Copyright is owned by the Author of the thesis. Permission is given for a copy to be downloaded by an individual for the purpose of research and private study only. The thesis may not be reproduced elsewhere without the permission of the Author. - ------- Behavioural Ecology and Management of Hi hi (Notiomystis cincta), an endemic New Zealand honeyeater. A Thesis presented in partial fulfilment of the requirements for the degree of PhD in Ecology at Massey University Isabel Castro 1 995 To Isabella, who first entered the house and into our hearts. Randy, who hid from me, but not from any of the other females. Tongue face, who showed me how strong a hihi can be. Mafiefia, who taught me that hihi are born to be free. Baby, who bred with another male and two females, and Geegee, who is a real character. Abstract This thesis is concerned with the release techniques, post-release survival and behavioural ecology of hihi (Notiomystis cincta), a rare New Zealand honeyeater. It aims at offering management strategies for translocated populations. The only self? sustaining population of hihi exists on Little Barrier I sland. The New Zealand Department of Conservation is trying to establ ish self-sustaining populations e lsewhere. In 1 99 1 and 1 992 hihi transfers to Kapiti Island were approached in an experimental way. Experiments provided four main conclusions: ( 1 ) immediate-release birds survived better than delayed-release birds; (2) there was no difference between the survival of birds released in pairs or in a group; (3) hihi released in the absence of resident conspecifics survived better than those released in their presence; and ( 4) birds released in the absence of resident conspecifics moved to an area with residents in three days. The breeding system of hihi is highly variable, including monogamy, polyandry, polygyny and polygynandry. Males have physical features found in other species with highly variable mating systems. Male and female hihi benefit from a mixed reproductive strategy where a female hihi can solicit copulations from males other than her partner and male hihi can perform extra-pair copulations both with wi l l ing females or by forced copulation. Field tests aimed at determining the influence of the distribution of food and nest sites on the choice of mating system by hihi are proposed. The phenology of a selected group of plants, important as honeyeater food, was fol lowed from 1 992 to 1 994. The onset and length of the flowering and fruiting periods for particular plant species varied between the years. The number of fruits and flowers per tree also varied. Hihi egg laying periods coincided with the period of greatest flowering. Hihi breeding success was low every year. In 1 993-94 there was great competition for nest sites with kakariki Cyanoramphus novaezelandiae . It is suggested that hihi fai lure to establish self-sustaining populations on Kapiti Island is mainly the result of competition for nest sites and food l imitation. It is recommended that feeding stations and nesting boxes are established in three different areas of the island. Food should be provided during the breeding season. The quality of nectar in some small flowers, and the rate of flower visitation by hihi, tui Prosthemadera novaezelandiae and bellbirds A nthornis melanura to those flowers were measured. The estimated nectar consumption rate for all flowers was enough to sustain hihi and bellbirds' energetic requirements. It is suggested that honeyeaters might play a previously unrecognised but important role in pol lination. Forest regeneration on the New Zealand mainland could be hampered by the loss of hihi and serious reduction in the abundance of tui and bel lbirds. Necessary studies to e lucidate the role of honey eaters in pol lination are offered. Acknowledgments It was Ed Minot, my "chief' supervisor, who is mostly responsible for me taking over the study of hihi . I did not want to work with such small bird, but he pointed out to me that after having experience with large birds, the best thing in my C.V. would be to have experience also with small ones. I am very thankful to him because thi s proj ect i s one of the best things that have ever happened in my l ife . Not only Ed started the whole process, but he has been a major support a l l the way through. Ed, Robin Fordham, my second supervisor, and Brian Springett, my third supervisor, have helped me with advice, interesting discussions, and enticing questions about my research. They have spent long hours making sure my English is right in my papers and my thesis, and I can only say mil gracias. I wish all students could have such a great team as advisers for their thesis. Chapter 4 would have never been possible without the intervention of Alastair Robertson. Although he was never part of the team of supervisors he has been one of them, teaching me all I know about the old story of the flowers and the bees. Thank you AI, I think this t ime I got the story right. Thanks as wel l to Tim B irkhead who although far away has inspired me for the last years. Chapter 3 would have been a more timid chapter without Tim's encouragement. This project was initiated by the Department of Conservation and within the Department I want to thank Raewyn Empson, from the Wellington Conservancy for her unconditional support for the project. Raewyn has been a source for ideas and inspiration throughout these study. Shaarina Boyd, from the Auckland Conservancy, and head of the Stitchbird Recovery Group, has also been supportive of this project. Both Raewyn and Shaarina have also helped during field trips. I am thankful for all their help. The members of the Stitchbird Recovery Group also provided ideas and support to the project. The research that led to this thesis could have never been so complete and interesting without Julie Alley's help . Julie was my field assistant for the first year of my research and has been my friend and support for the length of the study and I hope for many years to come . I really appreciate her patience, determination and love for nature. Field work was also made easier and fun thanks to the help of Karen Mason, and Vaughan Keesing. Although Karen was stil l a bachelor's student at Massey, she came to help me during every break. Not only did she provid labour, but excellent conversations and ideas. Vaughan was the "climbing man" . He i s responsible for al l the very high hihi nest boxes on Kapiti. I am really grateful to him for stopping every thing to come place boxes up in trees for me. Al l of us, who have worked on Kapiti Island, know that an important part of it i s the ranger's house, where we get together t o dine, have a good wine, conversation and company. Thank you Peter Daniel, Shona Pengely and Marisol Pengely Daniel for many lovely evenings and for being my family during my years on Kapiti (and now). A great number of people assisted me in the field (col lecting data and keeping me sane). It is fair to say that this thesis would have not been possible without their help. In no particular order I wish to thank: Alejandra Mejia, Brigitte Bakker, Colin Miskelly, David Bell, David Sutherland, Ellis Udy, Emma, Gill Rapson, George Bradd, Gretchen Rasch, Gustavo Bula, Gwen Evans, Ian Stringer, Jonathan Mi les, Karen Campbell , Karen Creedy, Keith Woodley, Liz Watson, many members ofOSNZ Waikato, Mark, Lucy and Steven the Canadians, Murray Potter, Murray Willans, Paul Barret, Peter Griffen, Robert Hickson, Rochelle Constantine, Roger Ell iot, Sarah Gibbs, Steve and Charlotte Ful ler, Steve Pi lkinton, Suzie Brow, Tania Waghom, Tertia Thureley, the two DoC teams that help catching hihi on Little Barrier Island, Tony Whiters and Yvette Cotam. Sandra Anderson kindly took blood samples from hihi. I am thankful to Maurice Alley, Susan Cork and Peter Stockdale for determining the causes of dead of several hi hi. Tim Lovegrove, Gretchen Rasch, and Ron Moorhouse, generously shared their data on the ecology of Kapiti and their knowledge of hi hi and kaka. The guys from the possum team, track workers and all visiting scientists provided many good times at the Whare, and my thanks go to them too. Thanks to my family : their encouragement has kept me going. To Ellis, Bud, Rook and Chispa for being there, waiting for me every time after my field trips. To my mum (lsabel Acacio) and brother (Carlos) who share with me this amazing love for nature. To Hart, Elsie and Craig Udy, and the Buchanan's, for their faith in me and for not giving up hopes that one day I wi ll get a "real" job. I also want to thank the Department of Ecology for all those little things that made my l ife easy. The graduate students all kept me going during the hard times, gave me reasons to laugh and celebrate. Petra van Kan and Erica Reid, our secretaries, sent faxes, answered calls, and kept the administrative part of the project rolling. Barb Just, my "treasurer", made sure there was always money in my account, I sti l l do not know how she did it!!!. Jens Jorgensen for building nest boxes and designing portable mist? netting poles. J i l l Rapson spent hours helping Julie and I to get ready for our talks in Australia, she helped editing Chapter 4, taught me how to set vegetation quadrats on Kapiti and fed me many lol lies over these years. Kirsty Bell had just been with us, as secretary, for a week when she undertook the task of reading this whole thesis looking for editorial mistakes. Thank you all!. Thank you to Liz Grant and Keith Woodley for making all the drawings of hihi in this thesis. Particularly thanks to Liz for drawing hihi mating postures from my exited descriptions. I would have never been able to do it myself!. The Federation of University Woman provided economic support for my first year and I want to thank Mary Skipworth for encouraging me to apply. Massey University gave me the V ice-Chancellor's Award, which provided funds for the second and third year. Trust Bank, The Royal Forest and B ird Protection Society, the Ornithological Society ofNew Zealand, and Massey University also provided funds for various other parts of the project. The Department of Conservation provided transport, nest boxes, feeders, transmitters, free accommodation on Kapiti Island, and food during the transfers. Ross Le ger , owner of Kapiti I sland Charter, took me to and from the island many times during this study. Hart and Elsie Udy also granted logistic support (Mum and Dad, thanks for DR HIHI, I couldn't have done it without it! . ). This thesis no sooner wil l go to the binders than I wil l probably remember several other people who made important contributions. I apologise for any such OmiSSIOnS. P. S . Despite the length of the l ist of those who assisted my efforts, I managed to make al l mistakes completely on my own. ? ,, . . .1?:' "" . ? PREFACE Hihi is the name given by Maori to Notiomystis cincta. I use this in preference to the more usual common name, stitchbird. Hihi in Maori has several meanings all related toN cincta. The word means "ray of sun", which refers to the yel low shoulders of the male hihi . It also means a head decoration used by Maori warriors which had white feathers on the sides of a black helmet, representing the white ear tufts of the male hihi. There are some derivative words such as whahihi which means "cheeky" in relation to the active curious behaviour of the species. As a Maori noun it keeps the same form in both the singular and the plural. Hence "many hihi" (also tui, kiwi, kakariki, kaka). Each chapter of this thesis has been written in the form of a paper. Several of the chapters have already been published or have been submitted for publication. The objective of this preface is: - to give the ful l references for those chapters already published, - whenever the papers have more than one author to explain the role of each author in the making of the paper and - to give information about the chapters submitted or to be submitted for publication. Chapter 1 and Chapter 2 were presented at the Conference on Reintroduction Biology of Australasian Fauna held in Healesvil le, Victoria, Australia in April 1 993. Papers published from the Conference were refereed by two people chosen by the Editor from those scientists attending the Conference. The full references for those papers are : Castro, 1., Al ley, J .C. , Empson, R.A. & Minot, E.O. 1 994. Translocation ofhihi or stitchbird (Notiomystis cincta) to Kapiti Island, New Zealand: transfer techniques and comparison of release strategies. Pp. 1 1 3- 1 20 In Reintroduction Biology of Australasian Fauna. Surrey, Beatty and Sons. Australia. Castro, 1., Minot, E.O. & Alley, J.C. 1 994. Feeding and breeding behaviour of hihi or stitchbirds (Notiomystis cincta) recently transferred to Kapiti Island, New Zealand, and possible management alternatives. Pp. 1 2 1 - 1 28 In Reintroduction B iology of Australasian Fauna. Surrey, Beatty and Sons, Healesvil le, Australia. I presented the second paper in the spoken session of the Conference and Julie Alley presented the first one. Julie Alley was my field assistant during the first release of hihi and later from March 1 992 to April 1 993. She helped me with data collection. Ed Minot, as my supervisor, was involved in the design of the experimental releases and later helped tracking hihi during the first week of each release. Raewyn Empson, from the Department of Conservation, was in charge of the logistics of the translocation, capture of hihi on Little Barrier I sland, and she also helped tracking the birds once released on Kapiti I sland. The manuscript gained from conversation with all of them as wel l as from their editing. Chapter 1 , has not been changed from the published fom1. Chapter 2 has been changed considerably particularly the information about feeding. I have added the information I collected from 1 992-94 to that published from 1 99 1 -92. Chapter 3 has been accepted for publication in Ibis. The ful l reference is: Castro, I . , Minot, E .O. , Fordham, R.A., and Birkhead, T.R. In press. Polygynandry, face-to-face copulation and sperm competition in the hihi Notiomystis cincta (Aves: Meliphagidae). Ibis. Ed Minot and Robin Fordham have both supervised my work on Kapiti I sland, helped with mistnetting the birds, and in Ed's case watching and finding nests. Tim Birkhead kindly dissected and described the male hihi cloacal protuberance and counted the spem1 in the seminal glomera. He also applied M0ller's formula to my testes measurements and drew the c loacal protuberance and its parts. The three of them also contributed to the manuscript with editing and valuable discussions both in person and via e-mai l . I have included more theoretical information in the Discussion section of the Chapter than that allowed in the publication. Chapter 4 has been recently submitted to Functional Ecology. The reference for this paper is : Castro, I . & Robertson A . W. (Submitted). Honeyeaters and the New Zealand forest flora: The util isation and profitabi lity of small flowers. Alastair Robertson taught me the techniques to collect nectar from the flowers and the processing of the samples in the laboratory. He was also helpful in providing references on the subject of pollination ecology and evolution, and advice on statistics. The paper gained from valuable discussions with him. He also helped to edit it. Chapter 5 will be submitted to the New Zealand Journal of Ecology . Chapter 6 was written as a report t o the Department of Conservation. Title page Abstract Aknowledgements Preface Table of Contents List of Figures List of Tables Chapter 1 Translocation of hihi or stitch bird (Notiomystis cincta) to Kapiti I sland, New Zealand transfer techniques and comparison of release strategies . ........ . . ... . .... . .. . . . ......... .. . ........................................ ................................. 1 1 . 1 introduction .... . ................ . .. . . .. ... . ..... . . .. ...... .......... . .. .. . ......................... ........... 1 1 .2 Methods ........ . . ....... ......... . . .. . . ........... . . ....... . .............. ......... ............................. 2 1 .2 . 1 Transfer Techniques ..... . . ........... .... .. .... ..... ..... ............... . ................. 2 1 .2.2 Preparation of release sites . .. . ........ ? ....... . . . ... . . .. . . . . ......... ........ . . ....... . 4 1 .2 . 3 Experimental Release Strategies ...... ............................ . ..... .......... .. 5 1 .2 . 3 . 1 Immediate vs delayed release ................ ..... . .... .. .. . ..... . . . . . . 5 1 .2 . 3 . 2 Paired vs grouped release ........... . ... ... ...... ...... . ................. 5 1 .2 . 3 . 3 Presence vs absence of conspecifics . . . ... .... . . . . . . .. ..... ........ 5 1 .2.4 Monitoring Techniques .. . .... . . . . . . . . .. . . ...... . . . . .. . .... . ............................. 5 1.3 Results . . . . .. . . . . . . . . . . . . . . . .. ... . . . . . . . . . . . . . .. . . . . . .... . ... . . . . . .. . . . .... .. ..... . .. ...... . . . .. . ...... ........... 6 1 .3 .1 Survival capture to release . .... . ... . . . . .. . . . ......... . ... .. ... . . . . ... . . . ............ .... 6 1 .3 .2 Post- release survival in 1 99 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1 .3 .2 . 1 Immediate vs delayed release .. . .... . .. . ............. ............ ...... 6 1 .3 .2 .2 Paired vs. grouped release . . . ....... .. .. . .. .. . . ........ . ..... . . .......... 6 1 .3 . 3 Post-release survival in 1 992 . ..... .... ........ .. . .... . ..... . ......... .... . .. .... ...... 7 1 .3 .4 Movement Patterns in 1 99 1 .. . .. . . . ..... .. . ..... ....... . ............. . .... .... . ........ 7 1 .3 .5 Movement Patterns in 1 992 ... .... . .. . .. : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . . . . . . . 8 1 .4 Discussion .... . . .. . . ......... .. .. .. .. . .. . . . . . ..................... . ..................... ....... .. .... .......... 9 1 .4. 1 Transfer teclmiques . .......... .. . . ... . . ............... . . . .. ................................ 9 1 .4.2 Experimental Release Strategies ... ........... . .... . . ... ........... ................ . 9 1 .4.3 Monitoring-Techniques . . . . . . . . . .... . ... .......... ... . ... . ..... .. ...... ... . ........ ...... 1 0 1 .4.4 B ird Movements ... ... . . .. . .. ...... . .. ........... ....... . . ... ............................... . 1 0 1 .4.5 Recommendations ........... ............... . .. .. . ................................... . ...... 1 1 1 . 5 References .. . ... ... . .. .... .. . ...... ................ .. . .......... . .............. .... ................. . .......... 1 1 Chapter 2 Feeding and breeding behaviour of hi hi or stitchbirds (Notiomystis cincta) recently transferred to Kapiti I s land, New Zealand, and possible management alternatives .. ..... . ...... ................ . ........ ................... ................... ........ . 1 3 2 . 1 Introduction ...... . ..... .. .. ....... ........ . . ................ ........ . . ............................... . ........ 13 2.2 Methods . . . ........................ . .... .... . . . . ..... .. ................. ..... ... ........................... .. .... 1 5 2 .2 . 1 Feeding ............... . ... . . .... . . . . ........... .... . . .. .. ........................ ................. 1 6 2 .2 .2 Breeding .. . . .. ....... . . . . . . .. . ..... . ..... ... . . .... ............................................... 17 2 . 3 Results .... . . .. .. ... .. ................ .. . ................ . . .. . . ................ . .... .............................. 1 7 2 . 3 .1 Feeding ......... .. . . . . . . . ................ . . .. . . .. . ... ........................................ ..... 17 2 .3 .2 Breeding ......... . . .. . . . ............................. ....................................... ..... 1 8 2.4 Discussion .................. . .......... . .... .............. . ........ .................. ...... ........ .. ..... . .... 21 2 .4. 1 Feeding . .................. . . . .... . ..... . .......................................... ........ ........ 2 1 2 .4.2 Breeding .............. ................ . ........................................ . ........ ......... 22 2 .5 References ................... . ..... . ...... ..... ..... ....... ........................................ . . .......... 24 Chapter 3 Polygynandry, face-to-face copulation and sperm competition in the Hihi Notiomystis cincta (A ves Meliphagidae) ... . ........ ... .. ..... . . . ......... ............ . . ... 26 3. 1 Introduction . . . . . . .. . . .. .. . . .. . . ... .. . . .. . . . . . . . . . . . . . . . . .... . . . . . ..... . ... . .. ........ .. . . . . . .. .... . . ..... . .... 26 3.2 Methods . . .. .. . . . . . . . . . . . . . . . . . . .. . . . .. .. . . . . . . . ... . . . . . . . ... . . .. . .. ......... ...... . . . . ... . . . . ....... .. . ......... 27 3.2 . 1 Study site ... . . .. . . .... . . . . . . . . . . . .. . .. . . . ... . . . ..... ... .. .. . . ...... . . .... . ..... . .. ......... . ..... 27 3.2.2 Study bird . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . .. . . . . . ...... . . . . . . . . ...... .... . . ... ....... . .... .... .. .. 27 3 .2.3 Copulation behaviour of Hihi . . ... . .... .................. .......... ... . ............... 27 3.3 Results ..... . . . . . ........ . . .. ...... .. .......... ... . . .. .... . . ....... . . .... . . . .................... . ..... . ........... 2 8 3.3 . 1 Mating system ...... . . . . ... ... .. . . . ...... . . . . ... .... ... . ... ..... . . ...... .... . . ...... . ......... 2 8 3.3 . 2 Reproductive anatomy . ...... .. . . . . . ..... . . . . .... . . . . ....... . . ..... ..... . . .... . . . ......... 3 2 3.3 1 Male Cloacal protuberance ??????????? ?n?? ? ? ????????????H?????????????? 32 3.3.2.2 Testis mass, seminal glomera mass, and spermatozoa numbers ....... . . . ... .. . ........ . . ......... . .......... ..... . . 32 3.3 .2.3 F e1nale cloaca . ...... .. . . . .. ..... . . . ... ... ... . ......... .. . . ........... . . ........ 34 3.3.3 Copulation behaviour of Hihi . . .......... . . .. . .... . . ... ... . . . ... . . ..... ... . . .. .. . ... . . 35 3.3.3. 1 Face-to-face Copulation .. . .... . . . . . ......... . . .... . .... . . .. .. . . .......... 35 3.3.3.2 Copulation with the male on the female's back . . ..... .... .... 35 3.3 .3 .3 Copulation sites . . . . .. . . . . . . . .. . .... . . ... . . .................... .... .. . .... ..... 36 3.3 .3 .4 Copulation at the feeder ... . . . . . . ..... . . . . . .. ..... . . .... . . . . . .. ... .. .. . .... 3 7 3.4 Discussion . . . . . ..... . . . . ..... . . . .. . . ..... . .. . . . .. . . . . . . . . . . . ..... .. . . . . . ......... . . . . .. .. . . . . . . . . . . .... . . . . . . . . 37 3.4. 1 Mating system . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . .. .. . . . .. . .. .. . .. . . . . .... . . .... . . . . . . . .. . . . . . . . . 37 3.4.2 Copulation behaviour and reproductive anatomy . . . . .... . . . . . .. . . . . .... .. . 38 3.5 References . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . .. .. .. . . . . . . . . . . . . . . .. . . . . . . . . . . .... .. . . . . .. . . .... ..... . . . . . . .... . . . . 40 Chapter 4 Honeyeaters and the New Zealand forest flora: The util isation and profitabi lity of small flowers . . . . . . . . . . . . . . .. . . . . . . . .. . . . . . . . . . ... . . . . . . . . . .. . ... .. .. . . .. . . . . . . . . . . .. . .. . ... 42 4 . 1 Introduction . . . . . . . . . . . . . . . . . . .. . . . .. . . . . . . . . . . . . . . . . . . .. . . .. . . . . . . . . . . . ... .. . . . . . . . . . . . . . . . ............ . . . . . . . 42 4.2 Methods . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . .. .. .. . . . . . . . . . ... . . . . . .. . . . . . ... .. . . ..... . . ... .... . . . 44 4.2. 1 Study area . . . . . . . . .. . . . . . .. . . . . . .. . . . . . . . . . . . . . . . . ... . . .. . . . . . . ... ... . . .... . . ... . . . . .. . . . ..... .. . 44 4 .2 .2 Nectar sampling . . . . . . .. . . . . . . . . . . . . . . . .. . . . . . . . . . . ..... . . . . . . . . . . ...... . . . . . ..... . . .......... 44 4.2 .3 B ird observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. ..... . . . . . .. ...... 45 4 .2 .4 Honeyeaters energy requirements .. . . . . . . . ... ...... . . . . . . . . . . . . . . .. . . .. . ... . . . .. . . . 45 4.3 Results . . . . . . .. . . . . . . . . . . . . . . . . . . .... .... .. . . . . . . . . . .. . . . . . .... .. ..... . . . . . . ......... ... . . . .... .. . ... ........ . . ... 46 4.4 Discussion . . . . . . . . . . . . .. . . .. .. . . .. . . . . .. .. . . . . . . .. . . . . . . .. .. . . . . .... .. . ... . . .... .. . .. . . . . . .... . . ... . ... . . . .. . . . 48 4 . 5 References . . .. ... . . . . . . . . . . . . . .. .. . . . . .. . . . . . . . . . . . . . . . . . . . ... . . . . . . ..... .. . . .. .. .. . . . . ..... .. . ...... . .. . . . .... 53 Chapter 5 Hihi (Notiomystis cincta) feeding and breeding behaviour in relation to flowering and fruiting phenology on Kapiti Island .. .. ..... ..... . . . .. ...... ...... ... ....... 57 5 . 1 Introduction . .. . . . . . .. ... . . . .. . .... ... .... . . ........ . . .. .... ........ ................... .. ... ........ . .......... 57 5 .2 Methods . .... . . . . . .. . . . . . . . . . . . ..... . .. ... .... . . ........ . . .............. .... .......... .... . . ........ . . . .. . .... ... 58 5 .2 . 1 Study area ... . . . .. . . . . . . . . . . . . . . . . . . . . . . . . .. . . . .. . . . . ..... . . ...... . . . . . . . . .. .. . . . ... . . ........ ... 58 5 .2 .2 Phenology .. . . . ... ....... . ..... ..... .. . . . . ............ . . .. ..... . ... ... ..... . ......... ............ 60 5 .2 .2 . 1 Criteria for selecting plants .. ............. . . ...... . . . . .... ....... ....... 6 1 5.2.2.2 Data collection ....... . . . .. .. ....... . .... .. ..... ............. .......... .. .. . .... 6 1 5 .2 .3 Cl imatic records ............ .... ... . . . .. .. ...... .... .. . .. ............ .......... ... ........... 63 5 .2.4 Feeder visitation ......... .. . . . . . ... .......... . .................. . .............. ........ .... .. 63 5.2 .6 Hihi behaviour and energy requirements ................. . . ....... ......... .... 63 5.3 Results ...... . ... . .. .. . . . . . . . . . . . . . . . .. . .. . . . .. . ..... . .. .... . .......... . . . . ......................... . ........ . .. .. 64 5 .3. 1 Phenology . . . . . . ............... ..... ...... . . . ......... . . ... . ..... ...... .. ...... .................. 64 5 .3. 1 . 1 Flowers ... .... ....... . . . ..... ..... . . ... ... ................. ..... . ................. . 64 5 .3 . 1 . 2 Fruits ... . . .. .. . . . ...... . . .......... . . . . . ................... ......... ............... .. 66 5 .3 .2 Influence of cl imate on Phenology ...................................... ..... . .. . . . 67 5 .3 .3 Hi h i behaviour and energy requirements .... . . . . .................... ........... 69 5 .3 .4 Use of the feeder . ...... ..... ...... . ....................... .... . . ............. . ...... ......... 70 5 .4 Discussion . . ...... ... . . . . . . . . . . . ...... .. . . . . . . . . ......... . . ...... .. . .... . ............... .......... . . ...... . . . . . 70 5 . 5 References . . .... . . . .. . . . . . . . . . . . . . . ... . . ... . .. . . ... .. . . . . .. . . . . . . . . . . . . . . . ...... . ... . . . . . ......... . . ........... 74 Chapter 6 Monitoring and management of hihi (Notiomystis cincta) on Kapiti Is land, New Zealand ..... ............ . . . . . . . ....... . . . . . . . ........ ........... . . . .... . . . . . . .... .. . ....... . . . . . . . . 75 6. 1 Introduction .... .. ... . . . .. . .. .. .. . . . .. . . . . .. . . . ........ . . . .. . . .. . . . . . . . . .. .. . .. . . ....... . . ... . . . . .. . ...... ..... 75 6 . 1 . 1 The hihi : a case study . . . . . . . .......... . . . . . . . . .. . . ...... .... ..... . . . . .................... . 76 Supplementary Feeding .. . . . . . .. . . . . . . . . . .. . . . .... . . . . . . . . . . . . . . . .. .. . . . .. . . . . ............ . . . . . . ... ..... ...... 77 6 .2 . 1 Methods . . . . . . . . . . . . . ...... . . . . . . ... . . . . . . .... .. . . . . . . . . . .... . .. . . . . ..... . . . . . . . ................ . . 77 6 .2 .2 Results . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . .. .. . . . . . . .. . .. .... .. . . . . . . . . .. .. . . .. . .. . ..... . .... . . . . . . . . 77 6.2 .3 Discussion . . . . .. . . . . . . . . . . . . . . . .. . . . . . . .. . . . . . .. . . . . . . . . . . . . . . . . . ....................... . .. . ... . 79 6.2 .4 Recommendations . . . . . . . . . ... . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . . . . . . . ... . . . . . . ... . . . . . . 79 Measures of Survival and Breeding . .. . . . . . . . . . . . . . . . . . .. .. .. . . . . . . . . . . . . . . . . .. ... .. . . . . . ....... .. . . . . . 8 1 6 .3 . 1 Methods . .. . ....... .... . . ... . . . .. . . .. . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . ... .... . . . . ....... . ... . . . .. . .. . 8 1 6 .3 .2 Results .. . . . . . . . . . .. . . . . . . . .. . ... . . . . . . .. . . . .. . . . . . . . . . . .. . . . . . . . . . .. . . . . .. ... . . . . ... . .. . . . . . . . . . . . 82 6.3 .2 . 1 Survival . . . . . . . . . . . . .. . . . . .. . . . .. . . . . . . . . . . .. . . . . .... . . . . . . . . . . . . .... ... .. ... . . . . . . 82 6 .3 .2 .2 Predation and Disease . . . . . . . . . . .. . . . . . . . . . .. .. . . . . . . . . . . . . . . . . . . . . . .. . . ... 83 6 .3 .2 .3 Hihi Distribution . . .. ... . ..... . . . . . . . . . . . .. . . .. . . .. . . . .... . ... . . . .. . . . . . . . . . . . . 84 6 . 3 .2.4 Visabil ity of Hihi . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . .. . . . . . . . . . . . . . . .. ..... 85 6.3 .2 .5 Breed ing . . . . . .... . . . .. . . . .. . . . .. . . ... . . . . . . . . . . . . . .. . . . . . . . . . . . . .. . ... . . . . ..... . . . 85 6 .3 .2 .6 Boxes . .. . . . .. ... . . ..... . . . . . . ... . . . . ... . . ... . . . . . ..... . . .. . . . . .. . ........ .... . . . .... 86 6 .3 .3 Discussion . . . . . . . . . . ... . . . .... . . . . . . . . . . . . ... . . . . .. . . ... . . . . . . . . . . . . . ... .... . . . . . . . . . . .. . . . . . .. . 86 Plant Phenology Assessment . .. . . . . .. ... .. . . . . . . . . . . . . ... . . . . . . . . . . . . . . . . .. . . . . .. . ... . .... . . . . . ... . . .... ... 87 6.4. 1 Methods . .. . . . . . . . . .. . . . . .. . . . . . . . . . . . . .... . . . . . . . .. .. . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . ......... 87 6.4.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . .. . .. . . .. . . . . . . . .. . . . . . . . . . ...... . . . . . . . . . . . . . . . . . . ..... . . . 87 6.4.3 Discussion .. . . . . . . . . . . . . .... .. . . . . . . . . . . . . . . . . . . . . . . . . . . .... . . ... .. ..... . . . . . ...... . . . . . . . . . . . . . 88 6.4.4 Recommendations . . . . . . . . . . . . . . . .. .... . . .. . . . . . . ..... .. . . . . . . .. . . ..... . . . .... .......... ... . 88 6 .5 Monitoring program . .... . . .. .. . ... .. . . . . . . . .. . . . . . . . . . . . ............ .. . . . . . . . ..... ... . . .... . . . .. . ........ 88 6.5 . 1 Hihi monitoring from April to August ..... . .. . . . . . .... .. . . . . . . . . . . . . . . . ......... 89 6 .5 .2 Hihi monitoring from September to March . . . . . . . . .... ..... . . . ........ ........ 89 6 .5 .2 . 1 Breeding Activity . . . . . . . . ............. . . . . . . . .. ............ . ........... ...... 89 6 .5 .2 .2 Nest boxes . . . . . . . .. . ...... . . .............. . . . . . . . . . .. ...... ... . .. ....... . . . ...... 90 6 .5 .2 .3 Monitoring the use of feeders by hihi . . . . . . . . .. . .................. 90 6 .6 References . . . . . . . . . . . . .. . .. . . . . . . . . . . .. .. .... .... .. . .. .. . . . . ... .. . . .... . . . . . . . . . ... ... . . ...... . ..... ........... 92 Chapter 7 Concluding Remarks . . . . .......... . . . . . . . ..... .... ............. .. ...... . . . . . ........................ . . . . . 94 7 . 1 Managen1ent . . . . .. . . .... . . ..... . . . . . . . . . . . .... . . . .. . . . . . . . . ... . ........ . . .. .. ........... .. . ........... . ....... 94 7 .2 Breeding biology .. .. . .......... . ....... . . . . .......................... . ......... . ............. .............. 95 7 .3 Honey eaters and pollination .... ..... . . . . . . . . . ... . . ...................... . ............ . . . ......... .... 96 References . . . . .... . . .... . ... .... . . . . . . . . . . . . . .... .. . . . . . . .. . . . . . . . . ........ . . . . . . .. .. .. . .................. . .......... 98 List of Figures Figure 1.1. Map of New Zealand showing the location of the islands to which hihi have been transferred from Little B arrier I sland since 1 983 . . . . . . . . . . . . . . . . . 2 Figure 1.2. Design of the cat-transporting box modified to carry hihi . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Figure 1.3. M ap of Kapiti Is land showing the 1 99 1 and 1 992 release sites . . . . . . . . . . . . . . . . . . . 4 Figure 1.4. Movements of radio-tagged birds in the first 15 days after release on Kapiti Is land in 1 99 1 . ...... . . .. .. . . . ....... .. .. . . . . .... . . .... ... ...... .... . . ..... . ... . . ............ .... . 7 Figure 1.5. Movements of radio-tagged birds in the first 15 days after release on Kapiti I sland in 1 992 . . . . . . . . . . . . . . . . . . . . . . : . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Figure 2.1. Map of New Zealand showing the distribution of hihi before the 1 800's . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . l 4 Figure 2.2. P ercentage of observations of hihi taking different types of food on Kapiti I sland ...... . . . . .. . . . .. .... . ..... . . . . . .. . . .. .. . . . . . . . . . ..... . . . .... . . ....... . ...... . . ..... . ............. IS Figure 2.3. Percentage of birds using the Rangatira feeder each month, and p ercentage of observations of hihi feeding on flowers on Kapiti I sl and ( 1 99 1 to 1 994) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 9 Figure 2.4. Comparison between the feeding preferences of hihi on Little B arrier I sland and Kapiti Is land . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Figure 3.1. Map showing male hihi home ranges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Figure 3.2. Number of calls per hour at hihi nest sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . 30 Figure 3.3. Male hihi cloacal protuberance . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Figure 3.4. Cross-section of male hi hi cloacal protuberance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 4 Figure 3.5. View o f female hihi cloacal protuberance . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 4 Figure 3.6. Hihi mating postures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 6 Figure 4.1. E stimated number of calories in the flowers and inflorescences of selected New Zealand forest plants on Kapiti Island . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Figure 4.3. Estimated number of calories per minute obtained by birds foraging of selected species of flowers on Kapiti Is land . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 2 Figure 5.1. Map showing the different vegetation zones present o n Kapiti I sl and . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 9 Figure 5.2. M ap o f Kapiti I sland showing the maximum elevations . . . . . . . . . . . . . . . . . . . . .. . . . . . . . 60 Figure 5.3. Map of Kapiti I sland showing all tracks used for the phenology transect. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 Figure 5.4. Number of species in flower each month on Kapiti I sland based on I 7 species of plants used by hihi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . 64 Figure 5.5. C omparison between the number of species in flower and the number of such species visited by hihi on Kapiti I sl and - 1 992- 1 994 . ...... . . ..... ..... . . .... . ...... . .... . . . . .. . .. ....... . . .. . . . ...... . . ........... .. .. . . .. . ... .... ............... 65 Figure 5.6. Comparison between the number of species in flower in 1 992, 1 993 and 1 994 on Kapiti I sland . ... . .. ............. . . . .. ...... ........... . . . . ....... . .......... .... . ....... . 66 Figure 5. 7. Comparison between the number of species in fruit in 1 992, 1 993 and 1 994 on Kapiti I sland . . ............... ......... ......... ... . . ... . . . ........ .. .......... . ... ...... . 66 Figure 5.8. Comparison between the number of plant species in fruit and the number of plant species used by hihi on Kapiti Island, 1 992- 1 994 . . . . . . . . . . . . . . 67 Figure 5.9. Lag period between the flowering of most species studied and the warmest months . .. . . ... . . . . . .. . . . .......... . ........ ..... . . .. . . . . . . . . . . .. . ....... . . ... . . . . . ................ 68 Figure 5.10. Lag periods between the largest number of species flowering and the total sunshine in hours. Top figure sows the best fit for 1 992 . . . . . . . . . . . . . . . . 68 Figure 5.11. Percentage of observations of hihi feeding on flower nectar, fruits and invertebrates on Kapiti Island from 1 99 1 - 1 994 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 Figure 5.12. Monthly percentage of hihi visiting the feeder on Kapiti I sland. 1 99 1 - 1 994 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 Figure 6.1. Feeders used to provide supplementary food for hihi on Kapiti Is land until 1 994 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 Figure 6.2. Hummingbird feeders made by Perky Pet Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 Figure 6.3. Feeder covered with a wire netting cage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 1 Figure 6.4. Number of birds known to be al ive on Kapiti Island from 1 99 1 to 1 994. ???? ??? ????? ???? ? ?????????? ?? ? ? ? ??? ? ??? ??????? ? ??? ?? ? ? ? ?? ? ? ? ? ? ? ?? ? ???? ? ?? ? ?? ? ?? ? ? ? ? ??? ?? ? ? ? ? ? ? ?? ??? ?? ? ? ? ? ? ??? 82 Figure 6.5. Location of all known hihi nest sites on Kapiti I sland . . . .. . ..... . . . .. . . . .... . . .. . . . .. 84 Figure 6.6. Number of species in fruit or flower on Kapiti Island from 1 992- 1 994 . . . . .. ..... . .... . ...... ... ....... ......... .. ... .. . . . . . ....... . . . . . .... . . . . . . . . . .... . . . . . . . . . . . . ... .... . . . .. . . . 87 Figure 6 . 7. Tracks suggested for hihi monitoring from April to August. . . . . . . . . . . ... . . . .. . ... 89 Figure 6.8. Suggested route to monitor hi hi from September to March . . . . . . . . . . . .. . . . . . ...... 90 Figure 6.9. Map showing the location of nesting boxes on Kapiti Is land . ......... . .. . . . .. . . . . 9 1 Figure 6. 1 0. Map showing the location of the feeders on Kapiti I sland . ......... . . . . . .... . .. . . 9 1 List of Tables Table 1.1. Comparison of survival of birds released as a group or in pairs on Kapiti I sl and in 1 99 1 . . . . ... . . . . ... .... . . . .. . .... . ... . . .... . . . . ... ... . . . . . . . . . .... . . . . .... . . . . . . . . . . . . . . . . 6 Table 2.1. Hihi breeding results for Kapiti I sland . .... . . .. .. . . . . . ... . . .. . . . . .. ..... . . ... . . . . . . . . . . . . . . . . . . . 2 1 Table 3.1. Comparison of male presence at nest sites during nest building and laying with male nest attendance during chick rearing . .. . . . . . .. . . . . .. . . . . . . . . . . . . . . . . . 3 1 Table 3.2. Review o f information about testes and cloacal protuberance sizes and number of sperm in seminal g lomera of polygynandrous species . . . . . . . . .. . . .. . . .. . . . .. .. .. .. . .... ... ..... . . . ..... . .. .. . . . . . . . . : . . . . . .. . . . . . . . . . . .. . . .. .. . . . . . . .. . .. . . . . .. . 38 Table 4.1. List of p lant species on Kapiti Island whose flowers are v isited by honeyeaters . .. . . . ..... . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . .. . . . .. . . . . . . . . . . . .. . . .. .. . . . . . . . . . . . .... . .... ... . . . . . . .. . . 4 7 Table 4.2. Results of ANOV A . . . . . . . . . .... . . . . ..... . . . . . . ..... .... . . . . . . . . . . ...... ..... . ... . . . . ... .. .. . . . . . . . . . . . . . . . 4 8 Table 4.3. Average number o f flowers visited by honeyeaters while foraging for one n1inute . . ... . . ...... . . . . . . . .. . . . . . . . . ... .... ... ...... .. . . . . . . . . . . .. .. . . . . . . . . . . . .... . .. .. . ...... . . . . . . . . ... . 49 Table 5.1. S ampling scheme to determine the phenology of plants important as hi hi food on Kapiti I s land . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1 Table 5.2 Additional species used by hihi as food plants . . . . . . .. . . . . . . . . . . . . ... .. . . . . .. . . . . . . . . . . . . . . . . 65 Table 5.3. Results from eight hihi nests on Mokoia Island . .. ....... . ... . . . . .... . . . . . . . .. . . . . . . . . . . . . . 72 Table 6.1. P ercentage annual surv ival of hihi on Kapiti I s land from 1 99 1 to 1 993 . . .. . .... . ....... . . . . .... . . .. . . ... . ... . ..... . . . . . .. ........ ...... . . . ............ . . . . .... ...... . . . . . . . . ...... . .. 83 Table 6.2. Home ranges, wintering areas, and known breeding sites of surviving hihi on Kapiti Island . . . . . . . . . . ..... ..... . . ... ...... . . .. . . ............ . . . . .... . . ... . . .... . ..... ... . ... . . . . . 85 Table 6.3. Monitoring plan for h i hi on Kapiti I sland . . . ... . . .. . . . ... ..... . .. . ..... .... . .... . .. . . . . . . . . .. . . 88 Chapter 1 =--?u? Chapter 1 Translocation of hihi or stitchbird (Notiomystis cincta) to Kapiti Island, New Zealand: transfer techniques and comparison of release strategies. ABSTRACT Little Barrier Island, New Zealand, currently supports the only self-sustaining population of hihi (Notiomystis cincta). The New Zealand Department of Conservation is trying to establish populations of this species elsewhere to ensure its survival. In 1 99 1 and 1 992 hihi transfers to Kapiti Island were approached in an experimental way. In 1 99 1 , four release strategies were tested: immediate, delayed, paired and grouped releases. In 1 992, the effect of the presence or absence of conspecifics was tested. Comparison of the different strategies based on bird sightings and movements during the first 4 weeks after release supported four main conclusions: ( 1 ) immediate-release birds survived better and travelled over a greater area than delayed-release birds; (2) there was no difference between the survival of birds released in pairs or as a group; (3) hihi released in the absence of resident conspecifics survived better than those released in their presence (not statistically significant); and (4) birds released in the absence of resident conspecifics appeared in an area with residents about three days after being released. 1 . 1 INTRODUCTION THE conservation of many ofNew Zealand's threatened species depends on the establishment of viable populations on islands with suitable habitat (Merton 1 975 ; Flack 1 978; Nillson 1 978; Rasch 1 988; Roberts 1 99 1 ) . However, translocations are seldom wel l monitored and the results often remain unpublished (Young 1 990). Conant ( 1 988) and Griffith et al. ( 1 989) have pointed out that the success of translocations can be enhanced by monitoring outcomes, improving transfer techniques and release strategies, and by making the results available to wildlife managers. The hihi (Notiomystis cincta) is one of three species of honeyeaters endemic to New Zealand. I t is considered vulnerable (IUCN 1 979) because the only self-sustaining population is restricted to Little Barrier Island, New Zealand. A recovery program for hihi began in 1 980, but previous attempts to establish self-sustaining populations of the species on Hen, Cuvier, and Kapiti Islands (Figure 1 . 1 ) have been unsuccessful (Angehr 1 984; Rasch 1 99 1 ). The aim of this Chapter is to describe the techniques used to transfer hihi from Little Barrier I sland (3053 ha) to Kapiti Island ( 1 968 ha) in August 1 99 1 and August 1 992, and to compare the success of different release strategies for translocated hihi in the first four weeks following release. 2 ) South Island t Little Barrier Island t o ? ? ..,.. Cuvier Island Paraparaumu Beach 0 100 Wellington 300 Km N t 500 Figure 1.1. Map of New Zealand showing the location of the islands (Hen, Cuvier, Kapiti) to which hihi have been transferred from Little Barrier Island s ince 1 983 . In 1 99 1 and 1 992, two transfers took place to Kapiti I s land. For these two transfers the birds were carried by hel icopter from Little Barrier Is land to Auckland A i rport, by schedu led commercial airplane to Wel l ington A i rport, by car to Paraparaumu Beach and by hel icopter to Kapiti I s land. 1 .2 METHODS 1 .2 .1 Transfer Techniques H ihi were captured with mist nets in three different catchments on Little Barrier Island (Empson 1 992). In August 1 99 1 , 54 birds were caught, of which 48 birds (24 males and 24 females) were transferred to Kapiti Island. In August 1 992, of 50 birds caught, 4 7 birds (24 males and 23 females) were transferred. Immediately after capture the birds were weighed, measured and fitted with coloured and metal leg bands. In 1 992, blood samples for DNA analysis were also taken from the brachial vein of all birds. After capture and handling, all birds were fed glucose solution before being released into an aviary on Little Barrier Island. The maximum length of time a bird was held in this aviary was 1 5 days in 1 99 1 and 1 0 days in 1 992. The aviary (5 x 3 x 2 m) was divided with shade cloth into four compartments, each being accessed through a shade cloth "door" . In 1 99 1 , two compartments held 1 4 birds each and a larger third compartment, comprising two compartments with the door between them open, contained 20 birds. Individuals in the third compartment, which was near to the entrance of the aviary, appeared particularly stressed when the aviary was serviced. For this reason, in 1 992 all the doors between compartments were left open, so that al l birds had access to the entire aviary. Artificial and natural foods, and water were supplied ad libitum in the aviary. Food consumption was c losely monitored (Castro 1 992). Fresh leaf l itter was put on the ground each week and the feeders were cleaned at two-day intervals. Feeders ( in 1 99 1 and 1 992) and nest/roost boxes (in 1 992) identical to those provided on Kapiti Island were placed in the aviary to fami liarize the birds with them. The birds were recaptured in the aviary using a hand net within 2 hours of being translocated to Kapiti I sland. They were then reweighed and placed in transfer boxes. In 1 99 1 , modified cardboard cat-transporting boxes were designed to carry the birds (one male and one female) to be released in pairs (Figure 1 .2). Those birds destined for group release were placed in wooden transfer boxes ( 1 x 0.4 x 0.4 m) originally designed for saddleback (Philesturnus carunculatus) transfers. In 1 992, the cardboard boxes were used to carry birds to be fitted with radio transmitters. The remaining birds were carried in wooden boxes in groups of five. In both years, the cardboard boxes were placed in specilllly constructed crates to avoid their being crushed or significantly shifting during transportation by plane. The inter-island transfer involved three modes of transport in al l : helicopter, commercial plane and car. Figure 1.2. Design of the cat-transporting box modified to carry hihi. I. A small entrance was cut on the side of the box. A metal pin held it shut. 2. Velcro attachments on tins and box were enough to hold the tins in place while at a stop. Tins were removed while in transit. 3. The floor of the box was covered with newspaper. A piece of chicken mesh was placed over the paper. 4. Apple halves were attached to the box with a piece of string. 5. Perches were attached to the box. 6. Shade cloth was placed inside the box to cover the small breathing holes. 7. The top of the box was flattened and taped to avoid unwanted opening during transportation. 4 The birds were provided with both water and l iquid food when not travelling. Apple halves were securely attached to the inside of each transport box to provide the birds with food throughout the transfer. 1 .2 .2 Preparation of release sites On Kapiti the 1 99 1 release sites were chosen from several areas which had aviaries previously used for the release of saddlebacks (Figure 1 .3 ) . The aviaries chosen for the delayed-release birds were provided with fresh vegetation, small water baths, perches and feeders. In 1 99 1 , feeders were also p laced outside the aviaries for the immediate-release birds. Because there was no restriction on the birds that could use the feeders outside the aviaries there was considerable disturbance, mainly by kaka (Nesto?r notabilis), a New Zealand parrot. In 1 992, the release sites were selected according to habitat suitabi l ity and the presence or absence of resident conspecifics. The feeders in 1 992 were designed to exclude b irds larger than the hihi but could not exclude the smaller bellbird (A nthornis melanura). Six feeders were placed in catchments between and including the Kaiwharawhara and the Te Rere streams. Nesting/roosting boxes were also placed in those catchments. m Figure 1.3. Map ofKapiti Island showing the 1991 (black circles) and 1992 (asterisks) release sites. An aviary and feeder were present at each release site. 1 .2.3 Experimental Release Strategies 1 .2.3 . 1 Immediate vs delayed release- In 1 99 1 , 1 2 pairs of birds were released as soon as possible after arrival at Kapiti Island (immediate-release), and 1 1 of the 1 2 remaining pairs after two weeks in captivity on Kapiti (delayed-release). Food was provided for both immediate-release and delayed-release birds for a period of 4 weeks. 1 .2 .3 .2 Paired vs grouped release - In the 1 99 1 delayed-release group, six pairs were kept together as a group in a large aviary (5 x 3 x 2.5 m) and then released together. The six remaining pairs were each taken to a different small aviary (2 x I x 2 m) at one of six locations. Similarly, in the 1 99 1 immediate-release group, six pairs were released together from the large aviary, while the remaining six pairs were released from each of the small aviaries. 1 .2 .3 .3 Presence vs absence of conspecifics- In 1 992, survival of birds released into areas supporting resident hihi was compared with that of birds released into areas without resident hihi. Twelve pairs were immediately released in the Kaiwharawhara, a catchment which held no resident hihi, and 1 1 pairs plus a male bird were immediately released at Rangatira. Surviving birds from the 1 99 1 release ( 1 8 adults) and their offspring ( 1 2 birds) had established between the Te Rere and Kahikatea catchments. Food was provided at each of six aviaries for 4 weeks. 1 .2.4 Monitoring Techniques Data on the survival and movements of transferred birds were obtained by tracking radio-tagged birds, observing hihi at the feeding stations, and undertaking dai ly searches of the Te Rere, Paripatea, Kahikatea, and Kaiwharawhara catchments (Figure 1 .3 ). Adjacent areas were also searched occasional ly. Different release strategies were evaluated for the first 4 weeks after release. After this time it was assumed that factors other than those directly associated with the release would be responsible for the disappearance of birds. Survival was monitored by sighting of colour-banded individuals. Birds not seen during the 4 weeks after release but seen subsequently were counted as survivors. Radio-telemetry was used to locate hihi and map their movements during the first 2 weeks after release on Kapiti I sland. The small size and high mobility of hihi made it necessary to use a small transmitter glued to the cl ipped bases of the contour feathers on the back of the birds (Castro 1 992). The transmitter's battery lasted on average 14 days. In 1 99 1 , two birds (a male and a female) in each of the delayed/grouped, immediate/grouped, and the immediate/paired treatments were fitted with transmitters, while only a male was fitted with a transmitter in the delayed/paired treatment because the eighth transmitter was faulty. A transmitter was also attached to a male, one of four birds that were on the island prior to the 1 99 1 release. In 1 992, transmitters were attached to four birds in the conspecifics-present group and four birds 5 6 in the conspecifics-absent group (two of each sex in each group). 1 .3 RESULTS 1 .3 . 1 Survival: capture to release Of 1 04 birds held captive on Little Barrier I sland, only one died, probably as a result of stress. The inter-island transfer took a maximum of 22 h including a 1 2 h stopover. A l l birds survived the transfer and were released on Kapiti between 09:00 and 1 2 :00 h. On Kapiti a single female from the delayed/paired treatment group died prior to release, apparently of a stress-related i l lness (P. Stockdale, pers. comm.). 1 .3 .2 Post- release survival in 1 991 1 .3 .2 . 1 Immediate vs delayed release- Delayed-release hi hi disappeared faster than hihi that were released immediately. Of the 24 hihi released immediately, 75% were known to be al ive 4 weeks after release, whereas only 46% of the 24 delayed? release birds were known to be alive 4 weeks after their release (Fisher exact test, 2 - tailed, P=0.075) . The greatest difference in survival in 1 99 1 was recorded between immediate? release females (83% resighted 4 weeks after release) and delayed-release females (50% resighted after release). A single female from the delayed-release group was found sick after release (due to possible pneumonia), recaught, and returned to captivity. However, before she recovered, she was ki lled and eaten by a Norway rat (Rattus norvegicus). Delayed-release females kept in pairs apparently suffered greater stress than those kept in groups. Dominant males in the wild were observed displaying aggressively to paired caged birds. In addition, when caged in pairs males were aggressive towards females, monopol izing the food tray for long periods of time. This kind of behaviour was not observed in the birds kept in a group. 1 .3 .2 .2 Paired vs. grouped release- There was no difference between the survival of birds released in pairs and birds released as a group (Table 1 . 1 ) . Table 1.1. Comparison of survival 4 weeks after release of birds released as a group or in pairs on Kapiti Island in 1 99 1 . Immediate Paired Female Male 516 4/6 Grouped Female Male 5/6 4/6 Delayed Paired Female Male 3/6 3 /6 Grouped Female Male 3/6 3 /6 1 .3.3 Post-release survival in 1 992 Although 7 1% of the 24 birds released in the absence of conspecifics were known to be alive at the end of the first 4 weeks after release, only 56% of the 23 birds released in the presence of conspecifics were sighted. However, this difference in survival was not statistical ly significant ( x: =1 .4, P > 0.05, NS). One male and one female from the group released in the absence of conspecifics and a male from the group released in the presence of conspecifics showed symptoms of a respiratory condition 2 to 3 weeks after release. A post mortem examination of one of the males fol lowing euthanasia suggested an air sac mite (Sternostoma tracheacolum) infection to be a possible cause of the condition (S. Cork pers. comm.). 1 .3.4 Movement Patterns in 1 991 Figure 1 .4 shows the areas used by the different groups of radio-tagged birds. Immediate-release birds travel led over a greater area than delayed-release birds. The behaviour of the latter group resembled that of the radio-tagged Kapiti Island resident male in that they stayed mostly in one relatively small area. However, after the first week all birds showed a preference for a particular area. Rangat?ra Point m 1 000 Figure 1.4. Movements of radio-tagged birds in the first 1 5 days after release on Kapiti Island in 1 99 1 . The arrows indicate general movements and do not correspond to individual birds. The dots represent release sites. Horizontal lines show the known area used by immediate release birds and the stippling shows the area used by delayed release birds. Immediate-release females were the only birds to be often sighted at different feeding stations on the same day. This was true for both radio-tagged and colour? banded birds. This suggests that immediate-release females moved over a greater area than males or delayed-release females. For surviving males immediate-release birds were sighted on 1 9 ofthe 30 search days on average(64%), delayed-release males were 7 8 sighted on 1 3 of the 1 6 search days on average (80%). This suggests that delayed? release males moved out of the monitored areas less often than the immediate-release males. 1 .3.5 Movement Patterns in 1 992 All but one of the four radio-tagged birds released at Rangatira moved to the Te Rere catchment immediately after the release and stayed there during the four weeks of intensive observations. The fourth bird moved in the opposite direction, establishing himself in an area of the upper Kahikatea catchment where there had been no sightings of hihi during the research in 1 99 1 (Figure 1 .5) . Of the 20 non-radio-tagged birds released in Rangatira, five moved to the Te Rere Catchment and five to the Kahikatea. The remaining ten were not seen after release. Release site without 500 1 000 m N / con specifics Figure 1.5. Movements of radio-tagged birds in the first 1 5 days after release on Kapiti Island in 1 992. The arrows indicate general movements and do not correspond to individual b irds. Dots represent the release sites. Horizontal l ines show the known area used by birds. The four radio-tagged birds released in the Kaiwharawhara moved into the lower Kahikatea and Te Rere catchments within 4 days of their release, and establ ished themselves in areas frequented by resident hihi. Of the non-radio-tagged birds released in the Kaiwharawhara, nine remained within the catchment and were seen using the feeding station during the four weeks of intensive observation, six moved to the Te Rere Catchment within the first six days of release and one moved into the Kahikatea Catchment. The remaining three birds were not found. 1 .4 DISCUSSION 1 .4 . 1 Transfer techniques Holding the hihi in a single large aviary fol lowing their capture on Little Barrier Island did not seem to affect the birds particularly. However, it is not clear whether the possible air sac mite infection in 1 992 developed on Little Barrier I sland prior to the transfer, or on Kapiti Island after release. This infection could have occurred as a consequence of weather conditions in 1 992 and may not be a problem in future transfers. However, if the disease was caught on Little Barrier Island, it is possible that keeping the birds in the same large cage could have encouraged the transmission of the parasite (Fi ll ipich 1 99 1 ) . The hihi appeared to travel wel l with no obvious problems. Wooden transfer boxes and modified cardboard cat-transporting boxes proved equal ly suitable for transporting the birds. When transporting the birds on foot, the cardboard boxes were considerably easier to carry than wooden boxes. However, wooden boxes allow birds to be carried in groups and provide better security for air travel. 1 .4.2 Experimental Release Strategies Siege! ( 1 980) pointed out that chronic stress may cause suppression of the immune system and predispose birds to disease caused by opportunistic pathogens. Previous post mort em examination reports of captive hihi indicate that this species is particularly susceptible to respiratory disorders (e.g. mycotic pneumonia) when held in captivity (S . Cork, pers. comm.) . The yellow-tufted honeyeater (Lichenostomus melanops melanops) from Australia is also very susceptible to respiratory diseases, notably mycotic pneumonia (Smales et al. 1 992) . Mortality of birds after release could be attributed to reduced competitive abi l ity brought about by the period of captivity and exacerbated by confrontation with a new environment. Hihi are at the bottom of the honeyeater hierarchy in New Zealand (Craig, Stewart and Douglas 1 98 1 ; Craig 1 984; Angher 1 984; Rasch 1 985 ; pers. obs.). S ize, sex and physical condition are important determinants of social dominance (Angehr 1 984). Delayed-release birds remained captive for a longer period of time and, in addition, when released had to compete with the immediate-release birds as well as other resident honey eaters. Kapiti Island has large populations of both tui (Prosthemadera novaeseelandiae) and bellbirds, the other two New Zealand honeyeaters. This potentially stressful situation could explain the lower survival of the birds. Hihi released into an area not inhabited by resident conspecifics had an only slightly greater rate of survival . Birds released in the absence of conspecifics found other hihi relatively quickly and sometimes established themselves in areas adjacent to or in the residents' home ranges. 9 1 0 It is uncertain how significant supplementary feeding was to the survival of the hihi. The regular and sometimes frequent use of the feeders by hi hi immediately after release indicates that it is an important source of food in this period. Having an easily accessible food supply may have alleviated some of the stress the birds could have experienced while famil iarizing themselves with a new environment. 1 .4 .3 Monitoring Techniques Transmitters were of great value in fol lowing the movements of birds and their behaviour after release. Because the transmitters fel l off natural ly or could be removed from the birds soon after the batteries expired, the birds did not have to carry them for a long period of time. There was no significant difference in the survival of birds with and without transmitters (for 1 99 1 and 1 992 combined, X? =0.005, NS). Surviving birds were sighted on average 95% of the time, so I am confident that the survival data presented here reflects the real survival rate of hihi. Feeding stations proved to be of considerable value in assessing survival of hihi during the 4 weeks following the release. Hihi readily learnt to use the feeders and although birds were also seen elsewhere, most sightings of hi hi were made at the feeders. 1 .4.4 Bird Movements The wide movements of immediate-release birds can be explained by the time of the year when they were released. Males and females have been seen searching for cavities on Little Barrier Island in August (Angehr 1 984). Females compete for nesting holes (Castro et al. 1 994; Chapter 2) and it would be an advantage for a female to know the location of all possible sites, so that she can select the best site or move to another if her first choice is taken by another bird. Male movements on Kapiti may therefore have been influenced by the need to establish a breeding territory, while female movements may have been influenced by the need to choose both males and an appropriate nesting site. I am not certain why the delayed-release birds moved less than the immediate? release ones. I t is possible that delayed-release birds stayed in the areas where they were released because their physical condition was relatively poor. By staying around the release site they had access to feeders and therefore to easy food. They also could have been prevented from leaving the release areas by immediate-release birds. However, the mere presence of resident conspecifics does not seem to prevent movement. In 1 992, birds were released in the presence of conspecifics and stil l they moved over areas as large as the immediate release birds in 1 99 1 . Thus, the difference in movements in 1 99 1 was a result of the delayed release rather than the presence of conspecifics. 1 .4.5 Recommendations 1 ) Hihi should be released as soon as possible after transfer, preferably in areas free of conspecifics, and be provided with supplementary food at more than one feeding station, that excludes larger competitors. 2) Hihi should be provided with water and food as often as possible during the transfer. Although nectar mix is an ideal food, as a liquid it can only be used when not moving. Apple halves, securely attached to the lower part or base of the transfer box, are more practical during transportation. 3 ) For hihi, stress in captivity can be minimized by not caging birds in areas inhabited by dominant conspecifics, by caging birds in groups rather than in pairs, and by keeping them in captivity for as short a period as possible. 4) The possibility of disease transmission during captivity should be minimized whenever possible. The birds should be kept in small groups of about a dozen birds, and isolated from each other rather than housed all together in a single large aviary. 5) Transfer boxes should be sturdy and l ight-weight. The modified cat-transporting boxes are useful if they need to be carried for substantial distances. Otherwise wooden boxes are suitable. 6) All releases should be closely monitored to determine their success. Monitoring also allows problems, some of which could be corrected immediately, to be identified. 7) Further work needs to be done to assess the influence on survival of ( 1 ) release of birds in groups of the same sex, (2) release of birds at different times of the year, (3 ) release of birds of different ages, and ( 4) use of nest boxes as roosts after transfer. 1 .5 REFERENCES Angehr, G.R. 1 984. Ecology and behaviour of the stitchbird. Internal report to the New Zealand Wildlife Service, Well ington. Castro, 1., Minot, E.O. and Alley, J.C .. 1 994. Feeding and breeding behaviour of hihi or stitchbirds (Notiomystis cincta) recently transferred to Kapiti Island, New Zealand, and possible management alternatives. In Reintroduction Biology of Australian and New Zealand fauna. Melody Serena Ed. Surrey Beatty and sons. Chipping Norton. Australia. Castro, I . 1 992. Stitchbird transfer: Little Barrier I sland to Kapiti I sland, August 1 99 1 . Internal report to the New Zealand Department of Conservation, Wellington. Conant, S. 1 98 8 . Saving endangered species by translocation. Bioscience 38: 254-25 7. Craig, J .L. , Stewart, A.M., Douglas, M.E. 1 98 1 . The foraging of New Zealand honeyeaters. New Zealand Journal of Zoology. 8 : 87-9 1 . Craig, J .L. 1 984. Wing noises, wing slots and aggression in New Zealand honeyeaters (Aves: Meliphagidae). New Zealand Journal of Zoology. 1 1 : 1 95-200. Empson R. 1 992. Report on release of stitchbirds on Kapiti I sland, August-September 1 1 1 2 1 99 1 . Internal report to the New Zealand Department of Conservation, Well ington. Fillipich, L.J. 1 99 1 . D iseases in finches. Pp 369-38 1 in Avian Medicine ed by T.G. Hungerford. University of Sydney: Sydney. Flack, J.A.D. 1 978. Inter island transfers ofNew Zealand black robins. Pp. 362-372 In Endangered B irds : Management Techniques for Preserving Threatened Species ed by S .A. Temple. The University of Wisconsin Press: Madison and Croom Helm Limited, London. Griffith, B . , Scott, J .M. , Carpenter, J. W. and Reed, C. 1 989. Translocation as a species conservation too l : status and strategy. Science 245: 477-480. International Union for the Conservation of Nature and Natural Resources (IUCN). 1 979. Red data book. Vol 2 Aves. IUCN: Morges, Switzerland. Merton, D.V. 1 975 . The saddleback: its status and conservation. Pages 6 1-74 In Breeding Endangered Species in Captivity ed by R.D. Martin. Academic Press: London. Nillson, R.J . 1 978 . The South Island saddleback. Pages 32-36 In Wildlife-A Review (Vol . 9). New Zealand Wildlife Service, Department of Internal Affairs : Well ington. Rasch, G. 1 988 . Draft recovery plan for North Island Kokako. Internal report to the New Zealand Department of Conservation, Wellington. Rasch, G. 1 985 . The behavioural ecology and management of the stitchbird. Pp. 6-2 1 . MSc Thesis, University of Auckland. Auckland, New Zealand. Rasch, G. 1 99 1 . A recovery p lan for the stitchbird (Notiomystis cincta). Internal report to the New Zealand Department of Conservation, Wellington. Roberts, A. 1 99 1 . A recovery plan for the South Island saddleback. Internal report to the New Zealand Department of Conservation, Wellington. S iege!, H .S . 1 980. Physiological stress in birds. Bioscience 30: 529-534. Smales, 1., Miller, M., M iddleton, D. , and Frankl in, D . 1 992. Establ ishment of a captive-breeding program for the helmeted honey eater. International Zoo Year Book. 3 1 :5 7-63 . Young, E.C. 1 990. Workshop: translocations (Summary) . Pp.29 1-292 In Ecological Restoration ofNew Zealand Islands ed by D .R. Towns, C.H. Daugherty and I .A.E. Atkinson. Conservation Sciences Publication No. 2. New Zealand Department of Conservation, Wellington. Chapter 2 Chapter 2 Feeding and breeding behaviour of hihi or stitchbirds (Notiomystis cincta) recently transferred to Kapiti Island, New Zealand, and possible management alternatives ABSTRACT The breeding and feeding behaviour of hihi differ between their source population on Little Barrier Island and Kapiti Island, where they have been transferred. In contrast to Little Barrier Island the hihi on Kapiti Island use artificial feeders during spring and adopt a variety of polygamous mating systems. The change in behaviour reflects both environmental differences between the islands and the abi l ity of the species to adapt to a new environment. Although Kapiti may be a suboptimal habitat for hihi, the behavioural plasticity of the species could be used, in combination with habitat management, to establish a population on Kapiti . 2. 1 INTRODUCTION THE hihi or stitch bird (Notiomystis cincta) is one of three species of honeyeater in New Zealand. Although the other two species are sti l l relatively common, the hihi is extinct throughout its former range with the exception of Little Barrier Island (Buller 1 967)(Figure 2. 1 ) . This severe contraction of range was probably caused by introduced mammalian predators, disease. removal of forests and col lection of birds for museums (Mil ls and Wil l iams 1 979; Rasch 1 99 1 ). The roosting and breeding behaviour of hihi could also have contributed to their extinction. Because hihi roost and nest in cavities they could be easy prey for nocturnal c l imbing predators. During the breeding season both mothers and chicks could be taken on the nest. Hihi roost in groups and this behaviour could faci litate the transmission of diseases and al low nocturnal predators to e l iminate several birds at one time. The O'o (Moho braccatus) from Hawaii is the only other species in the family Mel iphagidae that nests or roosts in cavities and it i s now nearly extinct, presumably for similar reasons ( IUCN 1 979). In addition, in social interactions hihi are subordinate to the other two New Zealand honeyeater species (Angher 1 984; Craig 1 984; Rasch 1 985). The destruction of mainland forests has resulted in a diminished food supply for honeyeaters. I f as a result there was increased competition among the honeyeaters, the subordinate status ofhihi could have been a factor contributing to their extinction. Several studies on Little Barrier Island have looked at the feeding and breeding ecology of hihi (Gravatt 1 970; Angher 1 984; Rasch 1 985). The three studies suggest that the birds feed on three main categories of food, namely fruits, flower nectar and insects. The proportions of these foods in their diet vary throughout the year according 1 3 1 4 to availability. These proportions also vary from year to year depending on p lant phenology. Hihi appear to breed monogamously with a single male and female attending a nesting cavity. Management directed to ensure the survival of hihi has included attempts to translocate the birds to suitable islands with reduced predation from mammalian predators and to establish a captive breeding program. Since 1 980 there have been two transfers to Hen I sland, two to Cuvier Island and five to Kapiti Island. The transfers to Hen and Cuvier I slands failed to establish viable populations (Rasch 1 99 1 ) . The first three transfers to Kapiti I sland ( 1 983, 1 985 and 1 990) were also unsuccessful (Rasch 1 99 1 ) and the last two ( 1 99 1 and 1 992) form part of the present study. After every transfer, the birds bred successfully at least during their first season (Angehr 1 984; Lovegrove 1 986) . South Island /) Hen I s l and t Litt le Barr ier Is land . t 0 1 00 o (\ ? ....,.. Cuvier Is land 300 Km N t 500 Figure 2.1. Map of New Zealand showing the distribution ofh ih i before the 1 800's (shaded area). The species i s now reduced to a population on L ittle Barrier Island. Birds have been transferred to Kapiti, Cuvier and Hen Is lands but no self-sustaining population has been estab lished to date. The suitability of each recipient island was evaluated prior to the transfers. The likely impact of predators, avai labi lity of foods and habitat quality were assessed in comparison with Little Barrier Island (Angehr 1 984). Kapiti, Hen and Cuvier I slands all had more than 1 00 ha of broad leaf forests, considered to be generally suitable habitat for hihi. Throughout the year, Hen Island has a minimum of 1 2 plant species preferred by hihi either in fruit or flower in any month, while Kapiti has nine and Cuvier eight species. By comparison, Little Barrier I sland has 1 5 species (Angehr 1 984). Polynesian rats (Rattus exulans) are present on Hen Island, Cuvier Island and Little Barrier Island. Kapiti I sland supports both Norway (Rattus norvegicus) and Polynesian rats. Angehr ( 1 984), Lovegrove ( 1 986) and Rasch ( 1 99 1 ) suggested that the failure of hihi to establish populations on other islands may be related to food avai labil ity. Rasch ( 1 99 1 ) also suggested that the number of birds transferred prior to 1 99 1 could have been too small, after post-release mortal ity, to reach the threshold population size needed to maintain a social structure. The availability of nest sites has never been assessed as a possible explanation of the failures, probably because the transferred birds bred well during the first couple of seasons. Predation has not been reported for any release prior to 1 99 1 , when a Norway rat kil led a sick bird held in captivity on Kapiti Island (Castro et al. 1 994; Chapter 1 ) . This incident shows that predation by Norway rats could occur, even if only on sick individuals. I n 1 992, Polynesian rat tooth marks were found on the remains of a dead radio-tagged bird on Kapiti Island (B.M. Fitzgerald, pers. comm.). However, the rat may have scavenged, rather than kil led, the bird. The purpose of this Chapter is to describe the feeding and breeding behaviours of hihi transferred to Kapiti Island in 1 99 1 and 1 992 and to compare them to the feeding and breeding behaviours of the hihi population on Little Barrier Is land. Behavioural differences are analysed in relation to differences in vegetation, food and nest site availability between the two islands. 2.2 METHODS Ninety-five hihi (48 males and 47 females) were transferred from Little Barrier Island to Kapiti Island between August 1 99 1 and August 1 992 (Castro et al. 1 994; Chapter I ) . The birds were individual ly marked with coloured and metal leg bands. At the time of the 1 99 1 transfer there were four resident hihi found on Kapiti Is land. Two of them were banded in the week after the 1 99 1 release. A third one had been banded prior to its release in 1 990, and the fourth, a female, disappeared soon after the 1 99 1 release. No other hihi were found surviving on Kapiti I sland (Empson 1 992; Castro 1 992; Castro et al. 1 994; Chapter 1 ) . From September 1 99 1 to August 1 992 and from April to September 1 993, data were collected during 36 field trips (two per month) of 7 to 1 0 days each. On every trip 1 5 1 6 the Kahikatea and Te Rere catchments were searched for birds at least four times. These are the catchments where hihi became established after the 1 99 1 release (Castro et al. 1 994; Chapter 1 ) (Figure 1 .3 , Chapter 1 ) . After the release of birds in August 1 992, some birds established in the Kaiwharawhara catchment, possibly as a result of the release strategy (Castro et al. 1 994; Chapter 1 ) . For this reason, the area of the Kaiwharawhara catchment was included in the areas searched for birds, and it was visited at least one day each trip. From October 1 992 to April 1 993 and from October 1 993 to March 1 994 (the breeding season) I stayed on the island permanently. During this period of time, the Kahikatea and Te Rere catchments each were visited at least three times per week and the Kaiwharawhara once every 2 weeks. Occasionally I checked other catchments to make sure that the birds were not dispersing elsewhere. 2.2 . 1 Feeding Birds were located by sight and/or sound. When possible, an individual was fol lowed unti l its sex, band combination, and feeding activity were recorded. Feeding activities were classified as fruit eating, nectar feeding, and invertebrate feeding, which included both gleaning (picking invertebrates from leaves and bark or perhaps l icking insect and plant exudates) and hawking (catching insects in the air) . Whenever possible specific food items were also identified. Because the number of plants available for hihi at a given time is small and because the birds tend to move from one plant to another within a short period of time the observations were recorded as visits to feed on a plant species, rather than units of food consumed. It was assumed that the number of times birds were sighted feeding on a particular plant represents the true contribution of such a plant (fruits or flowers) to their diet. Similarly, the times birds were observed hawking or gleaning every month was counted. It was assumed that if birds were observed hawking or gleaning more times in a particular month this meant that invertebrates were contributing to a larger extent to their diet. I n other words, it was assumed that the percentage of invertebrates in the diet is proportional to the frequency of observed hawking or gleaning. Data on diet is presented as a percentage of feeding observations. This percentage was obtained by adding all observations of a particular feeding activity for al l known individuals each month and dividing this value by the total number of observations of the three different feeding activities. For the first 4 weeks after each release, food was provided in several feeders at each release site (Castro et al. 1 994; Chapter 1 ). After the first 4 weeks only one of the feeders, the one at Rangatira, was left open and provisioned with food daily throughout the year. Other feeders were provisioned occasionally to attract birds to a mist net or to observe interactions among them. The main feeder was observed every 2 weeks for 1 - 1 6 hours. This feeder was used by bellbirds (Anthornis melanura), one of the two other New Zealand honeyeater species, as wel l as by hihi. Tui (Prosthemadera novaeseelandiae), the third species, were too large to enter the feeder. I recorded the sex and band combinations of hi hi present at the feeder area and noted any interactions between birds. The proportion of birds using the feeder was calculated by dividing the total number of individual birds visiting the feeder each month by the total number of birds known to be alive on the island. 2.2.2 Breeding Once breeding sites were discovered they were visited regularly and birds were observed for intervals ranging from 1 -5 h. During hihi courtship, nest building, incubation and brooding I monitored activity on the territory or at the nest for an average of 6 h per week per nest site. Males within 50 m of the nest were considered to be associated with the nest. When such males called frequently in the vicinity of the nest and entered the nest cavity, they were considered to be the "owners" of the nest. This "ownership" was corroborated by the observation of such male(s) feeding nestl ings . I used the term "polygamy" to refer to nests where two or more birds of one sex were caring for the eggs (females) or young (females and males). True polygan1y, rather than helping behaviour, was established after finding that male hihi have morphological traits only found in polygynandrous species (Castro et al. In press; Chapter 3 ) . During the 1 99 1 breeding season several female hihi showed interest in nesting boxes placed on Kapiti Island for saddlebacks (Philesturnus carunculatus), an endangered New Zealand passerine, and kakariki ( Cyanoramphus novaezelandiae) a resident parakeet species. Two recently released hihi nested that season in saddleback boxes of a design not accepted by saddlebacks. In 1 992, 48 of these nesting boxes were erected on Kapiti I sland, mainly along the Taepiro (5), Te M imi ( 1 0) and Kaiwharawhara ( 1 8) streams, because these catchments have fewer large trees where cavities could be present. The remaining boxes were placed in areas of Rangatira (5), Kahikatea (5) and Te Rere (5) catchments that were covered by small trees or bushes. 2.3 RESULTS 2.3.1 Feeding The percentage of observations of hihi feeding on the three categories of food changed through the year (Figure 2.2). Observations of birds feeding on flowers were most common during spring (46% for the three years combined) and winter (56%), while observations of insect-feeding activities were most common during spring ( 42%) and summer (49%). Fruit feeding was most common during the autumn (37.5%). I n summer the birds were observed feeding mostly on invertebrates (49%), while in autumn the birds fed similarly (invertebrates 37.5%, flowers 25%, fruits 37 .5 %) on each of the three categories of food. In the winter 56% of observations were in flowers 1 7 1 8 (mainly on Dysoxylum spectabile). (,f) c 0 ?.;:::; ro 1 00 80 2: 60 (I) (,f) ..0 0 - 0 1:: 40 (I) u ,_ (I) a.. 20 0 . ?? ?? . ? . . ? . . .. ;. ?-: .: ? -? ?- ? . ?? ?? ... ? ?: ?:? ... ? .. .. . ... ... . ;. ?:: r::. .. . . . .. . ... :. :_-; -:: :.? -: ::. ? 'f_? ?? ?? ? li' ? .. ,. . .. .. . :;:.:-... ?-: -::, ... -: .... . ... ... . .. . .. . . .. . .. . ... ... . .. . : .;. ;.. : ;. ?: : .;. ;.. : ;. ?: .. . . . ... :. ;.;C "? -: .:_ . ... ,.. .. . : ;. ;.. ;. ?: -: ?? j. i? ? ?j???? ? G? j.. ?? ?? c;.: ii-': .. .... -? -.: ... ? S O N D J F M A M J A S O N D J F M A M J J A S O N D J F M A t 92 t93 t 94 Figure 2.2. Percentage of observations of h i h i taking d ifferent types of food on Kapiti Is land from Sept 1 99 1 to Aug 1 994. Dotted areas = flowers; black area fru its; wh ite area invertebrates. The number of observations (n) per month varied from 1 5 to 1 20 with an average of 5 1 .3 observations per month. The use of the main feeder also changed throughout the year (Figure 2 .3) . From April to June birds did not use the feeder but were seen feeding on flowers and fruits. During the spring the use of the feeder was high: up to 1 00% of the birds used it i n November 1 99 1 . 2.3.2 Breeding In 1 99 1 , 1 992, and 1 993, from zero to six males were seen around each nest site during the period of nest building and egg laying (Oct-Nov). At this time there was much chasing of males by other males, and females by males and other females. Later in the season, males and females, other than those feeding chicks, also made short v isits to the cavities. S imilar behaviour has been reported from L ittle Barrier Island (Rasch 1 985) . There is no obvious explanation for this investigative behaviour but it could be an assessment of potential breeding sites for future use. Alternatively, visiting males may be investigating the breeding status of the resident female . There is no record of hihi lay ing more than two clutches in the same breeding season (G. Rasch pers. comm. for Little Barrier Island; P. Morton pers. cornrn. for captive birds; pers. obs. on Kapiti Island). In 1 99 1 , two nests (first clutches) were found during the early stages of nest building. More than one female was present at each nest site and all females participated in the building of the nest platform. However, only one female built the cup. This female was the dominant female and she often chased the other females (not males) and eventually became the one associated with the nest for the rest of the season. In 1 992, three of the seven nests found during early nest building had more than one female (three, three, and four) participating in the building of the nest platform. Again, for these three nests only the dominant female remained in the area. In 1 993, three of 1 3 nests found had two females bui lding the nest. In one case both females incubated while in the other two cases only the dominant female stayed. 1 00 I - - ? - - Feeder ? ? ? I 80 ? Flowers I Vl ft ', c 0 I ? -+-' ? ro > ...... 60 Q) Vl ..0 0 '+- 0 -+-' 40 c Q) u ...... Q) Q.. 20 ., S 0 N D J F M A M J A S 0 N 0 J F M A M J J A S 0 N 0 J F M A ?2 ?3 ?4 Figure 2.3. Percentage of birds using the Rangatira feeder each month (broken l ine), and percentage of observations of hihi feeding on flowers (solid l ine) on Kapiti Island ( 1 99 1 to 1 994). Spring = Sept, Oct, Nov; Summer = Dec, Jan, Feb; Autumn = Mar, Apr, May; Winter = Jun, Jul, Aug. In each nesting season two ( 1 99 1 and 1 992) or three ( 1 993) females built complete or partial nests in boxes. No birds used the new boxes placed in 1 992 until the breeding season of 1 993 when two boxes were used. In 1 99 1 , one female nesting in a box moved to a different box for her second clutch, and the second female abandoned her nest during incubation. In 1 992, one of the two females nesting in a box renested in a nearby box, and the other disappeared during brooding (presumed dead). In 1 993, one female built two nest platforms in boxes but nested in a nearby natural cavity. Two other females built two nests in boxes but only used one each. Females using natural 1 9 20 cavities renested at the same site. With the exception of two birds, all females from the 1 99 1 release nested at the same site as in the previous year. One female moved to a new site in 1 992, but her 1 99 1 nesting area was used by one of her offspring. The other female nested in a box 1 5 m from the cavity used in the 1 99 1 -92 breeding season. Males were also seen at the same sites as the year before. In 1 993, two females moved from their nest site in 1 992 to new sites. The two abandoned sites were used in 1 993 by first-year females, possibly the offspring of the former resident females. t.f) c: 0 -..::; rn c: Q) t.f) ..0 0 Q) Cl rn ...... c: Q) u ..... Q) 0.. 1 00 ?----------------------?----? F l owers 80 60 40 20 0 1 00 ?----------------------------? Fru its 8 0 60 1 00 .-----------------------------? 80 60 40 20 0 I n ve rteb rates Summer Autumn Winter Spring Figure 2.4. Comparison between the feeding preferences ofhihi on Little Barrier Island and Kapiti Island. The white bars represent data collected on Little Barrier Island. From left to right, Gravatt ( 1 970; data for 1 969); Angehr ( 1 984; data for 1 982); and Rasch ( 1 985 ; data for 1 984). The black bar corresponds to the data collected on Kapiti Island between Sept I 99 1 and Aug 1 994. While no polygamous nests were found in 1 99 1 , three were found in 1 992 and three in 1 993 (Table 2 . 1 ). In 1 992, two of the nests were polyandrous (two males attending each nest) and one nest was polygynous (two females sharing the same nest). In 1 993, one was polygynous, one was polyandrous and one polygynandrous (two males and two females feeding fledglings). Table 2. 1 . Hih i breeding results for Kapiti I sland. 1 For 1 99 1 first clutches and 1 993 breeding season, an additional two broods fledged, but the nests were never located. For these nests only the females were considered successful as I never found males feeding the young. 2"Involved" means the number of birds of e ither sex part ic ipating in breeding activities. 3The successful males are those known "owners" of a cavity that successfully fledged young. See text for definition of"owner". 4Successful nests are those which fledged young. 5Nesting males are those males involved in breeding that fed young. 6 No second c lutches found in 1 993 . Year/Clutch 1 991/1 1 991/2 1 992/1 1 992/2 1 993/1 6 Females on Kapiti 1 1 1 1 27 27 20 Males on Kapiti 1 3 1 3 25 25 1 6 Nests found 5 1 4 1 5 7 1 31 Females involved2 (%) 9 (82) 5 (45) 20 (74) 7 (26) 1 9 (95) Males involved2 (%) 1 1 (85) 8 (61 ) 1 8 (72) 1 2 (40) 1 5 (94) Females nesting (%) 7 (64) 4 (36) 1 6 (59) 7 (26) 1 4 (80) Males nest ings (%) 6 (46) 4 (31 ) 1 4 (56) 0 (0) 1 3 (8 1 ) Successful females (%) 6 (86) 1 (25) 1 1 (69) 1 ( 1 4) 1 0 ( 1 2) Successfu l males3 (%) 4 (67) 1 (25) 1 2 (86) 0 (0) 8 (61 ) Successful nests4 (%) 6 1 (86) 1 (25) 1 0 (67) 1 ( 1 4) 1 01 (77) Polygynous nests 0 0 0 Polyandrous nests 0 0 2 0 2.4 DISCUSSION 2.4. 1 Feeding Although variations in the seasonal use of different food items from year to year is expected, there is a consistent difference in autumn time feeding patterns between Little Barrier I sland and Kapiti Island (Figure 2.4). In autumn, hihi on Kapiti I sland fed mainly on invertebrates while hihi on Little Barrier I sland fed mostly on fruits. Fruits and flowers seem to be the greatest contributors to feeding observations on Little Barrier 1.. On Kapiti I. the greatest percentage of feeding observations are on invertebrates and flowers. Lovegrove ( 1 986) also found insects to be an important food category for hihi throughout the year on Kapiti I sland. The higher consumption of invertebrates on Kapiti I sland during autumn can be 2 1 22 explained in several ways. The greater percentage of observations of birds feeding on invertebrates could be the result of a shift in diet due to shortage of fruits or flowers and the competition for this resource with the other honeyeaters. There was a weak negative correlation (r = 0.6 1 ) between the percent of observations in fruits and invertebrates (Chapter 5 ) . Alternatively, it could be that the birds increased their consumption of invertebrates because the nutritional quality of the insects present on Kapiti was relatively low, and therefore they needed to consume many to obtain the minimum amount of a l imited nutrient. Support for the first option comes from data from a phenology transect run on Kapiti from April 1 992 to August 1 994 (Chapter 5) . During the year there were fewer species of plants in flower or fruit on Kapiti I. when compared to Little Barrier I . . During the spring the species in flower were plants with a restricted distribution on the island or the tree stands were young (Chapter 5) . Because hihi are at the bottom of the honeyeater dominance hierarchy, competition for this limited food resource could affect hihi more than the other honeyeaters. In autumn hihi seem to have shifted to invertebrates to compensate for a lack of fruits. In contrast, during the spring hihi increased their use of the feeder, a place where they could obtain nectar without having to compete with other species except for bellbirds. Thus, the large proportion of observations of birds feeding on invertebrates could be an overestimate because I am not considering the use of the sugar water as a substitute for nectar feeding. In conclusion, seasonal scarcity of food could be hampering the successful establishment of hihi on Kapiti Island. Competition with tui and bellbirds probably contributes to this problem. This may be particularly important during spring when the birds are preparing for breeding. Data on the availability and abundance of the different foods and the feeding of the two other species ofhoneyeaters should be collected to test these hypotheses more rigorously. Meanwhile, the use of a feeder during the spring, when nectar is scarce on Kapiti Island is recommended. 2.4.2 Breeding I believe there are two significant differences between the breeding of hihi on Little Barrier Island and Kapiti Island. F irstly, on Little Barrier Island breeding is apparently monogamous (Angehr 1 984; Rasch 1 985) whi le on Kapiti Island polygamy is observed. Secondly, hihi on L ittle Barrier Island tended not to reuse the same nest cavity for second clutches (Angehr 1 984; Rasch 1 985) , while on Kapiti I sland females used the same nests for first and second broods in 1 99 1 and again in the 1 992 breeding season (there were no attempts to renest in 1 993). Both of these differences are probably the result of intense competition for nest sites which could be reflecting l imited availability of suitable cavities on Kapiti Island. Kapiti holds healthy populations of three other cavity nesters: kakariki, saddlebacks and kaka (Nestor meridionalis) . The New Zealand kingfisher (Halcyon sancta) is also present but in small numbers. Of these cavity nesters kakariki and kingfishers are known to use the same cavities as hihi on Little Barrier Island, however there are no records of either species interfering with hihi nesting there. On Kapiti kakariki have been observed fighting for cavities with hihi every year. In 1 993, competition for cavities was very high: in two of 1 3 nests known in 1 993 hihi nesting was hampered by the constant fights for the nest cavity. In seven of the remaining 1 3 nests kakariki nested on hihi cavities two days after hihi chicks had fledged. This competition for nesting cavities could be the result of a shortage of nesting cavities for hihi on Kapiti Island or could result from either kakariki or hihi having a behavioural incapacity to find suitable nest sites. Hihi have found suitable nest sites after each of the five transfers, indicating that the species has the behavioural capacity to locate sites. However, it is possible that hihi are constrained behaviourally when choosing nesting sites. Van Balen et al. ( 1 982) studied the availabil ity and occupancy of natural nest sites by various species of passerines. They reported that different species have specific cavity preferences (e .g. entrance size and shape, depth and height of cavity, height of cavity from the ground). Hihi preferences for nest sites are not fully understood. While Kapiti I sland may have many cavities, the number of cavities that actually fit hihi requirements could be very small . Recently released birds and juveniles probably have difficulty locating a suitable nest site . I t should be even harder for birds released just before the breeding season. This is supported by the fact that birds released in 1 992 reproduced later than birds already established from 1 99 1 . The possibi lity of a shortage of suitable nests sites should be considered. Kapiti Island forests are not older than 1 00 years (Fuller 1 985). The propensity of trees to form natural cavities increases with age and varies from species to species. I n the young forests on Kapiti I sland the majority of trees of species that eventual ly produce hollows are not of cavity-forming age. I n addition, tree cavities are not permanent and deteriorate over time. For example, Sedgwick and Knopf ( 1 992) found that 46% of cavities found in a cottonwood bottomland in Colorado were no longer useable after 4 years. Because there are no primary cavity nesters on the island, cavities are only produced by the natural decay of particular tree species. It is l ikely that suitable nest sites on Kapiti Island are so few and hard to fmd or defend thus it is better for a male or a female to share a nest cavity than not to nest at al l . According to Emlen and Oring ( 1 977), some species can change their breeding strategy depending on the environment. This possibility should be studied for hihi. There are insufficient data to test the difference in fledging success between the polygamous nests and monogamous nests. Although males may frequently engage in extra-pair copulations, feeding of young is l ikely to increase their long-term reproductive success and therefore encourage them to co-operate with other males 23 24 rather than depend totally on extra-pair copulations to achieve reproductive success. An immediate way to overcome nest site limitation in this case is to provide nest boxes. However, more permanent measures would require gathering information about the availabil ity of suitable cavities, and cavity turnover on Kapiti Island, with the objective of planting of species that wil l eventually develop hollows. Erecting nest boxes does not guarantee that the birds will use them. The ability of the box to exclude competitors and predators, its particular orientation, height above the ground, internal shape, and the familiarity of the birds with its presence are important factors to be considered if boxes are to be used (Snyder 1 978 a, b; Zeleny 1 978). It is possible that hihi did not use the nesting boxes provided in 1 992 because the boxes were not located in positions or places that were attractive to hihi or because the boxes have not been in the area long enough. Experiments varying the height of the boxes from the ground are under way, but it is possible that other variables should be also tested. 2.5 REFERENCES Angehr, G.R. 1 984. Ecology and behaviour of the stitchbird. Internal report to the New Zealand Wildlife Service, Wellington. Buller, W.L. 1 967. Stitchbirds. Pp 55-59. In 'Buller's birds ofNew Zealand' ed by E.G. Turbott. Whitecomb and Tomb Ltd: Christchurch. Castro, I. 1 992. Stitchbird transfer: Little Barrier Island to Kapiti Island, August 1 99 1 . Internal report to the New Zealand Department of Conservation, Wellington. Castro, I . , Alley, J .C. , Empson, R.A. and Minot, E.O. 1 994. Translocation of hihi or stitchbird (Notiomystis cincta) to Kapiti Island, New Zealand: transfer techniques and comparison of release strategies. In Reintroduction Biology of Australian and New Zealand fauna. Melody Serena Ed. Surrey Beatty and sons. Chipping Norton. Australia. Castro, I . , Minot, E.O. , Fordham, R.A., and Birkhead, T.R. In press. Polygynandry, face-to-face copulation and sperm competition in the hihi Notiomystis cincta (A ves: Meliphagidae ) . Ibis Craig, J .L . 1 984. Wing noises, wing slots and aggression in honeyeaters (Aves: Meliphagidae). New Zealand Journal of Zoology. 1 1 : 1 95-200. Emlen, S .T. and Oring, L.W .. 1 977. Ecology, sexual selection and the evolution of mating systems. Science 197: 2 1 5-222. Empson, R. 1 992. Report on the release of stitchbirds on Kapiti Island, August to September 1 99 1 . I nternal report to the New Zealand Department of Conservation, Wellington. Fuller, S?.A. 1 985 . Kapiti I sland vegetation: a report on a vegetation survey ofKapiti Island 1 984/ 1 985 . Department of Lands and Survey: Wellington. Gravatt, D.J. 1 970. Honeyeater movements and the flowering cycle of vegetation on L ittle Barrier Island. Notornis 17: 96- 1 0 I . International Union for the Conservation of Nature and Natural Resources (IUCN). 1 979. Red data book. Vol 2 Aves. I UCN: Morges, Switzerland. Lovegrove, T. 1 986. Stitchbirds on Kapiti Island 1 983- 1 985. I nternal report to the New Zealand Department of Conservation, Well ington. M il ls, J .A. and Will iams, G.R .. 1 979. The status of endangered New Zealand birds. Pp. 1 47- 1 68. 1n Proceedings of the Royal Society of South Australia: The Status of Endangered Australasian Wildlife ed by M.J . Tyler. Graphics Services Pty . Ltd . : Northfield, South Australia. Rasch, G. 1 985 . The behavioural ecology and management of the stitchbird. M Sc. Thesis, University of Auckland. Rasch, G. 1 99 1 . A recovery plan for the stitchbird (Notyomistis cincta). Internal report to the New Zealand Department of Conservation, Wellington. Sedgwick, J .A. and Knopf, F .L. 1 992. Cavity turnover and equil ibrium cavity densities in a cottonwood bottomland. Journal of Wildlife Management 56: 477-484. Snyder, N.F .R. 1 978a. Increasing reproductive effort and success by reduci ng nest-site l imitations: a review. Pp 27-3 3 . In Endangered Birds: Management Techniques for Preserving Threatened Species ed by S.A. Temple. The University of Wisconsin Press. Madison and Croom Helm Limited, London. Snyder, N.F.R. 1 978b. Puerto Rican parrots and nest-site scarcity. Pp 47-5 3 . In Endangered B irds: Management Techniques for Preserving Threatened Species ed by S.A. Temple. The University of Wisconsin Press. Madison and Croom Helm Limited, London. Van Balen, J .H . , Booy, C.J .H. , Van Franeker, J .A. and Osiek, E.R. 1 982. Studies on hole-nesting birds in natural nest sites: 1 . Availabi l ity and occupation of natural sites. Ardea 70: 1 -24. Zeleny, L. 1 978. Nesting box programs for bluebirds and other passerines. Pp 5 5-60. I n Endangered Birds: Management Techniques for Preserving Threatened Species ed by S .A. Temple . The University of Wisconsin Press. Madison and Croom Helm Limited, London. 25 Chapter 3 Chapter 3 Polygynandry, face-to-face copulation and sperm competition in the Hihi Notiomystis cincta (Aves: Meliphagidae) ABSTRACT The H ihi Notiomystis cincta breeding system is highly variable, including monogamy, polyandry, polygyny and polygynandry. Males have large testes (4 .2% of body mass), large numbers of sperm and an enlarged cloacal protuberance. These features are found in other species with highly variable mating systems where males are under intense sperm competition. Hihi mate in two different positions: face-to-face, and male on-the-female's-back. This face-to-face copulation is unique among birds and appears to be a form of forced copulation. The very large cloacal protuberance in males, and the cloacal protuberance in females could aid the transfer of sperm. Both male and female H ihi benefit from a mixed reproductive strategy (Trivers 1 972) where a female Hihi can sol icit copulations from males other than her partner and male Hihi can perform extra-pair copulations both with wi ll ing females or by forced copulation. 3.1 INTRODUCTION ALTHOUGH the majority of bird species are social ly monogamous, extra-pair copulations and sperm competition are frequent and result in extra-pair paternity (B irkhead & M0ller 1 992 ; Birkhead 1 994). A number of male morphological correlates of sperm competition have been recorded in birds. In species where sperm competition is especially intense, such as those with polyandrous or polygynandrous mating systems, testes are relatively large and, in passerine birds, the male sperm store (seminal glomera) and cloacal protuberance are also relatively large (B irkhead et al. 1 993; Briskie 1 993 ) . These features have been recorded in passerine birds such as the Dunnock Prunella modularis (Davies 1 98 3 ; B irkhead et al 1 99 1 ; Davies 1 992), Alpine Accentor Prunella collaris (Nakamura 1 990), Smith's Longspur Calcarius pictus (Briskie 1 992, 1 9.93 ), Superb Fairy-wren Malurus cyaneus (Mulder & Cockburn 1 993) and Aquatic Warbler Acrocephalus paludicola (Schulze-Hagen, et al. 1 995) which all have highly variable mating systems as wel l as large testes and large c loacal protuberances. Large testes and seminal glomera are probably essential to sustain high copulation rates, anc:!Jor to produce relatively large ejaculates, both of which play a central role in paternity assurance in birds (Birkhead & M01ler 1 992) and other taxa. In this study I describe the breeding system, male morphology and unique copulation behaviour of another passerine, the Hihi or Stitchbird Notiomystis cincta (Meliphagidae ), in which sperm competition is also intense. 26 27 3.2 METHODS 3.2 . 1 Study site This study took place on Kapiti I s land Nature Reserve, on the west coast ofNew Zealand' s North Island during the 1 99 1 to 1 994 breeding seasons. The island was largely cleared for agriculture in the 1 800's but by 1 93 0 all browsing animals with the exception of Brush-tailed Possums Trichosurus vulpecula were eradicated . Kapiti I s . was freed o f possums i n 1 986 (Cowan 1 992). The forests o n the is land are divided into scrub and low forests (approx. 1 04 1 ha), and tall forests (rata forest; tawa-hinau-kamahi forest; and kohekohe forest, approx. 550 ha) (Fuller 1 985) . The remainder of the island is covered by low scrub, grass, tussock and flax. With the exception of Norway Rat Rattus norvegicus and Polynesian Rat R. exulans, Kapiti Is . is free of mammalian predators. 3.2.2 Study bird The Hihi i s one of three honeyeaters endemic to New Zealand. The species original ly inhabited the whole of North Island and its offshore is lands. It is now restricted to Little Barrier Island in the northern part of its original range. Translocations to various islands within the former range of the species with the aim of establishing further self-sustaining populations have taken place s ince 1 98 3 . In 1 99 1 and 1 992, there were two translocations of Hihi to Kapiti Is . (Castro et al 1 994a) where this study took place. The Hihi is a forest species, nesting in tree cavities during the southern hemisphere summer. Nesting activity starts in October and lasts until April . Females usually produce two clutches, each of 2-5 eggs, which are incubated by the female alone. The male helps with chick feeding. There is marked sexual dimorphism. Males have more striking, colourful plumage and are larger in size (at 40g, males are 3 0% heavier than females 3 0g; Craig et al. 1 982 ; Castro et al. 1 994b ). Females are subordinate to males in feeding areas (Craig et al. 1 982) . At nest sites on Little Barrier Island, females were dominant over males (Angher 1 984; Rasch 1 985a). On Kapiti Is . , however, females were subordinate, although they sometimes displayed aggressively to their mates. On Little Barrier Is . a l l nests found to date have been attended only by a single male and a single female although more than one male has been observed visiting several different nest sites during the breeding season (Angher 1 984; Rasch 1 985a) . 3.2.3 Copulation behaviour of Hihi The only published description of copulation is for captive Hihi (Anderson 1 993) . All of the copulations seen in this situation were in a face-to-face pos it ion that i s unique in birds. These observations left open the question of whether the behaviour was an aberration of captive birds or is normal for the species. Hihi numbers on Kapiti I s. since 1 99 1 have ranged from 25 to 79 with an average of 40 birds. Sex ratios have stabil ised at a 1 : 1 level . At any time during the period of study at least 90% of the hihi on the island were colour ringed for individual identification. Furthermore, some males could be identified by feather pattern or cal l . Observations of Hi hi breeding behaviour started each year in October and continued through the breeding season. As many nest sites as possible were located and monitored. Each nest was observed for at least six hours every two weeks. Nest observation periods ranged from one to five hours. Data on male and female visits to the nest ( incubation and brooding times), interactions between all conspecifics present during the watches and between Hihi and other species, and fledgling success were recorded. Since most of the study birds used inaccessible cavities, fledging success was measured as the number of chicks leaving the nest. Several feeding stations were established on Kapiti Is . to provide Hihi with supplementary food for four weeks after release (Castro et al. 1 994a). After these four weeks, a single feeder was retained. This feeder was observed for at least one hour per day throughout the breeding season to assess bird survival and to observe matings. Birds were mistnetted at this location, where their weight, general condition, brood patch development and presence of cloacal protuberances were recorded. In 1 994, the diameter and height of the cloacal protuberances were also measured. The testes of three males that died during the breeding season on Kapiti Is . were measured ( length and breadth) and their mass calculated using M0l ler's ( 1 99 1 ) equation. The cloacal protuberance from one o f the males was examined by T.R. Birkhead (The University, Sheffield, England) and the seminal glomera and spermatozoa described and counted following B irkhead et al. ( 1 99 1 ) . 3.3 RESULTS 3.3. 1 Mating system The Hihi mating system included social monogamy (22 of 3 3 nests or 67%), polyandry (3 or 9%), polygyny (4 or 1 2%) and polygynandry ( 1 or 3%). Four nests ( 1 2%) fai led during incubation or early brooding before the mating system was determined. Mating systems were assigned to each nest according to behaviour during i .'1cubation and brooding. Nests were classified as monogamous where a single male and a female or only a female alone, cared for the young. Polygyny included nests where two females shared the same cavity or where a male attended the nests of two different females . A nest would be classed as polyandrous if a single female p lus more than one male fed the chicks. Although this could be interpreted as helping behaviour, Hihi have physical characteristics associated with variable mating systems (see below) . Furthermore, males in polyandrous nests were present at the nest site, and call ed there 28 2 9 0 during nest building and egg laying. Some males guarded their females before and during egg laying by fol lowing them closely, but the presence of many males at a nest site during the female's ferti le period suggests a high degree of promiscuity . On Kapiti I s . , Hihi established home ranges in three catchments. Each single stream catchment was shared by a group of birds with overlapping home ranges (Figure 3 . 1 ) . Their nest sites were located within the home range. Hihi did not defend a territory, but males called within 3 0 m radius of their nesting tree (nest site) and behaved aggressively towards any other male within thi s area. Female hihi were also protective of the cavity itself, defending it against other female Hihi and other species. 200 m : s I ; s 5 \ 500 \ \ Figure 3. 1. Map showing male h ih i home ranges. The dots represent nest cavities and the circles around them a l 00 m diameter area where the females were found feeding during the breeding season. Males called within 30 m of the cavity. A l l birds v isited the feeder regularly, but unless a male was cal l ing or copulating at a nest s i te near the feeder, the area was not considered part of his home range. In the polyandrous and polygynous nests there was a dominant male and/or a dominant female, which was aggressive towards the other bird of the same sex. The minimum distance between nest sites was 1 1 0 m. and the largest distance was 1 620 m. From October, males called within nest sites. A male (s) calling consistently from an area with a nesting cavity was considered the owner (s) of that site. The number of calls per hour at a nest site changed during the season, being greatest a few days before the first egg was laid (Figure 3 .2) . Although males spent a great deal of time calling at their nest site, they also visited other nesting sites. Between nest building and egg laying, up to six males were seen visiting a particular nest site (Table 3 . I) . Calling at the nest site and visits to other sites continued until a clutch was completed at each site. After clutch completion male visits to that nest were less frequent. Cal ling virtually ceased during incubation and chick feeding (Figure 3 .2). ' '- ? 1 40 1 20 (f) 1 00 CJ) c 80 60 40 20 0 Nest site ----- Station 5 - - E - - lsabel la's ? 0 ? - Loveg rove - - - -T- - - - T43 -25 -20 - 1 5 - 1 0 0 I I 'J' : I . . ? . , ' ,? - . . I 1 ? - ,? -5 I ' I ' 0 5 1 0 1 5 Days from clutch initiation (0 = day of fi rst egg) Figure 3.2. Number of cal ls per hour at h ih i nest sites . Data comes from observations at four different nest sites. T43 and Lovegrove were po lyandrous nests whi le Jsabel la and Station 5 were social ly monogamous. 30 3 1 Table 3. 1. Comparison of male presence at nest sites during nest bu i lding and laying with male nest attendance during chick rearing. Only nests found during nest bui lding are included. Nest name Matai OH 56 (92-93) OH 56 (93?94) Big Rata 1 /4 Way Station 5 (92- 93) Station 5 (93- 94) T43 Box 3 Box 4 Lovegrove (93, First clutch) Lovegrove (93, Second clutch) 40/43 (92-93) 40/43 (93-94) lsabella's (92- 93) lsabella's (93- 94) Gretta's (92- 93) Gretta's (93? 94) R2 (92-93 A) R2 (92-93 A) R2 (93-94) Males present at n.b.'and laying. BA-R WA?YY G-RA A-BW A-WR WW-A WA-YY AB-Y A-WR BA-R AR-G WA-YY AB?Y G-RA R-WA AIWR Unb. male ? AB-Y R-WA AR-G A-WR AB-Y R-WA AR-G A-WR AR-G A-WR GW-A WW-A GW-A YG-A AB-BY GG-WA WW-A WA-BB A B-Y GW-A YG-A WW-A WA?BB AB-BY YG-A GG-WA AR-BR YG-A AR-BR A-BW A-WR AB-Y A-SW YG-A A-RR GG-WA GG-WA AB-BY AW-RR GG-WA A-GY GG-WA Unb. male ?? A-RR AB-BY Unb. male ?? AB-BY Unb. male ?? AB-BY ? Recognized by feather patern Female (s) r and B' A-WW BA-B AB-BB GY-A A-YY A-WY WA-RR WA-RR YR-A WA-W WA-W Y-A Y-A A-GG Unb. female RW-A Unb. female GR-A GR-A AY-YR W-BA A-RB A W-BA AG-GR ?? Recognized by call. Same male for all nests. Male (s) feeding chicks Failed None Failed. None A-WR Unb. male Failed R-WA GW-A AR-G YG-A Failed YG-A AR-BR None A-BW A-SW GG?WA GG-WA AW-R R A-GY Failed Unb. male .. Unb. male ?? AB-BY " This female dissapeared during incubation, might have died. ? Incubation ' Brooding ? Nest building Number of young fled ed 0 0 2 0 2 2 0 0 2 ? 2 2 3 2 2 0 Females took up to two weeks to build their complex nest, including a platform made of sticks, which in one case was 40 cm high. Nest cups were l ined with feathers and the scales ofponga ferns (Dicksonia and Cyathea spp.). The average period between nest completion and the start of egg laying was 9.6 days (range 1 -23 ; s.e. = 3 ; n = 7). During this period, and again during egg-laying, copulations were observed. For instance at Nest 1 in 1 992, the first copulation was observed on 24 November and the first egg on 29 November. In 1 993, the first copulation was on 1 December and the first egg on 5 December. However, copulations at the feeding area in 1 993 were first observed on the 2 December whi le most egg laying took place from 1 1 December to 1 9 December. Copulations were seen infrequently and were the result of observing nest sites rather than individual females. I observed copulations mainly in the morning and in the late afternoon. Breeding synchrony was low and the degree of asynchrony varied within years (egg laying for first clutches in 1 992 lasted from I 6 November to 9 December while in 1 993 laying started on 5 December and finished around 1 9 December) . Female Hihi laid one egg each day until the clutch was completed. Laying took place in the morning but 1 h (s.d. = 0.6) later each day. First eggs were laid around 0630 to 0700 h, and last eggs were laid around 1 I 00 or I 200 h. Female Hihi remained in the cavity for an average of 1 .4 h (s.d . = 0.4; n= 1 2 . Data relate to 9 different females, 2 from Kapiti Is. and 7 from Mokoia Is . ) whi le laying. Incubation started the day the last egg was laid. Females can raise and successfully fledge young alone (Table 3 . 1 ) . Hihi males provide parental care by feeding chicks. For socially monogamous pairs on Kapiti Is. the male contribution was about one third that of the female (Castro, in prep). In I 99 1 192 there were on average 2 .3 fledglings per nest, 1 . 1 3 fledglings per nest in 1 992- 93, and 1 .0 fledgling per nest in 1 993-94. Second clutches on Little Barrier Is. are usual ly successful (Angher 1 984 ). On Kapiti Is. , however, of four nests where second clutches were attempted in 1 99 1 -92, only one was successful. In 1 992-93 , of seven attempted second clutches only one was successful . In 1 993/94 there were no attempts to re-nest (Castro el al. 1 994b ) . 3.3.2 Reproductive anatomy 3.3.2 . 1 Male Cloacal protuberance (F igures 3.3 A and B) In males the cloacal protuberance started to develop in early September 1 994, and acquire its maximum size in early November. Average cloacal protuberance dimensions (height X width ? s.e.) at their maximum size for each of n ine males sampled was 1 2 .6 (? 0.28) X 1 1 .2 (? 0.26). 3.3.2.2 Testis mass, seminal glomera mass, and spermatozoa numbers The mean l inear dimensions of testes from three males found dead during the breeding season were used to estimate the mass of the two testes combined. The mean linear dimensions of left and right testes (length x breadth), respectively measured 1 0 X 1 5 mm, 1 1 X 1 3 mm, and 1 0 X 1 4 mm, and the testes were estimated to have a combined mean mass of 1 . 726 g. With an adult male body mass of about 4 1 g the testes 32 33 represents 4.2% o f adult mass. Using M0ller's ( 1 99 1 ) equation for the relationship between body mass and testes mass across birds in general, the observed mass of Hihi testes is three times greater than that predicted (0.52 g) from body mass alone. The cloacal protuberance and seminal glomera of a single specimen were dissected and weighed. The seminal glomera were relatively large ( left: 0. 1 499 g, right: 0. 1 5 1 8 g, total mass = 0.30 1 7 g). However, in contrast to other species examined (e.g. the Dunnock Prune !la modularis, Birkhead et al. 1 99 1 ) the seminal glomera did not form the main part of the cloacal protuberance (see above). Instead, the cloacal protuberance contained relatively loosely coiled tubes of the seminal glomera, the main part of which was a cigar-shaped structure al igned antero-posteriorly (Figure 3 .4). The seminal glomera contained very large numbers of spermatozoa ( left: 856 x 1 06 and right 606 x 1 06, total : 1 462 x 1 06) . These values underestimate the true total since the right seminal glomera was damaged and a small part had been removed before I examined it. Assuming that the intact right seminal glomera contained the same number of spermatozoa as the left, then the total number of spermatozoa would have been c 1 700 X 1 06. Figure 3.3. Male h ih i cloacal protuberance. (A) V iew from the top; (B) V iew from the s ide (Drawings by Elizabeth (Liz) Grant). V a s deferens Semina l g lomera I Cloacal P'otu b e , a o c e ? feat h er s a t e n t r a n c e to c l o a c a Figure 3.4. Cross-section of male h ih i c loacal protuberance (drawing by T .R . B irkhead). 3.3.2 .3 Female cloaca (F igure 3.5) The size of the cloaca varied between pre-laying and laying periods. Average cloaca size for pre-laying females (? s.d.) was 4.9 1 (?0.2 1 ) mm in breadth X 4.49(?0.24) mm in height (n = 6) . In laying females the cloaca was distended to an average breadth of 8 .07(?0.47) mm by 5 . 1 (?0 .76) mm in height. Figure 3.5: View of female hih i c loacal protuberance (Drawing by Liz Grant) 34 35 3.3.3 Copulation behaviour of Hihi Wild Hihi copulate in two radical ly different positions: face-to-face, and male? on-the female's back. In both positions the male was on top ofthe female for six to eight seconds. Castro (pers. obs. , Mokoia Island) has observed pairs performing multiple copulations in succession. Each mounting was separated from the next by a display where the male and female rubbed their necks and touched bills while vibrating their wings. The whole process of display and copulation lasted 7- 1 5 seconds. 3.3.3.1 Face-to-face Copulation (Figures 3.6 B, C, and D) When a s ingle male was present, the female was chased to the ground by the male and held on her back with her wings outspread. During copulation, the male spread his wings, or pointed his bil l straight up and extended his wings slightly. The female made calls similar to the distress calls (rapid repeated "pew") produced when females are chased by other birds, or when a female tries to enter a nest other than her own. Male vocal isations (low volume, short, high pitched squeaky sounds) resembled those produced when a male is chased by conspecific males during the breeding season, or when a male enters a nest that is not his own. After face-to-face copulation the female and male flew away separately. Most face-to-face matings, however, started with up to five males chasing one female to the ground. I observed an average of four attempted matings per hour in 7.3 hours of observation (maximum 8 attempts in I h; minimum 3 attempts in 1 .2 h. ) . Attempted copulations were those chases where I could not see the end result because the birds were too far away or visibil ity was impeded by vegetation. One male would start the chase and the others would join in. The female always called the "distress call" described above for face-to-face copulations. This call ing continued until the end of the chase or until the birds were on the ground. This call seemed to attract other males to the area. Once forced to the ground by the male, the females tried to fly, or kicked the male, as if attempting to escape. Often the females would be forced to the ground several times before being mated. In the chase, all males flew in l ine after the female and only the first male would "grasp" the female once she was on the ground. Al l other males then remained close to the pair watching from a distance of 1 0 cm to 1 m but not interfering. In all cases the female was upside down with the male on top of her (face? to-face). Once mating was over the male hopped away while the female rol led over and stood for a few seconds before flying away. 3.3.3.2 Copulation with the male on the female's back (Figure 3.6 A) In the male-on-the-female's-back copulations both birds went to the ground. In one of four instances the male and female made a neck-rubbing display including bi l l? touching before he c limbed on her back. There were no vocalisations from the female and, on one occasion, the male produced a repetitive "pew" call very quietly . After copulation the female either went back to her nest box, or preened, or flew away with the male. On one occasion the male and female performed the neck-rub bill-touching display after mating. For one copulation of this type the male landed on the female's back while she stood on a branch about 50 cm from the ground. His wings were outstretched, apparently for balance, and both birds were silent until another male arrived and landed on the mating male's back. At that point the female gave a distress cal l similar to the cal l produced during the face-to-face copulations. The birds did not separate immediately. On another occasion a male chased a female to the ground. The female produced the same distress cal ls heard during face-to-face copulations. At first the female lay on her back with the male on top, but she soon turned herself to an upright position, the male climbed up and mated in a male-on-female's-back copulation. Vocalisations stopped during this mating position. The affiliation of this male to the female was unknown. B D Figure 3.6. H ihi mating postures. (A) Copulation with the male on the female's back. (8, C, & D) Face? to-face copulation (Drawings by Liz Grant). 3.3.3.3 Copulation sites Hihi copulations occurred on the ground near nest sites or at feeding areas. At nest sites copulations occurred during the pre-laying and egg-laying periods. For instance, at site 1 the female completed two clutches in 1 992-93 . Six copulations were clearly observed during 25 h of observation (0.24 copulations per hour). These involved the female and each of three males. Only one of the males later helped in rearing the 36 3 7 chicks and therefore was considered her partner. O f these copulations two were face-to? face (involving her partner and another of the males) and four male-on-the-female's? back copulations (involving all three males). All three males were often seen at the nest site calling and chasing each other although her partner and another of the males had other nests of their own. On two occasions copulations were observed within 1 0 minutes of the female l eaving the box after laying an egg. On one of these occasions the partner male copulated with the female in the male-on-the-female's-back position and briefly left the area. When he came back he found another of the males in the area and, after chasing him away, forced the female out of the cavity and copulated in the face-to? face position with her. At site 2, males waited outside the cavity where presumably, the female was laying (3 observations). The female was chased to the ground when she left the cavity but observation of copulation was prevented by vegetation. 3.3.3.4 Copulation at the feeder During the observations at the feeding station, 22 birds ( 1 1 females and 1 1 males) used the feeder. I observed 3 1 mating attempts. All males and at least six females were involved with one or more of these attempts. Five of the 3 1 attempts were observed in detai l . Of the five females known to have mated face-to-face only one was regularly in the company of a single male . The other females were apparently unpaired. The large number of copulations observed at the feeders could be the result of having an unusual concentration of birds there . However, a similar situation, of smaller scale , could happen at nest sites when a female is fertile (Table 3 . 1 ) . 3.4 DISCUSSION 3.4. 1 Mating system The hihi breeding system and associated male reproductive structures are simil ar to those found in other polygynandrous species (Table 3 .2) . Based on its breeding system, the mating rate for the species should be very high. However, observations of matings are difficult to make because hihi l ive in dense forest located in steep terrain. More observations of hihi matings are required to analyse the role of large testes and the large cloacal protuberance in this species. Copulations at nest sites started five days before the female laid the first egg and continued until the last egg was laid. This suggests that the female fertile period is about 9 days. However, copulations at the feeding area started up to 1 7 days before laying, suggesting that the females might be able to store sperm for longer. The shorter period of copulations at the nest area could reflect the time spent at the nest site by the female, rather than her ferti lity status. Copulations were observed during the insemination window (ie . j ust after the female had l aid an egg) suggesting intense sperm competition in this species (Cheng 1 983, B irkhead & M0ller 1 993). Males would benefit from copulating with females at the feeding areas, but those males able to copulate with females at the nest site, and particularly during the insemination window, would be most l ikely to achieve paternity. DNA fingerprinting studies could be useful to identify the role of copulations during the insemination window for this species. Table 3.2. Review of infonnation about testes and c loacal protuberance sizes and number of sperm in seminal glomera of polygynandrous species. Data taken from Nakamura 1 990, Mulder &Cock burn 1 993, Briskie 1 993, Schulze-Hagen 1 995, Birkhead et. al. I 99 1 , and this study. Species A lpine Accentor Superb Fairy-wren Smith's Longspur Aquatic Warbler* Dunnock Hi h i Testes size 7.75 4 .9 4.2 4.1 3.36 4.2 Cloacal protuberance d imensions Number of sperm (breadth X height) in mm (? s.e. ) . in g lomera 1 4. 1 8 (? 0 .31 ) X 1 2 .78 (? 0.49) Not available Not available Not available 1 1 .49 ? (0. 1 8) X 9 .89 (? 0 .22) 2 1 7 X 1 06 6. 7 4?(0. 92)X7. 76(?0. 73)X5.62(?0.88) 1 97.8 X 1 06 8.50 (? 0.39) X 6.50 (? 0.3) 1 060 X 1 06 1 2.6 (? 0 .28) X 1 1 .2 (? 0.26) 1 462 X 1 06 * The measurements for this species incl udes breadth front to back, and breadth side to side, by height. Although a proportion of males guard their partner by close fol lowing during the breeding season, most male hihi do not spend much time with one female during the period from pre-laying to the end of incubation. Often the identity of the male partner (s) is not known unti l the chicks hatch and he (they) begins feeding them. 3.4.2 Copulation behaviour and reproductive anatomy For many species it is difficult to determine if a particular type of copulation is forced. Because of the contexts in which the two types of copulation occur I believe that face-to-face copulation is a form of forced copulation. Forced copulation in other species is also accompanied by distress calls and is often violent (reviewed in B irkhead & M0ller 1 992). In Bearded Tits Panurus biarmicus (Birkhead & Hoi 1 994) and White? fronted Bee-eaters Merops bullockoides (Emlem & Wrege 1 986), up to 1 2 males have been reported chasing the females to the ground to copulate with them in much the same way as in hihi. However, whi le several male White-fronted Bee-eater might force copulations on the female, only one male does in Bearded Tits and hihi . When mating face-to-face, hihi attract the attention of the observer and potential predators by making considerable noise and movement. Weka Gallirallus australis, opportunistic ral l id predators, often approached the calling birds when they were on, or close to, the ground. Because face-to-face copulation is so overt, it may be observed out of 38 3 9 proportion t o male-on-female's-back copulation' s . B i rkhead & M0ller ( 1 992) review the benefits and costs of forced copulations and extra-pair copulations. There are several possible explanations for face-to-face copulations in hihi . An interesting possibi lity is that females might test a male's quality through his ability to perform a face-to-face copulation. The face-to-face copulations of hihi are not only common, but also very energetic. Males chase females in succession for many hours. Possibly female hihi determine the quality of males in this way. However, both injury, and death by predation, could occur during this type of copulation. Hihi are sexual ly dimorphic and females may be aware of male fitness by the state of their external features . Another possible explanation is that a male-biased sex ratio could be leading to competition for females. Although the sex ratio on Kapiti I . is 1 : I the long term reproductive success (LTRS) of first-year females could be low compared to that of older birds. Males paired to young females might increase their own L TRS by forcing copulations with older females while still holding a nest. In 1 993, two o lder males paired with first year females provided less parental care for their chicks than males paired to older females. They also spent more time at the feeder, where there were opportunities for them to achieve copulations with older females. It is difficult to know if sperm are transferred from the male to the female during a face-to-face copulation. However, the cloaca of both male and female is unusually prominent making it easy to establish contact during a face-to-face encounter. The weight difference between the sexes would help a male to hold a female on the ground, especially on her back. Although the function of a cloacal protuberance for sperm storage in males is clear (Birkhead et al. 1 993 , Briskie 1 992, Mulder & Cockbum 1 993) , litt le is known about the importance of an enlarged c loaca in the females of some species. Recently Nakamura ( 1 990) described the use of the cloaca for breeding displays in female accentor, and Wilkinson & Birkhead ( 1 995) reported the use of enlarged cloacas in both male and female Greater vasa parrots, Coracopsis vasa to form a tie during copulation. I suggest that the protruding cloacas of both male and female hihi help to achieve successful c loacal contact for mating birds during copulations, and might facilitate insemination during a face-to-face copulation. In polygynandrous birds the relative sizes of the testes, cloacal protuberance, seminal glomera and the numbers of spermatozoa in the seminal glomera are large (Table 3 . 1 ), almost certainly because the intensity of sperm competition is high in these species. The relative sizes of these structures and spermatozoa numbers in the hihi are similar to, or greater than, that recorded in other polygynandrous species (Table 3 .2), providing additional evidence that sperm competition is intense in the hihi . 3.5 REFERENCES Anderson, S. 1 993 . Stitchbirds copulate front to front. Notomis 40: 1 4 . Angher, G.R. & Rasch, G . . 1 984. Breeding biology of the stitchbird (Notiomystis cincta) a hole nester Meliphagid. In Ecology and behaviour of the stitchbird. Unpublished report to the New Zealand Wildlife Service, Appendix A. Wel lington. Birkhead, T.R. 1 994. Enduring sperm competition. Journal of Avian Biology. 25: 1 67- 1 70. Birkhead, T.R. & M0ller, A.P. 1 992. Sperm competition in birds: evolutionary causes and consequences. Academic Press. London. Birkhead, T.R. & M0ller, A.P. 1 993 . Why do male birds stop copulating while their partners are still ferti le? Animal Behaviour 45: 1 05- 1 1 8 . Birkhead, T.R., Briskie, J .V. & M0ller, A.P. 1 993 . Male sperm reserves and copulation frequency in birds . Behavioural Ecology and Sociobiology 32: 85-93 . Birkhead, T.R., HatchwelL B .J. & Davies. .B . 1 99 1 . Sperm competition and the reproductive organs of the male and female dunnock Prunefla modularis . Ibis 1 33 : 306-3 1 1 . Birkhead, T.R. & Hoi, H. 1 994. Reproductive organs and mating strategies of the bearded tit Panurus biarnicus. Ibis 1 36 : 356-360. Briskie, J.V . 1 992 . Copulation patterns and sperm competition in the polygynandrous Smith's longspurs. The Auk 1 09 : 563-575 . Briskie, J.V . 1 993 . Anatomical adaptations to sperm competition in Smith's longspurs and other polygynandrous passerines. The Auk 1 1 0 : 875-888 . Castro, 1 . , Minot, E.O. & Alley, J.C. 1 994a. Feeding and breeding behaviour of hihi or stitchbirds (Notiomystis cincta) recently transferred to Kapiti Island, New Zealand, and possible management alternatives. Reintroduction Biology of Australasian Fauna. Surrey, Beaty and Sons, Healesville, Australia. Castro, I . , Alley, J .C . , Empson, R.A. & Minot, E.O. 1 994b. Translocation of hihi or stitchbird (Notiomystis cincta) to Kapiti Island, New Zealand: transfer techniques and comparison of release strategies. Reintroduction Biology of Australasian Fauna. Surrey, Beaty and Sons. Australia. Cheng, K.M., Bums, J.T. & McKinney, F. 1 983. Forced copulation in captivr mallards. I l l . Sperm Competition. The Auk 1 0 0 : 302-3 1 0 . Craig, J .L. , Douglas, M.E. , Stewart A.M. & Veitch, C .R. 1 982. Specific and sexual differences in body measurements ofNew Zealand honeyeaters. Notornis 28: 1 2 1 - 1 28. Cowan, P.E. 1 992. The eradication of introduced Australian brushtai l possums, Trichosurus vulpecula, from Kapiti Island, a New Zealand nature reserve. Biological Conservation. 6 1 : 2 1 7-226. 40 4 1 D avies, N .B . 1 983 . Polyandry, cloaca pecking and sperm competition in dunnocks. Nature 302 : 334-336. Davies, N.B. 1 992. Dunnock behaviour and social evolution. Oxford University Press. Oxford. Emlen, S .T. & Wrege, P.H. 1 986 . Forced copulations and intra-specific parasitism: two costs of social l iving in the white-fronted bee-eater. Ethology 7 1 : 2-29. Fuller, S. 1 985 . The vegetation of Kapiti Island. Department of Lands and S urvey. Miscellaneous publications. Wellington. M01ler, A.P. 1 99 1 . Sperm competition, sperm depletion, paternal care and relative testis size in birds. The American Naturalist 137: 882-906. Mulder, R.A. & Cockbum, A. 1 993 . Sperm competition and the reproductive anatomy of male superb fairy-vvTens. The Auk 1 10: 5 88-593 . N akamura, M. 1 990 . C loacal protuberance and copulatory behaviour of the alpine accentor (Prunella collaris). The Auk 107:284-295 . Rasch, G. 1 985a. The B ehavioural Ecology and management of the stitchbird. Unpubl ished MSc Thesis, University of Auckland. Auckland, New Zealand. Schulze-Hagen, K . , Leisler, B . , Birkhead, T.R. & Dyrcz, A. 1 995 Prolonged copulation, sperm reserves and sperm competition in the aquatic warbler Acrocephalus paludicola. Ibis 137: 85 -9 1 . Trivers, R.L. 1 972. Parental investment and sexual selection. In Sexual selection and the descent of man, 1 87 1 - 1 97 1 . B . Campbell (ed). Pp. 1 36- 1 79. Aladine-Atherton, Chicago. Wilkinson, R. & Birkhead, T.R. 1 995 . Copulation behaviour in the vasa parrots Coracopsios vasa and C. nigra. Ibis 137: 1 1 7- 1 1 9 ----------- ----- ------- Chapter 4 Honeyeaters and the New Zealand forest flora: The utilisation and profitability of small flowers ABSTRACT New Zealand flowers are frequently considered unspecialised having a simple often symmetrical morphology al lowing easy access to pollen and nectar by a wide range of visitors. Most New Zealand flowers conform with a syndrome of insect poll ination (entomophily), being smal l in size, l ight in colour and scented. However, pol l ination of forest flowers by birds has been described for a range of species whose flowers conform with omithophily, being large, red or yellow and unscented. On Little Barrier Island and Kapiti Island, all three species of New Zealand honeyeaters have been described feeding on flowers whose pollination is currently assumed to be by insects (entomophi lous) or where the pol l ination system is unknown. The persistence and regularity of visits suggests that the birds are obtaining suitable rewards in the form of nectar and could be serving as pollinators . The nectar production of three ?'omithophi lous" species and five '?entomophilous" species was measured to determine the value of the nectar rewards to birds. It was found that the estimated nectar consumption rate for all flowers sampled was enough to sustain the two smaller New Zealand honeyeaters. It is suggested that because (a) the nectar produced by "entomophilous" flowers provides sufficient calories to sustain the energetic requirements of birds, and (b) these plants flower in the cooler months when insect activity is reduced, birds might be play ing an important pol l ination role not previously considered. If this is the case, forest regeneration on the New Zealand mainland could be hampered by the loss of one species of honeyeater (hi hi) and the serious reduction in the abundance of the other two species. 4.1 INTRODUCTION THE flowers of many New Zealand native plants are structurally simple, clustered and small, scented, mostly green, white or cream in colour (Cockayne 1 928 ; Godley 1 979), and a large percentage of species have separate sexes (reviewed in Lloyd 1 985) . Moreover, most New Zealand flowers are regarded as unspecialised and frequently rather promiscuous, receiving visits from several kinds of insects. New Zealand's pol l inating fauna is considered depauperate and advanced pollinators are rare (Webb & Kelly 1 993 ; Godley 1 979). Long-tongued bees and hawkmoths are absent and butterflies few, their role in pol l ination possibly assumed by some species of Diptera (Thompson 1 927; Heine 1 93 8) . Some poll ination ofNew Zealand plants has also been ascribed to reptiles (Whitaker 1 987), birds (Godley 1 979; Delph & Lively 1 989), and 42 43 bats (Daniel 1 976; Ecroyd 1 993) . Flowers pollinated by birds are usually red or orange in colour, large in size, containing l arge amounts of dilute nectar, and often unscented (Faegri & van der Pij l 1 97 1 ) . Of the 3 0 New Zealand native species ofplants Godley ( 1 979) reported as being visited by New Zealand birds, only some display this typical ornithophilous syndrome. Godley suggested that the remaining species are probably visited by birds only incidentally, and that true ornithophily is not involved. He suggested that birds may foster self poll ination rather than cross pollination (except for the species with sexually dimorphic f1owers) . Stiles ( 1 978) states that the optimal pollinator i s one which most effectively transfers pollen producing the maximum seed set over a blooming season. However, specialisation for a particular pol l inator may not occur if the most consistent pol lination results from having visits from diverse poll inators ( Stiles 1 978) . To ensure the attraction of pollinators, plants need to offer some reward. Most birds poll inate flowers as they collect the sugar-rich nectar they contain (Simson & Neff 1 983) . N ectar is expensive to produce and a plant should produce sufficient nectar to ensure pollinators will repeatedly visit it, without being too easily satiated after one or a few visits (Heinrich & Raven 1 972). There has been very l ittle research on nectar production or availability in New Zealand species, other than for flowers used by bees in honey production (Cockayne 1 9 1 6 ; Palmer-Jones & Line 1 962; Fisher 1 965 ; Walsh 1 967 ; Cl inch, Palmer-Jones & Foster 1 972). Honeyeaters (family Meliphagidae) are common birds throughout Australasia, particularly Australia, where they pollinate many plant species (Paton & Ford 1 976). New Zealand has three endemic species of honeyeaters - the tui Prosthemadera novae? zelandiae 1 , the bellbird Anthornis melanura, and the hihi (stitchbird) Notiomystis cincta. All three species feed on flower nectar, fruits and insects in different proportions according to availability (Gravatt 1 970; Angehr 1 984; Castro, Minot & Alley 1 994) . The abundance of honeyeaters in the New Zealand mainland has decreased since human colonisation (Diamond & Veitch 1 98 1 ; Gill & Martinson). Predator-free or predator-reduced islands have greater densities of honey eaters when compared to the mainland, raising questions regarding the effects of this decrease on seed dispersal and pollination. In this paper I investigate nectar production and the visitation of the flowers of forest p lants by the three species ofNew Zealand honeyeaters on Kapiti Is land. Kapiti I . i s one of two remaining places in New Zealand where the three species of honeyeaters co-exist (the other place is Little Barrier I sland in the Hauraki Gulf). Most flowers used by the birds on Kapiti I. do not display the typical ornithophi lous syndrome described above and so I set out to determine the profitability of birds foraging on these and other 1 Bird names follow Turbott 1 990 flowers. Finally, I discuss the possible role ofNew Zealand honeyeaters in poll ination of these plants. 4.2 METHODS 4.2 . 1 Study area Observations were made on Kapiti Island ( 1 965 ha, Lat. 40?50 ' , Long. 1 74?55 ' ) on the south western coast of the North island of New Zealand from August 1 99 1 to August 1 994. The island was largely cleared for agriculture in the 1 800s but by 1 930, all browsing animals with the exception of brush-tai led possums (Trichosurus vulpecula) had been eradicated (Cowan 1 992). Kapiti was freed of possums in 1 986 (Cowan 1 992). Kapiti I . forests are divided into scrub and low forest (approx. I 04 1 ha) and tall forest (Metrosideros robusta forest; tawa Beilschmiedia tawa -hinau Elaeocarpus dentatus -kamahi Weinmannia racemosa forest; and kohekohe Dysoxylum spectabile forest, total l ing 550 ha). The remainder of the island is covered by low scrub, grass, tussock and flax (Ful ler 1 985) . With the exception ofNorway rats (Rattus norvegicus) and polynesian rats (Rattus exulans), Kapiti I. is free of mammalian predators. In the last ten years there have been several bird transfers to Kapiti I . aimed at reducing their vulnerability to extinction. Species transferred include hihi, saddlebacks Phylesternus carunculatus, and kokako Calleas cinerea. 4.2.2 Nectar sampling Phenological observations were made of the most important honeyeater plant species, measuring the presence and availabil ity of fruits and flowers (Chapter 5 ) . During 1 993 and 1 994 nectar samples from flowering species used by honeyeaters were taken by absorption onto filter paper "wicks" (McKenna & Thomson 1 989). Occasionally it was necessary to redissolve dried nectar on the flower with a small drop of water (Cresswell 1 990) prior to the application of the wick. Three "ornithophilous" flowers were sampled: Metrosideros fulgens *, Metrosideros excelsa and Fuchsia excorticata; and five '?entomophilous" flowers : Pittosporum umbellatum, Pseudopanax arboreus, Dysoxylum spectabile, Pittosporum eugenioides and Geniostoma rupestre. For each species, the nectar in the flowers of three to eight separate p lants was sampled. The number of flowers sampled per plant varied according to species (Table 4.2). Some of the flowers had poll inators excluded by means of mesh bags ("bagged") and some were left open to possible visitations ( "unbagged"). The mesh bags were left on the flowers for 24 hours before sampling. In the case of Dysoxylum spectabile flowers from the same plants were sampled both early and late in the flowering season. The nectar contained in the flowers is referred to as the "standing crop". Plant names follow Allan 1 961 ; Moore and Edgar 1 970; and Connor and Edgar 1 987. 44 45 Nectar concentrations (sucrose equivalent) per flower were calculated using an anthrone colorimetric assay of the sugar redissolved from the filter paper wicks (McKe1ma & Thomson 1 989). The calorific values of flowers were calculated assuming that 1 mg sucrose = 4.2 calories (Ford 1 979) . The average number of calories per inflorescence of each species was estimated from the product of the average calorific value of the standing crops and the average number of flowers in an inflorescence. To calculate the average number of flowers per inflorescence I counted the flowers in ten inflorescences of each species. Similar values were not obtained for Dysoxylum spectabile and Geniostoma rupestre because the flowers are not arranged in a compact inflorescence but are rather loosely arranged (Dysoxylum spectabile) or arising directly and spread along the branches ( Geniostoma rupestre ). 4.2.3 Bird observations Feeding by honey eaters on flowers or fruits was recorded fortnightly during the non-breeding months (April-September), and daily during the breeding season (October-March). From August 1 993 to August 1 994 the number of flowers of a particular plant species visited by each honeyeater species per unit of time vvas recorded as flowers became avai lable. Observations were done on randomly chosen trees whenever the trees were in flower. I aimed to fol low at least five individual birds of each species and to collect at least 1 0 minutes of observation per species of bird. This was not always possible. Most hihi on Kapiti are colour banded for individual identification, but there is no way to individually identify tui or bel lbirds. To minimise the possibil ity of collecting all the data on the same individuals for these two species, the observations were done on trees located in several different areas. 4.2 .4 Honeyeaters energy requirements The m etabolic requirements of New Zealand honeyeaters were estimated using a fonnul a that scales the requirements of another species Lichmera indistinct a by a factor proportionate to body mass (Collins & Newland 1 986). Winter conditions were assumed (nine hour day length; 1 3 ?C day temperature; 9?C night temperature) . It was also assumed that the New Zealand honeyeaters were, l ike L. indistincta, spending 93 . 8 % of their time in activities other than flying . The energetic requirements for hihi (average 3 5 .2 5 g) were estimated to be 29 cal/min; for bellbirds (28 g), 24. 56 cal/min; and for tui ( 1 04. 1 ), 63 .2 cal/min. When calculating the energy obtained by the honeyeaters feeding in the different species of p lants, I assumed (a) the birds were taking all the nectar in each flower visited; (b) the birds were not selective when foraging on these flowers, and (c) nectar content and handling time per flower were independent. The nectar reward gained is therefore the product of flowers visited per unit of time and the average calorific content of each flower. As each parameter was collected independently, each with its own degree of error, the error associated with the resulting product is compounded. To provide some idea of the possible range of energy obtained whilst feeding, a Montecarlo randomisation procedure was used. From the range of observed bird feeding rates one value was randomly chosen and that number of entries was chosen from the observed array of flower calorific values. These values were summed to give an estimated return to the bird for one minute 's foraging. This procedure was repeated 1 000 t imes to obtain a range of calorific returns. These randomisations were performed for both bagged and unbagged flowers. Nested ANOV A's were used to partition the variance in the amount of sucrose per flower into a treatment effect (bagged vs. unbagged); tree effects; the tree by treatment interaction; and in some cases a season effect (early or late); and/or an inflorescence effect. Type Ill , or partial sums of squares, are presented in the analysis of variance table . 4.3 RESULTS A large number of species with small flowers are visited by the honeyeaters in addition to the larger flowered. typical ornithophi lous flowers (Table 4 . 1 ). Many of these flowers were visited consistently each year (Chapter 5 , Appendix 1 ) whenever they became available (eg. Geniostoma rupeslre, PseudopanCLt arboreus, Pittosporum eugenioides, Pittosporum crassifolium, Piuosporum tenuifolium , and Weinmannia racemosa ) . Birds of al l three species of honeyeater were seen on both male and female flowers of PseudopanCLt arboreus and Pitlosporum crassifolium. In general, the entomophilous flowers were less rewarding than the omithophilous species (Table 4.2) although there is considerable overlap particularly in the unbagged flowers. There were significant differences between the unbagged and bagged standing crops of Metrosideros robusta, Dysoxylum spectabile, Fuchsia excorticata and Pseudopanax arboreus flowers (Table 4.2). There were no significant differences in bagged and unbagged flowers of the remaining species (Table 4.2). There was a significant difference in nectar production per tree in Pittosporum crassifolium. This difference was due to the presence of a female tree in the sample. This tree' s flower nectar had significantly more calories (average 1 3 .65, n= 1 2) than the male trees (6.53 , 6 .46, 3 .96, and 3 .24; n= 1 1 for each tree). 46 47 Table 4. 1. List of plant species on Kapiti Island whose flowers are visited by honeyeaters. To describe abundance I used a scale of: 1 = very rare; 2 = rare; 3 = occurs; 4 = abundant; 5 = common; 6 = very common (Angher, 1 984 ); y = young plants; 0 = not present. To describe frequency I used: 3 = frequent ( if every time the flowers were available there were visits by honeyeaters); 2= occasional (if the flowers were vis ited sometimes only);O = I have not recorded the birds feeding on such flowers on Kapiti; 1 = rarely (if I rarely saw birds on the flowers). T = tui . Flower sizes: Largest diameter > 20mm; Medium= 1 0-20 mm; Small= 5 . 1 -9.9 mm; Minute= < 5mm. SPECES AE3L.N:lAN:E FF?a..El'C(CfVSffA1rna.JKAPm R..Oi\ffi sc:u=rr Hi hi Bel lbird Tui size Fuchsia excorticata 2 3 3 3 Large Godley 1 979 Knightia exce/sa 5, y 3 3 3 Large Godley 1 979 Metrosideros exce/sa 2 3 3 3 Large Godley 1 979 Metrosideros fulgens 5 3 3 3 Large Godley 1 979 Metrosideros robusta 3 3 3 3 Large Godley 1 979 Phormium tenax 6 3 3 3 Large Godley 1 979 Sophora spp. 2 0 3 0 Large Godley 1 979 Vitex lucens 2 3 3 3 Large Godley 1 979 Dysoxylum spectabile 4 3 3 3 Medium Godley 1 979 Earina autumnalis 6 1 0 0 Medium This Study E/aeocarpus dentatus 5 3 3 3 Medium Craig 1 981 Hoheria popu/nea 1 3 0 0 Medium This Study Metrosideros perforata 5 3 0 0 Medium Godley 1 979 Myoporum laetum 3 3 3 0 Medium This Study Passiflora tetranda 6 2 0 0 Medium This Study Pittosporum cornifolium 2 3 3 0 Medium This Study Pittosporum crassifolium 1 3 3 3 Medium Godley 1 979 Pittosporum umbel/atum 2 3 3 3 Medium Godley 1 979 Rhopalostylis sapida 6-y 3 0 0 Medium This Study Ripogonum scandens 4 3 0 0 Medium This Study Syzygium maire 1 3 0 0 Medium This Study Aristotelia serrata 3, y 3 0 0 Small This Study Geniostoma rupestre 4 3 3 3 Small This Study Kunzea ericoides 6 0 3 0 Small This Study Laurelia novae-ze/andiae 3,y 0 3 0 Small This Study Melicytus ramiflorus 4 2 0 0 Small Godley 1 979 Personia toru 5 0 0 0 Small Godley 1 979 Pittosporum eugenioides 5-y 3 3 3 Small This Study Pittosporum tenuifolium 5 3 3 3 Small This Study Pseudopanax arboreous 4 3 3 3 Small Godley 1 979 Pseudopanax crassifolius 3, y 2 0 0 Small Godley 1 979 Rubus cissoides 4 3 0 0 Small This Study Shefflera digitata 5 2 0 0 Small Godley 1 979 Aestelia spp. 4 1 0 0 Minute Ths Study Beilschmiedia tawa 6 2 0 0 Minute Craig 1 981 Corynocarpus laevigatus 5 0 2 Minute Godley 1 979 Cyathodes spp. 2 0 0 0 Minute Godley 1 979 Griselinea littoralis 2 2 2 2 Minute This Study Hebe spp. 2 0 0 0 Minute Godley 1 979 Myrsine australis 5 1 0 0 Minute This Study Myrsine sa/icina 5 2 0 0 Minute This Study Nestegis lanceolata 3 0 0 0 Minute Godley 1 979 Weinmannia racemosa 5 3 3 3 Minute Godley 1 979 Plants not present on Kapiti Clianthus spp. 0 Large Godley 1 979 Metrosideros umbellata 0 Large Godley 1 979 Rhabdothamnus so/andri 0 Large Godley 1 979 Peraxila spp 0 Medium Godley 1979 Cordyline spp. 5 Medium Craig 1 981 Alseuosmia macrophylla 0 Small Godley 1 979 Dracophyllum spp. Small Godley 1 979 For Dysoxylum spectabile, the number of calories per flower increased as the season progressed (Table 4.2). The number of visitors to the flowers also increased, and 48 observations of kaka (Nest or meridionalis), a large omnivorous parrot, on D. spectabile only occurred later in the season ( I . C . pers. obs.) . The entomophilous species studied (except Geniostoma rupestre and Dysoxylum specLabile) had multiple numbers of flowers presented in compact inflorescences and so the inflorescence as a unit accessible from one perch may be a relatively rich source of food. When the standing and bagged nectar calorific content for inflorescences of the entomophilous species were calculated the differences in calorific content with omithophilous flowers were markedly reduced (Figure 4. 1 ) . This feature of many compactly arranged flowers allows for rapid turnover of flowers by a probing bird which often spend less than a second on each flower (Table 4.3) . Table 4.2. Results of ANOY A Plant Species Unbagged Bagged Probability Time effect Tree?trt Tree effect lnflorescenc X (cai)?SE n X (ca i)?SE n Unb/Bg p effect p e p p G. rupestre 0.43:':0.05 94 0 .52?0.06 78 NS NA NS NS NA D. spectabite (early) 0.94::0.2 23 1 . 16?0.21 38 < 0.0001 < 0.0001 NS NS NA D. spectabile (late) 1 .74?0.34 2 1 4 07?0.59 1 5 P. arboreus (male) 0.77:':0 . 16 38 0.83?0 . 13 23 < 0.007 NA NS NS NA P. eugenioides 0.67:':0.63 1 3 0.59?0.52 78 NS NA <0.0001 NS NA P. crassifolium 5.5::: 1 .35 26 8. 1 ? 1 . 1 7 30 NS NA NS < 0.0006 NS M futgens 1 .60?2. 1 5 22 1 8.4?2.22 1 7 < 0.0001 NA NS NS NS F. excorticata 0.42?0. 1 1 25 2.09?0.27 1 9 < 0.0001 NA NS NS NA M excelsa 5.25+1 .09 28 30. 1 4+4.58 1 2 N A NA NS < 0.005 * We did not compare unbagged and bagged flowers for this species because the samples from bagged flowers came from different plants than those from unbagged flowers. The average bagged and unbagged standing crops produced by the all of the species sampled when obtained by the birds at the ayerage visitation rate appeared sufficient to satisfy the estimated energetic requirements of bellbird and hihi, but not tui (Table 4 .3 and Figure 4.2). Honeyeaters feeding on Pittosporum crassifolium could be obtaining up to six times the required amount of nectar necessary to sustain this type of activity. In many cases the 95% confidence intervals are large, due in the most part, to the large variability of calorific content of flowers as determined by the Montecarlo procedure. 4.4 DISCUSSION It is c lear from Table 4. 1 that New Zealand honeyeaters are investigating small flowers as a regular foraging activity. This trait could be related to their evolution in an 49 environment where small , rather than l arge, flowers are the rule (Godley 1 979) . In many cases i t appears they are obtaining sufficient energy from the nectar to justifY this behaviour, but the importance of other components of the nectar that satisfY needs in the diet other than energy can not be discounted. The latter could explain the persistence of tui on Geniostoma rupestre, Pseudopanax arboreus, and Dysoxylum spectabile, which seemingly do not fulfil energetic demands. However, I have assumed that the birds do not forage selectively. Nevertheless, it is known that, at l east for some species of plants, tui and bellbirds can differentiate between flowers containing nectar and empty flowers on the basis of colour (Peraxilla spp . . A.W.Robertson pers. obs. ; Fuchsia excorticata, Delph & Lively 1 98 9) . Hummingbirds have been seen selectively visiting the flowers of Ribes sanguineum, a plant in which the small corona-like petals change colour (Weiss 1 995) . If honeyeaters forage selectively, the standing crops obtained for all the plants in Figure 4.2 could be sufficient to sustain the energetic needs of the honeyeaters (top of 95% confidence interval for unbagged standing crops). Moreover, the birds may potentially forage when flowers have the greatest amount of nectar eg. in the early morning. Bagged crops of all of the plant species on Table 4.2 provide excess reward to a l l three species of honeyeaters (top of 95% confidence interval for bagged standing crops) and may be better approximations to the true value of flowers to selectively foraging birds. Table 4.3. Average number of flowers visited by honeyeaters whi le foraging for one m inute. Values were obtained by counting the number of flowers v is ited by members of each species per unit of time whi le in s ight. B i rds were observed for as long as poss ible. E = early; L = Late. Plant G. rupestre D. spectabile E D. spectabile L Bellblfd n X+SE 47 35?6.29 8 37.2?5.35 42.0?5.84 P arboreus (male') 85.65?3.27 5 P. crassifolium 28.5?3.35 8 Obs.t1me 1 0 .22 6.0 8.6 3.1 1 7.48 * No data are avai lable for female flowers. Hihi X?SE n Obs.time Tui X?SE 70.45?8.35 5 48.5?9.13 2 25 .9?5.30 2 m in) 17 .28 1 .4 2.9 72.43?12 61 4 1 .3?8.49 43. 1 ?5. 1 1 79. 1 ?5.97 38?9.08 n Obs.time imin) 8 1 0 . 1 8 5 2.6 2 1 2.7 4 3.8 5 25.28 The l ack of difference between the calories in the nectar of bagged and unbagged flowers of Pittosporum eugenioides, Geniostoma rupestre, and Pittosporum crassifolium (Table 4.2), may be due to e ither a low visitation rate to the flowers or to the plant continuing to produce nectar as it is cropped by visitors. The mechanisms and time of nectar production for New Zealand plants i s unknown and more research i n this area IS necessary. (fJ ,_ Q) 5 0 - - 0 ,._. c Q) ,._. c 0 u u -?;:: 0 cu u (fJ Q) u c Q) u (fJ Q) ,_ 0 - c - 0 ,._. c Q) ,._. c 0 0 u -?;:: 35 t 30 25 20 400 350 0 50 EJ Bagged D Unbagged 8 o??????????-4?????????==??? Gen-rup Dys-spe Pse-arb Pit-eug Pit-era Met-ful Fuc-esc Met-exe "Entomoph i lous" "Orn ithophi lous" Figure 4. 1. Estimated number of calories in the flowers and inflorescences of selected New Zealand forest plants on Kapiti Island. Flowers of Ceniostoma rupestre. Dysoxylum spectabile and Fuchsia esconicata are not arranged in inflorescences. The flowers of some plants might have evolved alternative mechanisms to ensure pollination by birds other than large corolla size or specific calyx shape. The nectar on female Pseudopanax arboreus flowers is spread over the surface of the flower in such a way that to obtain it the flower has to be licked (I . C. , pers. obs. ) . This takes 50 5 1 some time and in the process the birds head has the opportunity to contact the stigma of the flower. This "lolly-pop" syndrome is shared by the entire inflorescences of Pittosporum eugenioides . Nectar collects on top of the flowers and the birds appear to lick the nectar from the surface of the inflorescence rather than from individual flowers. New Zealand's generally windy conditions (Kuschel 1 975) may make l arge flowers with wide petals suceptible to damage by wind. Smaller flowers arranged in inflorescences could produce a similar visual impact and can provide a large reward (Figure 4 . 1 ) , while resisting the effects of wind. Inflorescences could also make it easy for birds to brush their heads against flowers and be covered by pol len while stil l al lowing insect visitation. Tui have been observed brushing their white neck feathers and lower breast against Pseudopanax arboreus flowers while feeding on them and a female hihi captured in an area of Pseudopanax arboreus was carrying this species pollen on her head (I . C. pers. obs. ) . I have also collected Geniostoma rupestre pollen from the heads of hihi (I . C. pers. obs). New Zealand weather is unpredictable and cold spells are common especially during spring when many plants flower. New Zealand plants may benefit from being visited by birds as well as insects if this trait ensures the most consistent pol l ination. as suggested by Stiles ( 1 978) . Anna hummingbirds Calypte anna existing in cold, rainy conditions (Stiles 1 97 1 , 1 973 ), as well as many other birds living at high elevation areas (Cruden 1 972), are active during unfavourable weather conditions, so becoming important pollinators when insects enter torpor. Several of the species investigated flower in the cooler months. Pseudopanax arboreus (July-December) and Dysoxylum spectabile (May-July) flower in late Autum and winter while Pittosporum eugenioides (October), Geniostoma rupestre (August-November) and Pittosporum crassifolium (September-October), flower in spring. Temperatures and general weather conditions during winter and spring could favour bird activity over insect activity, at least in some years. Honey eaters tend to feed in discrete areas of forest and, because they have the same feeding interests, it is common to find the three species feeding together. While the larger, more dominant tui (Craig 1 984; Rasch 1 985 ; I . C . pers. obs .) can defend a whole tree and visit most flowers i n it, bellbirds and hihi are often displaced and move from tree to tree more often, increasing the possibility of cross pollination. This may provide higher quality poll ination (Stephenson 1 982) u 'V cu - Ci E' ::J g '- cu u -Q) -r. / !:::. 800 .------------------------- A. Tu i 600 BOO .-------------------------- 6. H lh i cOO LOO 2'JO ' ,- c I I 0 ??-?_?_' ?------- ----????- ? --? ?_?_?_? ?_? ?_? ?_? __ __ 800 I I I r I 600 ? I 400 [ 200 . I - . ? s Cj B Q c.: C. 6e ! lb 1 rd I . - ? - .? . . - . - J_ . ? . ? ? ? ? --! ? ? :;; u ('f. g ? 2 g g Vi .c ? .c c-;; r. 2 ;:; ;:; g g 0 V: V: 0 0 - c ! - - - - - c. ? "' "' :9 e CJ Vi "C V: Vi (1: e u c.: c:; 2 Figure 4.3. Estimated number of calories per m inute obtained by birds foraging of selected species of flowers on Kapiti Is land. Bars at either side of the mean indicate 95% confidence intervals by a Montecarlo randomisation procedure (see text for details). Dashed l ine indicates the estimated number of calories required per minute by each honeyeater species. Darked dots indicate unbagged flowers; circles indicate bagged flowers. 52 53 N ew Zealand forests have lost a large number of birds species and suffered a severe reduction in the abundance of others (Holdaway 1 989). Offshore islands free of mammalian predators are the only examples of what the New Zealand mainland forests may have been l ike (Diamond & Veitch 1 98 1 ) . Diamond and Veitch ( 1 98 1 ) show that although the number of species of birds sti l l present on the mainland ofNew Zealand is similar to those on islands, the relative abundances of such species i s much lower in the mainland forests . The three New Zealand honeyeaters coexist today only on L ittle Barrier I s land and on Kapiti I sland because hihi was never present in the South Island and became extinct from the North Island mainland in the late 1 800s (Oliver 1 955) . If New Zealand honeyeaters are contributing to pollination on offshore islands, the l ack of these species or the reduction of their numbers on the mainland could be hindering forest regeneration. This study demonstrates the value of small flowers to honeyeaters as complements to the larger showy "ornithophilous flowers". Descriptions of these visits as incidental is not justified. Rather these flowers are an important component of the honeyeater diet. There is a need for more research in this area specifically to measure the effectiveness of birds as pollen vectors, and to compare the efficiency of insects and birds as pollinators. Flowers such as Corynocarpus laevigatus and Griselinia spp. which are not visited intensively might not be actively pollinated by birds, but the possibil ity should be studied. Comparative mainland and island studies could help us to understand the importance of pollination by birds. The generalized nature of the New Zealand flowers could allow introduced insects, such as bees, to carry on poll ination, so alleviating the effects of reduced bird abundances, this possibility should also be investigated . The results of such studies wil l help us to understand forest dynamics and wil l be useful in the management of natural areas. 4.5 REFERENCES Allan, H.H. 1 96 1 . Flora of New Zealand, Vol.J. Government Printer, Wellington, New Zealand. 1 085 pp. Angehr, G .R. 1 984. Ecology and behaviour of the stitchbird. Internal report to the New Z ealand Wildlife Service, Wellington. Castro, I . , M inot, E .O . and Alley, J .C. 1 994. Feeding and breeding behaviour of hihi or stitchbirds Notiomystis cincta recently transferred to Kapiti Is land, New Zealand, and possible management alternatives. Pp. 1 2 1 - 1 28 In Reintroduction B iology of Australian and New Zealand fauna. Ed. by M. Serena. Surrey B eatty and Sons. Chipping North. Clinch, P .G . ; Palmer-Jones, T. ; Foster, I .W. 1 972. Effect on honeybees of nectar from the yellow kowhai (Sophora microphylla Ait) . New Zealand Journal of Agricultural Research. 1 5 : 1 94-20 1 . Cockayne, A.H. 1 9 1 6 . Present and future sources of honey in New Zealand. New Zealand Journal of Agriculture. 13 : 27-35 . Cockayne, L . 1 928. The vegetation of New Zealand. 2nd. Edition. Die Vegetation der Erde, XIV Wilhelm Englemann. Leipzig. 456 Pp. Coll ins, B .G . ; Newland, C. 1 986. Honeyeater population changes in relation to food availabil ity in the jarrah forest of Western Australia. Australian Journal of Ecology. 1 1 : 63-76. Connor, H .E . , and Edgar, E. 1 987 . Name changes in the indigenous New Zealand Flora, 1 960- 1 986 and Nomina Nova IV, 1 983- 1 986. New Zealand Journal of Botany. 25: 1 1 5- 1 70. Cowan, P .E . 1 992. The eradication of introduced Australian brushtail possums, Trichosurus vulpecula, from Kapiti Island, a New Zealand nature reserve. Biological Conservation. 6 1 : 2 1 7-226. Craig, J .L . , 1 984. Wing noises, wing slots, and aggression in New Zealand honeyeaters (Aves: Meliphagidae). New Zealand Journal of Zoology. 1 1 : 1 95-200. Cresswell, J .E . 1 990. How and why do nectar-foraging bumblebees initiate movements between inflorescences of wild bergamot Monardajistulosa (Lamiaceae) . Oecologia. 82: 450-460. Cruden. R. W. 1 972. Poil inators in high elevation ecosystems: the relative effectiveness of birds and bees. Science 1 76 : 1 439- 1 440. Danie l , M .J . 1 976. Feeding by the short-tailed bat (Mystacina tuberculata) on fruit and possibly nectar. New Zealand Journal of Zoology 3 : 3 9 1 -398 . Delph, L.F. ; L ively, C.M. 1 989. The evolution of floral colour change : poll inator attraction versus physiological constraints in Fuchsia excorticaw. Evolution 43 : 1 252- 1 226. Diamond, J .M . ; Veitch, C.R. 1 98 1 . Extinctions and introductions in the New Zealand avifauna: cause and effect?. Science 2 1 1 : 499-50 1 . Ecroyd, C . 1 993 . In search of the wood rose. Forest and Bird Journal 267 : 24-28. Faegri, K. ; van der Pijl , L. 1 97 1 . The principles of pollination ecology. 2nd. Ed. Pergamon Press. Oxford. Fisher, F .J .F . 1 965 . The alpine Ranunculi of New Zealand. New Zealand. Department of Scientific and Industrial Research. Bulletin 1 65 . 1 92 pp. Ford, H.A. 1 979. Interespecific competition in Australian honeyeaters - depletion of common resources. Australian Journal of Ecology 4: 1 45 - 1 64. Fuller, S . 1 985 . The vegetation of Kapiti Island. Department of Lands and Survey. M iscel laneous publications. Wellington. Gil l , B . and Martinson, P. 1 99 1 . New Zealand's extinct birds. Random Century New Zealand Ltd. Auckland, New Zealand. Pp 1 09. Godley, E.J . 1 979. Flower biology in New Zealand. New Zealand Journal of Botany 17: 54 5 5 44 1 -466. Gravatt, D.J . I 970. Honeyeater movements and the flowering cycle of vegetation on Little Barrier Island. Notornis 17 : 96-1 0 1 . Heine, E.M. 1 93 8 . Observations on the pollination of New Zealand flowering plants. Transactions and Proceedings of the Royal Society of New Zealand 67: 1 33 - 1 48 . Heinrich, B . ; Raven, P .H . 1 972. Energetics and pollination ecology. Science 1 76: 597- 602. Holdaway, R.N . ( 1 989) New Zealand' s pre-human avifauna and its vulnerabi li ty . Moas, mammals and climate in the ecological history of New Zealand (ed. M.R. Rudge). 1 1 -2 5 Supplement to the New Zealand Journal of Zoology, Vol 1 2. New Zealand Ecological Society, Christchurch. Kuschel , G. Biogoegraphy and Ecology in New Zealand. The H ague. Lloyd, D .G . 1 98 5 . Progress in understanding the natural history of New Zealand plants. New Zealand Journal of Botany 23: 707-722. McKenna, M . ; Thomson, J .D. 1 989. A technique for sampling and measuring small amounts of floral nectar. Ecology 69: 1 306- 13 07 . Moore. LB. ; Edgar, E. 1 970. Flora of New Zealand, Vol. 11. Government Printer. Well ington, New Zealand. 354 pp. Oliver, W.R.L. 1 955 . Ne?v Zealand birds. A.H. and A.W. Reed Eds. 2nd Edition. Wellington. Palmer-Jones, T . ; Line, L .J .S . 1 962. Poisoning of honeybees by nectar from the karaka tree . New Zealand Journal of Agricultural Research 5: 433-436. Paton, D . C . ; Ford, H.A. 1 976. Pollination by birds of native plants in South Australia. Emu 77: 73-85 . Stiles, F .G. 1 97 1 . Time, energy and territoriality of the Anna hummingbird (Calypte ann a). Science 1 73 : 8 1 8-82 1 . Stiles, F .G. 1 973 . Food supply and the cycle of the Anna hummingbird. University of California Publications in Zoology. No. 97. 1 09 Pp. Stiles, F .G . 1 97 8 . Ecological and evolutionary implications of bird pollination. American Zoologist 18 : 7 1 5-727. Stephenson, A.C. ( 1 982) When does outcrossing occur in a mass-flowering p lant?. Evolution 36: 762-767. Thomson, G.M. 1 927 . The pollination of New Zealand flowers by birds and insects. Transactions and Proceedings of the New Zealand Institute 57: 1 06- 1 25 . Turbott, E .G. 1 990 . Checklist ofthe birds of New Zealand. Ornithological Society. Well ington, New Zealand. 247 pp. Walsh, R .S . 1 967 . Handbook ofNew Zealand nectar and pollen sources. N ational B eekeepers Association ofNew Zealand. Upper Hutt. New Zealand. 5 5 Pp. ------ ------------------- Webb, C.J. ; Kelly, D. 1 993 . The reproductive bio logy of the New Zealand flora. TREE 8 : 442-447. Weiss, M.R. Floral color change: a widespread functional convergence. American Journal of Botany. 82 : 1 67- 1 85 . Whitaker, A.H. 1 987. The roles o f l izards i n New Zealand plant reproductive strategies. New Zealand Journal of Botany 25: 3 1 5-328 . 56 Chapter 5 Hihi (Notiomystis cincta) feeding and breeding behaviour in relation to flowering and fruiting phenology on Kapiti Is land ABSTRACT The phenology of a selected group of plants, important as honeyeater food, was followed from 1 992 to 1 994 on Kapiti Island. The onset and length of the flowering and fruiting periods for a particular plant species varied between the years. The total number of fruits and flowers per tree also varied. There were flowers present throughout the year with a peak during the spring. Of 1 7 species examined in 1 993-94 1 5 bore fewer flowers than in 1 992-93 . For most species the flowering season started a month later in 1 993 than in 1 992. Ripe fruits were present throughout the year, but there were no obvious peaks. The hihi Notiomystis cincta is a rare New Zealand honeyeater with a single self? sustaining population on Little Barrier Island. Translocation to other rat-free islands has been employed as a management option. Islands were chosen for translocation if they lacked predators, and there were similarities between their vegetation and that of Little Barrier Island. This chapter discusses the ways these comparisons were done and offers alternatives. 5.1 INTRODUCTION THE hihi (Notiomystis cincta) is a rare New Zealand native honeyeater. The species once occupied the whole of the North I sland and its offshore I slands. It became extinct throughout this range except for Little Barrier Island during the 1 800's. The causes of contraction are unknown but there is strong anecdotal evidence suggesting a combination of factors: predation by introduced mammalian predators (mainly ship rats Rattus rattus ), loss of habitat, and disease. From 1 980 there have been ten translocations of hihi from Little Barrier I . to other offshore islands in an attempt to establish new self-sustaining populations of the species (Castro et al. 1 994a; Chapter 1 ) . In all but the last two translocations to Kapiti Island the numbers of transferred birds diminished rapidly over the first two years after release, until the populations became extinct. The last two transfers to Kapiti I . have produced a small , but sel f-sustaining population of about 30-40 birds (Chapter 6). The main difference between previous transfers and the last two transfers to Kapiti I . is that food was provided throughout the year after the last two. Hihi feed on three main food types: nectar, fruits and invertebrates (Gravat 1 970; Angher 1 984; Rasch 1 985 ; Lovegrove 1 986; Castro et al. 1 994b; Chapter 2). Although this omnivorous tendency could be favourable to hihi establishing in a new 5 7 58 environment, the hihi is subordinate to its food-competitors - tui Prosthemadera novae - zelandiae and bellbird Anthornis melanura . The diets of these honeyeaters overlap almost totally with that of the hihi. Moreover, tui and bel lbirds are dominant over hihi at feeding sites (Craig 1 984; Angher 1 984; Rasch 1 985) . Kaka (Nestor meridionalis ), a New Zealand parrot, shares part of its diet with hihi (Moorhouse 1 995) and it is also dominant over hihi (pers. obs.). The islands to which hihi were translocated were chosen primarily for having no ship rats, and secondarily on the general similarity of their vegetation to that of Little Barrier I (Angher 1 984 ). Hihi have survived on Little Barrier I . in the presence of kiore Rattus exulans and cats Fe lis catus (Rasch 1 99 1 ) . All of the islands where hihi have been translocated, excepting Mokoia Island, had kiore and/or Norway rats Rattus norvegicus at the time of the translocations. The similarity of vegetation was measured as the number of flowering or fruiting species in common between each island and Little Barrier I . (Angher 1 984 ) . Food limitation is one reason offered to explain hihi failure to establish self? sustaining populations after translocation. Food limitation can occur because food is short or as a result of direct competition. This chapter looks at the first possibility . I examine the phenology of the most important flowering and fruiting species on Kapiti I . in relation to hihi feeding and breeding behaviour. 5.2 METHODS 5.2.1 Study area The study took place on Kapiti I. Nature Reserve on the southern part of the west coast of the North Island (Figure 1 . 1 ). Kapiti I. vegetation is mainly composed of regenerating forests and shrublands (Fuller 1 985) . The distribution of the different vegetation areas is shown in Figure 5 . 1 . The highest point on Kapiti I . is 5 1 0 m a.s . l . Figure 5.2 shows the main elevations on the island. Hihi were released in 1 99 1 and 1 992 in the Kahikatea and Te Rere catchments. In 1 992 birds were also released in the Kaiwharawhara catchment (Castro et al. 1 994b; Chapter 2). Today, however, all surviving birds live in the area between the margins of the Kahikatea and Te Rere catchments. While most of the nesting sites ofhihi are located in the valleys, where the largest trees are, hihi feed throughout the area (Chapter 6). According to Fuller ( 1 985) the vegetation of Kapiti I. has been considerably affected by the cloud that forms over the forest at altitudes above 400 m a.s . l . . Also, on coastal and south facing slopes the vegetation has been affected by both salt deposited during southerly gales and the wind itself. The Kahikatea, Taepiro and Te Rere catchments are those most affected by cloud, with increased precipitation, colder temperatures and fewer sunshine hours than the forests below this zone. Grasslands ? Shrub/Grasslands ? Muehlenbeckia vinelands Flax/Tussocklands Shrublands and Scrub Leptospermum dominated communities Pseudopanax arboreus dominated communities Melicytus ramiflorus forest Corynocarpus laevigatus forest ? Dysoxylum spectabile forest Beilschmiedia tawa!Elaeocarpus dentatus forest ? Weinmannia racemosa/Beilschmiedia tawa forest Ifl Metrosideros robusta forest Figure 5. 1. Map showing the different vegetation zones present on Kapiti Is land . Map from the New Zealand Land Inventory ( 1 987). 59 60 ? m il/111 1 1 1 1 1 Ca t c hmen t B oundar i es M a i n S treams o f each Catchment Tri g s and Spot H e ights C l o ud Z o n e Coa s t a l Z o n e Wha r e k o h u M etres Figure 5.2. Map of Kapiti Is land showing the max imum elevations. From Fu ller ( 1 985). 5.2 .2 Phenology Fruiting and flowering information for 28 forest plants was recorded from April 1 992 to July 1 994 (Table 5 . 1 ) . Additional information was obtained from direct observation of birds fee ding on flowers and/or fruits for the same period of time. Table 5. 1. Sampling scheme to detennine the phenology of plants important as h ih i food on Kapiti I . Further infonnation was gathered from observations ofhih i feeding. Plant name Elaeocarpus dentatus T Geniostoma rupestre SS Knightia excelsa T Melicytus ramiflorus ST Metrosideros excelsa T Metrosideros robusta T Myoporum laetum T Myrsine australis T Myrsine salicina T Pittosporum eugenioides T Pittosporum tenuifolium T Pseudopanax arboreus T Pseudopanax crassifolius T Vitex lucens T Weinmannia racemosa T Carpodetus serratus ST Hedycarya arborea ST Beilschmiedia tawa T Prumnopitys ferruginea T Entelea arborescens T Alectryon excelsus T Kunzea ericoides T Corynocarpus laevigatus T Coprosma grandifolia LS Oysoxylum spectabile T Fuchsia excorticata T Metrosideros fulgens T Pittosporum crassifolium T Month sampled AMJJASONDJFMAMJJASONDJFMAMJJ Part used Fl FI ,Fr Fl F I ,Fr Fl Fl FI , Fr Fr Fr Fl Fl F I ,Fr Fr Fl Fl Fr Fl None None None Fl FI ,Fr Fr F I ,Fr FI ,Fr F l Fl LS = large shrub; SS = small shrub; ST = small tree. T = tree. Fl = Flowers. Fr = Fruits. 5.2.2.1 Criteria for selecting plants To ensure that all the plants were capable of producing flowers and fruits selection fol lowed these criteria: I . Trees: dbh > I 0 cm. 2. Small trees: dbh > 5cm. 3. Shrubs: height > 2 m. 4. Small shrubs. Geniostoma rupestre: height > 1 m. 5.2.2.2 Data collection In 1 992 I obtained the phenology data by fol lowing a transect established by Ron Moorhouse (Figure 5.3) . Several trees of each species were marked along this transect. Although this allowed continuity in the sampling, I felt the transect did not 6 I 62 represent completely the vegetation of the area. To improve this, in 1 993 and 1 994 I ran transects fol lowing any combination of a series of tracks between T20 and T60 (see Figure 5 .3 ) . I walked the tracks stopping every 1 0 minutes to sample one plant of each selected species, if present, until at least 1 5 plants per species were sampled. I searched for plants in a 5 m radius from the sample point. When more than one plant was present I measured the one closest to the sample point, but also looked at neighbouring plants and recorded any differences in their phenologies. The tracks used to study the phenology ran through all forest types and all altitudes (Figure 5 .2). 0 Figure 5.3. Map of Kapiti Island showing all tracks used for the phenology transect. Dotted l ines represent tracks. Lines represent streams. The thick l ine corresponds to the transect established by Ron Moorhouse (In prep.). During sampling sessions from 1 992 to 1 994, I estimated for each plant the number of flower buds, flowers, green fruits and ripe fruits using the following scale: 0; 1 -5 , 5 - 1 0; 1 0-50, 5 0- 1 00; 1 00-500, 500- 1 000; 1 000-5000, 5000- 1 0000; 1 0000-50000, 50000- 1 00000; 1 00000-500000, and 500000- 1 000000. To provide an objective basis for these estimates I selected three plants of each species and counted the numbers of fruits or flowers per inflorescence, umbel, or branch (depending on the structure of each particular species). This gave a total number of fruits or flowers per tree. Subsequently I estimated the numbers by sight. Binoculars were used to sight fruits and flowers when it was not possible to see them with the naked eye. When calculating the monthly total number of flower buds, flowers, green and ripe fruits per species I used the arithmetic means for each interval, ie. 3 (for interval from 1 -5) , 7 .5 (interval from 5- 1 0) , etc . . 5.2.3 Climatic records Weather data were obtained from the New Zealand Meteorological Service, Wel l ington. Records are from Paraparaumu Beach, a small town located on the mainland 5 .6 Km from Kapiti I. Average monthly records were used for all calculations. 5.2.4 Feeder visitation From September to Apri l each year there were up to three feeders suppl ied with food (Chapters 2 and 6). These feeders were watched for at least one hour per day during this period and the presence of individual birds was recorded. I obtained a percentage of visitation by dividing the number of birds visiting the feeder by the total number of birds known to be alive and then multiplied that by 1 00. During the rest of the year there was a single active feeder at the ranger's house. Information on feeder use by hihi was provided by the ranger. Data were treated in the same fashion as for September to Apri l . 5.2.6 Hihi behaviour and energy requirements Hihi were located by sight and sound. When possible, an individual was fol lowed until its sex, band combination, and feeding activity were recorded. Feeding activities were classified as fruit eating, nectar feeding, and invertebrate feeding, which included both gleaning (picking invertebrates from leaves and bark or perhaps l icking insect and plant exudates) and hawking (catching insects in the air). Whenever possible specific food items were also identified. Because the number of plants available for hihi at a given time is small and because the birds tend to move from one plant to another within a short period of time the observations were recorded as visits to feed on a plant species, rather than units of food consumed. I assumed that the number of times birds were seen feeding on a particular plant represented the true contribution of such a plant (fruits or flowers) to their diet. Similarly, every month I counted the times birds were observed hawking or gleaning. I assumed that if birds hawked or gleaned more often in a particular month invertebrates were making a relatively larger contribution to their diet. I assumed that the percentage of invertebrates in the diet is proportional to the frequency of observed hawking or gleaning. I present the data on diet as a percentage of feeding observations. I obtained this percentage by adding all observations of a particular feeding activity for all known individuals each month and dividing this value by the total number of observations of the three different feeding activities. 63 64 5.3. 1 Phenology 5.3. 1 . 1 Flowers 5.3 RESULTS There were species flowering in all months of the year (Figure 5 .4). The largest number of species in flower occurred from August to January in 1 992-93 and from July to January in 1 993-94. This period corresponds to late winter, spring and early summer, with the peak of species number in October 1 992 and September and October 1 993, the equivalent to mid spring. The lowest number of species occurred from February to May ( late summer to autumn) with lowest numbers in April or May (autumn). 1 0 9 .._ 8 Q) ? 7 .2 - c 6 ({) Q) 5 "(3 Q) 0.. 4 ({) 0 3 0 z 2 0 A M J J A S O N D J F M A MJ J A S O N D J F MA M J J 1 992 t 1 993 t 1 994 Figure 5.4. Number of species in flower each month on Kapiti Is land based on I 7 species of plants used by hih i . Hihi visited the flowers of a larger number of species than those selected for the phenology study. Figure 5 . 5 includes all species used by hi hi during the period of study. The number of plant species in flower visited by hihi was usually less than the number of species available (Figure 5 . 5 ) but these two variables were positively correlated (r=0.686). Hihi seem to favour some flowers over others. The choice of these flowers was consistent from one year to the next (Appendix 1 ) . I recorded hihi feeding on a further 1 9 plant species not included as main food items in the phenology study (Table 5 .2) . 1 2 1 0 (f) 8 .._ Q) .0 E 6 :J z 4 2 0 ? F lowers e Birds A M J J A S O N D J F M AM J J A S O N D J F M A M J J 1 992 t 1 993 t 1 994 Figure 5.5. Comparison between the number of species in flower and the number of such spec ies visited by h ih i on Kapiti Is land - 1 992- 1 994. Abundance information comes from Angehr ( 1 984). Table 5.2 Additional species used by hihi as food plants. Kapiti I . 1 99 1 to 1 994 . When used by How common Species Part of plant used hi hi on Kapiti Aristotelia serrata Flowers and Fruits Each year Rare Astelia spp. Flowers and Fruits Each year Abundant Chamaecytisus palmensis Flowers Each year Unknown Clematis vitalba Flower Once Unknown Coprosma lucida Fruits Each year Common Coprosma robusta Fruits Few times Occasional Earina autumnalis Flowers Once Unknown Griselinea littoralis Flowers and Fruits Each year Rare Hoheria popu/nea Flowers Each year Rare Laurelia novazelandiae Flowers Each year Unknown Macropiper exe/sum Fruits 1 992 Abundant Maire tawaki Flowers Each year Rare Metrosideros perforata Flowers Each year Unknown Passiflora tetranda Flowers and Fruit Each year Unknown Phormium tenax Flowers Each year Common Pittosporum cornifolium Flowers Each year Rare Rhopalostris sapida Flowers 1 992 Common Ripogonum scandens Flowers and Fruits Each year Unknown Rubus cissoides Flowers and Fruits 1 993 Common Overall, the flowering season was one month earlier in 1 992 than in 1 993 (Figure 5 .6). Hihi breeding seems to follow this pattern as wel l : egg laying started in November in 1 992, while in 1 993 it started in Qe._f'-""'b:!-I(Castro et al. in press. ; Chapter 3 ) . Of the 17 species examined, seven flowered later in 1 993 than 1 992, two flowered earlier and the remaining species flowered during the same month (Appendix 2). The length of the flowering season also varied from year to year. In sum, Pseudopanax arboreus and Elaeocarpus dentatus had standing flowers for a month longer in 1 993 than in 1 992. On the other hand Metrosideros excelsa, Pseudopanax crassifolius, Pittosporum crassifolium, Pittosporum eugenioides, Pittosporum ------ ---- --- 65 66 tenuifolium, and Weinmannia racemosa had shorter flowering seasons in 1 993 when compared to 1 992 (Appendix 2) . 1 0 .... 8 (!) $ 2 c 6 ({) (!) '[5 (!) Cl.. 4 ({) - - - , ' ' "' ' ' "' ' ? "' ' ' ' ' ' ' ? 0 ' ci z 2 "' 0 "' "' --A-- Flower-92 - - ? - .. Flower-93 ? Flower-94 Figure 5.6. Comparison between the number of species in flower in 1 992, 1 993 and 1 994, out of a selection of 1 7 species preferred by hihi on Kapiti Island. 5.3. 1 .2 Fruits Ripe fruits were present in all months of the year (Figure 5 . 7). The lowest number of species in fruit was in June 1 992 and March to June 1 993 . The largest number of species in fruit was in January 1 992 and September 1 993 . 7 6 ... ....., ::J ? ,_ 5 ?. ? .. - c ? ? ? ? ? ({) .. ? ? (!) 4 ? -. ?u (!) .. Cl.. 3 ?? ({) 0 ? ? ? ? ? 1 992 0 2 z -o 1 993 1 994 0 A M J J A s 0 N D J F M Figure 5. 7. Comparison between the number of species in fruit in 1 992, 1 993 and 1 994 on Kapiti Island. The eight species were selected from those preferred by hihi . Hihi took fruits of plants other than those selected for the phenology study. Figure 5 .8 includes all of the species hihi used during the study. Hihi fed on fruits each month with the exception of May and November 1 992 and January 1 994 (Figure 5 . 8). There was a positive correlation between the number of species available and the number visited by hihi (r=0.799) . 1 0 (/) 8 Q) u Q) Cl.. 6 (/) ...... 0 ..... Q) 4 ..0 E ::J z 2 1 993 ? Fruits o Birds Figure 5.8. Comparison between the number of plant species in fruit and the number of plant species used by h ih i on Kapiti Island, 1 992- 1 994 . The number of fruits or flowers borne by a species varied greatly from year to year (Appendix 3) . Overall , there were more flowers on the trees of most species in 1 992 than in 1 993. In 1 994 I sampled only until July, but by then there were obviously many buds in some of the species, suggesting a good flowering season ahead. There were three species that had more flowers in 1 993 than in 1 992 (ie. against the general trend) - Pseudopanax arboreus, Elaeocarpus dentatus and Metrosideros excelsa. Melycitus ramiflorus produced very few flowers in both 1 992 and 1 993, but had a large number of flowers in 1 994. Fruits often represent the success of the previous flowering season. Both Melycitus ramiflorus and Pseudopanax arboreus had large number of fruits in 1 992, suggesting a good flowering season in 1 99 1 . 5.3.2 Influence of climate on Phenology There was a Jag between each c limatic parameter measured and the number of species flowering. The largest number of species in flower was found in both years four months before the hottest period (Figure 5 .9). The largest number of species in flower was recorded five months before the greatest number of sunshine hours was registered in 1 992, but only four months earlier in 1 993 (Figure 5 . 1 0). 67 68 -o-- Flower ) 1 8 ? Temperature (C) I 1 2 1 6 z 1 0 ? Q: 1 4 8. 8 (/) c .Q 60 '" ? 50 Q) V> .D 0 40 Q) Ol "' 30 c Q) u Q; 20 a.. 1 0 0 s 0 N 1 992 1 00 80 60 40 ?:. 20 1.1 ?? :: :;: ?: ?i ?: ??; n ''' ?= .:? :;. 0 s 0 N 0 1 993 11( J F t ? ??? . , ?, .. J F t ? 1: .:: ;:? k Flowers Fruits Invertebrates ?-ll, "0 Cl> Cl> ...... +-' 60 ro (f) "0 I- :0 40 +-' c Cl> u I- 20 Cl> 0... 0 S ON D J F MA M J J A S OND J FM AMJ J A S ON DJ F MA M J J A 1 99 1 ? 1 992 ? 1 993 ? 1 994 Figure 5. 12. Monthly percentage ofhihi visiting the feeder on Kapiti Island. 1 99 1 - 1 994. 5.4 DISCUSSION This phenology study shows considerable variation in initiation and duration of fruiting and flowering for the species studied. I t also shows between-year differences in flower and fruit numbers. Both these factors could be of great importance to hihi survival ? on Kapiti I . Several studies have looked at the relationship between nectar avai lable or inflorescence density and abundance ofhoneyeaters in Australia. Most studies have found a positive correlation between them (Collins and Briffa 1 982; Coll ins and Newland 1 986; McFarland 1 986) although honeyeater abundance should also be related to other habitat factors ( eg. avai labil ity of nesting sites) . In continental areas honeyeaters can move long distances looking for food, but in an island situation they have a more l imited food supply. Ford ( I 99 1 ) found that honeyeaters either left the area or lost weight in periods of nectar shortage during the flowering season of Banksia spinulosa. To i l lustrate the effect of the difference in numbers of flowers per tree between the years I wi l l use Pseudopanax arboreus. In Chapter 4 I reported that an average flower (open to visitations) in a tree contains 0 .83 calories. In I 992 an average tree could have provided 1 24,500 calories in nectar while in 1 993, the average tree could have provided from 298,800 to I ,494,000 calories, ie. from 2 .5 to I 2 times more calories. In other words, in I 992 an average tree could have maintained the dai ly metabolic requirements (Castro and Robertson submitled; Chapter 4) of four hihi, but in 1 993 it would have maintained from nine to 4 7 hihi. These calculations do not take into consideration competition for resources, but they give an idea of the impact of a year with low flower numbers. Most plant species studied had more flowers in I 992 than 1 993, so the overall number of calories on Kapiti I. must have been much greater in I 992 and the birds feeding on these flowers should have benefited. In 1 992, the percentage of observations of hihi feeding on flowers was 40-80%, but only 20-60% in I 993, suggesting that hi hi might have responded to the flower bonanza (Figure 5 . 1 I ) . Species l ike Melicytus ramiflorus which provides food to hihi in the way of both fruits and flowers could have a marked effect on the birds' survival. In 1 992 and 1 993 M ramiflorus bore small numbers of flowers and as a result there were very few fruits in I 993 and 1 994. M ramiflorus fruits are a preferred food. Often by fol lowing hihi, I found isolated trees that had a few fruits. There is some evidence that fruits can be taken by honeyeaters as replacement for nectar. B . Lee (pers. comm.) has found that for some Coprosma species the calorific content of fruits is similar to that of flower nectar. Angher ( 1 984) found that on Little Barrier I . hi hi used sl ightly more species of plants for their fruits (34) than for their flowers (27). On Kapiti I. I found the opposite relation (33 species of flowers and I 7 species of fruit) . Also, in contrast with Little Barrier 1., Kapiti hihi do not have a clear seasonal pattern when feeding on fruits. Although Kapiti I. has 2 1 of the 27 fruit species l isted by Angher, 8 of them are rare. It is possible that on Kapiti hihi are more dependent on nectar because of the lack of preferred fruit species. Since tui and bellbird are more nectarivorous than hihi (Gravatt I 970; Angher I 984, Rasch 1 985) competition for this resource could be great. The vegetation on Little Barrier I sland comprises mature forests including large numbers of trees of nectar-rich species such as Metrosideros excelsa, Vitex lucens, Sophora spp. , Fuchsia excorticata, and Metrosideros robusta. Furthermore, the understorey of the Little 7 1 72 Barrier I. forests contains a large number of species where the fruits and flowers are used by hihi. This contrasts markedly with the situation on Kapiti I . , where the forests are relatively young, and do not have the diversity of species present on Little Barrier I . . Egg laying i n hihi coincided with a one-month shift in flowering for most plant species. Hihi visited the feeders only during the breeding season. Energetic demands for both male and female birds increase during the breeding season (Paton 1 982) . Hihi visitation to the feeder could be a reflection of an increased metabolic requirement that can not be fulfi l led by feeding on natural foods. Of the thirteen species that flower from September to January on Kapiti, five are rare (Fuchsia excorticata, Metrosideros excelsa, Myoporum laetum, Pittosporum crassifolium, Vitex lucens), two are restricted in distribution (Metrosideros robusta, Weinmannia racemosa), one provides low energetic rewards ( Geniostoma rupestre ), and for three most of the tree stands are young (Pittosporum eugenioides, Knightia excelsa, Pseudopanax crassifolius) and flowering and fruiting occur only in the older trees. This leaves only two species Elaeocarpus dentatus and Pseudopanax arboreus from which the three species of honeyeaters and kaka could obtain nectar during the breeding season. Tui, bellbirds and kaka breed during the same period as hihi, and competition for food should be very intense this time of the year. Furthermore, on Mokoia I sland I found that hihi nesting c lose to a feeder fledge more young and heavier young than hihi nesting more than 200m from a feeder, suggesting that food constrains nesting success (Table 5 .3) . Table 5.3 . . Resu lts from eight h ih i nests on Mokoia I sland. SM=Social Monogamy; Py=Polygyny; FA=Femele Alone. * * Nests heavi ly infested with ectoparasites. Nest Mating Py 5 SM 47 Py 45 Py Average 39 FA 54 FA 36 SM 24 SM Average Dist. to Age at Average No. of No. eggs No. eggs feeder weighing weight fledglings hatched 3 21 42.875 4 5 4 1 00 24 41 .4 5 5 5 20 21 42.625 4 5 5 5 23 37.84 3 ?? 5 5 41 .185 4 5 4.75 700 21 38.3 2 4 3 500 24 39.6 3 4 4 500 21 27.5 3** 4 3 250 24 33.5 2 5 4 34.725 2.5 4.25 3.5 When Angher ( 1 984) compared the vegetation of the islands proposed for translocations of hihi, he counted the total number of species known or suspected to be used by hi hi, and compared that with the number of those species present on each of the islands. For instance, he found a total of 1 25 species used by hi hi. Of those I 02 exist on Little Barrier I . and 75 on Kapiti I . . Although he offered estimations of the abundance of those species, the values were not used to qualify the habitat avai lable for hihi on each island. Thus Angher ( 1 984) estimated that all broad leaf forests provided good or medium quality habitat for hihi without considering the composition of such forests. Finally he determined the monthly number of plant species in fruit or flower on each island using phenological information from Little Barrier 1.. He assumed that phenology would be the same in al l locations. This study shows the great variabi lity in time and duration of fruiting and flowering within an island. It also shows relationships between temperature and total sunshine and flowering. Comparisons between areas should be made bearing in mind the difference in plant composition and latitude. Although Kapiti I. contains a relatively large number of species when compared to the other islands proposed for hihi translocations, its forests are young and many species preferred by hihi are rare. Tui and bellbird populations on Kapiti are large while kaka populations seem to be at carrying capacity levels (R. Moorhouse, pers. comm.). Hihi might never reach large numbers on Kapiti I . because the island might be at or near its carrying capacity for nectar/fruit eaters. This is accentuated by the highly variable food avai labil ity : years when fruits and flowers are short could be detrimental to the small hihi population. There could be negative effects from good flowering and fruiting years as wel l , if those years favour competing species. In 1 993-94 kakariki (Cyanoramphus novaezelandiae), a native cavity nesting parakeet, were found fighting for nest sites with hihi (Castro et al. 1 994; Chapter 2). Hihi laid only one clutch in 93- 94. I t is possible that a good fruiting season in 1 992 coupled with a very mi ld winter in 1 993 might have allowed a large number of kakariki to survive to the breeding season and this was reflected in a greater degree of competition for nest sites. There is need for experimental studies to determine the relationship between food availabil ity and hihi survival . Research on competition with other species for various resources would allow us to interpret information on food abundance. Such studies might be necessary on each island receiving translocated hihi . 73 74 5.5 REFERENCES Angher, G.R. 1 984. Ecology and behaviour of the stitchbird. Internal report to the New Zealand Wildlife Service, Wel l ington. Castro, I . , Minot, E.O. & Alley, J.C. 1 994a. Feeding and breeding behaviour of hihi or stitchbirds (Notiomystis cincta) recently transferred to Kapiti Island, New Zealand, and possible management alternatives. Reintroduction Biology of Australasian Fauna. Surrey, Beatty and Sons, Healesville, Australia. Castro, I . , Alley, J .C. , Empson, R.A. & Minot, E.O. 1 994b. Translocation of hihi or stitchbird (Notiomystis cincta) to Kapiti Island, New Zealand: transfer techniques and comparison of release strategies. Reintroduction B iology of Australasian Fauna. Surrey, Beatty and Sons. Australia. Collins, B .G. & Briffa, P. 1 982. Seasonal variation of abundance and foraging of three species of Australian honeyeaters. Australian Wildlife Research 9: 557-569. Collins, B .G. & Newland, C . 1 986. Honeyeater population changes in relation to food availability in the Jarrah forest of Western Austral ia. Australian Journal of Ecology 1 1 : 63-76. Craig, J .L. , 1 984. Wing noises, wing slots, and aggression in New Zealand honeyeaters (Aves: Meliphagidae). New Zealand Journal ofZoology. 1 1 : 1 95 -200. Ford, H.A. 1 99 1 . Coping with an erratic nectar source - eastern spinebi l ls Acanthorhynchus tenuirostris at New England National Park. Emu 9 1 : 53-56. Fuller, S . 1 985 . The vegetation of Kapiti Island. Department of Lands and Survey. Miscellaneous publications. Wellington. Gravatt, D.J. 1 970. Honeyeater movements and the flowering cycle of vegetation on L ittle Barrier Island. Notornis 17: 96-1 0 1 . Lovegrove, T. 1 986. Stitchbirds on Kapiti Island 1 983-1 985 . Internal report to the New Zealand Department of Conservation, Wellington. McFarland, D.C. 1 986. Seasonal changes in the abundance and body condition of honey eaters (Meliphagidae) in response to inflorescence and nectar availability in the New England National Park, New South Wales. Australian Journal of Ecology 1 1 : 33 1 -340. Paton, D.C. 1 982. The diet of the New Holland honeyeater, Phylidonyris novaehollandiae. Australian Journal of Ecology. 7 : 279-298 . Rasch, G . 1 985 . The behavioural ecology and management of the stitchbird. MSc. Thesis, University of Auckland. Rasch, G. 1 99 1 . A recovery plan for the stitchbird (Notiomystis cincta). Internal report to the New Zealand Department of Conservation, Wellington. Chapter 6 Chapter 6 Monitoring and management of hihi (Notiomystis cincta) on Kapiti Island, New Zealand ABSTRACT There were three major management strategies fol lowed to support hihi (Notiomystis cincta) translocated to Kapiti Island in 1 99 1 and 1 992. 1 .) Feeders were established to supplement the natural food of hihi, and the major nesting areas were provisioned with nesting boxes. 2. ) All released birds were individually marked and monitored to learn about the use of resources and survival . 3 . ) A phenological study was done to investigate the avai lability of foods. This paper presents the results of hihi management during 1 99 1 -94 on Kapiti Island and suggests a monitoring program. 6 . 1 INTRODUCTION THIS Chapter is aimed at constructing a rationale for monitoring as part of the management of translocated species. With the number of endangered and threatened species increasing worldwide, ecological studies are emphasising management, as much as biological, aspects of research. In the New Zealand archipelago, where the three main islands are inhabited and have several introduced species of predatory mammals, the only hope for most endemic species is to be translocated to offshore islands (Merton 1 975; Flack 1 978; Nil lson 1 978; Rasch 1 988 ; Roberts 1 99 1 ). Some islands are sufficiently small that unwanted predators can be eradicated (Taylor and Thomas, 1 990; Veitch and Bell 1 990). On larger islands predators can be successful ly control led until ski l ls and technology allow their removal (Veitch and Bell 1 990). Most of the small offshore islands have been modified by people, but their size also permits plant restoration. Translocation of species is not new, since most human colonists have introduced species from their land of origin. Translocations have also been done to restore native fauna extirpated by hunting (Griffith, et al. 1 989). However, only recently has it been used as a tool to save rare native species. Study of the ecology of a species can teach us about its general requirements at a particular site. This important knowledge allows decisions about possible translocation sites and informed criticism of shortcomings of translocation efforts. For most species, however, it is difficult to find translocation sites that are identical to localities where the species is known to do best. Furthermore, for most endangered and threatened species I can not be sure that the habitat it presently occupies is the one where it does best. L ittle is known about the effects of mobilisations on animal behaviour (Conant, 1 988) and until now there has been no monitoring of released organisms (Young, 1 990). This lack 75 76 of monitoring (or lack of report of the outcome) has led to situations where the success or failure of translocations can not be properly analysed. There are no standard ways to determine the success of translocations. In many cases the success of a translocation is known only if, after some time (often years), the released species has become common enough to be seen regularly or so rare it is not seen at all . In some cases translocation is successful if the translocated species breeds successful ly during the season fol lowing release. Even if breeding occurs post-reproductive events might impede the establishment of the species. These could include the depredation of young, cold weather, fire, or too few offspring to establish a population. Scott and Carpenter ( 1 987) consider that the ultimate measure of translocation success is the establ ishment of a self-sustaining population. Careful long term monitoring is, therefore, necessary to show the ways in which species respond to translocations (Scott and Carpenter 1 987). Careful monitoring aids a programme of species establishment by indicating problems the species face. I n extreme cases where the species does not survive in the new locality the monitoring helps in the review of future sites for translocation. Monitoring can also reveal behaviours of biological interest and provide ideas for scientific investigation. Monitoring of a species involves more than simply recording that individuals of the species are al ive and breeding. In the case of birds which depend on plants for food and shelter, monitoring seasonal plant production (phenology) and the weather, should be included. 6. 1 . 1 The h ih i : a case study In this chapter I first discuss the results of a three-year study of translocated hihi (Notiomystis cincta) to Kapiti Island (Castro et a/ 1 994;. Figure 1 .3 ; Chapter 1 ) . The results are applicable to long term monitoring of the species. Secondly, I describe a monitoring program designed to help managers evaluate the health of the population and make sound management decisions. Hihi is a species of honeyeater endemic to New Zealand. There is a single population of this species on Little Barri?r I sland in the Hauraki Gulf. The New Zealand Department of Conservation is trying to establish self-sustaining populations of this species elsewhere to ensure its survival. Hihi were released on Kapiti I sland in 1 983, 1 985 and 1 990 (a total of 72 birds) but in each case although the birds bred during their first season (and in some cases in later seasons) they slowly disappeared in subsequent years. Lack of monitoring resulted in no information about why the birds disappeared. I n 1 99 1 and 1 992 The New Zealand Department of Conservation arranged two further transfers, each of which was monitored intensively. The fol lowing document on monitoring and management is based on the three years of monitoring surrounding these transfers. I wil l now discuss in sequence three major strategies for supporting and monitoring translocated hihi. The strategies are : I . Supplementary feeding 2. Measures of survival and breeding 3 . Plant phenology assessment. 6.2 1 . Supplementary Feeding 6.2. 1 METHODS Feeders were placed on Kapiti Island at each hi hi release site in 1 99 1 , and at the release sites and in every major catchment from the Kaiawharawhara to the Te Rere in 1 992. Four weeks after each release all feeders were closed except the feeder near the ranger's house. The feeder at the ranger's house was provided with food year round. Observations were carried out at this feeder fortnightly during the winter months, and several times per week during the breeding season (Sept-Mar). Observation times ranged from 30 min to 1 2 hours. The food offered at the feeder for four weeks after the transfers comprised nectar mix, jam water, honey water and sugar water (Lovegrove and Veitch 1 994). After these periods the food used was mainly sugar water. On rare occasions the birds were offered honey water. 6.2.2 RESULTS The feeder structure used until 1 994 is shown in Figure 6. 1 . The protective wire mesh was necessary because kaka Nestor meridionalis and tui Prosthemadera novae? zelandiae are also attracted to sugar water, and are dominant over hihi . Food was provided in small glass jars. This type of feeder and jar proved hard to clean. The main problems were: - growth of fungus on the wooden base and frame caused by spilled food, - bird faeces accumulating on the wooden base and contributing to fungal growth, - birds perching on the edge of the jars to drink, after walking on the dirty wooden base. This could have caused some contamination of food, - birds spilled food on the wooden base because the jars were open, - al l birds used the feeder and jars at the same time, providing opportunities for stressful interactions and transmission of disease (M. Friend, pers. comm.), - dominant birds were able to monopolise the whole feeder. 77 78 Figure 6. 1. Feeders used to prov ide supplementary food for h ih i on Kapiti I s land until 1 994. Use of the feeder was strongly seasonal (Figure 5 . 1 2 ; Chapter 5) with frequent visits by birds during the breeding months, but no birds using it in late autumn and winter. Greater activity at feeders coincided with laying of eggs and fledging of first clutches. Second clutches have not been successful on Kapiti, with the exception of a female breeding near the feeder in 1 99 1 -92 and possibly a female in a polyandrous system at the Te Rere in 1 992-93 . In 1 99 1 -92 and 1 993-94, all known Kapiti pairs brought their first clutch offspring to the feeders. In 1 993-94 parents remained with fledglings in the nesting areas and there were no. attempts at second clutches. Observation of birds at feeders show a high level of interaction. Dominant birds monopolised the feeder at low bird densities, but defence became unprofitable once numbers were high. However, at high densities mating activity increased and, I believe, male stress as wel l . Females might also be stressed by this, because frequently several males chased a single female. At each feeder a different dominance hierarchy was found (ie. a dominant male or female at one feeder could be submissive at another) . The percentage of birds visiting the feeders has decreased between 1 99 1 and 1 994. In 1 99 1 -92 up to 1 00% of the birds known to be alive used the feeder, but in 1 993-94 only 7 1% used it. This difference might result because young birds, produced the season before, are less incl ined to use the feeder. 6.2.3 DISCUSSION The seasonal use of the feeder could reflect increased energetic demands and/or low avai labi lity of food of sufficient quality or quantity. The phenology transect shows clearly that the food available on Kapiti during this period is produced by plant species that are relatively rare, but are also preferred by breeding tui and bellbird Anthornis melanura. The only successful second clutch known to have been produced was raised near the ranger's house, which suggests that supplementary feeding might be important to reproduction. More significant however, is that in both seasons where second clutches were laid the first brood of fledgl ings was brought to the feeder. Possibly in 1 99 1 -92 and 1 992-93 there was sufficient food for incubating females, but not enough to al low them to lay a second clutch and rear them. By bringing chicks to the house, however, parents have a better chance. In 1 993-94 food was low and no females attempted to lay a second time. Work on Mokoia Island indicates that food limits breeding (Table 5 .3 ) . 6.2.4 RECOMMENDATIONS I recommend the use of the feeders during the breeding season with the following modifications: 1 . Three to four feeding stations should be established in each of the three main catchments where hihi are found. With this strategy the pressure on individual feeders will decrease as fewer birds will use each one. Also birds will be able to feed closer to their breeding grounds, and thus enhance their reproductive success. Finally, juvenile birds might be more likely to use the feeders because there are more of them around and they are closer to breeding sites which are often visited. 2. Feeders should be opened in September. Several studies have shown the importance of good quality food during the breeding season (Lack 1 950; Paton, 1 982; van Riper I l l 1 984; Davies and Graham 1 99 1 ; Castro pers. obs. Mokoia Island). 3 . Feeders should be closed when no more breeding activity is detected. An exception to this is the feeder at the ranger's house. Here food should be available all year. 4. Feeders and nest sites should be monitored. Without monitoring it would be impossible to determine the effect of supplementary feeding on the birds. Only through monitoring will it be possible to determine when it is safe to close the feeders ( ie. birds are not breeding any more). Moreover, as with the phenology transect, constant monitoring will help to identify years when the birds are making greater use of the feeder. 5 . Rough estimates of food consumption should be recorded. This wil l allow comparison of use between years, and a measure of how much the project costs. 79 80 Feeders are good places to mistnet and band birds. They are also key places to observe disease in birds and assess survival. Supplementary feeding is, however, expensive and not sustainable in the long run. It should be done only if necessary. To evaluate the effect of supplementary feeding on Kapiti I recommend the continuation of supplementary feeding and strict monitoring of the birds for three more years (ie. until 1 998) . After that time feeding should be suspended, but monitoring of breeding continued, and data compared. I f the effects of the suspension of food are very clear in the first year, (ie. lots of birds die or the birds fai l to breed) feeding should be resumed immediately. I n 1 994 the feeders (described in the Methods section) were changed for the model obtained from Perky Pet Products (Figure 6.2). Figure 6.2. Hummingbird feeders made by Perky Pet Products. These feeders, designed for hummingbirds, were modified to make them suitable for the larger hihi. The benefits of these feeders include: I . No spil lage of food since the birds have a very small opening through which they can drink. 2. The food is protected from the outside and birds' feet do not contact it. 3 . They can be located separately, thus avoiding crowding and monopolisation of food by dominant birds. To prevent tui and kaka from using the feeders, a cage of coarse wire netting could be built around them. The final feeder (Figure 6.3) would be bottomless, so that no faeces accumulate in it. The whole station can also be moved easily so that faeces do not accumulate on the ground under it. Because the feeder is used only during the breeding season better quality food is recommended. The Wellington Conservancy of the NZ DoC has already started to use Womberoo?, which is complete food mix for birds. For information on nutritional contents of this product and where to obtain it see Appendix 4. During the non-breeding season sugar water would be sufficient. Figure 6.3. Wire netting around the feeder can prevent kaka and tui from using the feeders. 6.3 2. Measures of Survival and Breeding 6.3. 1 METHODS Survival of hihi on Kapiti Is land has been measured by closely following the birds on foot, observing the feeders, and by searching and monitoring breeding sites. I searched the area between tracks 20 and 60 every other week from April to October recording the colour bands of all birds observed. From October to March I also observed the feeders recording all birds visiting it. I summarised the data as monthly survival. Each year in October I began the search for nest sites. To find a nest site, I first located cal l ing males. I found that males defending a cavity would, on average, call a three note territorial call every 1 7 to 20 seconds. This call was sung from perches located about 30 metres from the actual cavity. Sometimes the male would approach his cavity and briefly go inside it. However, most of the times it was a female who indicated where the nest site was. Females visit several nest cavities while males advertise them so, by fol lowing females, the hole can be found. Females were also less l ike ly than males to be discouraged from entering the cavity because an observer was there. Once the cavity was located I watched it for one to six hours to determine the type and duration of activities, and to identify the birds using the site. Nest boxes were erected in 1 992 prior to the release, to increase the number of nest sites available. In total 1 8 boxes were placed in the Kaiwharawhara, five in the Kahikatea, eight in the Te Rere and five at Rangatira. In 1 993 some nest boxes were moved near nest areas as part of an experiment This experiment was designed to find out if the birds would use boxes in areas already chosen for nesting and at which of two 8 1 82 heights. Two identical boxes were located at I 0 nest sites . One box was placed at one to one and a half metres and the other at least five metres from the ground. 6.3.2 RESULTS 6.3.2. 1 Survival Between 1 99 1 and 1 992 95 birds were translocated to Kapiti Island (Castro et al. 1 994. Chapter 1 ) . The largest number of birds known to be alive on the island (79) was in August 1 992 after the second release, and the minimum number was 25 in December 1 99 1 (Figure 6.4). 80 Releases 0) 70 .;::;: eo Fledglings c 60 ? ? 0 c Fledglings .::.:: 50 :.c :.c 0 40 0 z 30 20 A SO N D J FMA MJ J A SON D J F MAM J J A SON D J F M 1 99 1 + 1 992 t 1 993 + 1 994 Figure 6.4. Number of birds known to be al ive on Kapiti Is land from 1 99 1 to 1 994 Survival of birds is highly variable (Table 6. 1 ) . On average 5 1 .8% of adult birds and 5 5 .34% of their offspring have survived from one year to the next. Juvenile survival for the first year may be under- or over-estimated since many juveniles are unbanded until they first visit the feeders where they can be caught and banded. Table 6. 1. Percentage annual surv ival of hi hi on Kapiti Is land from 1 99 1 to 1 993 . Recruitment is written in ital ics and surv ival from one year to next in bold. Percentage survival from one year to next in parenthesis. Survival of fledg l ings: Fledge (year) 1 991 9 1 /92 1 6 92/93 93/94 Survival for b i rds released in 1 991 : August 91 48 November 91 23 (48) Survival for birds released in 1 992: August 92 47 November 92 26 (55) 6.3.2.2 Predation and Disease 1 992 9 (56) 18 November 92 1 6 (69) November 93 1 3 (50) 1 993 5 (55) 1 0 (56) 15 November 93 6 (37) Disease and predation are both factors that are associated with bird survival on other offshore islands in New Zealand. Kapiti I . has both Norway (Rattus norvegicus) and Polynesian (Rauus exulans) rats. Although during this study I did not find hard evidence of predation by any rat, the remains of a sick female who died in a cage on Kapiti, show the tooth marks of a Norway rat. Furthermore, a bird carrying a transmitter was found dead with the tooth marks of a Polynesian rat. The first case suggests that a Norway rat might be able to prey on sick or weakened individuals. In the second case either predation or scavenging can be suspected. Hihi transferred to Mokoia Island on Lake Rotorua in August 1 994 have had a high survivorship. Mokoia I. has no mammalian predators with the exception of mice (Mus musculus) . A single hihi on Mokoia I. even nested in a box left on the ground and successfully fledged three young. In 1 992, a few days after the second release had taken place, I found a male with the symptoms of a respiratory disease. He had partially lost his voice, looked fluffed up, and was less active. A week later he sat close to the feeder gaping (opening and closing his beak) continuously and coughing periodically . During the next few weeks I found two more males and two females with the same symptoms. All birds died or disappeared with the exception of one of the females who survived, but never recovered her voice completely. Post-mortem examination of two of the males showed that the disease was produced by air-sac mites Sternostoma tracheacolum (S. Cork, pers. comm.) . In 1 993 , I found two females and a male with similar symptoms. All three disappeared during the following winter. 83 84 6.3.2.3 Hihi D istribution Hihi have been seen in many places on Kapiti at one time or another. However, the birds have wel l establ ished home ranges where they are relatively easy to find. Figure 6 . 5 shows the location of all known nesting areas in relation to the streams and tracks. The four main catchments where hihi have establ ished on Kapiti are the Kahikatea, the two main streams at Rangatira (R l and R2) , and the Te Rere. Each nest site has a number preceded by the first letter of the name of the stream where it is located. Each nest site can have more than one cavity. For example, nest site K l , in the Kahikatea catchment consists of a natural cavity and two nest boxes all used by hihi. 0 Figure 6.5. Location of al l known hih i nest sites on Kapiti Island. Each dot indicates a nest site. Dotted l ine = tracks. Solid l ine = streams or shore l ine. Male and female differ i n their use of home ranges. A particular male hihi wil l have a stream as a home range including a small preferred area which normally is where he nested the year before. Female hihi also have a home range centred on a stream, but they have preferred "winter" areas that can be quite far from their nesting area. Table 6.3 l ists all known home ranges, preferred areas and nesting areas for each bird. Juveniles become independent two to three weeks after fledging, form groups with adults, and remain in the same area as adults. Table 6.2. Home ranges, wintering areas, and known breeding sites of surviving hihi on Kapit i Island Bird Sex Home range Wintering area Nest s ite A-WY F Kahikatea Bottom Kahikatea K1 A-WR M Wilkinson's Lookout K1 Unb M Unknown K1 WA-RR F Tracks S 1 8,53 & 54 K2 R-WA M SS Feeder Area K2 A-YY F Tracks 56,Trig & 40 K3 AR-G M Top Track 56 K3 BA-R M All over Kahikatea K8 & K6 Unb F Tracks 68 & 55 K6 GY-A F Track 56 K5 & K4 AB-Y M Upper Kahikatea,Whare K4 & K7 Y-BA M Tracks 56,60,55 Unknown WW-A M Ranga ra Track 43,S5 area,Whare R 1 ,2,4,5 AW-G F Track 43, Whare R1 Unb F Track 40 R3 A-BW M Te Rere All over T e Re re, T 40 R3 AB-BY M Lower Te Rere, T 44 R4,5,T1 BW-A F Junction, T 44 Unknown Unb M Tracks 20 & 44 T2 A-WG F T 45,Mck/Wi lk,Kaiwhara Unknown -BA M Tracks R4 & 45 Unknown BA-G M Tracks 44,45 and R4 Unknown GG-WA M Track 45/R3, Whare T4 A-GY M Summit, Track 47 T3 G-BA F Track 47 T5 A-BR F Track 20 Unknown Y-A F Upper Te Rere T6 AW-B F Track R7b Unknown A-GG F Te Rere & Track 47,S5 area K7,R3,R5 Kahikatea AG-GR F Te Rere & Track 56 T2, K6 Kahikatea W-BA F Unknown Unknown T2 6.3.2.4 Visability of H ihi Male hihi were easier to see than females because males gave a territorial call (three note call) or "chip" at any disturbance and had brighter plumage. However, when a male was found, often a female and/or juvenile bird was also sighted. By remaining in an area for at least 1 0 minutes after hearing a male or other birds, I usually observed all birds present. Squeaky callers (obtainable from the Audubon Society or any DoC shops) attracted hihi, enhancing the probabil ity of seeing them. Juvenile and female hihi were particularly attracted to callers during the autumn and winter. 6.3.2.5 Breeding During nest building one or more females brought nesting materials (sticks, ponga scales and feathers) to the cavity. Most activity took place in the morning. After the nest was finished there was a period of about 7 days where the females were rarely seen at the site, except for early morning and late afternoon. After this time the females spent several hours in the morning at the site. From nests in boxes I know this was the 85 86 period of egg laying. Female hihi laid 2 to 5 eggs. Hihi laid 1 egg per day, an hour later each day. For instance female W A-W's first egg was laid at 0700 h on day 1 and her last egg at 1 1 00 h three days later. Incubation started the day the last egg was laid and continued for about 1 5 days. Females did al l incubation alone. In polygynous nests the two females took turns to incubate. Incubation stints lasted 20 min. while the average absence periods were 5 min. Hatching at natural sites was determined by the presence of egg shells beneath the cavity. Brooding (chick rearing) was initiated by the female (s). Males started brooding one or more days after hatching. Brooding times were similar to those of incubation for the first week, but after that absences were longer than visitations. In average, males visited the nest at a lower rate than females. Brooding lasted from 28 to 34 days. Chicks fledged in the morning usually about 1 000 h. In 1 992 and 1 993 fledgl ings remained near the nest cavity (within 1 OOm) for the first few days after which they were moved to the house feeder. In 1 994 the chicks and parents remained near the site (within 200m) for two weeks. Fledglings were fed by both parents for one week. After this time the males alone continued feeding them while the females started second clutches (except for 1 994 when females did not attempt another clutch and they remained with the chicks as well) . This year there was great competition for nest sites between hihi and kakariki Cyanoramphus novaezelandiae (Castro et al. 1 994b; Chapter 2) . Chapter 2 (Table 2. 1 ) summarises the results of the breeding seasons from 1 99 1 to 1 993 . Hihi used the same cavities for nesting every year. 6 .3.2.6 Boxes None of the boxes erected in 1 992 were used that season, or until 1 994. However, of the boxes placed at the nest sites in 1 993, one was used during that breeding season and another one in 1 994. Nest building material was found in a further box at a nest site, and in both 1 993 and 1 994 sticks were found in many kakariki boxes. 6.3.3 DISCUSSION At present most birds are colour banded. It i s important to careful ly record the band combination of each bird sighted because these records enable the survival of b irds of different ages, and the distribution, movements, mate choice, and breeding output of individuals to be determined. Colour banding should continue whenever possible. Banding should be done at feeders early in the season, to avoid mistnetting females while they carry eggs. Survival of hihi on Kapiti I. is low. Recruitment of juveniles in the population ( 1 992-95) is enough to maintain the numbers at around 40 individuals, but the population does not seem to be increasing. I believe that a combination of: 1 . Low food availab i lity during late spring and early summer, 2 . Cavity limitation, 3 . Predation and disease, probably contribute to the low numbers of the species on Kapiti I . Providing food close to nest sites during the breeding season, adding boxes at nest site areas and eradicating the rats from the Island wil l all be important in the management of hi hi. 6.4 3. Plant Phenology Assessment 6.4. 1 METHODS From 1 986 to 1 990, a phenology transect to assess seasonal changes in plant production was designed by Ran Moorhouse (pers. comm.) as part of a feeding study on kaka. I continued this work with few modifications from 1 99 1 to 1 994. The transect was run once a month, usually taking two days and following any combination of tracks between Track 20 and Track 60. I monitored 22 plant species which are important hihi food (Table 5 . 3) . The methodology followed is described in Chapter 5 . 6.4.2 RESULTS The number of species in either fruit or flower varied throughout the year and from year to year (Figures 5 .6 and 5 .7 ; Chapter 5). The fewest species were in fruit or flower during the colder months with sl ight variation between the years (Figure 6.6). The greatest number of species in flower and in fruit was recorded in the spring and early summer (Figure 6.6). There were marked differences between years in the level of flower production (Chapter 5) . For instance Pseudopanax arboreus flowered during the same months in 1 992 and 1 993 with tens of thousands of flowers in 1 993, but only thousands in 1 992. The entire flowering season was a month later in 1 993 when compared to 1 992. 1 2 1 0 Vl Q) u 8 Q) a. Vl 6 .._ 0 0 4 z 2 0 . . ? ? I I 1 1 ? ? --- Fruit ? ? ? ? ? Flower . . . . . ? I ?o ? ? ? ? ? ? I I ? Figure 6.6. Number of species in fruit or flower on Kapiti Island from 1 992- 1 994. 87 88 6.4 .3 DISCUSSION Even from this very short period of time, the results of the phenology transect point to the important aspects of f1ower and fruit availability . On the one hand, the results show that there is a source of food for hihi through the year. On the other hand it shows that the food available varies both in quality (different species available) and quantity (different number of plants and different number of flowers and fruits). Although there is always food present, the quality and quantity wil l determine its use by hihi and other honeyeaters. For instance, in 1 992 hihi had more access to Elaeocarpus dentatus flowers than in 1 993 because the flowers were present in large quantities (Appendix 3 ) and therefore were harder for tui and bellbirds to defend. Larger quantities of flowers probably also meant greater number of calories per bird (Chapter 5) . Another important aspect is the onset of the flowering for the various species. A good example is kamahi. In 1 992 kamahi began flowering in October, but in 1 993 the first flowers were seen in August, two months earl ier. The influence of the availability of this preferred nectar on the birds physical condition prior to the breeding season could be important. The importance of assessing plant phenology can not be overemphasised. Long term results of phenology associated with nesting success and bird survival can help in understanding the dynamics of communities and wil l help in making important management decisions. 6.4.4 RECOMMENDATIONS I recommend the continuation of the phenology transect. The species used in the transect figure in the diet of many birds and the results of the transect will be of wide use on the Island. 6.5 Monitoring program Based on my own monitoring, the following is a monitoring p lan to determine the survival and breeding success of the hihi on Kapiti I. Table 6.3 gives a summary of the plan with the priority I think should be given to each activity. Table 6.3. Monitoring plan for h ih i on Kapiti Is land Activity Monitor breeding (Sep-Mar) Monitor feeders (Sep-Mar) Monitor activity (Apr-Aug) Monitor boxes (Sep-Mar) Priority 2 3 No. people 2 1 2 1 Time involved 1 day / 2 wk's 2-3 hrs I 3 days 2 days I month 1 day / 2 wk's 6.5. 1 Hih i monitoring from Apri l to August To monitor the survival and activity of birds during the non-breeding season (Apr to Aug), I suggest that once each month two people walk the tracks highlighted on Figure 6. 7 . The observers should start at different ends of the transect ie. crossing each other. The dots on the map are places where the observer should stop for a period of 1 0 min. This period o f observation maximises the possibility of seeing a bird. When a bird is heard or seen at any point on the transect its band combination and activity should be recorded in Data Sheet 1 (Appendix 5) . This transect should take a maximum of two days to complete and could be done at the same time as a phenology transect. N 0 I 500 I Figure 6. 7. Tracks suggested for h ih i monitoring from Apri l to August. Dotted l ine = tracks. Solid l ine = streams. Dots indicate suggested sites for ten minute watches. 6.5.2 H ihi monitoring from September to March 6.5.2 . 1 Breeding Activity Every two weeks, from September to March, one observer should sit at each nest site for 30 minutes to monitor breeding activity. An effort should be made to record the band combination of al l cal l ing males and all visiting females because often there is more than one bird of each sex associated with a cavity (see Data sheet 2; Appendix 4) . Two people could complete this work in one day if each one does a different area as suggested in Figure 6.8. 89 90 N 0 \ Figure 6.8. Suggested route to monitor h ih i from September to March. 6.5.2.2 Nest boxes During 1 992 and 1 993 nest boxes were erected at 54 sites. Figure 6.9. shows the approximate location of all nest boxes. It is important to monitor the use of these boxes, especially those located near natural nest sites. While watching for nesting activity observers should take a few minutes to look into the lower nest boxes and also look from below into the high boxes for nest material sticking out. Activity should be recorded on Data Sheet 2 (Appendix 5) . 6.5.2.3 Monitoring the use of feeders by hihi From the end of September until the end of March one person should watch each of the feeders (Figure 6. 1 0) for at least 30 min every three days. Once hihi start using the feeders the observation time should increase to at least one hour. The band combination of birds visiting the feeders should be recorded (Data Sheet 3 ; Appendix 5) . If there are unbanded birds visiting the feeder they should be mistnetted and banded. Figure 6. 9. Map showing the location of nesting boxes on Kapiti Is land. Each dot indicates a nest box . Dotted l ine = tracks. Sol id Line = streams. N 0 Figure 6. 10. Map showing the location of the feeders on Kapiti Island. Arrows point at the location of feeders. Dotted line = tracks. Solid Line = streams . 91 92 6.6 REFERENCES Castro, I . , Alley, J.C., Empson, R.A. and Minot, E.O. 1 994. Translocation of hihi or stitchbird (Notiomystis cincta) to Kapiti I sland, New Zealand: transfer techniques and comparison of release strategies. In Reintroduction B iology of Australian and New Zealand fauna. Melody Serena Ed. Surrey Beatty and sons. Chipping Norton. Austral ia. Castro, I . , Minot, E.O. and Alley, J.C .. 1 994. Feeding and breeding behaviour of hihi or stitchbirds (Notiomystis cincta) recently transferred to Kapiti I sland, New Zealand, and possible management alternatives. In Reintroduction Biology of Australian and New Zealand fauna. Melody Serena Ed. Surrey Beatty and sons. Chipping Norton. Australia. Conant, S. 1 988 . Saving endangered species by translocation. Bioscience 38: 254-257 . Davies, W.J. and Graham, D.J . 1 99 1 . The influence of food on reproductive strategies in a monogamous kingfisher (Chloroceryle amazona) . The Auk 1 08 : 780-789. F lack, J.A.D. 1 978 . Inter island transfers ofNew Zealand black robins. Pp. 362-372 in Endangered B irds : Management Techniques for Preserving Threatened Species ed by S.A. Temple. The University of Wisconsin Press : Madison and Croom Helm Limited: London. Griffith, B . , Scott, J .M. , Carpenter, J. W. and Reed, C. 1 989. Translocation as a species conservation too l : status and strategy . Science 245 : 477-480. Lack, D. 1 950 . The breeding periods of European birds. Ibis 92: 288-3 1 6. Lovegrove, T.G. & Veitch, C.R. 1 994. Translocating wild forest birds. Ecological Management 2 : 23-35 . Merton, D .V. 1 975 . The saddleback: its status and conservation. Pages 6 1-74. In Breeding Endangered Species in Captivity ed by R.D . Martin . Academic Press: London. Nillson, R.J. 1 978 . The South I sland saddleback. Pages 3 2-36. In Wildlife-A Review (Vol. 9). New Zealand Wildlife Service, Department of Internal Affairs: Wellington. Paton, D.C. 1 982 . The diet of the New Holland honeyeater, Phylidonyris novaehollandiae. Australian Journal of Ecology. 7 : 279-298. Rasch, G. 1 988 . Draft recovery plan for North I sland Kokako. Internal report to the New Zealand Department of Conservation, Well ington. Roberts, A. 1 99 1 . A recovery p lan for the South Island saddleback. Internal report to the New Zealand Department of Conservation, Wel lington. Scott J .M. & Carpenter, J.W. 1 987. Release of captive-reared or translocated endangered birds: What do we need to know?. The Auk 1 0 4 : 544-545. Thomas, B. W. & Taylor, R.H. 1 99 1 . The possibili ties for eradication of rats from Kapiti I sland. New Zealand Department of Scientific and I ndustrial Research. Land Resources Technical Record No. 60. van Riper, C., I l l . 1 984. The influence of nectar resources on nesting success and movement patterns of the common amakihi (Hemignathus virens) . The Auk. 1 0 1 : 38-46. Veitch, C.R. & Bell, B. 1 990. Eradication of lntroduced animals from the islands of New Zealand. Pp. 1 3 7- 1 46. /n Ecological Restoration ofNew Zealand Islands ed by D.R. Towns, C.H. Daugherty and I .A.E. Atkinson. Conservation Sciences Publication No. 2. New Zealand Department of Conservation, Well ington. Young, E.C. 1 990. Workshop: translocations (Summary). Pp. 29 1-292. In Ecological Restoration of New Zealand Islands ed by D.R. Towns, C.H. Daugherty and I .A.E. Atkinson. Conservation Sciences Publication No. 2 . New Zealand Department of Conservation, Well ington. 93 Chapter 7 Chapter 7 Concluding Remarks This thesis grew out of a basic problem in the management of a rare New Zealand species. After seven unsuccessful translocations, a total of 1 92 hihi, and thousands of dol lars, the managers (New Zealand Department of Conservation) were eager to know "Why are hihi fai l ing to establish self-sustaining populations after translocations?" . The answer to a question of such magnitude can be found only by first l isting all possible causes of failure and then evaluating each one. Separate causes, however, can have synergistic effects. I went about answering this question in two ways. The first was to evaluate how translocations were done to find out if the failures were a consequence of the methodology . The second was to study the behavioural ecology of hihi to see if the failures reflected a conflict in the relationship between hi hi and their new environment. L ittle was known about hihi when I started research for this thesis. On the way to answering some basic management questions I discovered a bird with unique behaviour and ecology that make it an ideal subj ect for advancing several topical areas of evolutionary ecology and conservation biology. 7. 1 Management The basic translocation technique involves the capture of individuals and their release in an area where they are absent. To achieve success when translocating organisms, however, it is necessary to develop techniques to ensure their survival during the transportation process. It is necessary to know that the area where translocation is taking place is free of whatever caused the organism to disappear in the first p lace. It is indispensable to have a way of measuring the success of the translocation. It is not easy to meet these requirements because many of the species translocated in New Zealand are endangered. We can not be certain that sites presently occupied represent the best ha?itat for the species. Moreover, the reasons why species disappear are often unknown. Until recently, translocations were done uncritical ly, often as a result of near crisis situations. Information gathered through translocations was rarely published. The outcomes of translocations were impossible to analyse because of the lack of experimental design and post-release monitoring. It was difficult to rational ise an experimental design when, because of their rarity, the loss of a s ingle individual could represent the loss of the species. As a consequence, the procedures that seemed to work for some species became standard practice for all species without real data to support them. 94 95 My proj ect with hihi is one rare example where translocations are being carried out before the species is at real risk. Particular attention was given to the way the birds were treated during the translocation process and the methodology was carefully recorded. Releases were caiTied out as designed experiments. After release the hihi were followed to observe the effect of the release techniques and to determine the b irds' use of the environment. Most important of all , descriptions of the techniques and the results from monitoring the translocation were publ ished and are available to managers. The techniques developed during my research and the results of the experimental translocations now determine the way hihi are translocated to other islands (Stitchbird Recovery Group, pers. comm.) . Moreover, the project itself has become an example of the way the study of translocations should proceed. Hihi translocations have set a precedent as the first translocations carried out as designed experiments, having been monitored for several years, and having been publ ished. As a result of my experiments it has even been necessary to reverse the terminology applied to releases. The former designation of "hard" and " soft" release are not descriptions of the process but rather of the presumed outcomes. Soft releases, where the birds are kept in captivity for some time prior to release, were thought to help the animals by providing a buffer period between the stress of the translocation and the hardships of adapting to a new area. Hard releases were those where birds were l iberated directly on arrival at the release sites . Soft releases proved to be relatively stressful for hihi, and birds released hard survived better (Chapter 1 ) . This shows the benefit of more obj ective tem1inology . It is necessary, therefore, to rename "hard" and " soft" releases as " immediate" and "delayed" respectively . Immediate releases were made less stressful by providing food for the released birds. The results of monitoring hihi showed that disease, interspecific and intraspecific competition for nesting sites, and food availabi l i ty, were all factors affecting the survival of the birds on Kapiti Is land. I learned this by careful monitoring of hihi after the releases. The results show that management measures can improve the birds' survival and establishment. Such m anagement must be trialed and monitored to observe the effects on hihi. In Chapter 6 I provide a monitoring scheme that will allow managers to evaluate the management alternatives based on my research. 7.2 Breeding biology During my research I found in hihi a breeding system unusual amongst birds. Hihi have a highly variable mating system including monogamy (one male and one female), polygyny (one male and several females), polyandry (one female and several males), and polygynandry (several males and several females) . This finding opens the possibility for the study of the evolution of mating systems. B ased on the knowledge of hihi's breeding and feeding ecology acquired on ----- ------------------------------------------ Kapiti I have developed a testable model to explain the variabil ity of mating systems in hihi . The model suggests that hihi choose their breeding system on the basis of the distribution of food and nest sites in relation to one another. Emlen and Oring ( 1 977) first suggested that environmental conditions determine the degree to which mates can be defended or monopol ised, and determine the temporal avai lability of receptive mates giving rise to the different mating systems. For instance, monogamy occurs when food is evenly distributed spatially and males can defend areas only large enough to support a single pair of individuals. Polygyny occurs when food is spatially aggregated. Davies ( 1 983, 1 992) convincingly demonstrated that for dunnocks Prunella modularis, another passerine with a variable mating system. His experimental manipulation of the spatial distribution of food resulted in the predicted change in mating system. Specifically, an increase in food avai lability increased polygynandry. Hihi provide a unique opportunity to test these hypotheses about mating systems because they are restricted to islands, breed in nest boxes, and readily take food from feeders providing the opportunity to experimentally manipulate nesting sites and food avai labil ity . Furthermore, I have also proposed that manipulation of hihi breeding system through altering resource availabi l i ty could be used as a management tool . Some mating systems might prove less stressful than others, decreasing the incidence of stress related diseases and increasing survival and reproductive success. Polygyny could be desirable in some circumstances (with more females than males, chick production would be greater), while monogamy could be useful in some others (greater genetic diversity). Testing this model in the wild is the next step. It wil l put hihi in the international scene by testing some general concepts about the evolution of mating systems. However, on the way to obtaining the results, there will be other contributions specifically regarding sperm competition, female sperm choice and parental investment. 7.3 Honeyeaters and poll ination New Zealand honeyeaters feed on nectar from flowers that are very different from those honeyeaters use elsewhere. Very few New Zealand flowers have characteristics usually associated with bird pollination. New Zealand flowers are renowned for being small, rather than large, and mostly yellow or white, rather than brightly coloured. They are also quite simple in their structure, and are often scented. All those characteristics are c loser to those known from flowers pol l inated by insects in other parts of the world. For that reason, unti l now it was believed that the New Zealand honeyeaters used those small flowers only incidentally. Pollination by New Zealand birds was thought to be restricted to those species with flowers that resembled those poll inated by birds in other parts of the world. From my work on hihi I knew that visitation to some of the very small flowers 96 97 was anything but incidental . Hihi, tui and bellbird visited small flowers consistently. They spent most of their foraging activity time on them, and visited the flowers every year. I was interested to know the calorific value of the nectar in the flowers the birds were using, to relate it to the rate of visitation to flowers, and obtain an idea of the avai lability of food on Kapiti I . . When I obtained those values it became apparent that the birds were using the flowers because they could get enough energy from the nectar to sustain their metabolism. Some of the small flowers could provide as many calories as some of the larger flowers. I proposed in Chapter 4 that New Zealand flowers might have evolved to accept poll ination from a number of pollinators, including birds. This idea arises from two observations . First, the New Zealand insect fauna is poorly developed when compared to other parts of the world, even lacking some of the commonest insect pollinators. Second, I found the honeyeaters visiting Pseudopanax arboreus female flowers (they also v isit the male flowers) . Unlike the male Pseudopanax arboreus flowers, the female f1owers are very small and for most of their lives do not have petals. Closer observation showed me that the birds were l icking the f1ower body, which resembled a fruit and which had nectar all over it. I described this as a " lol ly-pop" syndrome. It is possible that New Zealand flowers have evolved other ways to ensure pollination by birds different from those known from other parts of the world. It takes a few seconds for a honeyeater to remove the nectar from the " lolly-pop" . During this time it can contact the reproductive parts of the f1ower and deposit pollen from the male f1owers. If the New Zealand honeyeaters have an important pollinating role, the effects of their decrease in numbers on the mainland could be detrimental to the reproduction of plants. New Zealand honey eaters not only feed on the nectar of f1owers, but often eat the fruits as wel l . Reduction on honeyeater numbers might cause a decrease in fruit set and in turn a decrease in the dispersal of seeds. The use of small flowers by the New Zealand honeyeaters challenge the current views on evolution of f1owers and poll inators . I propose research projects that clarifY the role of birds as pollinators and their effect on the plants they pollinate. The different topics presented in this thesis provide the basis for building a hihi management program. Moreover, they make a significant contribution to several theoretical problems. Finally, my work has provided ideas and base l ine data for addressing further critical questions. REFERENCES Davies, N .B. 1 983 . Polyandry, cloaca pecking and sperm competition in dunnocks. Nature 302: 334-336. Davies, N.B. 1 992. Dunnock behaviour and social evolution. Oxford University Press. Oxford. Emlen, S.T. & Oring, L.W. 1 977. Ecology, sexual selection and the evolution of mating systems. Science 197 : 2 1 5-222. 98 Append ixes Appendix 1 . Comparison between the plants in flower at each month of the year and hihi Date April 1 992 May 1 992 June 1 992 July 1 992 August 1 992 September 1 992 October 1 992 November 1 992 flower use. Plants Avai lable Dysoxylum spectabile Me/icitus ramiflorus Metrosideros fu/gens Pseudopanax crassifolius Dysoxylum spectabile Melicitus ramiflorus Metrosideros fulgens Pseudopanax arboreus Pseudopanax crassifolius Vitex /ucens Dysoxylum spectabile Fuchsia excorticata Melicitus ramiflorus Metrosideros fu/gens Pseudopanax arboreus Pseudopanax crassifolius Vitex /ucens Fuchsia excorticata Pseudopanax arboreus Pseudopanax crassifolius Weinmannia racemosa Fuchsia excorticata Geniostoma rupestre Pittosporum crassifolium Pittosporum eugenioides Pittosporum tenuifolium Pseudopanax arboreus Pseudopanax crassifo/ius Vitex lucens Weinmannia racemosa Fuchsia excorticata Geniostoma rupestre Metrosideros fu/gens Pittosporum crassifolium Pittosporum eugenioides Pittosporum tenuifolium Pseudopanax arboreus Vitex /ucens Weinmannia racemosa Fuchsia excorticata Geniostoma rupestre Metrosideros fulgens Myoporum /aetum Pittosporum eugenioides Pittosporum tenuifolium Pseudopanax arboreus Pseudopanax crassifo/ius Vitex /ucens Weinmannia racemosa Elaeocarpus dentatus Geniostoma rupestre Knightia excelsa Metrosideros robusta Myoporum /aetum Plants used Male hihi Female hihi '1 '1 " " '1 " " '1 '1 " " " " " " " " " " " " " " " " " '1 '1 " " 99 1 00 ,; Pittosporum eugenioides Vitex /ucens ,; Weinmannia racemosa ,; ,; December 1 992 Elaeocarpus dentatus 'J 'J Knightia excelsa ,; Melicitus ramiflorus Metrosideros robusta ,; Myoporum /aetum Weinmannia racemosa ,; January 1 993 Melicitus ramiflorus Metrosideros excelsa ,; Metrosideros robusta ,; ,; Myoporum /aetum ,; Weinmannia racemosa ,; February 1 993 Metrosideros exce/sa Vitex /ucens Weinmannia racemosa March 1 993 Melicitus ramiflorus Metrosideros fulgens ,; ,; Vitex /ucens ,; Weinmannia racemosa ,; April 1 993 Dysoxylum spectabile 'J 'J Metrosideros fulgens 'J 'J May 1 993 Dysoxylum spectabile Metrosideros fulgens ,; ,; Vitex /ucens June 1 993 Dysoxilum spectabile 'J 'J Fuchsia excorticata Geniostoma rupestre ,; ,; Pittosporum crassifolium Pseudopanax arboreus ,; ,; Vitex /ucens July 1 993 Fuchsia excorticata 'J Geniostoma rupestre ,; ,; Metrosideros fulgens ? Pittosporum crassifolium Pseudopanax arboreus ,; Vitex lucens August 1 993 Fuchsia excorticata Geniostoma rupestre ,; Metrosideros fulgens ,; Pittosporum crassifolium ,; ,; Pseudopanax arboreus ,; ,; Vitex /ucens September 1 993 Fuchsia excorticata Geniostoma rupestre Pittosporum crassifolium Pittosporum teniufolium Pseudopanax arboreus Vitex lucens Weinmannia racemosa October 1 993 Fuchsia excorticata Geniostoma rupestre ,; Metrosideros excelsa Pittosporum eugenioides -.J Pittosporum teniufo/ium Pseudopanax arboreus ,; Pseudopanax crassifolius Vitex lucens Weinmannia racemosa November 1 993 Elaeocarpus dentatus Geniostoma rupestre 1 0 1 Knightia excelsa Me/icitus ramiflorus Metrosideros robusta Myoporum laetum Pittosporum eugenioides Pseudopanax arboreus Weinmannia racemosa December 1 993 Elaeocarpus dentatus 'J Knightia exce/sa Metrosideros fulgens Metrosideros robusta Myoporum laetum Pseudopanax arboreus Weinmannia racemosa January 1 994 Elaeocarpus dentatus " Melicitus ramiflorus Metrosideros exce/sa " Metrosideros fulgens " Metrosideros robusta Myoporum laetum " Vitex lucens Weinmannia racemosa February 1 994 Melicitus ramiflorus Metrosideros fu/gens Pseudopanax crassifolius Vitex /ucens March 1 994 Metrosideros fulgens Vitex /ucens April 1 994 Oysoxilum spectabile 'J 'J Metrosideros fulgens Vitex lucens May 1 994 Oysoxilum spectabile 'J 'J Metrosideros fulgens Vitex lucens June 1 994 Oysoxilum spectabile No data Fuchsia excorticata Geniostoma rupestre Metrosideros fulgens Pseudopanax arboreus Vitex lucens July 1 994 Oysoxilum spectabile No data Fuchsia excorticata Metrosideros fu/gens Pseudopanax arboreus Vitex /ucens 1 02 Appendix 2: Flower phenology for 1 7 plant species selected from hihi preferred nectar sources on Kapiti Island - 1 992-1 994. Myoporum laetum 1992?93 1993?94 1 994 Pittosporum crassifolium 1992-93 1993?94 1994 Pittosporum eugenioides 1992 1993 1 994 Pittosporum tenuifolium 1992 1993 1 994 Pseudopanax arboreus 1992 1993 1994 Pseudopanax crassifolius 1 992 1993 1994 Weinmannia racemosa 1 992?93 1993?94 1994 Geniostoma rupestre 1992?93 1993?94 1994 Melicitus ramiflorus 1992?93 1 993?94 1 994 Dysoxylum spectabife 1 992?93 1993-94 1994 Elaeocarpus dentatus 1992?93 1 993?94 1 994 Fuchsia excorticata 1992?93 1993?94 1994 Knightia excelsa 1 992?93 1993?94 1994 Metrosideros excelsa 1992?93 1 993?94 1994 Metrosideros robusta 1992?93 1993?94 1 994 Metrosideros fulgens 1 992?93 1 993?94 1 994 Vitex lucens 1 992?93 1 993?94 1994 Flowered later jn 1993-94 than in 1992-93 A M J J A S O N O J F M A M J J A S O N O J F M A M J J A S O N D J F M A M J J A S O N O J F M A M J J A S O N O J F M A M J J A S O N O J F M A M J J A S O N O J F M Flowered earlier jn 1993-94 than jn 1992-93 A M J J A S O N D J F M A M J J A S O N O J F M Flowered at the same time in 1993?94 and in 1992-93 A M J J A S O N O J F M A M J J A S O N O J F M A M J J A S O N O J F M A M J J A S O N D J F M A M J J A S O N O J F M A M J J A S O N O J F M A M J J A S O N O J F M A M J J A S O N D J F M Appendix 3. Number of fruits and flowers for a selection of forest plants on Kapiti Island, 1 992-1 994. M ramiflorus flowers twice yearly with about a month separation between the flowering episodes. Notice the difference in scale between 1 992, 1 993, and 1 994. 600 500 ? ? 400 300 200 1 00 ? 0 A M J J A s 0 N D J F M A 1 992 ? 1 993 1 .6 1 os ? 1 .4 1 os 1 .2 1 os 1 1 os 8 1 04 6 1 04 4 1 04 2 1 04 0 M J J A s 0 N D J F M A M J J 1 993 A 1 994 1 03 1 04 Number of buds (as a representation of flower numbers) for Metrosideros robusta and M excelsa. 2.5 1 05 ? Metrosideros robusta 0 Metrosideros excelsa 0 2 1 05 en ? "0 1 .5 1 05 0 ::J .0 -? 1 1 05 e 0 z 0 5 1 04 0 0 A M J J A S O N D J F M A M J J A S O N D J 1 992 A 1 993 A Number of buds produced by Knightia excels a 3.5 1 04 3 1 04 2.5 1 04 en "0 ::J 2 1 04 .0 0 1 .5 1 04 0 z 1 1 04 5000 0 1 05 Number of fruits and flowers produced by Pseudopanax crassifolius 5000 4000 -? ::J 3000 ._ - 0 0 2000 z 1 000 0 A M J J A S O N D J F MA M J J A S O N D J F M A MJ J 1 992 A. 1 993 A. 1 994 2500 2000 ? Q.l 1 500 ? ..Q - 0 1 000 0 z 500 0 1 992 1 06 Number of flowers and fruits produced by Pseudopanax arboreus 7 1 04 6 1 04 5 1 04 2 ?s 4 1 04 ..... - 0 3 1 04 ci z 2 1 04 1 1 04 0 Comparison between 1 992-93 and 1 993-94 in the number of Elaeocarpus dentatus buds. 2 1 05 1 .5 1 05 (/) "0 :J ..0 1 1 05 0 0 z 5 1 04 0 1 07 1 08 Appendix 4: I nformation about Womberoo? nectar mix. J , < > R :t 1? 1': I?: T A N Q H O N E Y E t\ T I': .r-: F O O D l .o r i i-:f' P l s a nd horu? vp .1 1 e r s . u n l i k e:! n l l t, r b 1 rds , have a b rush l i Y.t'\ ;t ppf!rutap,f'! a t. t. ltc t!nd of t he to t t,r:?uP. . Th i ? h rush t o ny,ufl: i s ;?n ada p t. a t i n n f n r P-?1 l hi! r i nr, fo0d a nd a J J ow!; spet? i fl J i s? t i <' n i n t he Cl' l l P r. t i on n f po l l e n . p 1 .1 n l s?cre l i uns ( nr e l. a r , m;?nn., , honeydew ) , a tH I i nsc r? t sP.cre t i ons ( l e tT? ) Th("y rt l !;O i nc l ude f r?u i t aud i nsec ts i n t he i r r e s pn..: t i v? d i e t s . Tlu! s ? h i rt ls de t? i vf'!' prn l c..> i u ;uui f. 1 t f rom t he po l l r. n o .. i nsec t c om?'O n c n t ?:d t h, i r d i r. t nne I t ? rdum 1t lu? CII(J J i l it! 1 2 5mg . l ' l to:;phoru:c. 1 . R g . Con t a i n s wl u!y pr o t e i n i so l a t e , h yd r o l ys?d r. a se i n , 'JP.8 powde r , "?fle t "! h l ? o i I :eo , l ?c i t h i n? . dex t ??aos , m< d l od?x t r i ns , suc rose , a t i t u t. i o n . When U IHJpcnroO t h i s p r orl?J c t i s s t a b l r. ( o r 1 2 mon t hs i f s t. ( lfeft i rt a coo l t lr? y ,, l a c e . SIJPr i.EifENT ?-oR W I I.U n i H ilS . T h i ? f o(ld i s r\ fl i dron l :?:nr pp i P.IIH!nt , ? nd u i. l h t i rnr> , i t s uc;P. w i i J ::t t t r n (' t m:. ny U i rds j u t? o the ?anJen . E s t ab I i ? h t\n c l P.v a t. ed f eed s t :t t ion 1),? sccu ?? i ng a p l a t form t. c a po i P. o r su?pend i n R i t f r om the branch o f a t r f"? . IJ i s p?n?e the prep.:t rcd foorl i n ?" :c; ha l l o,.,. , n l a zed , bow l . Ra t i on da i l y f e ,.. ,i t. o no more t ha n 1 0m 1 p? r l n r i ke P. t or 5m l pe r huneycot f. e r . A l ways 1 i m i t t l l? amoun t o [ < Ht)' f ood o f J ? r ?d h??au!"> P. o t he r p?u p l e i n your? ne i ghbour hood ;.d l l prnh.1 ld y he feed i up, t llf? :c;ame h i r tJ:c; . R?move t he bo? l al t e r i e cd. i n? and tho ruugh l r c I ean i t . fOOD FOR CAPT I VE n l ROS . Th i ?; f oorl w i l l "rlequa te l y ma i n t .1 i n l o r i ke e l s o r hon<'yP.a t e rs i n capt i v i t y , ho??ev?r for breed i ng succP.ss t he fo 1 l ow i nn i s s t r?our. l y rP.commende d . 1 . I t i ? d?s i rab l e t o keep J o r? i k c? t. ? and hou?}'f'!' il t P. r s i n l arge ;n?ar i r.? p l an t ed lf i l h na t i ve f lowe r i ng shnJl!s . Th i s a C f orrls t he b i rds a source o f prJ l l etl , nec t a r , a r'd i nRP.c t s . ?? r r i or t o . il ttrl du r i nR t hf" hr t'??rt i nR :rirt1.!!:onn t hr fond s ht'td d lw ?nptd ??rHt? n l ?!tl w i l l1 l nd l :"111 ., m i )( l un? t ha t ;l l 'f l t? u x i.m:. t t? s t h ?? t?(Hnpu s i t i Pn n f po ( l t?n flll?l f nr.??c t s , 1\u ?x.?unp l t? n r f'Uf: l l tl m i x i. H r n \ !> t he f.tJ f l ? ? i l l f ' l t?? ?? , r? :t t i ?H ? d?sct i b?!d i 1 1 : I ? '-: r ?q t :! s 011 . "' tH-H1 1\Ht'tJ l ti ? ?.C I' I VOP I. f: l:t\ 1? 1 N t; ? I f \ . Appendix 5 : Work sheets used in the Monitoring of hihi Data Sheet 1 Name of Observer: Pate Place Bird Activity Comments 1 09 Data Sheet 2 Name of Observer: Nest Date Male(s) Female(s) Activity* Nest Comments Site Box ? Ex. mating, chasing, calling, incubating (female enters nest for 20 + minutes and is out for 5 min) , brooding (male and female enter nest for 1 to 1 0 minutes and stay out for 1 5 to 30), feeding fledglings, etc. Data Sheet 3 Name of Observer: Feeder: ?: Bird bands Comments