3 and 0.4mm. (Fig. 3b). Two major nerve branches run up through the bones of upper and Fig. 2. view. B: Lower mandible in dorsal view. (b) Tongue of Black- faced Spoonbill in dorsal view. gp: gular pouch, pr: parallel ridges, tu: tubercles. (a) Inside of the bill. A: Upper mandible in ventral lower mandible to the spoons and branch there. 2) “a ) Alimentary tract The alimentary tract from a fresh carcass found in Hong Kong had deep longitudinal folds in the about 27cm long oesophagus. The proventriculus was rather solid, about 5 cm long, 3cm wide, and could easily be expanded. The wall mainly existed of a layer of tubular glands (Fig. 3c D); the inside showed several (+30 per cm’) small openings of the glands. The ventriculus (gizzard) was about 7cm long, 5.5cm high and 4.5 cm wide (Fig. 3c A). The muscu- lar layer was 9 mm thick on the dorsal side and 2mm on the ventral and lateral sides (Fig. 3c C). Locally, it was a little thicker near the passage to the duodenum. The wrinkled touch lining of the lumen was about 5.9mm thick along the sides and about 3mm on the top and bottom. The intestine was 132 cm long; it had two short and narrow caeca with lumens of about 2.5mm deep. 3) les Tarsus and tibia are laterally compressed with di- mensions in cross sections of 6.5+0.2 mm (mean SD; N=4) and 12.6+0.1 mm, and 7.5+0.2 mm and 13.5+0.1 mm, respectively. The three frontal toes are C. SWENNEN and Y. YU J du >) 'g yt . 4 is}. (d) ° Fig. 3. (a) Distribution and relative density of pits in the bony skeleton of the mandibles of the Black-faced Spoonbill. The pits are shown as small dots. A: Upper mandible dorsal view. B: Upper mandible ventral view. C: Lower mandible dor- sal view. D: Lower mandible ventral view. (b) Detail photo- graph of pits in the bone of the ventral side of the upper mandible. The photographed part measures 4.1 X3.2 mm. (c) Stomach of a Black-faced Spoonbill. A: Ventriculus (gizzard). B: Proventriculus. C: Cross section of ventriculus. D: Cross section of proventriculus. The position of the cross sections is indicated by dotted lines. (d) Feet of Black-faced Spoonbill in ventral view. L: Left foot, R: Right foot. du: opening to duode- num, oe: oesophagus, tg: tubular glands, cl: cuticle layer, ml: muscular layer. partially webbed (Fig. 3d); the hind toe is relatively long and has a slightly raised base, but still makes a full print in muddy sediments. The span from the tip of the nail of the middle toe to the nail of the hind toe is 141+1.9mm (mean+SD; N=4), that between the tips of the nails of left and right toes about 119+ 1.0mm. In the standing leg, the distance between foot sole and ankle is 130.2+14.7mm and to the belly 225+22.6 mm (N=10). Inspection of preserved specimens in the collection of the National Museum of Natural History, Leiden (The Netherlands) demonstrated that the bills and legs of the Eurasian, Royal, Roseate, and Yellow- billed Spoonbill show the same details as described above for the Black-faced Spoonbill. They mainly differ in length, relative width of the distal part, and colour. DISCUSSION The main function of a bill is collecting, tasting, 122 and transporting food into the oesophagus. Foods of birds differ between species and so do the bills and methods of collecting food items. The other functions such as a structure for drinking, feather preening, nest building, and defence can be conducted with all dif- ferent kind of bills. Therefore we concentrate on the aspects of food collecting. The spoonbill usually feed by sweeping its bill in turbid water and preferentially at dusk and dawn (Yu & Swennen 2004a, b). We deduce that the extreme flatness of the mandibles will minimise drag and turbulence during the lateral movement while searching for prey in water. This will be mainly intended to avoid disturbing potential prey in advance because tactile feeding means that prey has to be touched by the bill and caught at once; a disturbed prey will flee from the source of distur- bance and will be difficult to locate without visual clues. The bill allows also grasping of visible food items, which is rarely noted in the Royal and Yellow-billed Spoonbill (Vestjens 1975) and Eurasian Spoonbill (Stienen & Brenninkmeijer 1993; Weihs & Katzir 1994). Feeding by probing is mentioned in the litera- ture for the African Spoonbill (del Hoyo et al. 1992) and Yellow-billed Spoonbill (Vestjens 1975), but it is not clear what is meant. The bill shape shows that probing into sediment is not an option for any spoon- bill, because the wide and curved tip would receive much resistance during penetration as well as during retraction, but when the frontal parts can be pushed into sediment the wide tip can not be opened or closed for grasping a burrowing prey, because the upper mandible has no flexible area that allows an ac- tively upward bending of the distal part (rhynchoki- nesis) such as occurs in snipes (Scolopacidae) and some other bird groups. The numerous pits in the skeleton are most likely spaces for the sensory receptors. Similar pits occur in the better-studied bills of Scolopacidae (snipes, sand- pipers), but in a slightly different position (Bolze 1968; Piersma et al. 1998). Scolopacid shorebirds make sewing movements and probe into sediments for food; they have the pits for sensory corpuscles densest on the dorsal and ventral sides of the tips of the mandibles. Spoonbills make lateral feeding move- ments with the bill open; they have the pits relatively most dense on the lateral and insides of the bill (Fig. 3b). The most common Herbst corpuscles have a length of 0.05—0.2 mm (Schwartzkopff 1973), the pits are large enough to lodge a whole series of them similar as has been found in Red Knots (Piersma et Feeding structures of the Black-faced Spoonbill al. 1998). Histological evidence is still lacking, but it is supported by the presence of major nerves embed- ded in the skeleton. The ordering of the pits differs from the rows of supposed sensory corpuscles that are visible in the epidermis of dry preserved bills. The occurrence of such a density of supposed sensory elements agrees with the tactile way of feeding (Kushlan 1978). The dense covering with low cuticular ridges on the insides of the distal parts of the mandibles may have a function in keeping hold of a slippery prey. All these details lead to the conclusion that the pecu- liar widening of the tips is necessary for improving the chance of catching moving food items that are sensed between the mandibles. This is correlated with their way of feeding in turbid water (Yu & Swennen 2004a) during twilight (Yu & Swennen 2004b) which prevent a visual detection of food items in the normal feeding situation. The tongue is too small for a function in the upward transport of food, its role may be in covering up the glottis during swallowing and perhaps in taste. The rudimentary tongues of large fish-eaters such as occurring in the Pelicaniformes and Ciconiiformes are often considered an adaptation allowing bulky foods to be swallowed whole and quickly (McLelland 1979). The transport of food through the bill will be conducted via catch-and-throw movements. The two rows of tooth-like projections more proximal in the mandibles seem suited for holding a prey during transport to the throat, and thus replace the function of the cutting edges which give hold on a prey in other fish-eating birds. Vestjens (1975) states that these tubercles in Royal and Yellow-billed Spoonbills are also used for chewing larger food items. This is unlikely, as the projections do not have a molar-shape in any spoonbill species, and their positions in the upper and lower mandible do not match (Figs. lc, 2a A, B). Allen (1942) suggests that these projections may contain specific sensory receptors; however, no indications of sensory elements are visible in the dried skin on and around the tubercles and there are also no sensory pits or perforations for nerves in the skeleton below the tubercles (Fig. 3a B, C). The rudimentary tongue, gular pouch and wide, folded oesophagus indicate that rather large food items can be swallowed. The spoonbill ventriculus (Fig. 3c C) is a wide pocket for digestion suited for feeding on fish or meat. It lacks the thick muscular layer of a gizzard for grinding or cracking hard foods such as occurs in birds swallowing hard seeds or shelled molluscs (McLelland 1979). This means that Black-faced Spoonbills can only digest meat of mol- luscs that have a thin shell or no shell at all. Food lists of spoonbills given in del Hoyo et al. (1992) contain in variable detail several taxa including mol- luscs and plants, which may largely be based on recognisable fragments in stomach contents. Shellfish is named in the second place as food of the Black- faced Spoonbill, but their anatomical structure does not support the eating of molluscs, which was also doubted for other reasons by Hsueh et al. (1993). Fragments of shells in a stomach may have been swallowed accidentally and do not represent food intake. The length of the legs up to the belly determines the water depth in which a spoonbill can feed (Yu & Swennen 2004a). The long partly webbed toes allow walking on rather soft mud, and the lateral flattening of the legs may reduce resistance and agitation when feeding in water which may be important for not-dis- turbing prey. ACKNOWLEDGEMENTS This study was funded by the Agriculture, Fisheries and Conservation Department of the Hong Kong SAR Govern- ment for the research in Hong Kong via WWF Hong Kong, and by the Dutch Society for the Protection of the Wadden Sea and the Van der Hucht Fund for the research in Taiwan and South Korea. We are indebted to private collectors in Taiwan and South Korea, and the Kadoorie Farm and Botanic Garden, Hong Kong for permission to examine the Black-faced Spoon- bill materials in their collections. Thanks are also due to C. Smeenk of the National Museum of Natural History, Leiden (The Netherlands) and O. Overdijk for permission to study materials of other species in their collections. This study could not have been conducted without their co-operation. We sincerely thank R. Corlett for his helpful comments and for improving our English. REFERENCES Allen RP (1942) The Roseate Spoonbill. Research Re- port 2. Natl Audubon Soc, New York. Bolze G ( 1968) Anordnung und Bau der Herbstlichen Korperchen in Limicolen Schndbeln im Zusammen- hang mit der Nahrungsfindung (Arrangement and structure of Herbst’s corpuscles in bills of Limicolae in relation to food finding). Zool Anz 181: 313-355 (in German). Cramp S & Simmons KEL (eds) (1977). The birds of the western palearctic, Vol 1, Oxford University Press, Oxford. C. SWENNEN and Y. YU del Hoyo J, Elliott A & Sargatal J (eds) (1992) Hand- book of the birds of the world, Vol 1, Lynx Edicions, Barcelona. Hancock JA, Kushlan JA & Kahl MP (1992) Storks, ibises and spoonbills of the world. Academic Press, London. Hsueh PW, Yen CW & Chou WH (1993) Food habits of Black-faced Spoonbills Platalea minor Temminck & Schlegel wintering in Taiwan. Bull Natl Mus Nat Sci (Taiwan) 4: 87-90. Kennerley P (1990) A review of the status of the Black- faced Spoonbill. Hong Kong Bird Report 1989. pp 116-125. Kushlan JA (1978) Feeding ecology of wading birds. In: Sprunt A, Ogden JC & Winkler S (eds) Wading Birds. pp 249-297. Natl Audubon Soc, New York. McLelland J (1979) Digestive system. In: King AS & McLelland J (eds) Form and Function in Birds, Vol |. pp 69-181. Academic Press, London. Melville DS, Leader PJ & Carey GJ (1999) Movements and biometrics of Black-faced Spoonbills Platalea minor at Mai Po, Hong Kong in spring 1998. In: Ueta M, Kurosawa R & Allen D (eds) Conservation and Research of the Black-faced Spoonbills and_ their habitats. Proceedings of the international workshop in Tokyo, 12-16 June 1997. pp 19-26. Wild Bird Society of Japan, Tokyo. Piersma T, van Aelst R, Kurk K, Berkhoudt H & Maas LRM (1998) A new pressure sensory mechanism for prey detection in birds: the use of principles of seabed dynamics? Proc R Soc Lond B 265: 1377-1383. Schwartzkopff J (1973) Mechanoreception. In: Farner DS & King JR (eds) Avian biology. Ill. pp 417-477. Academic Press, New York. Stienen EWM & Brenninkmeijer A (1993) Spoonbill Platalea leucorodia preying on gull eggs. Limosa 66: 29-30. Vestjens WJM (1975) Feeding behaviour of spoonbills at Lake Cowal, NSW. Emu 75: 132-136. Weihs D & Katzir G (1994) Bill sweeping in the spoon- bill, Platalea leucorodia: evidence for a hydrody- namic function. Anim Behav 47: 649-654. Yu YT & Swennen C (2004a) Habitat use of the Black- faced Spoonbill. Waterbirds 27: 129-134. Yu YT & Swennen C (2004b) Feeding of wintering Black-faced Spoonbills in Hong Kong: When and how long? Waterbirds 27: 135-140. ORIGINAL ARTICLE Ornithol Sci 3: 125-132 (2004) Microscopic structure and distribution of various elements in the eggshell of the Black-tailed Gull, Larus crassirostris, as revealed by scanning and transmission electron microscopy and X-ray compositional microanalysis Akira CHIBA* Department of Biology, Nippon Dental University School of Dentistry at Niigata, Niigata 951-8580, Japan ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2004 Abstract The present paper describes the microscopic structure and distribution of various elements in the eggshell from the Black-tailed Gull Larus crassirostris. Scan- ning and transmission electron microscopy (SEM and TEM) in combination with light microscopy demonstrated four major zones in the eggshell: the shell membrane (SM), the mammillary zone (MZ), the palisade region (PR), and the cuticle layer (CL). The SM was composed of a further three layers: i.e. the limiting membrane and the inner and outer SMs consisting of thin fibers. The MZ was a layer lined with coni- cal knobs, the mammillare, each of which had a core portion characterized by a dense matrix with a few vesicles and aggregated fine fibrils. The PR showed a spongy fea- ture depicted as numerous vesicles embedded in a calcified matrix. The CL appeared as a simple structure different from that seen in the eggshells from other birds, such as grebes, cormorants and domestic fowl. X-ray compositional microanalysis (XCM) re- vealed differences in the distribution patterns of certain elements (Ca, Mg, and P) in the radial face of the eggshell. The concentration of Ca was markedly high through- out the true shell (MZ and PR), whereas that of Mg was locally high in the MZ. The concentration of P was slightly higher in the surface crystal layer of the PR than else- where. Key words Eggshell, Fine structure Avian eggshells are morphologically complex structures and have long been objects of studies in poultry science, ornithology, and general biology. A considerable amount of knowledge on the morphol- ogy and physiology of eggshells has hitherto been ac- cumulated and reviewed (Romanoff & Romanoff 1949; Gilbert 1979). The greater part of our knowl- edge of eggshell structure was obtained from studies on the domestic hen Gallus gallus (Bellairs & Boyde 1969; Fujii & Tamura 1969; Tung & Richard 1972; Tan et al. 1992: Dennis et al. 1996; Fraser et al. 1999). However, there is a growing body of data from the birds of various taxonomic groups (Tyler 1964, 1965, 1966, 1969; Becking 1975; Tullet et al. 1976; Tullet 1984; Mikhailov 1995a, b). These studies have contributed not only to comparative morphology of (Received 10 February 2004; Accepted 13 May 2004) * E-mail: chiba@ngt.ndu.ac.jp Black-tailed Gull Larus crassirostris, Distribution (Ca, Mg & P), the eggshells, but also to the respiratory physiology of the developing embryos (Paganelli et al. 1975; Rahn & Paganelli 1981; Tullet 1984) and considera- tion of avian systematics (Mikhailov 1995b). These studies have also revealed variations in the micro- scopic structures of eggshells and discrepancies in terminology used to describe them (Gilbert 1979; Mikhailov 1995a; Dennis et al. 1996). For a better understanding of the structural details and functional properties of avian eggshells, comprehensive studies conducted with the aid of different techniques are needed. To our knowledge, however, such studies are scanty and seem to be limited to the eggshell of do- mestic fowl (Tan et al. 1992; Dennis et al. 1996; Fraser et al. 1998; Richards et al. 2000). The present study was conducted to expand our knowledge of the microscopic structure and the dis- tribution of elements or minerals in the eggshells of wild Japanese birds by the use of modern techniques. A. CHIBA For this purpose, the eggshell of the Black-tailed Gull ussirostris was selected for sampling and in- vesugation, j LUTUS ¢ MATERIALS AND METHODS 1) Materials Four fresh, but slightly broken, Black-tailed Gull eggs were used in this study. They were collected (by permission of the Ministry of Environment) from two abandoned nests, during a field survey of breeding colonies on the coast of the Japan Sea in the northern district of Niigata Prefecture. The nest abandonment and egg damage seemed to have been caused by mis- chievous intruders. The eggs were opened in the lab- oratory, and their contents were carefully examined and removed. The shells were then processed for mi- croscopic observation. The eggs used were diagnosed to have been fertile and in an early stage of normal embryonic development. 2) Scanning electron microscopy For SEM, parts of the eggshells were washed in tap water, rinsed with distilled water, cut into small pieces (about 55mm), and dried for 48 hr at 38°C in an oven. Then, they were mounted on an alu- minum stub, sputter-coated with platinum-palladium, and examined with a Hitachi H-800 scanning electron microscope. The rest of the eggshells were incubated overnight in a 1% solution of potassium hydroxide at 38°C to remove organic compounds from the shells, washed in tap water, rinsed with distilled water, and dried in the oven. Then, they were similarly processed and examined for comparison. 3) Light and transmission electron microscopy For TEM, pieces of the shells were appropriately trimmed, washed briefly in tap water, rinsed with 0.1M_ phosphate-buffered saline (0.01M _ sodium phosphate, 0.15 M sodium chloride, pH 7.2) and fixed in a solution of 0.1 M sodium cacodylate containing 1% glutaraldehyde and 0.1M sodium ethylened- imine-tetraacetate (EDTA), pH 7.6. For fixation and decalcification, the samples were placed in conical tubes and continuously shaken with an _ electric shaker. After 24hr, the fixative was replaced with freshly prepared 0.1 M sodium cacodylate containing 0.1M EDTA. The samples were then incubated in this medium overnight, with gentle shaking, to re- move any remaining calcium. Next, they were washed three times in 0.1 M sodium cacodylate, post- 126 fixed for | hr in 0.1 M sodium cacodylate containing 1% osmium tetroxide, and washed once in 0.1M sodium cacodylate and twice in distilled water. The samples were then dehydrated by passage through a graded ethanol series and embedded in Spurr resin. Semi-thin sections were stained with toluidine blue in borax and observed under a light microscope. Ultra- thin sections were cut with a diamond knife, stained with uranyl acetate and lead citrate, and viewed with a JEOL 1200 EX electron microscope. 4) X-ray compositional microanalysis Pieces of the shells were embedded in epoxy resin and a radial face was polished with successively finer grades of diamond paste (final grade, | um). A final polish was given with aluminum oxide (0.3 um grade) as a paste in velvet supported by a plate glass. The polished surface was coated in vacuo with gold (ca. 30nm) and examined with a JEOL JXA-50A electron probe microanalyzer. The electron beam (ac- celerating voltage, 20kV; beam current, 0.009 WA) was scanned along a line across the polished radial face, beginning at the shell membrane, continuing across the tip of a cone, and terminating at the plastic at the outer edge of the shell. The results on Ca, Mg, and P were recorded on a chart moving at a known speed. No examination was made of other elements, such as O or S. RESULTS 1) Scanning electron microscopy SEM examination of the Black-tailed Gull eggshell showed a complex architecture that differed within each of the major zones of the eggshell: the shell membrane (SM), the mammillary zone (MZ), the pal- isade region (PR), and the cuticle layer (CL; Fig. 1A). The SM, the innermost zone of the eggshell, was 139.7+2.4 um (mean+SD, N=4) thick and con- sisted of three layers, i.e. the limiting membrane (LM), inner SM, and outer SM. The LM that contacts directly with the albumin of intact eggs appeared as a thin film and could be partly detached from the inner SM by utilizing a pair of forceps (Fig. 1B). The inner SM, 51.4+5.3 um (N=4) thick, was composed of fibers that ran parallel to the eggshell surface and crossed each other at different angles. The individual fibers measured 1.2+0.2 um (N=10) in mean diame- ter and had minute granular processes on its surface. The outer SM, 90.6+2.3 um (N=10), was also com- posed of similar fibers, although each individual fiber Eggshell structure of the Black-tailed Gull Fig. 1. SEM micrograph of an eggshell of the Black-tailed Gull Larus crassirostris, showing its inner and sur- face fine structures. (A) low-power view of radially fractured face of the eggshell, showing the cuticle layer (CL indistinct due to low-power view), the shell membrane (SM), mammillary zone (MZ), and palisade region (PR). (B) the shell membrane consisting of the limiting membrane (LM), and inner (I) and outer (O) SMs. (C & D) de- tails of the inner (C) and outer (D) SMs composed of thin fibers. (E) inner face of the MZ showing the mammillare (M) and internal opening (arrows) of the pore canal after maceration with potassium hydroxide. (F) closer view of the PR, showing its spongy feature owing to the numerous vesicles embedded in the calcified matrix. (G & H) closer view of the external opening of the pore canal, with (H) and without (G) maceration. In the macerated spec- imen (H), the densely mineralized PR with crystal structure (asterisk), is disclosed and shows a rough surface due to the removal of the CL from the shell. In contrast, the non-macerated specimen (G) shows a rather smooth sur- face due to the presence of the cuticle layer. Organic or cuticle substance (arrow in G) is visible in the pore canal. Scale bar=100 um (A & B), 50 um (E), 5 um (C, D, G, & H), 1 um (F). 127 A. CHIBA the outer SM was thicker, 3.8+0.6 um (N=10) in diameter, than the fibers of the inner SM and had a th surface (Figs. 1C, D). The border between the inner and outer SMs was usually indistinct. The MZ occupied about 25-30% of the thickness of the min- eralized (calcified) portion of the eggshell, and was 79.8+2.4 um (N=4) thick. It consisted of numerous conical knobs, the mammillare, the apices of which were embedded in the outer SM and, via the apices, fibers from the outer SM were connected with the mammillare (Fig. 1A). This feature of the apices was clearly shown as irregular processes and grooves in the mammillare when the shell had been macerated with potassium hydroxide solution (Fig. IE). The PR external to the MZ had a spongy feature, being en- dowed with numerous vesicles ranging from 0.5 to 1.3 um (0.9+0.1 um; N=4) in diameter (Fig. 1F). Sometimes, two or more adjacent vesicles were con- nected to each other. Occasionally the PR showed a squamatic pattern. The external part of the PR was distinguished from the major (spongy) part of the PR; there were few vesicles, and the calcified matrix ap- peared to be very compact, showing an appearance of denser mineralization (Fig. 1A). This part was com- parable to the external zone or surface crystal layer recognized in other species. The CL, the outermost layer of the shell, was thin (4.1+0.4 um thick; N=4) and contained occasional vesicles or minute pores. It was devoid of any microglobular structures, which were recognized within the cuticle of other avian species. The presence of the CL was confirmed by comparison of the shells with or without pre-treat- ment with potassium hydroxide (Figs. 1G, H). ol \ 9, a ) Light and transmission electron microscopy Light microscopic examination of semi-thin sec- tions allowed characterization of the overall organic matrix architecture of the decalcified eggshells (Fig. 2A): the four major zones, SM, MZ, PR, and CL, of the eggshells were also recognized in the sections stained with toluidine blue, as already shown by SEM. TEM examination of the SM identified the LM, about 2 4m thick, and the fibrous elements in this zone (Figs. 2B, C). The LM appeared as a heteroge- neous structure consisting of electron-dense and less electron-dense materials. The LM contacted with the fibers of the inner SM without forming any special structure between them (Fig. 2B). The fibers of the inner SM showed a round or ovoid contour in cross section and had a homogeneous content of moder- ately electron-dense material although they were as- sociated with highly electron-dense particles (Fig. 2B). The fibers of the outer SM were thicker in diam- eter than those of the inner SM, but no essential dif- ference was found in ultrastructure between them. In the apices of the mammillare, the fibers of the outer SM connected to the apices were surrounded by a mantle-like layer of less electron-dense material, from which fine fibrils projected mainly to the PR (Fig. 2C). These fibrils, 27.8nm in mean diameter, were densely aggregated in the core portion of the mammillare, where they existed with occasional vesi- cles and electron-lucent flocculent material (Fig. 2D). The PR was characterized by the presence of numer- ous vesicles with an electron-dense fringe, 0.9 zm in mean diameter, and embedded in a matrix substance together with the flocculent material (Fig. 2E). The vesicles and the flocculent material became remark- ably less dense in the external part of the PR than in the major PR zone. The CL was composed of an elec- tron-dense amorphous material, and some parts of it were associated with a thin layer of the fibrillar ele- ments (Fig. 2P). 3) X-ray compositional microanalysis The distribution patterns of Ca, Mg, and P in the radial face of the eggshell were simultaneously re- vealed (Fig. 3). The concentration (or count) of Ca was markedly high throughout the true shell (MZ and PR), but was low or almost negligible in the SMs. In the inner zone of the shell, the concentration curve of Ca tended to be slightly reduced. On the other hand, a locally high concentration of Mg was noted immedi- ately beneath the inner surface of the shell, approxi- mately corresponding to the MZ. The peak Mg con- centration was found in the inner part occupying about 1/6 (ca. 17%) of the total thickness of the radial face of the true shell and consisted of two indistinct peaks. In the outer part of the shell and also through- out the shell membrane, the Mg concentration was clearly low. The concentration of P was slightly higher in the surface region, corresponding to the ex- ternal zone (or surface crystal layer), than in any other layers of the shell or in the SMs and tended to decline progressively toward the inner zone. DISCUSSION The present study on the eggshell of the Black- tailed Gull demonstrated four major zones in the ra- dial face: the shell membranes, the mammillary zone, Eggshell structure of the Black-tailed Gull Fig. 2. Light (A) and TEM (B-F) micrographs of the decalcified eggshell. (A) cross section of full thickness of the eggshell showing the shell membrane (SM), mammillary layer (MZ), palisade region (PR), and cuticle layer (CL). (B) part of the SM showing the limiting membrane (LM) and inner SM (i). (C) part of the outer SM (0) showing fine fibrils (f) in the matrix and less electron dense mantle layer (arrowhead) of filaments. (D) core por- tion of the mammillare, characterized by dense aggregation of fine fibrils (asterisks) and a few vesicles (arrows). (E) part of PR of the true shell proper, showing a spongy feature depicted by numerous vesicles with an electron- dense fringe. (F) the external part of the shell, showing a few vesicles in the matrix and the cuticle layer (CL) asso- ciated with discontinuously with a thin coat consisting of fibrillar material (double arrow). Insets (C and D) show the fine fibrils. Scale bar=100 um (A), 1 wm (B-F). 129 A. CHIBA —_ = ® £ 2 Lu = oO oO wi we |. ° i ~~ c {I t s Eo = : ——— 8 | : P- SS SSS == SSS SS ® if ay i T aoe =| ~ i ic - ® ~ Peary a Rare ear aLt a AvAOity Fig. 3. Different patterns of Ca, Mg, and P concentration (counts) in the radial face of the eggshell obtained by X-ray compositional microanalysis. MZ mammillary zone; R resin as embedding material, S true shell; SM shell membrane. the palisade region, and the cuticle layer, as recog- nized in the hen eggshell (Dennis et al. 1996). Ac- cording to Gilbert (1979), there are generally three parts to the eggshell; the shell membranes, the testa or calcified portion, and the cuticle. The shell mem- branes consist of the inner membrane (membranae testae interna) and the more complex outer mem- brane (membranae testae externa). Similarly, Tullet (1984) recognized three layers: the shell membranes, the true shell, and the cuticle. Furthermore, the limit- ing membrane as the third element of the shell mem- branes was distinguished between the inner shell membrane and the egg albumin. According to his view, the true shell includes three parts: the cone layer, the palisade layer, and the surface crystal layer. On the other hand, Mikhailov (1995a) applied similar or even identical terms, i.e. the shell membranes, the true eggshell, and the accessory material, to the gross structure of the eggshell. At the ultrastructural level, he recognized fiver layers in the eggshell: the basal plate groups, the radiating elements, the squamatic zone, the external zone, and the cuticle cover. In spite of minor differences in the usage or application of the terms, the present study prefers the view of Tullet (1984), for we identified the limiting membrane and demonstrated sublayers in the true shell proper, which correspond to the surface crystal layer and the cone layer (Tullet 1984). TEM data on the organic parts of eggshells are few and exclusively concerned with the eggshell from the domestic fowl (Tan et al. 1992; Dennis et al. 1996; Fraser et al. 1998, 1999). The present TEM data were obtained from decalicified samples, not from intact and decalcified samples as in the domestic fowl (Dennis et al. 1996). Nevertheless, no essential differ- ence was found in the fine structure of the shell mem- brane between the domestic fowl and the Black-tailed Gull. Concerning the structure and distribution of the vesicles embedded in the shell matrix, there were similarities and dissimilarities between the domestic fowl and the gull. In the gull, as in the domestic fowl, the vesicles were densely distributed in the palisade region, but scattered in the mammillary layer and the external part of the palisade region. Their size and fine structure, in particular the presence of the elec- tron-dense fringe, were almost the same in the do- mestic fowl and the gull. However, the ordered stack- ing of co-aligned sheets of vesicles reported to exist in the domestic fowl was indistinct or almost invisi- ble in the gull. Moreover, in the domestic fowl (Den- nis et al. 1996), two types of vesicles were discerned based on their topography and internal structure, i.e. 130 Eggshell structure of the Black-tailed Gull the palisade vesicles and the crown vesicles. The vesicles in the gull, with no relation to their topogra- phy in the shell, seem to be equivalent to the palisade vesicles in the domestic fowl. The biological signifi- cance of this difference remains unknown. The microscopic features and chemical composi- tion of the cuticle and/or cover layers of eggshells show a considerable variation among different taxo- nomic groups or species. In the domestic fowl, the cuticle has a vesicular structure (Simons & Wierts 1966) and consists of protein, fat and polysaccharides (Tyler & Simkiss 1959; Wedral et al. 1974). Such an organic cuticle was found in the eggshells of Guinea Fowl Numida meleagris, and the Greater Flamingo Phoenicopterus ruber (Tullet et al. 1976). However, among species belonging to the Gaviiformes, Pele- caniformes, and Podicipediformes, the external layer has a chalky appearance, is rich in inorganic material, and is more appropriately termed a cover rather than a cuticle (Tyler 1969; Tullet et al. 1976; Board et al. 1984). According to Tullet et al. (1976), this layer is rich in vaterite, a form of calcium carbonate. The va- terite-rich spherules in the eggshell cover of the grebes (Podicipidae) were investigated in details by X-ray compositional microanalysis and shown to contain O, Ca, P, and S (Board et al. 1984). The va- terite-containing spherules were present also in the cuticle of the hen egg (Dennis et al. 1996). Thus, the external layer of the eggshell of the Black-tailed Gull appears to be rather simple and differs considerably from that of members of the Podicipediformes, Pele- caniformes, and Galliformes. The reason for (or background causing) such a difference is obscure, but various factors may be concerned. These include, for example, the physiology of the oviduct, the life-style of the birds, and their breeding strategies in associa- tion with the microenvironment of egg-laying sites. The markedly high concentration of Ca throughout the true shell is not surprising, because it is generally known that the avian eggshell is a biomineralized composite ceramic consisting of calcium carbonate. The higher concentration of Mg in the cone layer is, however, of particular interest. It has been generally considered that the cones or mammillares are special parts, where calcification of the eggshell initiates and ultimately the porosity and thickness of the shell are determined (Tullet & Board 1977; Tyler & Fowler 1979). Previous studies have shown two peaks in the concentration of Mg in the eggshells of galliform birds, one in the cone layer and the other at the outer edge of the shell (Itoh & Hatano 1964; Quintana & 131 Sandoz 1978; Board & Love 1980). In contrast, only one peak (in the cone layer) was found in the shells of many other birds (Board & Love 1980). According to Board and Love (1980), the eggshells of charadri- iform birds are of the latter type, being consistent with the present results on the Black-tailed Gull. The functional significance of Mg in eggshell formation is not yet clear (Board & Love 1980), although its in- hibitory role for nucleation or calcite growth during shell formation has been postulated (Bernar 1975). ACKNOWLEDGMENTS I am grateful to Dr. M. Nakayama, Department of Material Science of Nippon Dental University, for his technical support for the X-ray compositional microanalysis. REFERENCES Becking JH (1975) The ultrastructure of the avian egg- shell. Ibis 117: 143-151. Bellairs R & Boyde A (1969) Scanning electron mi- croscopy of the shell membranes of the hen’s egg. Z Zellforsch Mikrosk Anat 96: 237-249. Bernar RA (1975) The role of magnesium in crystal growth of calcite and aragonite from sea water. Geochim Cosmochimi Acta 39: 489-504. Board RG & Love G (1980) Magnesium distribution in avian eggshells. Comp Biochem Physiol 66A: 667-672. Board RG, Perrott HR, Love G & Scott VD (1984) The phosphate-rich cover of the eggshells of grebes. J Zool (Lond) 203: 329-343. Dennis JE, Xiao S-Q, Agarwal M, Fink DJ, Heuer AH & Caplan AL (1996) Microstructure of matrix and mineral components of eggshells from White Leghorn chickens (Gallus gallus). J Morphol 228: 287-306. Fraser AC, Bain MM & Solomon SE (1998) Organic matrix morphology and distribution in the palisade layer of eggshells sampled at selected periods during lay. Br Poult Sci 39: 225-228. Fraser AC, Bain MM & Solomon SE (1999) Transmis- sion electron microscopy of the vertical crystal layer and cuticle of the eggshell of the domestic fowl. Br Poult Sci 40: 626-631. Fujii S & Tamura T (1969) Scanning electron mi- croscopy of the hen’s egg shell. J Fac Fish Anim Husb Hiroshima Univ 8: 95-98. Gilbert AB (1979) Female genital organs. In King AS & Mclelland J (eds) Form and function in birds. Vol 1. pp 237-360. Academic Press, London. Itoh H & Hatano T (1964) Variation of magnesium and A. CHIBA phosphorus deposition rates during eggshell forma- on. Poul Sci 43: 65-80. \iikhailov KE (1995a) Eggshell structure in the shoebill and pelecaniform birds: comparison with Hamerkop, herons, ibises, and storks. Can J Zool 73: 1754-1770. Mikhailov KE (1995b) The evolutionary implications of eggshell structure in falconiform and ciconiiform birds. Cour Forsch Senckenberg 181: 337-355. Paganelli CV, Ar A, Rahn H & Wangensteen OD (1975) Diffusion in the gas phase: the effect of ambient pres- sure and gas composition. Resp Physiol 25: 247-258. Quintana C & Sandoz D (1978) Coquille de l’oeuf de caille: étude ultrastructurale et cristallographique. Calcif Tiss Res 25: 145-159. Rahn H & Paganelli CV (1981) Gas exchange in avian eggs. Buffalo Press, New York. Richards PD, Richards PA & Lee ME (2000) Ultrastruc- tural characteristics of ostrich eggshell: outer shell membrane and the calcified layers. J S Afr Vet Assoc 71: 97-102. Romanoff AL & Romanoff AJ (1949) The avian egg. John Wiley and Sons Inc, New York. Simons PCM & Wierts G (1966) The ultrastructure of the surface of the cuticle of the hen egg in relation to egg cleaning. Poult Sci 45: 1153-1162. Tan CK, Chen TW, Chan HL & Ng LS (1992) A scan- ning and transmission electron microscopic study of the membranes of chicken egg. Histol Histopathol 7: 339-345. Tullet SG (1984) The porosity of avian eggshells. Comp. Biochem Physiol 78A: 5-13. Tullet SG & Board RG (1977) Determinants of avian eggshell porosity. J Zool (Lond) 183: 203-211. Tullet SG, Board RG, Love G, Perrott HR & Scott VD (1976) Vaterite deposition during eggshell formation in the cormorant, gannet and shag, and in “shell-less” eggs of the domestic fowl. Acta Zool (Stockh) 57: 79-87. Tung MA-& Richard JH (1972) Ultrastructure of the hen’s egg shell membrane by electron microscopy. J Food Sci 37: 277-281. Tyler C (1964) A study of the egg shells of the Anati- dae. Proc Zool Soc Lond 142: 547-583. Tyler C (1965) A study of the egg shells of the Sphenis- ciformes. J Zool (Lond) 147: 1-19. Tyler C (1966) A study of the egg shell of the Falconi- formes. J Zool (Lond) 150: 413-425. Tyler C (1969) A study of the egg shells of the Gavi- iformes. J Zool (Lond) 158: 395-412. Tyler C & Fowler S (1979) The size, shape and orienta- tion of pore grooves in the egg shells of Rhea sp. J Zool (Lond) 187: 283-290. Tyler C & Simkiss K (1959) Studies on egg shells XII. Some changes in the shell during incubation. J Sci Food Agr 10: 611-615. Wedral EM, Vadehra DV & Baker RC (1974) Chemical composition of the cuticle and the inner and outer shell membrane from egg of Gallus gallus. Com Biochem Biophys 47B: 631-640. ORIGINAL ARTICLE Ornithol Sci 3: 133-138 (2004) Amak Island Song Sparrows (Melospiza melodia amaka) are not evolutionarily significant Christin L. PRUETT*, Daniel D. GIBSON and Kevin WINKER* University of Alaska Museum, 907 Yukon Drive, Fairbanks, Alaska 99775, USA ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2004 Abstract The conservation status and evolutionary distinctiveness of the isolated, small, and endemic population of Amak Island Song Sparrows (Melospiza melodia amaka) have been equivocal. Coupled with a reassessment of phenotypic evidence for this taxon, we used mitochondrial cytochrome b sequences and eight microsatel- lite loci to evaluate the relationship of the Amak population to nearby Song Sparrow populations. Phenotypically, M. m. amaka is not a valid taxon, and we found that Amak Song Sparrows possess no unique haplotypes and have allele frequencies and heterozygosity values similar to those in other populations. Congruence between ge- netic and morphological evidence suggesting no diagnosable differences leads us to propose that this population is not an evolutionarily significant unit (ESU), not a valid subspecies, not a distinct population segment (DPS), nor a diagnosable conservation unit, but rather a sink colonized by regional source populations. Key words Cytochrome b, Evolutionarily significant unit, Melospiza melodia amaka, Microsatellite, Song Sparrow Little is known about the taxonomy, demography, or conservation status of many morphologically- based subspecies found in remote or inaccessible lo- cations, and endemic island taxa have proven particu- larly susceptible to extinction in historic times (BirdLife International 2000). Song Sparrows (Melospiza melodia) found on tiny Amak Island (55.4°N, 163.16°W), remote and difficult to reach in the Bering Sea (Fig. 1), represent such a problematic case. The putatively nonmigratory subspecies M. m. amaka, known only from Amak Island, was de- scribed based on the phenotype of six specimens (Gabrielson & Lincoln 1951). After examining the few specimens existing at that time, Gibson and Kessel (1997) tentatively submerged M. m. amaka in subspecies M. m. sanaka. A single individual from Amak was included in a genetic study (mitochondrial DNA restriction fragment length polymorphisms) of Alaska’s Song Sparrows (Hare & Shields 1992). These genetic data were equivocal, in that this bird possessed a haplotype found in the Aleutian Islands (Fig. 1.) (Received 11 March 2004; Accepted 23 June 2004) * Corresponding Author, E-mail: ffksw @uaf.edu * Present address: Department of Wildlife and Fisheries Sciences, Texas A&M University, College Station, Texas 77843-2258, USA 133 Demographic information on the Song Sparrows of Amak Island is limited. Given the size of the island (10km’) and limited survey data, the population is small, but some evidence suggests that it fluctuates markedly in size. Although survey details such as ex- tent of island surveyed, weather conditions, and time of day and year are insufficient to verify the conclu- sions, the population was considered to be extinct in 1980 (Williams & Novak 1993), and anecdotal infor- mation indicated that 25 birds were seen in 1987 and four or five in 1988 (NatureServe 2003). This sug- gests that Amak Song Sparrows may have experi- enced severe population reductions (with near-extinc- tion events) or that the Amak population may be in effect a sink, in that it might become extinct if not for immigration from nearby populations (Hanski & Simberloff 1997). Amak Song Sparrows are not listed as threatened or endangered by the U. S. Fish and Wildlife Service, but the Nature Conservancy consid- ers this population to be imperiled (NatureServe 2003). Resolution of the incongruous status of this popu- lation hinges on determination of the validity of the named taxon M. m. amaka and its evolutionary his- tory. Because genetic data constitute primarily neutral variation and would be unlikely to include tiny por- C. L. PRUETT, D. D. GIBSON and K. WINKER Unalaska Island Benng Sea re Shumagin Adak Island Islands + King Cove Seeeeeean sensay 52°N -~. os : Pacific Ocean Fig. 1. Map of Alaska with Song Sparrow (Melospiza melodia) collection locations used in this study. Alaska distribution of Song Sparrows shaded in gray. Boxes surround subspecies ranges with M. m. maxima outlined with a dashed black line and M. m sanaka with a solid black line. Table 1. Location, subspecies, number of individuals sequenced, number of individuals genotyped, expected (H.) and observed (H,) heterozygosities and Genbank accessions for Song Sparrows (Melospiza melodia) used in this study. Museum voucher num- bers provided in Appendix 1. Sequenced Genotyped Location Subspecies e s H, H, GenBank accession Amak Is., Bering Sea, Alaska amaka 4 4 0.49 0.38 AY 450608-61 1 King Cove, Alaska Peninsula, Alaska sanaka 6 10 0.49 0.40 AY 156406-41 1 Popof Is., Shumagin Is., Alaska sanaka 4 9 0.56 0.49 AY 156162-165 Unalaska Is., Aleutian Is., Alaska sanaka 0 2 — — — Adak Is., Aleutian Is., Alaska maxima 10 30 0.42 0.40 AY 156396-405 tions of the genome that might be under strong selec- _ distinct from nearby populations?; and 2) Does this tion in peripheral populations, such data alone are not population show genetic evidence of severe reduc- a reliable diagnostic tool for assessing subspecific va- _ tions in size? lidity (e.g., see Bulgin et al. 2003). Genetic data can provide invaluable insight into evolutionary history, MATERIALS AND METHODS however, and, when coupled with phenotypic assess- ments, congruent patterns between genotype and phe- Whole genomic DNA from 55 Song Sparrows notype can be both insightful and diagnostic. from Amak Island and four neighboring breeding We reassessed phenotypic evidence and used mito- populations (Fig. 1, Table 1) were extracted follow- chondrial (mt) DNA sequences and nuclear mi- ing Glenn (1997). Most of the mtDNA cytochrome b crosatellite loci to evaluate key questions about this gene (1,137 bp) was amplified and cycle-sequenced little-known population: 1) Are Amak Song Sparrows — using four primer pairs per individual for a subset of 134 Amak Song Sparrows are not distinct the extracted tissues (Table 1). Primers used included: L14851 (Kornegay et al. 1993), H16064 (Harshman 1996), L15350 (Klicka & Zink 1997), and H15424 (Hackett 1996). Amplified products were sequenced in both directions using an ABI 373A or 3100 auto- mated sequencer (Applied Biosystems Inc., Foster City, CA). All sequences were deposited in GenBank (Table 1). All birds used in this study were sampled during the breeding season, including the four indi- viduals from Amak. All sampled populations are con- sidered to be non-migratory (Murie 1959). Thus, it is likely that the birds examined represented the local breeding populations at these locations. Eight microsatellite loci were amplified for all in- dividuals using fluorescent dye-labeled primers de- veloped for Song Sparrows (Jeffery et al. 2001) and for two other bird species (Escu/, Hanotte et al. 1994; GF5, Petren 1998). Amplicons were screened for variation using an ABI 373A or 3100 automated sequencer. Average expected and observed heterozy- gosities for each population (except Unalaska) were determined using GDA (Lewis & Zaykin 2001). Phenotypic assessment was done using new mate- rial and traditional taxonomic methods of visual com- parisons of external phenotype—the same methods used in the original assessment of Amak Island Song Sparrows (Gabrielson & Lincoln 1951) and known to be effective in other drab-plumaged passerines (e.g., Winker 1997). All known existing M. m. amaka spec- imens (N=9) and several hundred each of M. m. sanaka and M. m. maxima were included in these ex- aminations. RESULTS Even with new material, we did not find M. m. amaka phenotypically separable from M. m. sanaka or M. m. maxima using the classic “75% rule” (Amadon 1949; Patten & Unitt 2002). This is concor- dant with the reassessment of Gibson and Kessel (1997), in which the original material used by Gabrielson and Lincoln (1951) did not seem to sup- port the original erection of a subspecies amaka. The assertion that Song Sparrows from tiny Amak Island are separable from nearby Song Sparrows at Unimak Island on the Alaska Peninsula and in the Shumagin Islands was founded on an alleged distinctness in several plumage and mensural characters in a type se- ries of only six “adult” specimens (Gabrielson & Lin- coln 1951:253). The plumage characters are equivo- cal: “Resembles maxima from the western Aleutians 135 in color and extensive brown markings, but some- what more heavily marked with brown than that race both on back and breast; in most available specimens the brown markings also somewhat brighter. Closer in color to maxima than to the geographically closer race sanaka.” And neither these plumage characters nor the average culmen measurement given for the four male and two female specimens will separate with certainty even one Amak specimen from a series of adjacent sanaka. This equivocal situation is not at all alleviated with the new material now in the Uni- versity of Alaska Museum. Only four of Gabrielson and Lincoln’s (1951) orig- inal six specimens are present today at the U. S. Na- tional Museum, and these include the adult male holotype, a second adult male, one immature male, and one immature female (fide R. C. Banks, in litt., 2004). Thus, the original plumage description con- flates both sexes and two age classes in the descrip- tion of six “adults.” We think the authors should have heeded their own caveat regarding problems separat- ing maxima from sanaka in the central-eastern Aleu- tians (in their description of subspecies maxima— published in the same article as their description of “amaka’): “The series of specimens in similar plumage is too limited to make a certain decision” and “The tail and wing measurements vary somewhat more than normal on account of wear and are not fully reliable” (Gabrielson & Lincoln 1951: 251- 252). With such small sample sizes statistical tests are not useful where there is considerable overlap in pu- tatively diagnostic characters. Despite the fact that plumage characters can be well assessed visually, and although we do support the subspecies concept, we find nothing phenotypically to justify continued recognition of this subspecies. Genetically, there are no unique cytochrome b hap- lotypes found on Amak Island. Haplotype A is shared with Adak Island, and haplotype B is found in high frequencies at King Cove and the Shumagin Islands (Table 2; Fig. 1). There is a single unique allele (locus Mme 12; Fig. 2) in the Amak samples; all other alleles are found at frequencies similar to those from other sampled locations (Fig. 2). Observed and expected heterozygosities are comparable to values in other populations (Table 1). DISCUSSION Gibson and Kessel (1997) tentatively lumped Amak Island Song Sparrows (M. m. amaka) into the C. L. PRUETT, D. D. GIBSON and K. WINKER an £ = skates Mme 1 Mme 2 > = = o ‘ | 5 0.6} 7 Sint ar | Sa = s05 —— £04 = o - cy 0.3} = = 0.2 = Shh! | 0.1 0 Amak King Cove Popof Unalaska Adak Amak King Cove Popof Unalaska Adak Location Location Mme 3 Mme 7 > >0.9 i 2 08} FS 08} S | $s 0.71 @ 067 5 06} = | = 05 / 2 04+ 204 |2 203 <02 | < 02 J 0.1 LJ — ~ — 0 Eom Amak King Cove Popof Unalaska Adak Amak King Cove Popof Unalaska Adak Location Location Mme 8 Mme 12 Ore - = 506) =I 05 | | 0 04 | » 03 o = 0.2 0.1 ; 0 Amak = _KingCove —_— Popof Unalaska Adak Amak KingCove —_— Popof Unalaska Adak Location Location Escu 1 > 0.8 |S 3 oy 2 |o 2 |< | Amak King Cove Popof Location Unalaska Adak Adak Amak King Cove Popof Unalaska Location Fig. 2. Frequencies of microsatellite alleles from eight loci among Amak Island Song Sparrows and the frequencies of these al- leles in other populations. Each allele is shaded differently to track its frequencies in other populations. Table 2. each collection location in Alaska. Frequency of mtDNA cytochrome b haplotypes at Amak Island, Bering Sea King Cove, Alaska Peninsula Popof Island, Shumagin Islands Haplotype Location A B C | 3 0 0 5 | 0 4 0 10 0) 0 Adak Island, Aleutian Islands 136 neighboring subspecies, M. m. sanaka, which is found on the Alaska Peninsula and the eastern Aleut- ian Islands. Our assessment, which includes substan- tial new material obtained since their evaluation, sup- ports this treatment. We recognize that sample sizes of amaka are small. Because this is a small popula- tion in an isolated and inaccessible location large sample sizes were not possible for this study. Given logistical and ethical constraints, these samples are not likely to be substantially increased. Only 11 spec- Se Amak Song Sparrows are not distinct imens of “amaka” have been collected over the last 80 years (Gabrielson & Lincoln 1951; Hare & Shields 1992; this study). We deem our evidence suf- ficiently conclusive to warrant reporting now for management considerations of Amak sparrows. Genetically, Amak Song Sparrows share haplo- types with populations from two nonmigratory sub- species (M. m. maxima from Adak Island and M. m. sanaka from our other sample locations) that have no haplotypes in common. This suggests that Amak Song Sparrows might be intergrades or that this pop- ulation was colonized by individuals from both sub- species. Interestingly, Gabrielson and _ Lincoln’s (1951) original description of M. m. amaka included phenotypic evidence of such intergradation. In the Aleutian Islands and on the Pacific coast of the Alaska Peninsula, Song Sparrows are found among rocky beaches and _ beachside _ grasses (Gabrielson & Lincoln 1951, Murie 1959). Thus, Song Sparrows can only inhabit a narrow ring around Amak Island. Even if optimal conditions existed around the entire perimeter of the island (which they do not; KW pers. obs.), this population would always be small. Considering these limitations together with anecdotal suggestions of fluctuations in population size, it is likely that Amak Song Sparrows would ex- hibit very little genetic diversity if this population were evolutionarily isolated. Although only four indi- viduals were sampled, heterozygosities were similar to those in other populations. This suggests that im- migration to Amak from other populations is ongo- ing. A similar pattern was found for Song Sparrows on Mandarte Island, British Columbia (Keller et al. 2001), in which very low numbers of migrants caused a rapid recovery in a genetically bottlenecked popula- tion. Amak Island is a small volcano that appeared above the sea about 6,700 years ago (Marsh & Leitz 1979). It is visible from the mainland coast, only 17km away. Volcanic activity was reported in the 1700s but had ended by 1867 (Dall 1870). Thus, Amak Song Sparrows probably colonized within the last several thousand years, and subsequently they may have experienced severe ecological distur- bances. Coupled with the harsh climatic conditions currently found in this region, cycles of extinction and recolonization have probably characterized this population since its founding. The overarching question about using small sam- ple sizes is whether larger sample sizes would alter the conclusions of the study. Phenotypically the an- 137 swer to this 1s no; new material does not bear out Gabrielson and Lincoln’s (1951) suggestion that amaka is a diagnosable subspecies. Nor does their original material suggest (phenotypically) that a for- merly endemic population has gone extinct and been replaced by new colonists. This is in contrast to the one other subspecies that Gabrielson and Lincoln (1951) described at the same time; maxima is a valid subspecies (Gibson & Kessel 1997, unpubl. data). Genetically, the four Song Sparrows from Amak shared haplotypes with adjacent populations. If more individuals were examined from Amak and were found to have other haplotypes, this would not dimin- ish the importance of the four sparrows that pos- sessed haplotypes found in other populations. In ad- dition, increasing microsatellite sample sizes would likely increase the heterozygosity levels found on Amak. However, given the probable demographic shifts and the current heterozygosity values for the Amak population, it seems improbable that we would have sampled the only sparrows that were genetically diverse. Thus, gene flow would be inferred regardless of sample size. Overall, our evidence suggests that “amaka” was a weak subspecies that does not hold up under scrutiny. Based on this evidence, we suggest that these birds are simply a peripheral extension of other, larger re- gional Song Sparrow populations. Recognition of Amak Song Sparrows as an evolutionarily significant unit (ESU), a distinct population segment (DPS), or a management unit (MU) is not warranted. ACKNOWLEDGMENTS We gratefully acknowledge support from the University of Alaska Museum, the U.S. Department of Agriculture, the Na- tional Science Foundation (DEB-9981915), an anonymous donor, and the Izembek and Alaska Maritime National Wildlife Refuges. We also thank G. M. Spellman for assis- tance in sample collection, R. C. Banks for rechecking speci- mens for us at the U.S. National Museum, and two anonymous reviewers for comments. REFERENCES Amadon D (1949) The seventy-five per cent rule for subspecies. Condor 51: 250-258. BirdLife International (2000) Threatened birds of the world. Lynx Editions and BirdLife International, Barcelona. Bulgin NL, Gibbs HL, Vickery P & Baker AJ (2003) Ancestral polymorphisms in genetic markers obscure detection of evolutionarily distinct populations in the C. L. PRUETT, D. D. GIBSON and K. WINKER endangered Florida Grasshopper Sparrow (Ammodra- mus savannarum floridanus). Mol Ecol 12: 831-844. Dall WH (1870) Alaska and its resources. Lee & Shep- ard, Boston. Gabrielson IN & Lincoln FC (1951) The races of Song Sparrows in Alaska. Condor 53: 250-255. Gibson DD & Kessel B (1997) Inventory of the species and subspecies of Alaska birds. West Bird 28: 45-95. Glenn T (1997) Genetic bottlenecks in long-lived verte- brates: mitochondrial and microsatellite DNA varia- tion in American alligators and Whooping Cranes. PhD thesis, Univ Maryland, College Park. Hackett SJ (1996) Molecular phylogenetics and _ bio- geography of tanagers in the genus Ramphocelus (Aves). Mol Phylogen Evol 5: 368-382. Hanotte O, Zanon C, Pugh A, Greig C, Dixon A & Burke T (1994) Isolation and characterization of mi- crosatellite loci in a passerine bird: the Reed Bunting Emberiza schoeniclus. Mol Ecol 3: 529-530. Hanski I & Simberloff D (1997) The metapopulation approach, its history, conceptual domain, and applica- tion to conservation. In: Hanski IA & Gilpin ME (eds) Metapopulation biology. pp 5-26. Academic Press, San Diego. Hare MP & Shields GF (1992) Mitochondrial-DNA variation in the polytypic Alaskan Song Sparrow. Auk 109: 126-132. Harshman J (1996) Phylogeny, evolutionary rates, and ducks. PhD thesis, Univ Chicago. Jeffery KJ, Keller LF, Arcese P & Bruford MW (2001) The development of microsatellite loci in the Song Sparrow, Melospiza melodia, and genotyping errors associated with good quality DNA. Mol Ecol Notes 1: 11-13. Keller LF, Jeffery KJ, Arcese P, Beaumont MA, Hochachka WM, Smith JNM & Bruford MW (2001) Immigration and the ephemerality of a natural popu- lation bottleneck: evidence from molecular markers. Proc Royal Soc Lond Series B 268: 1387-1394. Klicka J & Zink RM (1997) The importance of recent ice ages in speciation: a failed paradigm. Science 277: 1666-1669. Kornegay JR, Kocher TD, Williams LA & Wilson AC (1993) Pathways of lysozyme evolution inferred from the sequences of cytochrome b in birds. J Mol Evol 37: 367-379. Lewis PO & Zaykin D (2001) Genetic Data Analysis: Computer program for the analysis of allelic data. Version 1.0 (dl6c). (http:/Aewis.eeb.uconn.edu/ lewishome) Marsh BD & Leitz RE (1979) Geology of Amak Island, Aleutian Islands, Alaska. J Geol 87: 715-723. Murie OJ (1959) Fauna of the Aleutian Islands and Alaska Peninsula. North American Fauna No. 61, US Fish and Wildlife Service, Washington DC. NatureServe (2003) NatureServe Explorer: An Online Encyclopedia of Life. Version 1.8 (http://www. natureserve.org). NatureServe, Arlington. Patten MA & Unitt P (2002) Diagnosability versus mean differences of Sage Sparrow subspecies. Auk 119: 26-35. Petren K (1998) Microsatellite primers from Geospiza fortis and cross species amplification in Darwin’s finches. Mol Ecol 7: 1782-1784. Williams J & Nowak R (1993) Vanishing species in our own backyard. In: Kaufman L & Mallory K (eds) The last extinction. pp 107-139. MIT Press, Cambridge. Winker K (1997) A new form of Anabacerthia variegat- iceps (Furnariidae) from western Mexico. In: Dicker- man RW (ed) The era of Allan R. Phillips: A festschrift. pp 203-208. Horizon Comm, Albu- querque. Appendix. Voucher numbers for specimens used in this study. Subspecies Museum?* Catalog numbers Locality Melospiza melodia amaka UAM 13425-13428 Alaska: Bering Sea, Amak Island Melospiza melodia sanaka UAM 9321, 9328, 10091, 11230, 11362, Alaska: Alaska Peninsula, King Cove 11365—66, 11381, 11389, 11823 Melospiza melodia sanaka UAM 10090, 10171, 10187, 11276, Alaska: Shumagin Islands, Popof Island 11379, 11390, 11585, 11713, 12142 Melospiza melodia sanaka UAM 111238239 Alaska: Aleutian Islands, Unalaska Island Melospiza melodia maxima UAM 8460-61, 1004042, 10167-68, Alaska: Aleutian Islands, Adak Island 10170, 10172, 10179, 10188, 10942, 10946-47, 11048, 11175—78, 11267—-69, 11501, 11511, 11827, 11850, 12143, 13057, 13059, 13161 ‘UAM=University of Alaska Museum. 138 ORIGINAL ARTICLE Ornithol Sci 3: 139-144 (2004) Passerine bird pollination and fruiting behaviour in a dry sea- son blooming tree species, Erythrina suberosa Roxb. (Fabaceae) in the Eastern Ghats forests, India Aluri J. S. RAJU* and Srungavarapu P. RAO Department of Environmental Sciences, Andhra University, Visakhapatnam 530 003, India ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2004 ment. Key words birds, Pollination Bird pollination is as important as insect pollina- tion in the tropics and in southern temperate zones. About 100 families of flowering plants are known to have at least some members adapted for bird-pollina- tion (Meeuse & Morris 1984). The family Fabaceae includes species pollinated by insects and others pol- linated by birds; although the type of flower remains essentially similar, the bird-pollinated flowers show marked differences in structure related to this method of pollination. In insect-pollinated species, the flow- ers have wings and a keel forming an alighting plat- form for insects, which in so doing set off a remark- able trigger mechanism that delivers pollen onto the bodies of the insects. These parts of the flower are more or less suppressed when birds are the agents of pollination, and for this reason the trigger mechanism is absent in bird-pollinated species (Jaeger 1961). The genus Erythrina provides a typical example of bird pollination within the Fabaceae (Kumar 2000). The genus Erythrina is found in the tropics and subtropics of both the New and Old Worlds (Baker & Baker 1982). Old world species are visited and polli- nated by a wide range of passerine birds (Porsch 1924; Singh 1929; Ali 1932; Docters van Leeuwan 1932) whereas most New World species are probably hummingbird pollinated (Raven 1974; Toledo 1974), (Received 17 April 2004; Accepted 25 July 2004) * Corresponding author, E-mail: ajsraju@yahoo.com 139 Abstract Erythrina suberosa blooms during the dry-season. The flowers are large, papilionaceous, and partly self-compatible. The floral characteristics provide an ex- ample of ornithophilous pollination; all the flowers are pollinated exclusively by passerine birds. The fruit set rate was only 10%, but this was compensated for by a higher seed set rate. The flowers normally produce seven ovules, with those in the number two and five positions appearing to be the most preferred for seed develop- Breeding system, Erythrina suberosa, Fruiting behaviour, Passerine however, there are also reports of various passerines visiting native species of Erythrina in the New World (Skutch 1954; Timkin 1970; Snow & Snow 1971; Leck 1974; Raven 1974). In India, Erythrina species are ornithophilous, but they are mainly visited by par- rots and woodpeckers (Kumar 2000). Subramanya and Radhamani (1993) provided a list of flowering plants regularly frequented by birds based on previ- ous works. In this list, E. variegata is reported to be visited by passerine and non-passerine birds whereas E. stricta, E. cristagalli, E. suberosa and E. subum- brans are visited by passerines only. However, there is no information on any Erythrina species regarding the floral organization, functional events of flowers in relation to the visits of birds for pollination, or fruit- ing behaviour. In view of this, these aspects of E. suberosa a common prickly tree species in the East- ern Ghats, were studied. MATERIALS AND METHODS Fifty-three Erythrina suberosa Roxb. trees were studied in the Lambasingi-Lotugedda area (17°52'N; 82°21’E), a forest stretch of about 90 km in the East- ern Ghats located at an elevation of 900m in Visakhapatnam District, Andhra Pradesh, during Feb- ruary—May 2002 and 2003. Fifty flowers were sam- pled to record their floral morphometrics. The time of anthesis and anther dehiscence was noted by observ- A. J. S. RAJU and S. P. RAO arked mature buds in the field. The time and of anther dehiscence were noted by using a 10x hand-lens. Ten flowers were marked and bagged at the mature bud stage, opened after anthesis and nectar squeezed into a micropipette in order to meas- ure the volume of nectar per flower; the average vol- ume of nectar per flower was determined and ex- pressed in jl. Pollen grain number/anther per flower was determined for 25 flowers from different individ- uals following the procedure in Aluri and Subba Reddi (1994). Stigma receptivity was tested with H,O, according to Dafni (1992). Breeding behaviour by autogamy (bagged and __ hand-pollinated), geitonogamy and xenogamy, were tested through controlled pollinations following the detailed proce- dure in Aluri and Subba Reddi (1994). For each mode of pollination, 50 flowers (ten each) on five dif- ferent trees were used. Five hundred and fifty flowers (11 each) on five different trees were tagged and fol- lowed until fruit development to observe the rate of natural fruit and seed production. A sample of 43 fruits was used to note the seed set pattern from the pedicel end to the opposite end. Flower-visitors in- cluded only birds. They were observed with binocu- lars and also directly when they happened to forage at close range. Their mode of approach, landing, prob- ing behaviour, forage collected, contact with the es- sential organs to effect pollination, inter-tree foraging activity, damage to the flowers, if any, were all care- fully observed. Furthermore, the frequency of forag- ing visits for each bird species was also recorded. For this, five trees with full flowering were selected and the number of times each bird species visited the tree for nectar collection from 0600 to 1800 was noted. These observations were made on the same trees on four different days in 2002 and 2003. Then, the aver- age number of visits made by each bird species to a tree/day was calculated. ing m manne! RESULTS E. suberosa is leafless during flowering, which oc- curs from early March to mid-April (Fig. la). The flowers are produced in terminal racemes occupying only the distal half of the inflorescence axis, thus they provide suitable perches for foraging birds (Fig. 1b). The flowers mature and open acropetally. The flowers are large, 5.3cm long, bisexual and zygomorphic (Fig. Ic). The calyx is green, campanulate and bilipped. The corolla is composed of five unequal and variously coloured petals. The corolla is characteristi- 140 cally papilionaceous and the petals are free. The scar- let-red standard petal is larger (4.8cm long) and broader than the rest; it encloses the margins of the lateral pair of maroon wing petals. These are 7mm long and overlap the margins of the greenish-maroon keel petals. The keel petals (1.1 cm long) form a cari- nal-like structure in which nectar is well seated. The ten stamens are diadelphous, with nine of them united into a bundle, and the tenth one free. The filaments are free towards the apex of the staminal tube and bear dithecous anthers. The stamens vary slightly in length, five are one length, four another length, and the free stamen lies below the level of the bundled stamens. The stamens are almost at the same level with respect to making contact with a bird’s bill or breast as it takes nectar from a flower. The ovary is semi-inferior with one carpel having seven (rarely eight) ovules in it. The ovary is enclosed by the sta- minal tube, but the style protrudes through the stami- nal tube. The style is 11mm long and ends in a minute stigma, and lies behind the forward level of the anthers. The flowers open early in the morning from 0500-0600. Anthers dehisce via longitudinal slits about one hour before flowers open. Flowers produce an average of 33,870+27 (SD) pollen grains. Nectar is secreted prior to flower opening and amounts to 140+4.6 (SD) ul per flower. Nectar is well protected from direct exposure to sunlight by the keel and wing petals. The stigma is receptive to pollen from anthesis onwards and remains so until the evening of the 2nd day. The flowers remain in place for three days if not disturbed by flower visitors. Hand-pollination tests indicated 88% pod-set oc- curred through xenogamy and 40% _ through geitonogamy, with no pod-set with autogamy (Table 1). Each fruited flower matured without abortion. Fruit development is very fast, taking only about three weeks to reach maturity. Fruit maturity and de- hiscence is almost immediately followed by the initi- ation of leaf-flush. The natural fruit set rate was only 10% (Table 1). The fruited flowers produced 1-4 seeds with 63% produced two-seeded pods (Fig. 2). Seed set was more frequent from the ovules in num- bers two and five positions in the ovary (Fig. 3). Seed set rate was 26%. Mature pods dehisce explosively (Fig. 1d) to scatter their seeds. Twelve species of passerine birds in six families were identified as visiting flowers and taking nectar (Table 2, Fig. 1f—k). Of these, three species (Common Rosefinch Carpodacus erythrinus, Red-whiskered Passerine bird pollination and fruiting behaviour in E. Suberosa Fig. 1. Erythrina suberosa: a, flowering, leafless trees; b, inflorescence; c, flower; d, explosive pods; e, anterior part of standard petal cut by Black Drongo; fk: bird visitors; f, Black Drongo; g, Indian Myna; h, Red-vented Bulbul; i, Red-whiskered Bulbul; j, Magpie-Robin; k, Common Rosefinch (left, male; right, female). 141 A. J.S. RAJU and S. P. RAO Table 1. Fruit set rate in controlled pollination treatments in Erythrina suberosa Treatinenl No. of flowers No. of flowers Percentage Cc . 7 . pollinated/tagged set fruit of fruit set Autogamy (bagged without hand-pollination) 50 0 0 Autogamy (hand-pollinated) 50 0 0 Geitonogamy 50 20 40 Xenogamy 50 44 88 Open-pollination 550 55 10 70 n=43 20 n=43 ao] 60 3 —_ Ss = 50 B15 3 40 ° c no] $ 30 ® 10 2 me E20 ° 10 25 E 0 z 1 2 3 4 0 Seed number per pod 1 2 3 4 5 6 7 Ovule position Fig. 2. Frequency occurrence of seeds per pod in . “fh E sublerosen Fig. 3. Seed production rate in relation to ovule position, Table 2. Bird Family Scientific name Dicruridae Dicrurus adsimilis Sturnidae Sturnus pagodarum Acridotheres ginginianus A. tristis A. fuscus Pycnonotidae Pycnonotus melanicterus P. cafer P jocosus Muscicapidae Turdoides striatus Copsychus saularis Zosteropidae Zosterops palpebrosa Fringillidae Carpodacus erythrinus Bulbul Pycnonotus jocosus and Red-vented Bulbul P. cafer) made over 100 visits/tree/day; five species (Brahminy Myna Sturnus pagodarum, Black Drongo Dicrurus adsimilis, Jungle Myna Acridotheres fuscus, Indian Myna A. tristis and Magpie Robin Copsychus saularis) made 57-91 visits/tree/day; and the remain- ing four species made fewer than 20 visits/tree/day from the pedicel tip in E. suberosa. Passerine bird visitors to the flowers of Erythrina suberosa Average no. of _— Percentage Common name Be AN visits/tree/day of visits Black Drongo WS 9 Brahminy Myna 91 10 Bank Myna 16 2 Indian Myna 69 8 Jungle Myna 74 8 Black-headed Yellow Bulbul 1] I Red-vented | Bulbu 143 16 Red-whiskered Bulbul 151 17 Jungle Babbler 9 1 Magpie-Robin S// i White-eye 9 1 Common 182 20 Rosefinch (Table 2). All 12 species perched first on the inflores- cence axis, moved toward the flower, then thrust their bill and head into the flower. While probing the flow- ers the birds invariably contacted the stamens and stigma with their throat or breast. After the birds took nectar, they moved backward and raised their beak to swallow nectar. The Black Drongo, however, cut the Passerine bird pollination and fruiting behaviour in E. Suberosa anterior part of the standard petal to reach the nectar more easily and affected pollination (Fig. le). The birds visited the flowers throughout the day, but did so most often during the morning. The inflorescences were found to rock and swing when large birds such as rosefinches and mynas hopped from one inflores- cence to another, and as a result withered and wither- ing flowers fell off. Birds were found to move fre- quently between E. suberosa trees in search of more nectar, and mynas and bulbuls also visited the flowers of Bombax ceiba for more nectar. DISCUSSION Erythrina is a pantropical genus (Cruden & Toledo 1977) that shows remarkable variation in inflores- cence orientation and floral organization. In pale- otropical species, inflorescences are oriented horizon- tally and flowers are confined to the distal part of the inflorescence, providing a standing place for bird vis- itors. These species are described as being pollinated by passerine birds that require a standing place on the inflorescence for probing the flowers (Baker et al. 1983). In neotropical species, similar inflorescence orientation and floral organization have been reported in Erythrina species that have also been reported to be pollinated by passerine birds, however, there are also Erythrina species that are pollinated by hum- mingbirds. In these species, the inflorescence is al- most vertical in orientation and the flowers are placed in such a way that they can be reached by humming- birds hovering outside the flowers in order to take nectar (Baker et al. 1983). Erythrina suberosa is an Asian species and occurs in India. In this species, the inflorescence is oriented horizontally and the flowers occur in the distal half, providing a standing place for passerine birds for probing the flowers as reported by Baker et al. (1983) for paleotropical Erythrina species. Bird-pollinated flowers open during the day, are predominantly red, odourless, and larger than insect- pollinated flowers. Their inflorescences are few in number, but they are long-lived for long-term attrac- tion to birds. They normally have inferior ovaries, which confer protection to the ovules against the probing bills of birds. Their stamens and stigma brush against the breast or head of the visiting birds. The flowers produce copious sugary nectar, which is well protected from nectar thieves (Stiles 1978; Fae- gri & van der Pijl 1979; Meeuse & Morris 1984; Kumar 2000). E. suberosa is typically ornithophilous, 143 showing all of these characteristics. Furthermore, the flower of E. suberosa is papilionaceous with a well- developed standard petal and poorly developed wing and keel petals exposing the stamens, style and stigma to the aerial environment. The standard petal stands in an upright position and is the most conspic- uous part of the flower and it is this that attracts polli- nators. These characteristics have previously been re- ported for a range of bird-pollinated species in the Fabaceae (Jaeger 1961). In seasonal tropical forests, many trees are polli- nated by large, far-flying, systematic pollinators (Kumar 2000). E. suberosa trees bloom while still leafless during the dry season, when its red flowers are attractive to avian visitors. Subramanya and Rad- hamani (1993) reviewed information on bird and bat pollination in the Western Ghats forests of India and provided a list of birds that frequent the flowers of different plant species for nectar. They documented 15 species of passerine birds visiting the flowers of E. suberosa for nectar in the Western Ghats. The birds included Dicrurus adsimilis, D. paradiseus, Acridotheres tristis, Corvus splendens, Chloropsis cochinchinensis, Pycnonotus jocosus, P. luteolus, Hypsipetes madagascarensis, Turdoides striatus, T. affinis, Orthotomus sutorius, Turdus merula, Nec- tarinia zeylonica, N. lotenia and N. asiatica. All these except H. madagascarensis occur in the present study area (Krishna Raju 1985). Of these, birds such as D. adsimilis, A. tristis, P. jocosus and T) striatus were found to visit E. suberosa flowers for nectar. In addi- tion, other passerine bird species also visited the flowers to drink nectar. All these birds perch on the proximal region of the inflorescence and move for- ward while probing the flowers. Acropetal anthesis of the inflorescence is an added advantage for passerine birds to probe flowers. While probing for nectar, the birds contact the stamens, style and stigma with their head or breast, which result in them being dusted with pollen and so transferring pollen between flow- ers. Birds have considerable energy requirements and the nectar production of one Erythrina tree may be insufficient for the energetic requirements of even one bird. As a result, birds are forced to make fre- quent visits to different trees. All of the birds we ob- served frequently visit different trees to quench their thirst for nectar and in so doing they effect cross-pol- lination. Although hand-pollination tests indicate that E. suberosa is partly self-compatible, setting fruit only through geitonogamy, it is nevertheless highly cross- A. J.S. RAJU and S. P. RAO compatible and shows maximum fruit set through xenogamy. This breeding system is indicative of fac- ultative xenogamy (Cruden 1977), which essentially requires pollen vectors. Even though self-pollination can occur, most fruit set is through cross-pollination. Because, visiting birds make frequent visits to differ- ent E. suberosa trees and such foraging activity may result in more cross-pollen transfer. Despite the hec- tic foraging activity of birds, E. suberosa is able to produce only 10% fruit set, but this low rate is com- pensated for by higher seed set. Seed set pattern in the mature pods indicated that most of the pods pro- duce only two seeds although there are seven or rarely eight ovules in the ovary. Furthermore, the po- sitions of the developed seeds in the pods examined revealed that there is a strong positional preference for seed development; ovules in the number two and five position (from the pedicel tip to the stigmatic end) seem to produce most of the seeds. This obser- vation does not agree with Joshi et al. (1993) who re- ported that positions 3, 4 and 5 may be the most pre- ferred ones for seed development. Therefore, we sug- gest that further work is needed to confirm which po- sitions of ovules are the most preferred for seed de- velopment. ACKNOWLEDGEMENTS Financial support from the Ministry of Environment and Forests, Government of India, New Delhi through a Major Re- search Project (No. 30/12/97-RE) to AJSR is gratefully ac- knowledged. We thank several anonymous referees for their comments, which helped to improve the manuscript. REFERENCES Ali SA (1932) Flower-birds and bird-flowers in India. J Bombay Nat Hist Soc 35: 573-605. Aluri RJS & Subba Reddi C (1994) Pollination ecology and mating system of the weedy mint, Leonotis nepetaefolia R.Br. in India. Proc Indian Nat Sci Acad B60: 255-268. Baker HG & Baker I (1982) Chemical constituents of nectar in relation to pollination mechanisms and phy- logeny. In: Nitecki MH (ed) Biochemical aspects of evolutionary biology. pp 131-171. The University of Chicago Press, Chicago. Baker HG, Bawa KS, Frankie GW & Opler PA (1983) In: Golley FB (ed) Tropical rain forest ecosystems. Elsevier Scientific Publishing Company, New York. Cruden RW (1977) Pollen-ovule ratios: A conservative indicator of breeding systems in flowering plants. 144 Evolution 31: 32-46. Cruden RW & Toledo VM (1977) Oriole pollination of Erythrina breviflora (Leguminosae): Evidence for a polytropic view of ornithophily. Plant Syst Evol 126: 393-403. Datni A (1992) Pollination ecology: a practical ap- proach. Oxford University Press, New York. Docters van Leeuwan WM (1932) Vogelbesuch an den Bluten von einigen Erythrina-Arten auf Java. Ann Jard Bot Buitenz 42: 46-95. Faegri K & van der Pijl L (1979) The principles of polli- nation ecology. Pergamon Press, Oxford. Jaeger P (1961) The wonderful life of flowers. EP Dut- ton & Co, Inc, New York. Joshi NV, Uma Shaankar R & Ganeshaiah KN (1993) The ‘neighbour effect’ and its role in shaping the po- sitional pattern of seed development in fruits: an illus- tration from the pods of Erythrina suberosa. Curr Sci 65: 234-238. Krishna Raju KSR (1985) A contribution to the or- nithology of Visakhapatnam region. Kingfisher Spe- cial Issue. Kumar HD (2000) Plant-animal interactions. Affiliated East-West Press Private Ltd, New Delhi. Leck C (1974) Further observations of nectar feeding by orioles. Auk 91: 162-163. Meeuse B & Morris S (1984) The sex life of flowers. Facts on File, New York. Porsch O (1924) Vogelblumen-Studien I. Jahrb Wiss Bot 63: 553-605. Raven PH (1974) Erythrina (Fabaceae): Achievements and opportunities. Lloydia 37: 321-331. Singh TCN (1929) A preliminary note on the pollination of the coral tree (Erythrina indica Lamk.). J Bombay Nat Hist Soc 33: 460-462. Skutch AF (1954) Life histories of Central American birds. Pacific coast avifauna No 31. Cooper Ornitho- logical Society, Berkeley. Snow BK & Snow DW (1971) The feeding ecology of tanagers and honeycreepers in Trinidad. Auk 88: 291-322. Stiles FG (1978) Temporal organization of flowering among the hummingbird food plants of a tropical wet forest. Biotropica 10: 194-210. Subramanya S & Radhamani TR (1993) Pollination by birds and bats. Curr Sci 65: 201-209. Timkin RL (1970) Food habits and feeding behaviour of the Baltimore Oriole in Costa Rica. Wilson Bull 82: 184-188. Toledo VM (1974) Observations on the relationship be- tween hummingbirds and Erythrina species. Lloydia 37: 482-487. ORIGINAL ARTICLE Ornithol Sci 3: 145-153 (2004) Offspring size as an index of habitat degradation Ian G. WARKENTIN'*, J. Michael REED? and Susie M. DUNHAM2** ' Environmental Science-Biology, Sir Wilfred Grenfell College, Memorial University of Newfoundland, Corner Brook, NL A2H 6P9, Canada * Biological Resources Research Center, University of Nevada, Reno, NV 89557, USA ORNITHOLOGICAL SCIENCE © The Ornithological Society of Japan 2004 Abstract Disturbances that shift a community away from its potential natural state may also degrade the quality of that community for some species. Having an index to measure changes in habitat quality resulting from such disturbances would be useful in assessing the impact of human activities on native fauna. We propose that average egg mass per clutch and offspring size for a population in a particular habitat may be a useful index of habitat quality, and perhaps degradation, for that population relative to the status of populations occupying other similar habitats in that region. We studied American Robins (Turdus migratorius) breeding along streams in three canyons on the western side of the Toiyabe Mountains of central Nevada, USA. The level of habi- tat degradation associated with cattle grazing and other human activities was deter- mined a priori based on soil and understory vegetation characteristics. The density of adult birds and their body condition did not differ among canyons with differing habi- tat quality, nor did clutch size or brood size at day 8. However, nests containing larger eggs and chicks were associated with canyons assessed as having a higher quality, or lower level of degradation. Key words stock grazing Larger eggs typically result in heavier nestlings with greater growth rates than the young arising from smaller eggs, for at least a short period after hatching (Williams 1994). Egg size also may influence nestling survival for both precocial (e.g., Blomqvist et al. 1997) and altricial (e.g., Bolton 1991; Smith & Bruun 1998) species, although the extent of influence depends upon circumstance. Most of the variation in egg size is among rather than within clutches, and egg size can be strongly heritable (Christians 2002). Positive correlations also are seen between environ- mental factors and female condition and egg size (Smith et al. 1993; Potti 1999; Styrsky et al. 2002), although female size and mass alone typically explain 20% or less of the variation in the egg size within species (Christians 2002). It is not clear that there is a relationship between egg size and fitness of the re- (Received 5 June 2004; Accepted 15 September 2004) * Corresponding author, E-mail: iwarkent@swgc.mun.ca * Present address: Dept. of Biology, Tufts University, Medford, MA 02155, USA ** Present address: Department of Botany and Plant Pathology, Oregon State University, Corvallis, OR 97331, USA 145 Egg and chick size, Great Basin Desert, Habitat quality, Index, Live- sulting offspring (Williams 1994). However, if larger egg size leads to larger, faster growing chicks, then these traits could result in enhanced juvenile survival and recruitment (Tinbergen & Boerlijst 1990; Cichon & Lindén 1995; Saino et al. 1997). The role of habi- tat quality in determining the size and number of eggs and chicks produced has been examined for a wide range of species. Whereas studies of some species in- dicate no link between quality of the surrounding habitat and egg size (Smith & Bruun 1998) or chick size (Hinsley et al. 1999), many have found lower habitat quality associated with decreased offspring body mass (Lens & Dhondt 1994; Verhulst et al. 1997; Turner & McCarty 1998; Hinsley et al. 1999; Huhta et al. 1999). One of the major sources of habitat degradation in the Great Basin of the U.S.A., particularly in the ri- parian areas of this region, is overgrazing by live- stock (Knopf et al. 1988, Fleischner 1994; Knopf & Samson 1994; Brown & McDonald 1995). The con- centrated activities of domesticated animals can change vegetative structure and species composition, alter soil structure and porosity, and modifying I. G. WARKENTIN, J. M. REED and S. M. DUNHAM stream bank morphology (Smith 1940; Ellison 1960; Brown 1978: Kauffman & Krueger 1984; Baker & Guthery 1990; Smith et al. 1994). In addition, large stretches of riparian vegetation in this region have been destroyed or degraded by water diversions, agri- cultural development, mining activities, and road construction (Chaney et al. 1990; Brussard et al. 1998). The effects of these activities include reduc- tion of natural vegetation, stream channel widening, and lowered water tables (Kauffman & Krueger 1984; Platts 1991; Fleischner 1994; Belsky & Blu- menthal 1997). Consequently, the structurally diverse undisturbed native riparian flora is simplified by dis- turbance, which in turn modifies the native bird com- munities typically present (Dobkin 1994; Warkentin & Reed 1999). Although there is recognition of the links between habitat quality and size of the young produced, to our knowledge no one has proposed that features of off- spring size in a particular habitat might be a useful predictor of the quality, or extent of degradation, of one habitat relative to others in the same region. We test the hypothesis that average egg mass per clutch and offspring size are potential indices of relative habitat degradation for altricial songbirds. Our work was done in riparian forests of the Great Basin, with habitat degradation determined independently a pri- ori based on soil and understory vegetation character- istics. METHODS 1) Study area We studied American Robins (7urdus migratorius) breeding along three permanent streams in the rela- tively narrow and steeply-walled canyons on the west slope of the Toiyabe Mountains, Lander and Nye Counties, in central Nevada, USA (39°N, 117°W) in 1995 and 1996. Study areas along streams ranged in elevation from 2050-2300 m in San Juan Creek, and 2250-2600 m in Stewart and Clear Creeks. San Juan Creek lies 24km north of Stewart and Clear Creeks and is separated from them by several ridges reaching 2800-3000 m. Stewart and Clear Creeks lie adjacent to one another, separated by a single ridge 2700 to 3000 m high. For all three creeks, the dominant ripar- ian vegetation on the 50-200 m wide canyon floor is quaking aspen (Populus tremuloides) interspersed with willow (Salix spp.) and water birch (Betula occi- dentalis). Away from the riparian zone are steep, rocky slopes with sparse upland forests of single leaf 146 pinyon (Pinus monophylla), Utah juniper (Juniperus osteosperma), and scattered patches of curl leaf mountain mahogany (Cercocarpus ledifolius). Pri- mary understory shrubs include sagebrush (Artemisia sp.), wild rose (Rosa woodsii), and snowberry (Sym- phoricarpos albus). Each canyon contains grassy meadows characterized by dense cover of Carex ne- brascensis, C. aquatilus, Poa secunda, Juncus balti- cus, or Deschampsia cespitosa, and extended stream banks that are used by robins for foraging and as a source of mud for nests. Length of the area sampled within each canyon was estimated using 1 :50,000 topographic maps. 2) Assessing degradation Habitat degradation was determined based on soil and understory vegetation characteristics using a ranking scheme developed by Weixelman et al. (1996). These characteristics were examined at four locations along San Juan Creek for this assessment, as well as three in both Stewart and Clear Creek canyons (D. Weixelman, D. Zamudio & K. Zamudio, U. S. Forest Service, Sparks, NV, unpubl. data). Lo- cations were assigned an ecological status rank rela- tive to the potential natural riparian community that would be established if successional sequences were completed without human-caused disturbances (Weixelman et al. 1996). This ecological status rank reflected high, moderate or low levels of similarity to the potential natural community. Aerial photographs were used to estimate the proportion of the study area in each canyon represented by the sample locations. To develop a quantitative value for comparing can- yons, we converted these ranks for each location to a numerical score of 3, 2 or 1; a rank of 3 reflected a location with characteristics representative of the nat- ural condition, whereas a rank of 2 indicated less similarity to the natural condition and a rank of 1 suggested limited or no similarity to natural succes- sional patterns in the resulting community. We then calculated the average ecological status score for each canyon based on the four or three locations eval- uated within each drainage. We assessed the extent of habitat degradation in the three canyons using three additional parameters. Based on aerial photography, the percent of habitat within 100m of each stream that contained riparian vegetation was determined, as well as the percent of riparian habitat along our study streams through which dirt roads passed or where roads defined the edge of the riparian habitat (D. Weixelman, D. Zamu- An index of habitat degradation dio & K. Zamudio, U. S. Forest Service, Sparks, NV, unpubl. data). Based on the similar topography and soil types of the three canyons, we assumed that all drainages had the same initial potential to contain ri- parian vegetation within the 100 m strip of land along the stream (cf. Weixelman et al. 1996). We also as- sumed that differences in the amount of riparian veg- etation relative to other canyons were the result of different disturbance histories. Accurate grazing his- tories and recreational visitation rates were not avail- able for these drainages, so dirt road cover was used as an indirect measure of road-based disturbance. Fi- nally, we qualitatively assessed the presence or ab- sence of severe down-cutting (streambanks with non- vegetated vertical drops 20.3m) along significant portions of each stream. 3) Study species The American Robin is an abundant riparian spe- cialist in the Great Basin during the breeding season. It forages primarily on the ground, requiring a con- stant source of invertebrates to raise the 2—3 broods produced by each pair per year. The robin’s diet and foraging ecology are relatively well studied (Salla- banks & James 1999), and nests are easily located; adults tolerate multiple nest visits (Ortega et al. 1997). We censused adult robins and their nests in all suit- able habitats along each stream. To assess relative densities of adult robins in the three canyons, from late June through early July 1996 we placed sets of 10 standard 12-m mist nets at randomly selected low, medium, and high elevation stretches in each canyon for a total of 700 mist-net hours. This timing meant that the average successful nest (based on mean clutch initiation date in each canyon) had recently fledged young. To standardize for any potential sea- sonal effects, all trapping events at a given elevation (i.e., low, medium, high) took place within 7 d of one another. An index of adult density was calculated as the overall capture rate per 100 net hours divided by the area of riparian vegetation in that canyon. Upon capture, adults were sexed following the criteria of Pyle et al. (1987), color banded, and their mass and tarsal length measured; age could not be reliably de- termined based on the criteria of Pyle et al. (1987). We tested for differences in overall body size among canyons for each sex using a Kruskal—Wallis one- way analysis of variance (ANOVA). This test was based on examining PC-1 scores from a principal components analysis (PCA) of tarsus and mass (Free- 147 man & Jackson 1990). The adults captured were not necessarily the birds present at nests where reproduc- tive success variables were measured, so we could not correlate adult condition with success of a partic- ular nesting attempt. We assumed, however, that the condition of trapped birds represented the average condition for adults, including breeding birds, in each canyon. Nest searching was initiated on 1 June 1995 and 1 May 1996 with each canyon searched every 4—5d through the middle of July for both years. Because we could not distinguish between first and later broods, data for all nests were pooled for these analy- ses. The difference between years in the start of nest searching constituted a systematic bias that would not alter comparisons between canyons across years. Nests found before or during egg laying were re- visited every 3d until clutch completion so that clutch initiation dates, final clutch sizes, and esti- mated hatch dates could be determined. Nests located after clutch completion were checked every 3d to de- termine hatch date. We estimated clutch initiation dates for nests found after clutch completion by sub- tracting the average incubation time of 13d (Salla- banks & James 1999) from the date of hatch. When nests were discovered after hatching, we estimated clutch initiation date based on an average incubation time of 13d in combination with chick age deter- mined from an ageing key developed using known- age chicks (day 0 being when the first egg hatched; Warkentin et al. 2003). We compared initial clutch size and brood size at day 8 among canyons, once in- cluding all nests, and once including only those nests that had one or more chicks at day 8, using a Kruskal—Wallis ANOVA. For all nests found during incubation with at least 3 eggs, we measured the maximum length (L) and breadth (B) of each egg to the nearest 0.1 mm using digital calipers. We converted these values to fresh egg mass (W) to assess differences in the amount of resources invested in eggs by females in different canyon types. Fresh egg mass was calculated as W=K,,LB’, where Ky is a shape-specific mass coef- ficient reported for American Robin eggs by Hoyt (1979). Mean egg mass, as well as mass of the largest and smallest egg in each clutch were compared among canyons using Kruskal-Wallis ANOVAs. Robin eggs hatch asynchronously with broods of three typically hatching over two days and broods of four or five hatching over three days (Slagsvold 1997). As a result, there are marked differences in I. G. WARKENTIN, J. M. REED and S. M. DUNHAM body size among individuals in a nest at a given point in the nestling phase. Therefore, we based our assess- ment of body size for chicks in a nest on morphologi- cal measurements from the largest chick at day 8 of the nestling phase in each nest. Interestingly, the largest individual on day 8 was not always the first to hatch because sometimes a chick’s growth was re- tarded, apparently from extensive black fly bites. We measured mass, tarsus length, total head length, and bill length to the nearest 0.1mm _ using digital calipers. These measurements were combined using PCA to develop a single measure of body size that could be used to compare chicks among canyons using a Kruskal-Wallis ANOVA. Although American Robins fledge between day 14 and 16 (Sallabanks and James 1988), we avoided handling nestlings older than 8 days to prevent premature fledging. We defined successful nests as those having at least one chick alive on day 8. 4) Statistical analyses Clutch initiation date and fledging success data were collected and calculated for all nests encoun- tered regardless of nestling age. Nests used in egg and chick size comparisons represent a subset of the total number of nests sampled as not all nests found were at the same stage of the breeding cycle or could necessarily be visited on day 8 to collect morphologi- cal data. This is an exploratory assessment (sensu Steidl et al. 1997) of the predictive power of offspring size rel- ative to habitat degradation, thus accepting P<0.10 as statistically significant is appropriate. Where Kruskal—Wallis ANOVA indicated a significant dif- ference among canyons for a particular variable, we conducted a post-hoc examination of the main ef- fects. For the post-hoc test, we rank-transformed the data and then repeated the ANOVA (cf. Conover & Iman 1981) using Fisher’s LSD to make post-hoc multiple comparisons; for these post-hoc tests we used an alpha level of 0.1. We conducted all statisti- cal analyses using SAS version 8.12 (SAS Institute Inc. 2001). RESULTS San Juan Creek canyon was the most heavily de- graded of the three canyons (Table 1). This canyon had the lowest area of riparian vegetation (varying from 4 to 28m wide), in combination with the high- est percentage of riparian habitat covered or bordered 148 by roads, the presence of severe down-cutting along significant proportions of the stream, and the lowest average ecological status score. Based on the same measures, Stewart (with a 15 to 45m wide riparian zone) and Clear (with a 50-m wide riparian zone at all three locations examined) Creek canyons are simi- lar in their degree of degradation, but Clear Creek may be relatively less degraded based on the higher proportion of riparian vegetation present and its higher average ecological status score (Table 1). Data from mist-netting indicated that the density of adult robins was similar among canyons (Table 2). For adults captured, female and male size based on the PC-1 score, which explained 66.2% of the vari- ance, did not differ among canyons when examined with a Kruskal-Wallis one-way ANOVA. We found 90 active American Robin nests during two summers in the three canyons, of which 77 had known outcomes with eggs and/or chicks produced. There was a significant difference among canyons in clutch initiation date with nests in San Juan Creek canyon started on average 5.5 days before those in Stewart and 14.8 days before those in Clear (Table 2). Post-hoc testing suggested that nests in Clear Creek canyon were significantly later than those in the other two canyons which, in turn, did not differ from each other. The overall relationship probably reflected the significant correlation between clutch initiation date and nest elevation (F, ,=5.25, P=0.025, r>=0.06) across all three canyons. But since none of our pa- rameters of interest (mean egg mass, clutch size, and Table 1. Study site dimensions and characterization of degradation for riparian habitat in three canyons of the Toiyabe Mountain Range, Nevada. Criteria for percent habitat within 100m of each stream that contains riparian vegetation, percent riparian area covered or bordered by roads, presence/ absence of streambank down-cutting, and the ecological status score all suggest that relative degradation is least for Clear Creek and most for San Juan Creek canyons. Canyon San Juan Stewart Clear Length (m) 3505 4451 2713 Area of riparian vegetation 84,120 137,981 135,650 (m’) Riparian area (%) 24 31 50 Riparian area bordered by 54 28 35 road (%) Severe downcutting Present Absent Absent Ecological status score 1.68 2.00 2.10 An index of habitat degradation Table 2. Productivity and morphometric features of an American Robin population breeding in the riparian habitat of three canyons in the Toiyabe Mountain Range, Nevada. Values presented are mean+SE(n) along with statistics from Kruskal—Wallis analyses; those with different superscripted letters are significantly different from each other. San Juan Stewart Clear i P Adult Density 0.10 0.06 0.07 (per 100 net-hours per km’) Female PC-1 0.20 + 0.14 (6) —0.26 + 0.36 (8) 0.16 + 0.42 (9) 1.26 0.53 Male PC-1 0.23 + 0.44 (6) —0.36 + 0.32 (9) 0.09 + 0.28 (13) 0.67 0.71 Clutch Mean initiation date May 22 = 2 (Q7)? May 27 24 (26)°— Tunes.. =4_ C18)" 5.96 0.05 Initial size 3.8 £0.2 (17) 3.500) 3.6+0.1 (14) 1.32 0.51 Brood size at day 8 2.5 £0.3 (27) 1.9+0.3 (28) 1.7+0.4 (17) 2.51 0.28 Brood size at day 8 3.2+0.2 (21) 3.0 £0.2 (18) 3.3 £0.2 (9) 0.48 0.78 (successful only) Eggs Mean egg mass (g) 6.6 +0.1 (13)® 6.9+0.2 (6)48 7.2+0.1 (10)* 5.37 0.06 Smallest egg mass (g) 6.5+0.1 (13) 6.6 +0.2 (6) 6.9+0.2 (10) Dg 0.31 Largest egg mass (g) 6.9+0.1 (13)® 7.3 £0.2 (6)*8 7.4+0.1 (10)4 5.17 0.07 Chicks PC-1 (largest chick) —0.39 £0.21 (12)8 0.23 £0.39 (9) 0.37 + 0.39 (7)* 4.76 0.09 chick size at day 8) were correlated with clutch initia- tion date (F, »,=2.68, P=0.11, r°=0.09; F, 59=0.41, P=0.52, r°=0.007; and F, 5,=1.24, P=0.27, r°= 0.05; respectively), we assumed that the effects we report on egg and chick size were not due to differ- ences in elevation (as it affected clutch initiation date) among canyons. In addition, most research (e.g., Hochachka 1990) suggests that earlier hatching chicks are larger than those hatching later in the sea- son. This is contrary to the trends that we report below. Data on clutch and brood sizes suggested minimal differences among canyons in habitat quality. Initial clutch sizes, as well as brood size at day 8 for all nests and brood size at day 8 for successful nests only, did not differ significantly among canyons (Table 2). Partial-brood loss rates did not differ among canyons (Table 3; Fisher Exact Test, P=0.66), and likewise whole-brood loss rates did not differ among canyons (Table 3; Fisher Exact Test, P=0.34). In general, eggs and chicks from nests in San Juan Creek canyon, the most degraded canyon, were smaller than those in Stewart Creek canyon, which in turn were smaller (but not significantly so) than those in Clear Creek canyon, the least degraded of the three (Table 2). Mean egg mass per clutch was 8% greater in Clear Creek canyon than those in San Juan Creek canyon and mean egg mass for Stewart Creek canyon was intermediate. Although egg mass of the smallest 149 Table 3. Nest outcomes for a population of American Robins breeding in the riparian habitat of three canyons in the Toiyabe Mountain Range, Nevada. San Juan Stewart Clear Total number monitored 29 31 17 Outcome Failed 8 13 8 At least 1 young at day 8 21 18 9 Known partial brood loss 3 5 Brood size at day 8 equaled 8 3 3 number of eggs laid egg in the clutch did not differ among canyons, the largest egg in the clutch followed the same pattern as reported above for mean egg mass, with largest eggs from Clear being 7% heavier than those in San Juan (Table 2). We did not have morphometric data on specific individuals because we were not able to match eggs (i.e., egg mass) with individual nestlings, nor did we know hatching order of the young. We combined data for the four morphometric measures taken from the largest chick present in the nest at day 8 to create a PC-1 score, which explained 72% of the variance. This statistic indicated that nests in Clear Creek and Stewart canyons had the largest young and San Juan Creek canyon had significantly smaller young (Table 2). I. G. WARKENTIN, J. M. REED and S. M. DUNHAM DISCUSSION Anthropogenic disturbances that shift a community away from its potential natural state may also de- grade the quality of that community for some species. Having an index to measure changes in habitat qual- ity resulting from such disturbances would be useful in assessing the impact of human activities on native fauna. Fluctuating asymmetry has been proposed as an index of habitat degradation based on develop- mental responses by individuals inhabiting a particu- lar community (e.g., Badyaev et al. 2000; Leung et al. 2000; Lens et al. 2002; Zakharov 2003). These analyses of American Robins were complicated by limited sample size, a qualitative rather than quantita- tive assessment of habitat degradation, and differ- ences amongst canyons in elevation. However, our results suggest a tendency for mean egg mass and chick size to vary with the extent of habitat degrada- tion and thus such measures may be a useful index of habitat quality or degradation across suitable habitats within a region. For passerines that depend on the daily accumula- tion of energy reserves to produce eggs (Perrins 1996), decreased availability or quality of food may cause females to allocate fewer resources to egg pro- duction. Reduced habitat quality could also be due to an increase in species that use disturbed areas, such as some competitors or predators (Paton 1994). In lower quality habitats, this could result in the produc- tion of smaller clutches and fewer young fledged (Cowie & Hinsley 1987) or lower body-mass chicks (Lens & Dhondt 1994; Verhulst et al. 1997; Hinsley et al. 1999). Our data suggest that American Robins did not alter the size of their clutch, but rather allo- cated less energy to each egg. In the most degraded habitat (San Juan Creek canyon) egg mass was on av- erage 8% lower than for robin eggs from nests in the least degraded habitat (Clear Creek canyon). This dif- ference in mean egg mass is similar to the maximum increase in egg size attributable to dietary supple- ments (Christians 2002). Condition of breeding females (an index of mass relative to body size) and egg mass have been shown to be significantly correlated in a number of species (Slagsvold & Lifjeld 1989; Smith et al. 1993; Potti 1999; Styrsky et al. 2002). However, since female mass and size typically explain 20% or less of the variation in the egg size within species (Christians 2002), it appears to be the ability to translate environ- mental resources into egg mass that is important and 150 not necessarily overall body size. We trapped adults later in the season when females may have lost mass accumulated during egg laying, but we detected no difference in female size among canyons. Some studies suggest that the effect of egg mass on chick size disappears with nestling age and is gone by the time of fledging (Magrath 1992; Smith & Bruun 1998). Smith and Bruun (1998) also found that nei- ther egg nor nestling masses were related to the avail- ability of high quality foraging habitat for starlings, but that this habitat variable did influence nestling survival late in the nestling period. This contrasts with our results, which indicate that there were no differences among canyons in initial clutch sizes or brood sizes at day 8. Smith and Bruun (1998) also suggested that availability of high quality habitat may only influence the translation of large egg size into large nestling size when habitat availability is limited, although egg size did not vary with availability of high quality foraging habitat in their study. Bize et al. (2002) proposed that rearing conditions, reflected by the size of eggs laid by the care-giving parent (real or foster), were more important than initial egg mass in predicting survival. The latter result contrasts with the findings of Schifferli (1973) but concurs with Reed et al. (1999) who found that survival and growth of nestlings were largely influenced by factors other than egg size. Although measurable morphometric differences may disappear in the nestlings of some species by the time they fledge, it appears that the impact of this dif- ference during rearing may influence the ability of an individual to obtain a high quality nesting site in the future. Verhulst et al. (1997) found that Great Tits (Parus major) that were relatively heavier as nestlings bred in better quality habitat as adults, and other studies suggest that these individuals will have better reproductive performance than individuals who were lighter as nestlings (Green & Cockburn 2001; Perrins & McCleery 2001; Styrsky et al. 2002). In arid ecosystems, the presence of roads and over- grazing can adversely affect the quality of riparian habitat for some members of the biological commu- nity and lead to either population decline or loss (Fleischner 1994; Trombulak & Frissell 2000). Based on population density measures, studies of American Robins suggest that this species may actually benefit from moderate levels of habitat degradation (Page et al. 1978; Crouch 1982; Mosconi & Hutto 1982; Sedgwick & Knopf 1987; Schulz & Leininger 1991; Warkentin & Reed 1999), but examination of fitness An index of habitat degradation components in any habitat are lacking. While inten- sive grazing can simplify vegetation diversity and consequently cause decreased diversity of associated phytophagous insects (Lawton & Schroder 1977), moderate grazing levels may lead to enhanced forag- ing opportunities for robins and increased habitat quality through providing a more diverse vegetative community (Milchunas et al. 1988; Grime 1990; Collins et al. 1998) and greater diversity among those same insects groups (Eijsackers 1983; Morris 1990; Tscharntke & Greiler 1995; Oates 1995; Kruess & Tscharntke 2002). Although no canyon-level records of grazing activity are available for our study sites, we would argue that the physical features and vegeta- tion remaining indicate that grazing pressure had been very intensive in San Juan Creek canyon. There was a low level of similarity to the potential natural community as a consequence of both cattle grazing and high levels of human use for recreational activi- ties which together led to the lowering of habitat quality. Consequently, we propose that the inverte- brate community that forms the vast majority of the diet during the breeding season for robins in this area (see Sallabanks & James 1999) would be adversely affected by grazing and human recreation resulting in a reduced invertebrate fauna and less food available to invest the resources in eggs by female robins. This degradation was reflected in decreased quality of the young robins produced in this canyon. ACKNOWLEDGMENTS Field and logistical assistance was provided by J. Bury, L. Butcher, J. Dunham, C. Elphick, E. Fleishman, S. Fleury, G. Guscio, L. Hillerman, L. Morris, C. Mutembi, N. Olson, and J. Ramos. This research was funded by the U. S. Forest Service, the Center for Conservation Biology of Stanford University, a grant from the Wells Family Foundation, the Biological Re- sources Research Center of the University of Nevada, Reno, and the Principal’s Fund of Sir Wilfred Grenfell College, Memorial University of Newfoundland. Monetary support was increased and/or aided by C. Boggs, G. Grevsted, and M. Schwalbach. We thank the Yomba Shoshone Tribal Council, staff from the Austin Ranger District Office of the U.S. Forest Service, and the people of Austin, Nevada for their assistance and cooperation. REFFRENCES Badyaev AV, Foresman KR & Fernandes MV (2000) Stress and developmental stability: Vegetation re- moval causes increased fluctuating asymmetry in shrews. Ecology 81: 336-345. 151 Baker DL & Guthery FS (1990) Effects of continuous grazing on habitat and density of ground-foraging birds in south Texas. J Range Manage 43: 2-5. Belsky AJ & Blumenthal DM (1997) Effects of live- stock grazing on stand dynamics and soils in upland forests of the interior west. Conserv Biol 11: 315- 827k Bize P, Roulin A & Richner H (2002) Covariation be- tween egg size and rearing condition determines off- spring g quality: an experiment with the Alpine Swift. Oecologia 132: 231-234. Blomqvist D, Johansson OC & G6tmark F (1997) Parental quality and egg size affect chick survival in a precocial bird, the Lapwing Vanellus vanellus. Oce- cologia 110: 18-24. Bolton M (1991) Determinants of chick survival in the Lesser Black-backed Gull: Relative contributions of egg size and parental quality. J Anim Ecol 60: 949-— 960. Brown DE (1978) Grazing, grassland cover and game- birds. North American Wildlife Conference 43: 477— 485. Brown JH & McDonald W (1995) Livestock grazing and conservation on southwestern rangelands. Con- serv Biol 9: 1644-1647. Brussard PF, Charlet DA & Dobkin DA (1998) Great Basin-Mojave desert region. In: Mac MJ, Opler PA, Puckett Haecker CE & Doran PD (eds) Status and trends of the nation’s biological resources. pp 505— 542. US Dept. Interior, US Geological Survey, Reston. Chaney E, Ellmore W & Platts WS (1990) Livestock grazing on western riparian areas. Produced for the U. S. Environmental Protection Agency by Northwest Resource Information Center, Eagle. Christians JK (2002) Avian egg size: variation within species and inflexibility within individuals. Biol Rev 77: 1-26. Cichon M & Linden M (1995) The timing of breeding and offspring size in Great Tits Parus major. Ibis 137: 364-370. Collins SL, Knapp AK, Briggs JM, Blair JM & Stein- auer EM (1998) Modulation of diversity by grazing and mowing in tallgrass prairie. Science 280: 745- 747. Conover WJ & Iman RL (1981) Rank transformations as a bridge between parametric and nonparametric statistics. Am Stat 35: 124-129. Cowie RJ & Hinsley SA (1987) Breeding success of Blue Tits (Parus caeruleus) and Great Tits (Parus major) in suburban habitats. Ardea 75: 11-20. Crouch GL (1982) Wildlife on ungrazed and grazed bot- tomlands on the South Platte River, northeastern Col- I. G. WARKENTIN, J. M. REED and S. M. DUNHAM orado. In: Peek JM & Dalke PD (eds) Wildlife-live- stock relationships symposium. pp 188-197. Forest, Wildlife and Range Experiment Station, University of Idaho, Moscow. Dobkin DS (1994) Conservation and management of Neotropical migrant landbirds in the northern Rock- ies and Great Plains. Univ Idaho Press, Moscow. Eijsackers HJP (1983) Development of earthworm pop- ulations in abandoned arable fields under grazing management. In: Satchell JE (ed) Earthworm ecol- ogy: from Darwin to Vermiculture. pp 241-246. Chapman and Hall, London. Ellison L (1960) Influence of grazing on plant succes- sion of rangelands. Bot Rev 26: 1-78. Fleischner TL (1994) Ecological costs of livestock graz- ing in western North America. Conserv Biol 8: 629-644. Freeman S & Jackson WM (1990) Univariate metrics are not adequate to measure avian body size. Auk 107: 69-74. Green DJ & Cockburn A (2001) Post-fledging care, philopatry and recruitment in Brown Thornbirds. J Anim Ecol 70: 505-514. Grime JP (1990) Mechanisms promoting floristic diver- sity in calcareous grasslands. In: Hillier SH, Walton DWH & Wells DA (eds) Calcareous grasslands— ecology and management. pp 51-56. Bluntisham Books, Huntingdon. Hinsley SA, Rothery P & Bellamy PE (1999) Influence of woodland area on breeding success in Great Tits Parus major and Blue Tits Parus caeruleus. J Avian Biol 30: 271-281. Hochachka W (1990) Seasonal decline in reproductive performance of Song Sparrows. Ecology 71: 1279- 1288. Hoyt DF (1979) Practical methods of estimating volume and fresh weight of bird species. Auk 96: 73-77. Huhta E, Jokimaki J & Rahko P (1999) Breeding suc- cess of Pied Flycatchers in artificial forest edges: The effect of a suboptimally shaped foraging area. Auk 116: 528-535. Kauffman JB & Krueger WC (1984) Livestock impacts on riparian ecosystems and stream management im- plications: a review. J Range Manage 37: 430-438. Knopf FL, Johnson RR, Rich TD, Samson FB & Szaro RC (1988) Conservation of riparian ecosystems in the United States. Wilson Bull 100: 272-294. Knopf FL & Samson FB (1994) Scale perspectives on bird diversity in western riparian ecosystems. Con- serv Biol 8: 669-676. Kruess A & Tscharntke T (2002) Grazing intensity and the diversity of grasshoppers, butterflies, and trap- nesting bees and wasps. Conserv Biol 16: 1570-1580. Lawton JH & Schréder D (1977) Effects of plant type, size of geographical range, and taxonomic isolation on number of insect species associated with British plants. Nature 265: 137-140. Lens L & Dhondt AA (1994) Effects of habitat frag- mentation on the timing of Crested Tit Parus crista- tus natal dispersal. Ibis 136: 147-152. Lens L, Van Dongen S & Matthysen E (2002) Fluctuat- ing asymmetry as an early warning system in the crit- ically endangered Taita Thrush. Conserv Biol 16: 479-487. Leung B, Forbes MR & Houle D (2000) Fluctuating asymmetry as a bioindicator of stress: comparing effi- cacy of analyses involving multiple traits. Am Nat 155: 101-115. Magrath RD (1992) The effect of egg mass on the growth and survival of blackbirds: a field experiment. J Zool (Lond) 227: 639-653. Milchunas DG, Sala OE & Lauenroth WK (1988) A generalized model of the effects of grazing by large herbivores on grassland community structure. Am Nat 132: 87-106. Morris MG (1990) The effects of management on the invertebrate community of calcareous grasslands. In: Hillier SH, Walton DWH & Wells DA (eds) Calcare- ous grasslands—ecology and management. pp 128— 133. Bluntisham Books, Huntingdon. Mosconi SL & Hutto RL (1982) The effect of grazing on the land birds of a western Montana riparian habi- tat. In: Peek JM & Dalke PD (eds) Wildlife-livestock relationships symposium. pp 221-233. Forest, Wildlife and Range Experiment Station, University of Idaho, Moscow. Oates MR (1995) Butterfly conservation within the management of grassland habitats. In: Pullin AS (ed) Ecology and conservation of butterflies. pp 98-112. Kluwer, Dordrecht. Ortega CP, Ortega JC, Rapp CA, Vorisek S, Backensto SA & Palmer DW (1997) Effect of research activity on the success of American robin nests. J Wild! Man- age 61: 948-952. Page JL, Dodd N, Osborne TO & Carson JA (1978) The influence of livestock grazing on non-game wildlife. Calif Nev Wildl 1978. pp 159-173. Perrins CM (1996) Eggs, egg formation and the timing of breeding. Ibis 138: 2-15. Perrins CM & McCleery RH (2001) The effect of fledg- ing mass on the lives of great tits Parus major. Ardea 89: 135-142. Platts WS (1991) Livestock grazing. In: Meehan WR (ed) Influences of Forest Management and Rangeland Management on Salmonid Fishes and Their Habitats. pp 389-423 American Fisheries Society, Bethesda. An index of habitat degradation Potti J (1999) Maternal effects and the pervasive impact of nestling history on egg size in a passerine bird. Evolution 53: 279-285. Pyle P, Howell SNG, Yunick RP & DeSante DF (1987) Identification guide to North American passerines. Slate Creek Press, Bolinas. Reed WL, Turner AM & Sotherland PR (1999) Conse- quences of egg-size variation in the Red-winged Blackbird. Auk 116: 549-552. Saino N, Calza S & Moller AP (1997) Immunocompe- tence of nestling Barn Swallows in relation to brood size and parental effort. J Anim Ecol 66: 827-836. Sallabanks R & James FC (1999) American Robin (Tur- dus migratorius). In: Poole A & Gill F (eds) The birds of North America, No 462. Academy of Natural Sciences, Philadelphia, and American Ornithologists’ Union, Washington. SAS Institute Inc. (2001) SAS for windows, version 8.12. SAS Institute Inc. Cary NC. Schulz TT & Leininger WC (1991) Nongame wildlife communities in grazed and ungrazed montane ripar- ian sites. Great Basin Natur 51: 286-292. Schifferli L (1973) The effects of egg weight on the sub- sequent growth of nestling Great Tits Parus major. Ibis 115: 549-558. Sedgwick JA & Knopf FL (1987) Breeding bird re- sponse to cattle grazing of a cottonwood bottomland. J Wildl Manage 51: 230-237. Slagsvold T (1997) Is there a sexual conflict over hatch- ing asynchrony in American robins? Auk 114: 593-600. Slagsvold T & Lifjeld JT (1989) Constraints on hatch- ing asynchrony and egg size in Pied Flycatchers. J Anim Ecol 58: 837-849. Smith CC (1940) The effect of overgrazing and erosion upon the biota of a mixed-species prairie of Okla- homa. Ecology 21: 381-397. Smith HG & Bruun M (1998) The effect of egg size and habitat on starling nestling growth and survival. Oe- cologia 115: 59-63. Smith HG, Ottosson U & Ohlsson T (1993) Interclutch variation in egg mass among starlings Sturnus vul- garis reflects female condition. Ornis Scand 24: 311-316. 153 Smith PJ, Pressey RL & Smith JE (1994) Birds of par- ticular conservation concern in the western division of New South Wales. Biol Conserv 55: 325-338. Steidl RJ, Hayes JP & Schauber E (1997) Statistical power analysis in wildlife research. J Wildl Manage 61: 270-279. Styrsky JD, Dobbs RC & Thompson CF (2002) Sources of egg-size variation in House Wrens (Troglodytes aedon): ontogenetic and environmental components. Auk 119: 800-807. Tinbergen JM & Boerlijst MC (1990) Nestling weight and survival in individual Great Tits (Parus major). J Anim Ecol 59: 1113-1127. Trombulak SC & Frissell CA (2000) Review of ecologi- cal effects of roads on terrestrial and aquatic commu- nities. Conserv Biol 14: 18-30. Tscharntke T & Greiler H-J (1995) Insect communities, grasses and grasslands. Ann Rev Entomol 40: 535-558. Turner AM & McCarty JP (1998) Resource availability, breeding site selection, and reproductive success of Red-winged Blackbirds. Oecologia 113: 140-146. Verhulst S, Perrins CM & Riddington R (1997) Natal dispersal of Great Tits in a patchy environment. Ecol- ogy 78: 864-872. Warkentin IG & Reed JM (1999) Effects of habitat type and degradation on avian species richness in Great Basin riparian habitats. Great Basin Natur 59: 205- ADs Warkentin IG, Reed JM & Dunham SM (2003) Nest site characteristics of American Robins breeding in desert-riparian habitat. Wilson Bull 115: 16—23. Weixelman DA, Zamudio DC & Zamudio KA (1996) Central Nevada riparian field guide. USDA Forest Service, Sparks. Williams TD (1994) Intraspecific variation in egg size and egg composition in birds: Effects on offspring fit- ness. Biol Rev 68: 35-59. Zakharov VM (2003) Linking developmental stability and environmental stress: a whole organism ap- proach. In: Polak M (ed) Developmental instability: causes and consequences. pp 402-414. Oxford Univ Press, London. Is adh gol iE eebiaRA AUN AM fh Aten? Ai ASdsp-e? dine n-roaeos mtd ea) vet er djenc® lori dstd Actin det 1 ywedusis? 2 Tt teva Ss itdai2 (oewent- otitis aaeigarrewcey?/ i ome AAS Seia 1) 4D: aoehmodll ad DR edd igen Poke Sane )tieligoge woltetiasn 7 eyter, Uitneekyein? / rp-dter asohenadiatiaekcn » Sachel) Jf (eupeadaiatha den ites ve een eT ieee devbi vibe ab levine bre Mi + Wttosk 0p bottles cpeiaaty Lutnud Meypatewd (GI A oka Malodine) site Heit tas Honjegnc| neal ria Nuva ite vt Dad AE mam Ya) APaeinis tosent (2001) LH sina T sxe al O° fered] Jers ovon ee emcts ‘ = pee, Mee 740) ht (#01) 4 (re. "eM & MA sonuwT sath Taubes) Ant Uti oiiae Sits gyaibsood ebet> Link haigolesoG! ablliteall Begaite- ied | nei TOMAR hiledibhiA 2 WO anne? 2 elude dir pamleiiiadaans — Reig Oe dew beets DE ae. qt Waid: % ber PhoTAN A) DAL bund 5 sawainell aay omellevelsee Snip ad \neanhceso tur MO?) waelicdpel inv) « tt Oday cotineghetial (MLCT: cout Fe yederegd ee noe pe lird <> leet is TIN adie Cig s gun han (O2PCH 20 > slol8 yan rhs Shwhérwpoloy big wih ¢ G00 rev Yo yeni. (6 oc itevogne, biyptw aki vc ped iene Shue ne M1 ths) i dello cnt omer hap D4) Ob LCase Oy Laces eh ca ) nepow baa hae wingemmynpha: shcergg anew § : iM) CIT OF IRV Sheek Vy fr (RIL. AOpen 2909, 2a? ‘ome ness weriwn on che dung ctf ser7 DP Ok): FA Litnindin? | dM homgethveSbe testy coi Snbitelehyaat a9 ~peeryitineiry as an aft .GbtpaLe -squdlqeornatnal(h@eT) hy, Haat RH ooltels i ewollaw? monet i. on BFR hdd BE distAnb lakh ous}) anak meron bend dete LiFe tlh tes) odo lids Bavolelc yiint . widiis AeYtibbasl, Sah aff babe _ VaretiolntinO (addin balincéut — proweh and sry val Of bine sal etre Ceertind” Sai ehe STAT Milchiings DC. Sola DEae te Vitti ve! aenad yomoid T HOOR HD eq) vont baxaaiaast bore Sn 13: R7eepags 7e sua Video urfi'Ag 1diiiw 9 Pi ors saniuhh aN ry bos xaool boeewnesH | COB TT OD het] “fi sued 26 yiiglis z nol) 4 Co OC Oe ant Le OW OGULLICH Noss Seat W nolaivid v a~séosaib sed lo wiobewenlih (22¢1 ) in-voqqoh ik T lwwnnem bas (szonmqet af) Gtoqqne mT pene) ow gi it Woe Ate y Za Wises (FiisinR We oni 7 | At @) PAY AInT ipa Ar petri] 4? rile s | Seinen) ne) CORK war! ben Uderhiituels TemoicdbuTQRoarey yierhbweenM 1 MPD nihbBis? eevtaseeily than sf ebamg Bl + OU dexbwiesto hele hy revo hm etinodot? VAY & aie a aoe seiwe atest ts oar 1-44 nol .oterwen. Man ghve Aht4sdabsepeat)? 2 eel ae, ul - Saber onlen Mii, <73SR > rt t# ed D tiga Gly Cb. Loett etki). he LITRES SIC) sre AC Fosters Sh wet cic” ty Lae) pee) Paw vee Liye! ri oT LE Ue sere levy h (220!) of) doped ah sane WE (if) Dae adi bow yl tert pity ft werkt by se yut wd AT is Balt oF iw ole a? be itd Hig ye iv, AK) ; ‘ Ma) i : ’ a 45) Ai, 7 OV whiiwy { “9 : J ! ime (1 ee ROD erry 7 bay ' \ awe ical real ey rath i aa Wy rod Po Vil boewtiol peed, ao che gopisigk-non- dda» ome aulautotadh ogni nie the) yore OD, Aero lew. ox konenityqus ibid. Had rqsehn we canna inelaitidey + ict orensnt off tort lene bot chritagy to ayel aoe cuttin weasleind dati ly oGreeponnns beer nse eee e HEC) Mideiwsr 6:4) Sa).1215 -eRovi-ahke ie. Swiieviiin « dguodlA noi ald Hitlaup istidad gititeiinvs oi lemiedsd y jon srartw lo ETDS FURS Bed pee E ‘eal (bula Lnibw I ie 7 Le a shew ij, “atin Sciatk vi Team) et rafal Loe ni Viericw, ¥ Wrreeelrite Livailivize ( gaiiugl bos coitanillog trl online samt gormtoold nuanse vib. ar agetvarter (onsoednl) dank wrath gt sibal 2tectot 21 brrsten! M2 medaud 2 Mk boost MUMnEM JaNUaTY 2004 y sana o cagy intidedl To xebni (i ae oxie gato EPECTAL FEATOMS nodsebarget invasive hird species S2 amulo’ yyolodtinod OF VOLUMES 21 evOrTAOTRUM MO? OMe “tit * aderulta li no 2bidede no hoteborqine lm 274 Q t t apporltuoe brale noel sett ne Feuen! k o Atha ewctic firs th JGROn ayant ot) To i9dtie sAT vulseabeow hide. q0 aes vaganills ol 2 Htiay uv bin Liolt m vil fou (pedad crraatey JPteal } lagaeh »S& beach kM aja & A imadewedt 4 ono Y tnselY 3 A oho Nakamura S Of HOMNGD SFT to rT ayibpsyd 10, a / od ne an ra bene iain en nisitd a “teNyey RB. -Wolier, fe wr gig bem oa i ‘ich ie ANAL AIL Miagadlasrod evouginas 2 01 be ete Wes Hie. Columbely t Wi vcestty, lire KEI. The Cini verytey af Quesada ‘conti. «MPa aM [OE at Bese Lervter, Nad ane k- esoty Ri Chand Rant, prypert i t pe is sis t Phy Po AR tia La Nevetinghais, Ae. Virginu ene i Uptversity Atack