ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI CHRISTCHURCH, NEW ZEALAND 2-9 DECEMBER 1990 VOLUME II Museum of Comparative Zoology Library Harvard University ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI VOLUME II HRISTCHURCH, NEW ZEALAND 2-9 DECEMBER 1990 National Library of New Zealand Cataloguing-in-Publication data Congressus International^ Ornithologici (20th : 1990 : Christchurch, N.Z.) Acta XX Congressus International^ Ornithologici, Christchurch, New Zealand, 2-9 December 1990. Wellington, N.Z. : New Zealand Ornithological Congress Trust Board, 1990-1991. 4 v. + 1 supplement ISBN 0 - 9597975 - 1 - 3 (Vol I) ISBN 0 - 9597975 - 2 - 1 (Vol II) ISBN 0 - 9597975 - 3 - X (Vol III) ISBN 0 - 9597975 - 4 - 8 (Vol IV) ISBN 0 - 9597975 - 0 - 5 (supplement) ISBN 0 - 9597975 - 5 - 6 (set) 1 . Ornithology - Congresses. 2. Birds - Congresses. I. New Zealand Ornithological Congress Trust Board. II. Title. III. Title: Acta Twentieth Congressus International^ Ornithologici : Christchurch, New Zealand, 2-9 December 1990. 598 Reference to material in this volume should be cited thus: Author(s), 1991. Title . Acta XX Congressus International^ Ornithologici: pages. ISBN 0 - 9597975 - 5 - 6 (five-volume set) ISBN 0 - 9597975 - 2 - 1 (Vol. II) is/icZ- . Copyright © New Zealand Ornithological Congress Trust Board 1991 Published by New Zealand Ornithological Congress Trust Board P O Box 12397, Wellington, New Zealand Typeset, printed and bound in New Zealand by Hutcheson, Bowman & Stewart Ltd Wellington The New Zealand Ornithological Congress Trust Board acknowledges support for the publication of this volume from the Science & Research Division, New Zealand De¬ partment of Conservation; Victoria University of Wellington; the New Zealand 1990 Commission; and the New Zealand Lottery Board. ACTA XX CONGRESSUS INTERNATIONALE 0RNITH0L6GIC! 649 VOLUME I SYMPOSIUM 8 SYMPOSIUM 9 SYMPOSIUM 10 SYMPOSIUM 11 SYMPOSIUM 12 SYMPOSIUM 13 SYMPOSIUM 14 SYMPOSIUM 15 SYMPOSIUM 16 SYMPOSIUM 17 SYMPOSIUM 18 SYMPOSIUM 19 SYMPOSIUM 20 AUTHOR INDEX VOLUME III VOLUME IV CONTENTS 1-644 VOLUME II Ecology and Social Behaviour of Parrots and Parakeets 651 Bird Flight 699 New Aspects of Avian Migration Systems 749 Ecological and Evolutionary Consequences of Body Size 789 Ecological and Behavioural Adaptations of Southern Hemisphere Waterfowl 839 The Avian Feeding System * , 887 Parent-offspring Relationships 927 Brood Parasitism 999 Filial and Sexual Imprinting 1049 Nocturnality in Birds 1089 Social Organisation of Nectar-feeding Birds 1137 Social Behaviour in the Non-breeding Season 1 193 Acquisition and Functions of Avian Vocalisations 1241 1287 1293-1948 1949-2568 650 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI . ■ ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 651 SYMPOSIUM 8 ECOLOGY AND SOCIAL BEHAVIOUR OF PARROTS AND PARAKEETS Conveners D. A. SAUNDERS and P. C. ARROWOOD 652 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI SYMPOSIUM 8 Contents INTRODUCTORY REMARKS: SYMPOSIUM ON THE ECOLOGY AND SOCIAL BEHAVIOUR OF PARROTS AND PARAKEETS PATRICIA C. ARROWOOD and DENIS A. SAUNDERS . 653 THE EFFECT OF LAND CLEARING ON THE ECOLOGY OF CARNABY’S COCKATOO AND THE INLAND RED-TAILED BLACK COCKATOO IN THE WHEATBELT OF WESTERN AUSTRALIA D. A. SAUNDERS . 658 MALE-MALE, FEMALE-FEMALE AND MALE-FEMALE INTERACTIONS WITHIN CAPTIVE CANARY-WINGED PARAKEET BROTOGERIS V. VERSICOLURUS FLOCKS PATRICIA C. ARROWOOD . 666 NESTING BEHAVIOUR OF THE BAHAMA PARROT AMAZONA LEUCOCEPHALA BAHAMENSIS ON ABACO ISLAND, BAHAMAS ROSEMARIE S. GNAM . 673 SOCIAL BEHAVIOUR AND POPULATION DYNAMICS OF THE MONK PARAKEET E. H. BUCHER, L. F. MARTIN, M. B. MARTELLA and J. L. NAVARRO . 681 ANNUAL VARIATION IN PRODUCTIVITY OF NORTH ISLAND KAKA ON KAPITI ISLAND, NEW ZEALAND RON J. MOORHOUSE . 690 CONCLUDING REMARKS: SYMPOSIUM ON THE ECOLOGY AND SOCIAL BEHAVIOUR OF PARROTS AND PARAKEETS PATRICIA C. ARROWOOD and DENIS A. SAUNDERS . 697 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 653 INTRODUCTORY REMARKS: SYMPOSIUM ON THE ECOLOGY AND SOCIAL BEHAVIOUR OF PARROTS AND PARAKEETS PATRICIA C. ARROWOOD' and DENIS A. SAUNDERS2 ’ Department of Biology, Dept. 3AF, New Mexico State University, Las Cruces, New Mexico 88003, USA 2 CSIRO Division of Wildlife and Ecology, Western Australian Laboratory, Locked Bag No. 4, PO, Midland, WA 6056, Australia The popularity of psittacines as research subjects has never matched their popular¬ ity as pets and aviary birds. Because of their longevity and their natural inclination to form close, long-lasting relationships with a mate, they are ideally “preadapted” for human companionship. In addition, their mimetic abilities made many famous as good talkers, endearing them to a public intent on giving them human voices. Unfortunately, these characteristics that have made them so attractive to many people have prob¬ ably made them seem somehow unreal as birds to scientists. Other factors have also undoubtedly contributed to the paucity of studies on psittacines. For one, psittacines are not usually territorial but range widely and often erratically over a large area. In the tropical forests that many inhabit, their ranging habits make them exceedingly difficult to find and follow, or to catch for banding and marking. Nevertheless, a few species have been well studied, though primarily in captivity. The Budgerigar Melopsittacus undulatus, an Australian parakeet, is probably the best- studied of all psittacines. Studies in the wild by Rothwell & Amadon (1964), Schrader (1975) and Wyndham (1 978,1 980a, b,c,d, 1981, 1983) and in captivity by Cinat- Thomson (1926), Masure and Allee (1934), Ficken et al. (1960), Brockway (1962a. b, 1964a, b,c, 1965, 1967a, b,c, 1968, 1969a, b, 1974), Hinde & Putman (1973), Putman & Hinde (1973), Trillmich (1976a, b,c), Stamps et al. (1985, 1987, 1989, 1990), Dooling & Saunders (1975), Dooling & Searcy (1981), Dooling et al. (1 987,1 990, in press), Brown et al. (1988), Okanoya & Dooling (in press), and Kavanau (1987) have revealed many fascinating behaviours and abilities. The Australian cockatoos (Carnaby 1948, Pidgeon 1970,1981, Rogers & McCulloch 1981, Rowley 1980a,b, 1983,1990, Rowley & Saunders 1980, Saunders 1974a, b, 1977a, b, 1979a, b,c, 1980, 1982, 1983, 1986, 1989, 1990, Saunders & Curry 1990, Saunders & Ingram 1987, Saunders & Smith 1981, Saunders et al. 1982,1984,1985, Buckland et al. 1983, Adams et al. 1984, Mclnnes & Came 1978, Campbell & Saunders 1976, Schodde et al. 1979, Millam et al. 1988, Myers et al. 1988,1989, Yamamoto et al. 1989, Joseph 1982a, b,c, Clout 1989, Jones 1987), the Puerto Rican Parrot Amazona vittata (Snyder 1977, Snyder & Taapken 1977, Snyder et al. 1987, Wiley 1980), and the Monk Parakeet Myiopsitta monachus (Humphrey & Peterson 1978, Caccamise 1980, Bucher & Martin 1987, Martella & Bucher 1984, in press, Martella et al. 1985, 1987, Navarro & Bucher 1990) are the most studied in the wild. Noticeably lacking are studies of most Neotropical (South and Central American), Afro-Asian and Pacific (non-Australian) species. De¬ spite Dilger’s (1960, 1 962) earlier work with the African Agapornis lovebirds, little fur¬ ther scientific interest has been shown in that geographical group of psittacines. 654 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI We are very pleased in this symposium to be presenting papers covering Neotropical (Arrowood, Bucher, Gnam) and Pacific (Moorhouse, Saunders) species. With the exception of the Moorhouse study that has just begun, these papers also represent sustained work on a single species for periods up to twenty-two years. Psittacines are generally long-lived birds and many aspects of their ecology and behaviour will not become apparent without such long-term studies, particularly in areas subject to change due to human pressure. The iosses of wild psittacine populations for the pet trade and through habitat destruc¬ tion will make many species unavailable for study if these destructive trends continue. This will result in more studies like those of the very small population of the Puerto Rican Parrot in last-ditch efforts to forestall extinction. It is our hope that the threat of extinction is not the only stimulus that can awaken more scientific interest in this fascinating group of birds, and we hope that growing popular and scientific concern can save most parrot species from this danger. LITERATURE CITED ADAMS, M., BAVERSTOCK, P.R., SAUNDERS, D.A., SCHODDE, R., SMITH, G.T. 1984. Biochemi¬ cal systematics of the Australian cockatoos (Psittaciformes: Cacatuinae). Australian Journal of Zool¬ ogy 32: 363-377. BROCKWAY, B. F. 1962a. The effects of nest-entrance positions and male vocalizations on reproduc¬ tion in Budgerigars. The Living Bird 1: 93-101. BROCKWAY, B. F. 1962b. Investigations of the auditory stimuli for laying in Budgerigars ( Melopsittacus undulatus). American Zoologist 2: 71. BROCKWAY, B. F. 1964a. Ethological studies of the Budgerigar ( Melopsittacus undulatus ): Non-repro- ductive behavior. Behaviour 22: 193-222. BROCKWAY, B. F. 1964b. Ethological studies of the Budgerigar: Reproductive behavior. Behaviour 23: 294-324. BROCKWAY, B. F. 1964c. Social influences on reproductive physiology and ethology of Budgerigars ( Melopsittacus undulatus). Animal Behaviour 12: 493-501. BROCKWAY, B. F. 1965. Stimulation of ovarian development and egg laying by male courtship vocali¬ zation in Budgerigars ( Melopsittacus undulatus). Animal Behaviour 13: 575-578. BROCKWAY, B. F. 1967a. The influence of vocal behavior on the performer’s testicular activity in Budgerigars (Melopsittacus undulatus.) Wilson Bulletin 79: 328-334. BROCKWAY, B. F. 1967b. Social and experiential influences on nest box-oriented behavior and go¬ nadal activity of female Budgerigars (Melopsittacus undulatus). Behaviour 29: 63-82. BROCKWAY, B. F. 1967c. Interactions among male courtship warbling, photoperiodic and experien¬ tial factors in stimulating the reproductive activity of female Budgerigars (Melopsittacus undulatus). American Zoologist 7: 215. BROCKWAY, B. F. 1968. Influences of sex hormones on the loud and soft warbles of male Budgeri¬ gars. Animal Behaviour 16: 5-12. BROCKWAY, B. F. 1969a. Hormonal and experiential factors concerning nestbox-oriented behavior. Behaviour 34: 1 -26. BROCKWAY, B. F. 1969b. Roles of Budgerigar vocalization in the integration of breeding behaviour. Pp. 131-158 in Hinde, R.A. (Ed.). Bird Vocalizations. Cambridge, Cambridge University Press. BROCKWAY, B. F. 1974. The influence of some experiential and genetic factors, including hormones, on the visible courtship behavior of Budgerigars (Melopsittacus). Behaviour 51: 1-18. BROWN, S.D., DOOLING, R.J., O’GRADY, K. 1988. Perceptual organization of acoustic stimuli by Budgerigars ( Melopsittacus undulatus ): III. Contact calls. Journal of Comparative Psychology 102- 236- 247. BUCHER, E. H., MARTIN, L. F. 1987. Los nidos de cotorras (Myiopsitta monachus) como causa de problemas en lineas de transmision electrica. Vida Silvestre Neotropical 1: 50-51. BUCKLAND, S.T., ROWLEY, I., WILLIAMS, D.A. 1983. Estimates of survival from repeated sightings of tagged Galahs. Journal of Animal Ecology 52: 563-573. CACCAMISE, D.F. 1980. Growth and development of major body components in the Monk Parakeet Wilson Bulletin 92: 376-381. CAMPBELL, N.A., SAUNDERS, D.A. 1976. Morphological variation in the White-tailed Black Cocka- ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 655 too, Calyptorhynchus baudinii, in Western Australia: A multivariate approach. Australian Journal of Zoology 24: 589-595. CARNABY, I.C. 1948. Variations in the White-tailed Black Cockatoo. Western Australian Naturalist 1: 136-138. CINAT-THOMSON, H. 1926. Die geschlechtliche Zuchtwahl beim Wellensittich (Melopsittacus undulatus [Shaw]). Biol. Zbl. 46: 543-552. CLOUT, M. N. 1989. Foraging behaviour of Glossy Black Cockatoos. Australian Wildlife Research 16: 467-473. DILGER, W. C. 1960. The comparative ethology of the African parrot genus Agapornis. Zeitschrift fur Tierpsychologie 17: 649-685. DILGER, W. C. 1962. The behavior of lovebirds. Scientific American 206: 88-98. DOOLING, R. J., BROWN, S. D., PARK, T. J., OKANOYA, K. 1990. Natural perceptual categories for vocal signals in Budgerigars (Melopsittacus undulatus). Pp. 345-374 in Berkley, M., Stebbins, W. (Eds.). Comparative perception: Communication, Vol. 2. New York, John Wiley and Sons. DOOLING, R. J., PARK, T. J., BROWN, S.D., OKANOYA, K. In press. Perception of species-specific vocalizations by isolate-reared Budgerigars ( Melopsittacus undulatus). International Journal of Com¬ parative Psychology. DOOLING, R. J., PARK, T. J., BROWN, S. D., OKANOYA, K., SOLI, S.D. 1987. Perceptual organiza¬ tion of acoustic stimuli by Budgerigars (Melopsittacus undulatus ): II. Vocal signals. Journal of Compara¬ tive Psychology 101: 367-381. DOOLING, R. J., SAUNDERS, J. C. 1975. Hearing in the parakeet (Melopsittacus undulatus): Abso¬ lute thresholds, critical ratios, frequency difference limens, and vocalizations. Journal of Comparative and Physiological Psychology 88: 1-20. DOOLING, R. J., SEARCY, M. H. 1981. Amplitude modulation thresholds for the parakeet (Melopsittacus undulatus). Journal of Comparative Physiology 143: 383-388. FICKEN, R.W., TIENHOVEN, A. VAN, FICKEN, M.S., SIBLEY, F.C. 1960. Effect of visual and vocal stimuli on breeding in the Budgerigar (Melopsittacus undulatus). Animal Behaviour 8: 104-106. HINDE, R.A., PUTMAN, R.J. 1973. Why Budgerigars breed in continuous darkness. Journal of Zool¬ ogy 170: 485-491. HUMPHREY, P.S., PETERSON, R.T. 1978. Nesting behavior and affinities of Monk Parakeets of south¬ ern Buenos Aires Province, Argentina. Wilson Bulletin 90: 544-552. JONES, D. 1987. Feeding ecology of the Cockatiel Nymphicus hollandicus , in a grain-growing area. Australian Wildlife Research 14: 105-115. JOSEPH, L. 1982a. The Red-tailed Black Cockatoo in south-eastern Australia. Emu 82: 42-45. JOSEPH, L. 1982b. The Glossy Black Cockatoo on Kangaroo Island. Emu 82: 46-49. JOSEPH, L. 1982c. The origin of the population of the Glossy Black Cockatoo on Kangaroo Island. South Australian Ornithologist 56: 46-47. KAVANAU, J. L. 1987. Lovebirds, Cockatiels, Budgerigars: Behavior and evolution. Los Angeles, Cali¬ fornia, Science Software Systems. MARTELLA, M.B., BUCHER, E.H. 1984. Nesting of the Spot-winged Falconet in Monk Parakeet nests. Auk 101: 614-615. MARTELLA, M. B., BUCHER, E. H. In press. Vocalizations of the Monk Parakeet. Bird Behaviour. MARTELLA, M. B., NAVARRO, J. L., BUCHER, E. H. 1985. Vertebrados asociados a los nidos de la cotorra Myiopsitta monachus en Cordoba y la Rioja. Physis 43: 49-51. MARTELLA, M. B., NAVARRO, J. L., BUCHER, E. H. 1987. Metodo para la captura de cotorras Myiopsitta monachus en sus nidos. Vida Silvestre Tropical 1: 52-53. MASURE, R. H., ALLEE, W. C. 1934. Flock organization of the Shell Parrakeet Melopsittacus undulatus Shaw. Ecology 15: 388-398. MCINNES, R.S., CARNE, P.B. 1978. Predation of cossid moth larvae by Yellow-tailed Black Cocka¬ too causing losses in plantations of Eucalyptus grandis in north coastal New South Wales. Australian Wildlife Research 5: 101-121. MILLAM, J.R., ROUDYBUSH, T.E., GRAU, C.R. 1988. Influence of environmental manipulation and nest-box availability on reproductive success of captive Cockatiels ( Nymphicus hollandicus). Zoo Bi¬ ology 7: 25-34. MYERS, S.A., MILLAM, J.R., HALAWANI, M.E. El 1989. Plasma luteinizing hormone and prolactin levels during the reproductive cycle in Cockatiels (Nymphicus hollandicus). General and Comparative Endocrinology 73: 85-91. MYERS, S.A., MILLAM, J.R., ROUDYBUSH, T.E., GRAU, C.R. 1988. Reproductive success of hand- reared vs. parent-reared Cockatiels ( Nymphicus hollandicus). Auk 105: 536-542. NAVARRO, J. L., BUCHER, E. H. 1990. Growth of Monk .Parakeets. Wilson Bulletin 102: 520-525. OKANOYA, K., DOOLING, R.J. In press. Minimum detectable gap in noise as a function of intensity and frequency in Budgerigars (Melopsittacus undulatus) and Zebra Finches (Poephila gutata). Hear¬ ing Research. 656 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI PIDGEON, R.W.J. 1970. The individual and social behaviour of the Galah Kakatoe roseicapiila Vieillot. M.Sc. Thesis, University of New England, Armidale. PIDGEON, R. 1981. Calls of the Galah Cacatua roseicapiila and some comparisons with four other species of Australian parrot. Emu 81: 158-168. PUTMAN, R.J., HINDE, R.A. 1973. Effects of light regime and breeding experience on Budgerigar re¬ production. Journal of Zoology 170: 475-484. ROGERS, L. J., MCCULLOCH, H. 1981. Pair-bonding in the Galah Cacatua roseicapiila. Bird Behav¬ iour 3: 80-92. ROTHWELL, R., AMADON, D. 1964. Ecology of the Budgerygah. Auk 81 : 82. ROWLEY, I. 1980a. Parent-offspring recognition in a cockatoo, the Galah, Cacatua roseicapiila. Aus¬ tralian Journal of Zoology 28: 445-456. ROWLEY, I. 1980b. Social organisation and the use of creches in the Galah, Cacatua roseicapiila. Acta XVII International Ornithological Congress, Berlin, Volume II, Verlag der Deutshcen Ornithologen- Gesellschaft. ROWLEY, I. 1983. Mortality and dispersal of juvenile Galahs Cacatua roseicapiila in the Western Aus¬ tralian wheatbelt. Australian Wildlife Research 10: 329-342. ROWLEY, I. 1990. Behavioural ecology of Galahs. Chipping Norton, N.S.W., Surrey Beatty & Sons. ROWLEY, I., SAUNDERS, D. A. 1980. Rigid wing-tags for cockatoos. Corella 4: 1-7. SAUNDERS, D. A. 1974a. Subspeciation in the White-tailed Black Cockatoo Calyptorhynchus baudinii in Western Australia. Australian Wildlife Research 1: 55-69. SAUNDERS, D. A. 1974b. The function of displays in the breeding of the White-tailed Black Cocka¬ too. Emu 74: 43-46. SAUNDERS, D. A. 1977a. The effect of agricultural clearing on the breeding success of the White¬ tailed Black Cockatoo. Emu 77: 180-184. SAUNDERS, D. A. 1977b. Red-tailed Black Cockatoo breeding twice a year in the south-west of West¬ ern Australia. Emu 77: 107-110. SAUNDERS, D. A. 1979a. Distribution and taxonomy of the White-tailed and Yellow-tailed Black Cocka¬ toos Calyptorhynchus spp. Emu 79:215-227. SAUNDERS, D. A. 1979b. The availability of tree hollows for use as nest sites by White-tailed Black Cockatoos. Australian Wildlife Research 6: 205-216. SAUNDERS, D. A. 1979c. The biology of the short-billed form of the White-tailed Black Cockatoo Calyptorhynchus funereus latirostris Carnaby. Ph.D. Thesis, University of Western Australia. SAUNDERS, D. A. 1980. Food and movements of the short-billed form of the White-tailed Black Cocka¬ too. Australian Wildlife Research 7: 257-269. SAUNDERS, D. A. 1982. The breeding behaviour and biology of the short-billed form of the White¬ tailed Black Cockatoo Calyptorhynchus funereus. Ibis 124: 422-455. SAUNDERS, D. A. 1983. Vocal repertoire and individual vocal recognition in the short-billed White¬ tailed Black Cockatoo Calyptorhynchus funereus latirostris Carnaby. Australian Wildlife Research 10: 527-536. SAUNDERS, D. A. 1986. Breeding season, nesting success and nestling growth in Carnaby’s Cocka¬ too, Calyptorhynchus funereus latirostris over 16 years at Coomallo Creek, and a method for assessing the viability of populations in other areas. Australian Wildlife Research 13: 261-273. SAUNDERS, D. A. 1989. Changes in the avifauna of a region, district and remnant as a result of frag¬ mentation of native vegetation: The wheatbelt of Western Australia. A case study. Biological Conser¬ vation 50: 99-135. SAUNDERS, D. A. 1990. Problems of survival in an extensively cultivated landscape: The case of Carnaby’s Cockatoo Calyptorhynchus funereus latirostris. Biological Conservation 54: 111-124. SAUNDERS, D. A., CURRY, P. J. 1990. The impact of agricultural and pastoral industries on birds in the southern half of Western Australia: Past, present and future. Proceedings of the Ecological Soci¬ ety of Australia 16: 303-321. SAUNDERS, D. A., INGRAM, J. A. 1987. Factors affecting survival of breeding populations of Carnaby’s Cockatoo Calyptorhynchus funereus latirostris in remnants of native vegetation. Pp. 249- 258 in Saunders, D.A., Arnold, G.W., Burbidge, A.A., Hopkins, A.J.M. (Eds). Nature conservation: The role of remnants of native vegetation. Chipping Norton, N.S.W., Surrey Beatty and Sons. SAUNDERS, D. A., ROWLEY, I., SMITH, G. T. 1985. The effects of clearing for agriculture on the dis¬ tribution of cockatoos in the south-west of Western Australia. Pp. 309-321 in Keast, A., Recher, H.F Ford, H., Saunders, D.A. (Eds). Birds of Eucalypt forests and woodlands: Ecology, conservation, man¬ agement. Chipping Norton, N.S.W., Surrey Beatty & Sons. SAUNDERS, D.A., SMITH, G.T. 1981. Egg dimensions and egg weight loss during incubation in five species of cockatoo and the use of measurements to determine the stage of incubation of birds’ eaqs. Australian Wildlife Research 8: 411-419. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 657 SAUNDERS, D.A., SMITH, G.T., CAMPBELL, N.A. 1984. The relationship between body weight, egg weight, incubation period, nestling period and nest site in the Psittaciformes, Falconiformes, Strigiformes and Columbiformes. Australian Journal of Zoology 32: 57-65. SAUNDERS, D. A., SMITH, G. T., ROWLEY, I. 1982. The availability and dimensions of tree hollows that provide nest sites for cockatoos (Psittaciformes) in Western Australia. Australian Wildlife Research 9: 541-556. SCHODDE, R., SMITH, G.T., MASON, I.J., WEATHERLY, R.G. 1979. Relationships and speciation in the Australian Corellas (Psittacidae). Bulletin of the British Ornithologists Club 99: 128-137. SCHRADER, N. 1975. The breeding of Budgerygahs in western New South Wales. Australian Bird Watcher 6: 1 18-122. SNYDER, N. F. R. 1977. Puerto Rican Parrots and nest-site scarcity. Pp. 47-53 in Temple, S.A. (Ed.). Endangered Birds, Management techniques for preserving threatened species. Madison, University of Wisconsin Press. SNYDER, N. F. R. and TAAPKEN, J. D. 1977. Puerto Rican Parrots and nest predation by Pearly-eyed Thrashers. Pp. 113-120 in Temple, S.A. (Ed.). Endangered birds, management techniques for preserv¬ ing threatened species. Madison, University of Wisconsin Press. SNYDER, N. F. R„ WILEY, J. W., KEPLER, C. B. 1987. The parrots of Luquillo: Natural history and conservation of the Puerto Rican Parrot. Los Angeles, California, Western Foundation of Vertebrate Zoology. STAMPS, J., CLARK, A., ARROWOOD, P., KUS, B. 1985. Parent-offspring conflict in Budgerigars. Behaviour 94: 1 -40. STAMPS, J., CLARK, A., ARROWOOD, P., KUS, B. 1989. Begging behavior in Budgerigars. Ethology 81: 177-192. STAMPS, J., CLARK, A., KUS, B., ARROWOOD, P. 1987. The effects of parent and offspring gender on food allocation in Budgerigars. Behaviour 101: 177-199. STAMPS, J., KUS, B., CLARK, A., ARROWOOD, P. 1990. Social relationships of fledgling Budgerigars, Melopsittacus undulatus. Animal Behaviour 40: 688-700. TRILLMICH, F. 1976a. Spatial proximity and mate-specific behaviour in a flock of Budgerigars (Melopsittacus undulatus-, Aves, Psittacidae). Zeitschrift fur Tierpsychologie 41 : 307-331 . TRILLMICH, F. 1976b. Learning experiments on individual recognition in Budgerigars (Melopsittacus undulatus). Zeitschrift fur Tierpsychologie 41 : 372-395. TRILLMICH, F. 1976c. The influence of separation on the pair bond in Budgerigars ( Melopsittacus undulatus] Aves, Psittacidae). Zeitshcrift fur Tierpsycholgie 41: 396-408. WILEY, J. W. 1980. The Puerto Rican Parrot Amazona vittata: Its decline and the program for its con¬ servation. Pp. 133-159 in Pasquier, R.F. (Ed.). Conservation of New World parrots, ICBP Technical Publication No. 1. WYNDHAM, E. 1978. Ecology of the Budgerigar Melopsittacus undulatus Shaw (Psittaciformes: Platycercidae). Ph.D. Thesis, University of New England, Armidale, Australia. WYNDHAM, E. 1980a. Environment and food of the Budgerigar Melopsittacus undulatus. Australian Journal of Ecology 5: 47-61. WYNDHAM, E. 1980b. Total body lipids of the Budgerigar Melopsittacus undulatus (Psittaciformes: Platycercidae) in inland mid-eastern Australia. Australian Journal of Zoology 28: 239-247. WYNDHAM, E. 1980c. Aspects of biorhythms in the Budgerigar Melopsittacus undulatus Shaw, a parrot of inland Australia. Pp. 485-492 in Proceedings 17th. International Ornithological Congress. WYNDHAM, E. 1980d. Diurnal cycle, behaviour and social organization of the Budgerigar Melopsittacus undulatus Emu 80: 25-33. WYNDHAM, E. 1981. Breeding and mortality of Budgerigars Melopsittacus undulatus. Emu 81: 240- 243. WYNDHAM, E. 1983. Movements and breeding seasons of the Budgerigar. Emu 82: 276-282. YAMAMOTO, J.T., SHIELDS, K.M., MILLAM, J.R., ROUDYBUSH, T.E., GRAU, C.R. 1989. Reproduc¬ tive activity of force-paired Cockatiels (Nymphicus hollandicus). Auk 106: 86-93. 658 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI THE EFFECT OF LAND CLEARING ON THE ECOLOGY OF CARNABY’S COCKATOO AND THE INLAND RED-TAILED BLACK COCKATOO IN THE WHEATBELT OF WESTERN AUSTRALIA D. A. SAUNDERS CSIRO Division of Wildlife and Ecology, LMB No. 4, PO, Midland, WA 6056, Australia ABSTRACT. European settlement of south-western Australia brought about extensive changes over the last 160 years. Widespread clearing of the semi-arid zone (wheatbelt) for cereal cropping and sheep farming has removed about 93% of the native vegetation, leaving the remainder scattered in small remnants across the landscape. Before these changes, Carnaby’ s Cockatoo Calyptorhynchus funereus latirostris was the only cockatoo occurring throughout this area. It has not been able to adapt to the reduction and fragmentation of its habitat and has disappeared from over a third of its range within the last 20 years. By way of contrast, the Inland Red-tailed Black Cockatoo C. magnificus samueli occurred along the watercourses of the arid zone (pastoral area). With clearing for agriculture, this species has expanded its range into the semi-arid agricultural area, invading part of the range of Carnaby’s Cockatoo. The Red -tailed Black Cockatoo is slowly expanding its range, moving into areas of heavy infestation of Emex australis, an agricultural weed on which it feeds almost exclusively. The biologies of these two black cockatoos are compared in the light of their changing distributions and conservation status. Keywords: Carnaby’s Cockatoo, Calyptorhynchus funereus latirostris, Red-tailed Black Cockatoo, Calyptorhynchus magnificus samueli, distribution, breeding biology. INTRODUCTION The southwest of Western Australia has undergone rapid and extensive change since settlement by Europeans in 1829. In the semi-arid zone (defined here as the area receiving between 300 and 650 mm mean annual rainfall), now known as the wheatbelt (see the unhatched area in Figure 1), an area of 14 million ha has had 93% of the original vegetation removed, over half of it since 1945 ( Saunders et al. 1985, Saunders & Hobbs 1989). The remainder is scattered across the landscape in thou¬ sands of patches of varying sizes, shapes, vegetation associations, landuse histories and ownership. This change in the distribution and abundance of native vegetation has had major effects on the fauna of the region, the mammals being the first to suffer a wave of extinctions (Kitchener et al. 1980). The avifauna of the wheatbelt is under¬ going similar changes, with two species already extinct in the region and many more species changing in distribution and abundance (Saunders 1989; Saunders & Curry 1990). Before European settlement, only Carnaby’s Cockatoo Calyptorhynchus funereus latirostris occurred throughout the wheatbelt, with two other species of cockatoo oc¬ curring in parts of the wheatbelt; the Long-billed Corella Cacatua pastinator pastinator [corella nomenclature follows Schodde et al. (1979)] and Major Mitchell’s Cockatoo C. ieadbeateri. Another three species now occur in the wheatbelt; the Galah Eolcphus roseicapilla, Little Corella C. p . gymnopis and the Inland Red-tailed Black Cockatoo Calyptorhynchus magnificus samueli [ Black cockatoo nomenclature follows Adams et ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 659 al. (1984)] . Saunders et al. (1985) discuss the way in which clearing native vegeta¬ tion affected the distribution of all six of these species in southwestern Australia. FIGURE 1 - The past and present distributions of Carnaby’s and Red-tailed Black Cocka¬ toos. The hatched area represents uncleared native vegetation and the remainder repre¬ sents cleared land. The wheatbelt is represented by the unhatched area north and east of the line between Perth and Albany, (a) Distribution of Carnaby’s Cockatoo in 1968. (b) Carnaby’s Cockatoo 1987-89. (c) Red-tailed Black Cockatoo 1961-1965. (d) Red-tailed Black Cockatoo 1987-89. 660 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI Carnaby’s Cockatoo was the only species that formerly occurred over the entire semi- arid zone; a detailed ecological study of it began in 1969 and is still continuing (Saunders 1979, 1980, 1982, 1986, 1990; Saunders & Ingram 1987). This species nested in hollows in eucalypt trees and fed on seeds and flowers of the Proteaceae which dominated the scrub-heath vegetation. The Inland Red-tailed Black Cockatoo formerly occurred along the watercourses of the arid zone (defined here as the area receiving less than 300 mm mean annual rainfall), an area further inland from the wheatbelt or semi-arid zone (see hatched area to the north and east of the wheatbelt in Figure 1). It nested in hollows in eucalypt trees and fed on grass seeds and burrs distributed along the river floodplains. Neither species has taken to feeding on cereal crops, and they have demonstrated entirely different responses to the change in the distribution and abundance of native vegetation. This paper briefly examines these responses and looks at the future of these two species. METHODS The distribution of Carnaby’s Cockatoo in 1968 is based on the results of a survey conducted through schools throughout the southwest of Western Australia. The dis¬ tribution of the Red-tailed Black Cockatoo in the southwestern portion of its range between 1961 and 1965 is based on surveys made by Agricultural Protection Board staff. The present distributions of both species are based on weekly observations by residents throughout the wheatbelt who collected data on the avifauna of their areas between May 1987 and December 1989 (Saunders 1989). The distribution of the Red¬ tailed Black Cockatoo in the arid zone has not been assessed; however, Saunders & Curry ( 1990) list it as an uncommon nomad of riverine woodland in the arid zone. The breeding biology of Carnaby’s Cockatoo was studied at Coomallo Creek (Figure la) between 1969 and the present, and at Manmanning (Figure la) from 1969 until 1977, by which time it had become extinct in that area. Methods, described in Saunders (1982), involved the study of individually marked birds, with particular at¬ tention to their diet, behaviour, who they mated with, where they nested, the growth of their nestlings, nesting success and movements throughout the year. The breeding biology of the Red-tailed Black Cockatoo was studied at Three Springs (Figure la) between September 1974 and December 1981. The Red-tailed Black Cockatoo has two distinct breeding seasons each year: Autumn and Spring (Saunders 1977); this allowed 14 separate breeding seasons to be studied. Methods were the same as those used with Carnaby’s Cockatoo. RESULTS Until 1968 Carnaby’s Cockatoo occurred throughout the wheatbelt; however, 20 years later it had disappeared from over one third of its previous range and is now absent from the central and eastern wheatbelt (Figures la & 1b). In contrast, by the early 1960s, the Red-tailed Black Cockatoo had invaded the northern wheatbelt from the ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 661 adjacent arid zone and was moving into the northeastern edge of the central wheatbelt. Thirty years later this species continues its invasion of the central wheatbelt and is slowly extending its range southward (Figures 1c & Id). Carnaby’s Cockatoo and the Red-tailed Black Cockatoo are approximately the same size [folded left wing 367mm (CC) vs. 379 (RTBC)] and weight (650 g vs. 660 g). The eggs are roughly the same size and weight (32.8 g vs. 33.5 g), representing about 5% of the maternal body weight. Carnaby’s Cockatoo normally lays two eggs (clutch size 1.8, N=494), with about eight days between eggs; the second nestling usually dies within 48 hours of hatching. The Red-tailed Black Cockatoo only lays one egg (451 out of 460 clutches). Both incubation periods are about 29 days. The females of both species carry out all the incubation and brooding, being fed by the male during this time. Carnaby’s Cockatoo usually fledges only one young, however it may occasion¬ ally fledge two young if conditions are favourable (Saunders 1986). The period be¬ tween hatching and leaving the nest is 77 (SD ± 4; N= 1 88) days in Carnaby’s Cocka¬ too, regardless of season or breeding area, while the Red-tailed Black Cockatoo takes 84 (±8; N=147) days or 9% longer. Breeding success, defined as the percentage of nests that produce free-flying young, shows considerable difference between the two species. Carnaby’s Cockatoo at Coomallo Creek had an average breeding success of 64.5% (range 58-86%), based on 529 nests over eight years whose history was followed throughout the breeding season. At Manmanning this species had an average breeding success of 35% (range 7-50%), based on 102 nests over 7 years. The Red-tailed Black Cockatoo had a breeding success of only 37.6% (range 24-66%), based on 428 nests in 1 1 breeding seasons, which represents only 58% of the breeding success of Carnaby’s Cockatoo at Coomallo Creek but is similar to that at Manmanning. One hundred and fifty-one (35.3%) nesting attempts by Red-tailed Black Cockatoos failed at the egg stage and 28 (6.5%) of the total nesting attempts failed probably because feral cats climbed the nest tree and preyed upon the nest contents and/or the breeding female. Predation rates by cats varied and in Spring 1978 it reached a maximum of 17.2% of nesting attempts. In the early 1970s, the breeding population of Carnaby’s Cockatoo at Coomallo Creek averaged 69 breeding attempts per year; however, land clearing in the study area led to a reduction of the breeding population which, since 1977, has averaged 43 breed¬ ing attempts per season (range 38-52). The Red-tailed Black Cockatoo at Three Springs averaged 25 breeding attempts in Autumn (range 12-38) and nearly twice that at 47 in Spring (range 30-61). The occurrence of breeding females in the breeding areas, based on the resighting of individually marked birds is shown for a sample of both species in Table 1. Carnaby’s Cockatoo bred each Spring and returned to the same area to nest. In many cases they returned to the same hollow, or a hollow near the one used previously. Female UF nested regularly in nest 29 and in 1980 and 1986 she was seen in the study area and was almost certainly nesting somewhere but there was no time avail¬ able to search for her nest. The other point of interest is that Carnaby’s Cockatoos were usually successful in their breeding attempts, renesting if they were unsuccessful early in the season (see UC-1975, US-1976, Table 1). 662 ACTA XX CONGRESSUS INTERNATIONAL^ ORNITHOLOGICI o 0 ■ - • ~ 0 TJ c £ >4— 0 O v _ 0 0 0 0 0 0 Q. o 0 O J= JZ O o 0 0 c 0 0 0 0 C 0 O 0 0 0 ■*—> _Q 0 $ 0 0 E CL 5 \ E « c o £ S • - CC ^3 O IS si *2is < o c M“ • J5 ® s P CD -n g 03 ^ > =3^0)^ P P 0 [g > '(0 C T3 V P C "O "O — 0 C -o cc • O) c \_ CL CO o w <= c E &1 .+.-° | CD ~ _3 ca _£? ^ < E 0 » ° CD =J 03 £ C _ C ^ .- 03 - ^ C/3 TD CD -LJ CD 0 J- C 0 03 03 C -C -^T3 c/3 0 -C c w d 03 CD _ *2 CD -q O o £ CD C o ° =6 c — c E p o c 0 0 0 0 0 £ 0 i_ _Q 0 CO 0 0 -o O 0 O 0 Q 0 ^ _Q 0 m ^ G > ° . C33 -Q > c rn 0 E CD _ 0 n c O © a, ■g «T ~ - 0 — r- .= 0 0^3 P £ 0 _ C O -Q O 03 •- 0 E X3 J= 0 Q- Z3 c O 0 C - E -c 0 c £ o 0 0 P 0 0.-00 v_ T3 . „ 0 O o c o _Q 0 _Q 0 0 0 0 E 0 > o c 0 0 0 ZJ o o o 0 Q. 0 — "E P !5 P E 0 0 ES^- 0 o ^ 0 — |_L 0 ■*-’ C C (/) 0 P O 0 i _ 0 0 3^ + 03 0 _C 0) 0 ■*= 0 $ 0 ^ — T- ^ 0 0 D HI t; > 0 •- © j55.5?' < 3 O (- = C i C LL 0 E (J) 0 O O ■*-" CO -* o o O w > n co c i— ro O ET 03 CO 03 CO CO 03 03 e'¬ en CO r- 03 to r- 03 'Tf r- 03 P 0 E 0 C^- + + 03 C\J h- + + + 03 CO 03 CM LO C\J CO CD + + + + 1 + r- CO CO 00 03 CO CM 03 LO CO r- CM CO 1 — 00 00 + 03 03 oo * T_ CO cn CO c + o 00 0 ■M- 00 ■ + 03 0 03 03 LT3 o 0 T_ CM "M- < LO o 00 + + oo 03 03 r- 03 0 CM CO c CO 1 + + 03 c 00 1— 03 r-- 0 03 in CM LO 03 ■ i 0 T_ T— 0 < CO + + + 03 00 LO 03 r- + + 03 CO CM LO 03 o 00 T_ ■St LO CO CM + + + 00 i i 1 00 1 — 03 CO r- T - o r- 03 LO CM o 03 CM LO 00 -r_ T_ < T — c 00 00 + + 0 r- 03 CO 03 0 03 "r~ CO CM 0 T_ co o C c r-- i i c 1 00 + 0 0 h- 03 e-- 0 LO 03 CO 0 0 03 O r^- 0 o CO 0 0 1— T— 0 T— < r- r- 03 CO 0 0 0 0 c c c c I + + + + c m co n 00 O) cd CD Lfi CO N CM o t- i- 0 + co + + + + N 00 O) CM CO ^ CM Tf O Z) o o *-> co .* o o O o ra CD T3 0 CO 4— > I TJ 0 CC CO c c r- 1 0 0 03 CM 0 0 0 0 0 00 0 0 C c c < CO C e-. i + 0 03 Is- CO 0 0 0 0 T_ Is- 00 0 c c CO CO c/) ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 663 The Red-tailed Black Cockatoo demonstrated a different pattern. They rarely used the same hollow to nest in twice and were often unsuccessful. Female BE (Table 1) shows that individual females may nest successfully in both Autumn and Spring (1975). She was not seen in the breeding area the following Autumn, and then re¬ turned to breed unsuccessfully in the Spring. Red-tailed Black Cockatoo females that were known to be alive often did not return to the breeding area each season, and when seen outside the breeding area they sometimes had unmarked young with them, indicating they had bred successfully elsewhere. SI was an example of this pattern. Carnaby’s Cockatoo feeds on the seeds and flowers of a wide range of Proteaceous species, particularly the genera Grevillea, Hakea, Dryandra and Banksia (Saunders 1980), while the Red-tailed Black Cockatoo feeds almost exclusively on the seeds of an agricultural weed Emex australis, an annual and member of the family Polygonaceae. Of 238 observations of Red-tailed Black Cockatoos feeding, 219 (92%) were of feeding on Emex and only nine were of feeding on native vegetation (five spe¬ cies). Twenty-one Red-tailed Black Cockatoos examined post-mortem all had Emex seeds in the crop and five also had seeds of Raphanus raphanistrum, another agri¬ cultural weed, in the crop. DISCUSSION Comparative studies of the breeding biology of Carnaby’s Cockatoo at several loca¬ tions revealed that the birds’ breeding success was markedly lower in areas that had been extensively cleared: the growth rates of nestlings at Manmanning were de¬ pressed compared with birds in areas like Coomallo Creek, where extensive areas of native vegetation remained within sight of the nesting grounds. This mosaic of native vegetation over agricultural landscapes, linking remnants to each other, has allowed this species to survive, even though breeding population sizes have decreased (Saunders 1990). The range of Carnaby’s Cockatoo will continue to contract as it adjusts to the changes imposed on the landscape by humans. The final result may be as little as half its former range. The Red-tailed Black Cockatoo is, however, expanding its range into the semi-arid zone adjacent to its historical distribution in the arid zone. In the semi-arid zone. Eu¬ ropean activities have created a landscape of millions of hectares of riverine plain (Saunders et al. 1985). The Red-tailed Black Cockatoo is gradually invading the ar¬ eas which have heavy infestations of the agricultural weed Emex australis (Gilbey 1974). This plant can be prolific and heavy infestations can provide several million viable seeds to the hectare; furthermore, the seeds may remain viable on, and in, the soil for several years (Gilbey 1974). The Red-tailed Black Cockatoo feeds on this weed almost exclusively and has thus been provided with a constant and plentiful food supply. Nest sites are widely distributed as woodland was left in patches across the landscape. It is not known if the two annual breeding seasons are an adaptation to the semi-arid zone or its native arid zone, as there have been no comparable stud¬ ies of this species in other parts of its range. The birds are not site specific like Carnaby’s Cockatoo. They do remain with conspecifics in feeding flocks and, while breeding, will travel in excess of 40 km per day to join a foraging flock, overflying areas near their nest sites where food is plentiful. The wide fluctuation in numbers of birds breeding in the Three Springs study area each season (12-61) illustrates the 664 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI nomadic breeding pattern. This contrasts with the stable population of Carnaby s Cockatoo at Coomallo Creek, which has remained at just over 40 pairs since 1977. There are several areas of concern in relation to the breeding of the Red-tailed Black Cockatoo in the wheatbelt and these relate to the apparently low breeding success, the high failure rate of eggs and the apparently slow growth rates of the nestlings. Un¬ fortunately there have been no studies of the Red-tailed Black Cockatoo in the arid zone to allow comparisons with the wheatbelt population, so it is necesssary to make comparisons with Carnaby’s Cockatoo which has been studied in both favourable and unfavourable breeding areas (Saunders 1982). The Red-tailed Black Cockatoo with a breeding success of 37.6% of all nesting attempts is similar to that of Carnaby’s Cockatoo (35%) at Manmanning, an area where Carnaby’s Cockatoo became extinct. Particularly worrying is that a species which lays only one egg should have a failure rate of eggs of 35.3%. Observations at nests indicated that the females were not con¬ sistent in their incubation, readily leaving the nest for long periods, something that Carnaby’s Cockatoo did not do, except where food was limiting. In addition, weight increase by Red-tailed Black Cockatoo nestlings was considerably slower than in Carnaby’s Cockatoo and demonstrated the same pattern of growth that nestlings of Carnaby’s Cockatoo showed in areas where Carnaby’s Cockatoo became extinct. These facts seem to indicate that even though the Red-tailed Black Cockatoo is slowly increasing its range, something is affecting breeding success and nestling growth rates. The most obvious cause is the food supply, which is almost exclusively Emex australis. It is possible that a diet of Emex seeds is not suitable for normal growth of nestlings, however it is hard to see how this would affect the behaviour of the adults and result in elevated failure rate of eggs. Emex is often sprayed with a range of her¬ bicides and it is possible that ingestion of these may affect the birds adversely, for example, affecting their behaviour. This possibility has not been tested; however, it is the most likely explanation for the high failure on eggs and the slow growth rates of the nestlings. With a dependence on one species of weed for food, the position of the Red-tailed Black Cockatoo in the areas it has recently colonised is precarious. Agricultural sci¬ entists are working on methods to control and, preferably eradicate Emex. If they are successful, that will see the elimination of the Red-tailed Black Cockatoo from the wheatbelt, making the conservation of this species in the arid zone a high priority. ACKNOWLEDGEMENTS I would like to thank J. Ingram who carried out the field work on the cockatoos, P. de Rebeira who drew the figures, and E. Bucher, P. Arrowood, I. Rowley, E. Russell and G. Smith who provided constructive criticism on this paper. LITERATURE CITED ADAMS, M„ BAVERSTOCK, P. R„ SAUNDERS, D. A., SCHODDE, R., SMITH, G. T. 1984 Biochemi¬ cal systematics of the Australian cockatoos (Psittaciformes: Cacatuinae). Australian Journal of Zooi ogy 32: 363-377. GILBEY, D. J. 1974. Emex species in Australia with particular reference to Western Australia Jour nal of the Australian Institute of Agricultural Science 40: 114-120. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 665 KITCHENER, D. J., CHAPMAN, A., MUIR, B. G. & PALMER, M. 1980. The conservation value for mammals of reserves in the Western Australian wheatbelt. Biological Conservation 18: 179-207. SAUNDERS, D. A. 1977. Red-tailed Black Cockatoo breeding twice a year in the south-west of Western Australia. Emu 77: 107-110. SAUNDERS, D. A. 1979. The availability of tree hollows for use as nest sites by White-tailed Black Cockatoos. Australian Wildlife Research 6: 205-216. SAUNDERS, D. A. 1980. Food and movements of the short-billed form of the White-tailed Black Cocka¬ too. Australian Wildlife Research 7: 257-269. SAUNDERS, D. A. 1982. The breeding behaviour and biology of the short-billed form of the White¬ tailed Black Cockatoo Calyptorhynchus funereus. Ibis 124: 422-455. SAUNDERS, D. A. 1986. Breeding season, nesting success and nestling growth in Carnaby’s Cocka¬ too, Calyptorhynchus funereus latirostris, over 16 years at Coomallo Creek, and a method for assessing the viability of populations in other areas. Australian Wildlife Research 13: 261-273. SAUNDERS, D. A. 1989. Changes in the avifauna of a region, district and remnant as a result of frag¬ mentation of native vegetation: the wheatbelt of Western Australia. A case study. Biological Conser¬ vation 50: 99-135. SAUNDERS, D. A. 1990. Problems of survival in an extensively cultivated landscape: the case of Carnaby’s Cockatoo Calyptorhynchus funereus latirostris. Biological Conservation 54 : 111-124. SAUNDERS, D. A., CURRY, P. J. 1990. The impact of agricultural and pastoral industries on birds in the southern half of Western Australia- past, present and future. Proceedings of the Ecological Soci¬ ety of Australia 16: 303-321. SAUNDERS, D. A., HOBBS, R. J. 1989. Corridors for conservation. New Scientist 1649: 63-68. SAUNDERS, D. A., INGRAM, J. A. 1987. Factors affecting survival of breeding populations of Carnaby’s Cockatoo Calyptorhynchus funereus latirostris in remnants of native vegetation. Pp 249-258 in Saunders, D. A., Arnold, G. W., Burbidge, A. A., Hopkins, A. J. M. (Eds). Nature conservation: the role of remnants of native vegetation. Chipping Norton: Surrey Beatty and Sons. SAUNDERS, D. A., ROWLEY, I., SMITH, G. T. 1985. The effects of clearing for agriculture on the dis¬ tribution of cockatoos in the southwest of Western Australia. Pp 309-321 in Keast, A., Recher, H. F., Ford, H., Saunders, D. (Eds). Birds of Eucalypt forests and woodlands: Ecology, conservation, man¬ agement. Chipping Norton: Surrey Beatty and Sons. SAUNDERS, D. A., SMITH, G. T., ROWLEY, I. 1982. The availability and dimensions of tree hollows that provide nest sites for cockatoos (Psittaciformes) in Western Australia. Australian Wildlife Research 9: 541-556. SCHODDE, R., SMITH, G. T., MASON, I. J., WEATHERLY, R. G. 1979. Relationships and speciation in the Australian corellas (Psittacidae). Bulletin of the British Ornithologists Club 99: 128-137. 666 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICi MALE-MALE, FEMALE-FEMALE AND MALE-FEMALE INTERACTIONS WITHIN CAPTIVE CANARY-WINGED PARAKEET BROTOGERIS V. VERSICOLURUS FLOCKS PATRICIA C. ARROWOOD Department of Biology, New Mexico State University, Las Cruces, New Mexico 88003, USA ABSTRACT. Psittacine birds are known for their monogamous yet gregarious, flocking nature. Few studies, however, have investigated the nature of inter-individual interactions within psittacine flocks. This paper describes the patterns of interactions within captive Canary-winged Parakeet flocks. Affiliative interactions in these monogamously-paired birds are restricted to the mate and recently- fledged young. Within-pair interactions are egalitarian, with reciprocal allopreening, allofeeding and contact-seeking behaviour. Males and females do not pursue affiliative interactions with nonmates. Both males and females have agonistic encounters, in equal proportions, with the same- and oppo- sitely-sexed nonmate individuals. Canary-winged Parakeets seem different from most birds in the ex¬ clusiveness of their monogamy, in the egalitarian nature of their intrapair interactions, and in the equal roles each sex plays in interactions with same- and oppositely-sexed nonmate individuals. Keywords: Canary-winged Parakeet, Brotogeris v. versicolurus, affiliative behaviour, agonistic behav¬ iour, egalitarian behaviour. INTRODUCTION The study of social relationships and intra-specific organization in flocking species is still in its infancy. This is in contrast to the large and growing literature on the advan¬ tages and disadvantages of the phenomenon of flocking. Because of the far-ranging nature of most psittacine flocks outside the breeding season, studies of their behav¬ iour are difficult. This paper uses captive flocks of Canary-winged Parakeets Brotogeris v. versicolurus P.L.S. Muller (nomenclature after Forshaw 1977), a Neotropical parakeet occurring in the wild in flocks of 10-100 individuals (Rocha et al. 1988, Forshaw 1977, Arrowood pers. obs.), to document social relationships, to illus¬ trate what may be some general aspects of psittacine inter-individual behaviour, and, finally, to compare and contrast psittacine social relationships with those documented in other flocking birds. Psittacine birds are characteristically gregarious (having a positive tendency to join others in a more or less peaceful manner, Moynihan 1958), remaining in groups, if not year-round, then during the nonbreeding portion of the year (Forshaw 1977, Kunkel 1974). Maintenance of an exclusive pair relationship year round also characterizes many psittacines, with evidence of long-enduring pairs in some species (Dilger 1960, Rogers & McCulloch 1981, Serpell 1981, Snyder et ai. 1987, Saunders 1982'. Arrowood 1988). Furthermore, in many psittacines there is no sexual dimorphism or dichromatism, and no one has documented any intrasexual physical variability that is related to variability in dominance status as in some other flocking birds (Trivers 1985 Rohwer 1977,1982). For similarly-appearing, paired birds living in flocks, what is the nature of their inter-individual relationships, both within the pair relationship and out¬ side of it? ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 667 METHODS Forty-one individuals and 14 heterosexual pairs of Canary-winged Parakeets were observed in seven different captive flocks (N=1021 individuals/flock) at the University of California, Davis and Irvine, from 1978 through 1988. Some individuals and pairs resided in more than one flock; whenever this occurred, their values from the differ¬ ent flocks were averaged to give a single value for that individual or pair for data analysis. Unpaired individuals and pairs with differing lengths of pair association resided within the flocks. All subjects for this analysis were at least one year old. Mate changes unrelated to death of a pair member were rare. Whenever flock membership was al¬ tered (as in moving birds to a newly-constructed aviary), pairs were kept intact. Sexual identity of the parakeets was established by surgical laparoscopy, success¬ ful breeding, or morphometric measurements (Arrowood, unpubl.). A black dye (Melchior & Iwen 1965) applied to individually-distinct areas of each bird’s plumage made the birds instantly identifiable. All parakeets had numbered aluminum leg bands for permanent identification. The first captive flock was composed of wild-caught, imported birds (Arrowood 1988). Subsequent flocks contained additional wild-caught, imported birds as well as the captive-bred young of flock pairs. I never attempted to tame any of the birds, nor in¬ stituted any kind of domestication process with artificial selection for behavioural or morphological traits. All birds were kept in a captive environment behaviourally rel¬ evant (Price 1984) to this species, i.e., an environment where their normal flocking behaviour (Arrowood, unpublished data on a wild flock, Shroads 1974) could be and was maintained. Social interactions were not restricted and individuals could freely choose mates. Housed in large outdoor aviaries (range of 4.6 m wide x 4.6 m long x 2.2 m high to 4 mw x 12 ml x 3 mh), the birds experienced daily and seasonal environmental changes. Successful breeding in all flocks suggests that reproductive and parental behaviour was not detrimentally affected by captivity. Within the aviaries, the para¬ keets behaved as a flock; movements were often either coordinated or individual/pair movements resulted in the entire group residing in one small part of the aviary. Ter¬ ritoriality did not exist except during the breeding season when pairs became posses¬ sive of nest boxes. None of the data for these analyses includes pairs that had eggs or dependent young; thus, reproduction itself did not contribute to inter-individual or inter-sexual relationships (Erickson 1978). A few offspring maintained extended affiliative relationships with their parents; in these analyses, interactions between parents and their independent, but still affiliative, young are not included. The birds’ behaviour was sampled using five-minute focal animal observations (Altmann 1974). All birds were habituated to my presence, so I sat at one site in the open within each aviary. The 41 parakeets included in these analyses were observed on average 7.74 hours each (SD=5.47; range: 1.42 to 24.17 hours). 668 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI Affiliative (N=14) and aggressive (N=6) behaviours are listed and defined in Arrowood (1988); this analysis also included “nibble” and “move beside” as affiliative behaviours. All statistical tests used two-tailed probabilities. RESULTS Affiliative Interactions Affiliative interactions were almost entirely restricted to the mate. Affiliative acts (N=939) by 13 heterosexually-paired males to their female mates averaged 13.9 acts/ hr (SE=1 .8), compared to 27 acts or 0.1 3 acts/hr/individual (SE=0.1 ) to other females (Wilcoxon Matched Pairs Signed-Ranks Test [hereafter WMPSR], T=0, P<0.005). Twenty-two of the 27 acts to nonmate females were by one male who was the object of interest by four females; however, only four of his 22 acts to these females actu¬ ally involved physical contact (allopreening or allofeeding). The 13 heterosexually- paired males directed only two affiliative acts to other males (0.01 acts/hr/individual; SE=0.01 ). male 0.8 0.7 0.6 03 3 TD % 0.5 1 female nonmate female nonmate male mate Recipient FIGURE 1 - Average hourly rate of aggressive acts by Canary-winged Parakeets against mates and nonmates, x (SE). The rate of acts against nonmates is the average hourly rate per individual nonmate. ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 669 The 13 heterosexually-paired females similarly restricted their affiliative acts (N=1285) to their mates (x= 1 6.9 acts/hr.; SE= 1 .8) compared to other flock males (N=3 acts; x=0.01 acts/hr/individual; SE=0.01) or females (N=4 acts; 37=0.01 acts/hr/individual; SE=0.01). The difference in the hourly act rate to the male mate vs. other males is highly significant (WMPSR Test, T=0, P<0.005). All three affiliative acts by females to nonmate males were movements closer to another male. Thus, in contrast to the one paired male who allopreened and allofed nonmate females, no paired females allopreened or allofed birds other than their mates. Paired females interacted with nonmates of the opposite sex less than paired males did (3 vs. 27 interactions). However, since one male accounted for most of those 27 interactions by males and since most (9 of 13 males; 11 of 13 females) males and females had no affiliative interactions with nonmates of the opposite sex, an analy¬ sis of which partner is more likely to initiate affiliative interactions with the opposite sex is not meaningful. Overall, the Canary-winged Parakeet mates interacted with each other affiliatively at high rates; on average there was one affiliative act within the pair every 2.8 minutes. Although there was a tendency for the female of a pair to direct more affiliative acts to her male mate than vice versa, the difference was not a significant one (WMPSR Test, T=22.5, N=14, P>0.05). Aggressive Interactions Canary-winged Parakeets occasionally act aggressively toward their mates (Figure 1). Males directed aggressive acts toward their mates on average about once every two hours (x=0.42 acts/hr); the rate of aggressive acts by females to their mates was simi¬ lar (x=0.65 acts/hr). Out of 14 heterosexual pairs, no aggressive acts were ever ob¬ served between partners in three pairs, in one pair the male behaved aggressively toward his mate but she did not reciprocate, and in two pairs there were female to male aggressive acts but no male to female aggressive acts. In the remaining eight pairs there was reciprocal aggression. No significant difference exists between the rate of male aggressive acts to the female mate and female aggressive acts to the male mate (WMPSR Test, T=1 7, N=8, P>0.05). In contrast to the lack of affiliative acts to nonmates, aggressive acts against nonmates occurred moderately frequently (Figure 1 ). For this analysis, males (N=1 7) and females (N=24) were used regardless of their pairing status. The rate of male aggressive acts against nonmate females does not differ significantly from the rate of male aggressive acts against other males (WMPSR Test, T=63, N=17, P>0.05). Fe¬ males, similarly to males, behaved as aggressively against males as they did against females (WMPSR Test, T=100, N=23,P>0.05). Even though each gender distributes its aggressive acts equally to males and females within a flock, it is still possible for males and females to differ in the frequencies of their aggressive acts to a particular sex. Indeed, comparing the rate of male aggres¬ sive acts toward other males with the rate of female aggressive acts toward other (nonmate) males, there is a significant difference (Mann-Whitney U Test, z=1.932, n1=17, n2=24, P<0.0536), with males having higher rates of aggressive acts to other males. The converse is not true, however: females are not more aggressive to fe¬ males than males are (Mann-Whitney U Test, z=0.238, P>0.05). 670 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI TABLE 1 - Proportion (%) of nonmate males and females in Canary-winged Parakeet flocks receiving aggressive acts by males and females, ^ (SE). % in Flock Receiving Aggressive Acts Males Females Aggressor 42.1 (6.5) Male 44.0 (4.6) Female 29.3 (5.1) 34.7 (5.4) Males and females may also differ in the proportion of others in the flock to whom they direct aggressive acts (Table 1). That is, females might direct aggressive acts to a greater proportion of females than males in a flock. For each of 24 females and 17 males, I determined how many females and males, out of the flock total for each sex, each female or male had directed aggressive acts to. The frequency of acts to an individual was not considered. An average proportion value was calculated for any bird sampled from more than one flock. The proportion of flock males receiving aggressive acts from females was not differ¬ ent from the proportion of flock females receiving aggressive acts from females (WMPSR Test, T=156, N=23, P>0.05). Similarly, males did not deliver aggressive acts to a greater proportion of flock males than females (WMPSR Test, T=70, N = 17, P>0.05). Males, however, did have a tendency to direct aggressive acts to a greater proportion of flock males than females did to flock males (Mann-Whitney U Test, z=1 .707, P=0.087). Males and females did not differ significantly in the proportion of females in a flock to whom they directed aggressive acts (Mann-Whitney U Test, z=1 .27, P=0.204). DISCUSSION Among monogamous species in which the members of a pair stay together continu¬ ously (see Oring 1982), there may be few other groups in which affiliative interactions between partners occur at as high levels year round as they do in psittacines (but see Zann 1977). Since Canary-winged Parakeets have no courtship displays (Arrowood 1988), this record of frequent affiliative interactions is even more impressive. The benefits of frequent interaction within established pairs have not been tested. Arrowood (1988) found that among newly-established pairs, intrapair affiliative inter¬ actions were not high at the beginning of pairing; thus, frequent affiliative interactions did not seem to be required for the establishment of a pair relationship. In monoga¬ mous grassfinches Poephila , birds that formed pair bonds did not differ in the type or frequency of inter-individual interactions from birds that did not form bonds, but the most rapidly-formed pairs had reciprocally-directed interactions of the same degree (Zann 1977). Enduring season-to-season pairing may be reproductively efficient (Kunkel 1974), with continual affiliative interactions maintaining familiarity between partners, and. thus, making courtship displays unnecessary and/or enhancing behavioural and hormonal synchrony (Erickson 1978). Some captive parakeet pairs do seem to breed more “ef- ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 671 ficiently” than others, but the data are insufficient to determine if differing levels of breeding efficiency are correlated with intrapair affiliative interactions. Alternatively, the advantages to perennial monogamy might accrue outside the reproductive sea¬ son (Oring 1982), selecting, possibly, for increased affiliative interactions during non- reproductive periods compared to the breeding season. Frequent affiliative interactions, on the other hand, may be a mate-guarding strategy; proximity to the mate and behaviours that occupy the mate’s time might prevent both the mate’s gallivanting and incursions by others on the mate. Canary-winged Para¬ keet males, however, have shown no evidence of gallivanting, even though some females have pursued pairing opportunities with them. Adult females, on the other hand, seem to be the ones to leave established pair relationships to pursue pairings (not just copulations) with other males; even so, these cases have been rare in the captive flocks (Hammond, Yamamoto & Arrowood, unpubl.). In monogamous Bewick’s Swans Cygnus columbianus bewickii, Scott (1980) found that partners benefitted from continual spatial proximity; females that remained close to their mates spent more time feeding, were threatened less frequently and were more successful in aggressive interactions than when some distance from their mates. Separated males experienced less severe effects, but alone they were less successful in aggressive encounters and were threatened more by others. No data were given on intrapair interactions. And in Cockatiels Nymphicus hollandicus, Yamamoto et al. (1989) found that birds force paired at the onset of environmental conditions stimulatory to breeding had lower reproductive activity scores than other groups where partners had greater opportunities to exchange social interactions and thus enhance mate familiarity prior to breeding. Aggressive interactions in Canary-winged Parakeets were distributed in an egalitar¬ ian fashion (Hand 1986) to males and females in the flock. Even though parakeet partners sometimes acted together in directing aggressive acts to others, alone they were just as likely to act intersexually as intrasexually. The only significant trend was for males to direct more frequent aggressive acts to males than females did to males. Psittacines may again be unique in that males and females have equal roles in intraflock aggression. Given the psittacine flock organization as a grouping of persistent pairs, our notions of male-male aggression, female subordination to males, and male-female courtship, based mainly on temperate territorial species, may have to be modified. Even com¬ pared to other flocking species, psittacines may show many differences. ACKNOWLEDGMENTS I am grateful to the American Ornithologists' Union and to the Biology Department, New Mexico State University, for financial assistance to attend the 20th IOC. The manuscript benefitted from review by Roy Arrowood, Jr., Enrique Bucher, Katherine Dickerson, Ian Rowley and Denis Saunders. 672 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI LITERATURE CITED ALTMANN, J. 1974. Observational study of behavior: Sampling methods. Behaviour 49: 227-267. ARROWOOD, P. C. 1988. Duetting, pair bonding and agonistic display in parakeet pairs. Behaviour 106:129-157. DILGER, W. C. 1960. The comparative ethology of the African parrot genus Agaporms. Zeitschrift fur Tierpsychologie 17: 649-685. ERICKSON, C. J. 1978. Sexual affiliation in animals: Pair bonds and reproductive strategies. Pp. 697- 725 in Hutchison, J.B. (Ed.). Biological determinants of sexual behavior. New York, John Wiley and Sons. FORSHAW, J. M. 1977. Parrots of the world. Neptune, N. J., T. F. H. Publications, Inc. HAND, J. L. 1986. Resolution of social conflicts: Dominance, egalitarianism, spheres of dominance, and game theory. Quarterly Review of Biology 61: 201-220. KUNKEL, P. 1974. Mating systems of tropical birds: The effects of weakness or absence of external reproduction-timing factors, with special reference to prolonged pair bonds. Zeitschrift fur Tierpsychologie 34: 265-307. MELCHIOR, H. R., IWEN, F. A. 1965. Trapping, restraining and marking Arctic ground squirrels for behavioral observations. Journal of Wildlife Management 29: 674-678. MOYNIHAN, M. 1958. Some adaptations which help to promote gregariousness. Proceedings of the Xllth. International Ornithological Congress: 523-541. ORING, L. W. 1982. Avian mating systems. Pp. 1-92 in Farner, D.S., King, J.R., Parkes, K.C. (Eds). Avian biology, Vol. VI, New York, Academic Press. PRICE, E. 0. 1984. Behavioral aspects of animal domestication. Quarterly Review of Biology 59: 1-32. ROCHA, C.F.D. da, BERGALLO, H. de G., SICILIANO, S. 1988. Migragao circadiana em cinco especies de psitacideos em Parintins-AM. Acta Amazonica 18: 371-373. ROGERS, L. J., MCCULLOCH, H. 1981. Pair-bonding in the Galah Cacatua roseicapilla. Bird Behav¬ iour 3: 80-92. ROHWER, 5. 1977. Status signaling in Harris’ sparrows: Some experiments in deception. Behaviour 61: 107-129. ROHWER, 5. 1982. The evolution of reliable and unreliable badges of fighting ability. American Zoolo¬ gist 22: 531-546. SAUNDERS, D. A. 1982. The breeding behaviour and biology of the short-billed form of the White¬ tailed Black Cockatoo Calyptorhynchus funereus. Ibis 124: 422-455. SCOTT, D. K. 1980. Functional aspects of the pair bond in winter in Bewick’s Swans Cygnus columbianus bewickii. Behavioral Ecology and Sociobiology 7: 323-327. SERPELL, J. 1981. Duets, greetings and triumph ceremonies: Analogous displays in the parrot genus Trichoglossus. Zeitschrift fur Tierpsychologie 55: 268-283. SHROADS, C. V. 1974. Studies on a population of the Canary-winged Parakeet Brotogeris versicolurus (P.L.S. Muller) in Dade County, Florida (Aves: Psittacidae). Masters thesis, University of Miami. SNYDER, N. F. R., WILEY, J. W., KEPLER, C. B. 1987. The parrots of Luquillo: Natural history and conservation of the Puerto Rican Parrot. Los Angeles, California, Western Foundation of Vertebrate Zoology. TRIVERS, R. 1985. Social evolution. Menlo Park, California, Benjamin/Cummings Publ. Company YAMAMOTO, J.T., SHIELDS, K.M., MILLAM, J.R., ROUDYBUSH, T.E, GRAU,C.R. 1989. Reproduc¬ tive activity of force-paired Cockatiels Nymphicus hollandicus. Auk 106: 86-93. ZANN, R. 1977. Pair-bond and bonding behaviour in three species of grassfinches of the genus Poephila (Gould). Emu 77: 97-106. ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 673 NESTING BEHAVIOUR OF THE BAHAMA PARROT AMAZONA LEUCOCEPHALA BAHAMENSIS ON ABACO ISLAND, BAHAMAS ROSEMARIE S. GNAM Department of Ornithology, Amer. Mus. Nat. Hist., New York, NY 10024, USA and Department of Biology, The City College of New York, New York, NY 10031, USA ABSTRACT. Although the Bahama Parrot was once abundant and ranged throughout the Bahamas archipelago, it is endangered now and survives only on two islands- Abaco and Great Inagua. Bahama Parrots on Abaco nest in limestone solution cavities beneath the ground, a habit unique among New World psittacines. Between 1985 - 1988, I located and monitored 76 parrot nests in two nesting areas on southern Abaco. Eggs were laid in late May and early June and hatched asynchronously, 26-28 days after the female began incubation. While the female incubated the eggs, the male visited the nest on average four times per day to feed her. After the first week post-hatching, the female left the nest to forage with the male. Parents returned to their nests four to six times per day to feed the nestlings. Chicks fledged asynchronously in late August and early September, 56-58 days after hatching. Bahama Parrots exhibited low reproductive success; successfully nesting pairs fledged a mean of 1.79 ± 0.16 chicks per nesting effort. Keywords: Bahama Parrot, Amazona leucocephala bahamensis, Abaco Island, nesting behaviour, parental care, subterranean nesting. INTRODUCTION The Cuban Parrot Amazona leucocephala is a polytypic species with five recognized subspecies: leucocephala (Cuba), palmarum (Western Cuba and Isla de la Juventud), caymanensis (Grand Cayman), hesterna (Cayman Brae) and bahamensis (Bahamas) (Bond 1956). Although the Bahama Parrot was probably present on all major islands in the Bahama archipelago, historically it was recorded from Abaco, New Providence, San Salvador, Long Island, Crooked Island, Acklins and Great Inagua. Today, this species is listed as endangered and persists only on the islands of Abaco and Great Inagua. In recent years, the Abaco population has declined as a result of habitat de¬ struction, logging activities, development, Hurricane Betsy in 1965, and hunting pres¬ sures (Attrill 1981, Snyder et al. 1982). Bahama Parrots were studied by Snyder et al. (1982), who estimated the Abaco popu¬ lation to number less than 1000 birds. Current popufation estimates range from 860 to 1300 parrots (Gnam 1991). Unlike the Inagua population and other subspecies of leucocephala which nest in tree cavities, parrots on Abaco nest in limestone solution cavities beneath the ground, a habit unique among New World parrots (Forshaw 1989). The Abaco population of the Bahama Parrot is found in Caribbean pine Pinus caribaea and mixed broadleaf coppice (native, evergreen hardwood) areas of south¬ ern Abaco. As part of a larger investigation into the breeding biology of bahamensis, I studied the nesting behaviour of the Abaco population. With the notable exception of studies on 674 ACTA XX CONGRESSUS INTERNATIONAL^ ORNITHOLOGICI the Puerto Rican Parrot Amazona vittata (Snyder et al. 1987), the Kakapo • rig P habroptilus (Merton et al. 1984) and Australian psittacines (Row ey » Wyndham1981 , Saunders 1982), data on the nesting behaviour of parrots in the wild are often anecdotal, fragmentary and based on the observations of one or two nes ing pairs. The objectives of my study were to (1) investigate the behaviour o t e Bahama Parrot throughout the stages of its nesting cycle, (2) determine the pattern of parental care, and (3) compare its nesting behaviour with other Caribbean Amazona. STUDY AREA AND METHODS I studied Bahama Parrots in an area 64 km south of Marsh Harbour (Latitude 26° N, longitude 78° W), Abaco Island, from early May to September each year, 1985 through 1988. Since 1985, I located 76 nests but not all of these nests were active in a given year. Clutch size was recorded in all active nests and nests were checked at least once a week until chicks fledged or the nest failed. Observation blinds were placed 10-15 m from the nests of five pairs and the behaviour of these pairs was re¬ corded from sunrise to sunset (a 14-15 hour period). These nesting pairs were ob¬ served at least once a week from egg-laying until chicks fledged or the nest failed. Due to risks associated with tagging parrots (Saunders 1988), I did not try to capture and tag individuals, but instead recognized them by the usually considerable differ¬ ences in their plumage markings and other physical characteristics. These physical differences were consistent through a season and from year to year. Initially, the five nesting pairs to be observed were selected randomly from a pool of active parrot nests but in succeeding years, pairs which returned to the same nest cavity were given observational preference over newly found nests. The external and internal dimensions of nest cavities were measured to the nearest 0.5 cm. Nest depth was measured from the lip of the nest to the floor of the nest cav¬ ity. I used the SAS (1985) software package on an IBM mainframe system (at the City University of New York) for statistical analyses. Data were pooled over the four years of study and provide a general view of nesting behaviour in this unique population. Unless otherwise specified, mean values with their standard errors are reported. RESULTS Bahama Parrots on Abaco are monogamous and seasonally defend their nest site. Mates remain together throughout the nesting cycle to produce a single brood. Nest site characteristics Although Carraway and Carraway (1979) reported a Bahama Parrot nest on Abaco in the hollow of a Pond Top Palm Saba! palmetto and I searched for nests in this area, all of the nests which I located were in limestone-solution cavities beneath the ground, the normal pattern for this population. Nest cavity entrances measured on average ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 675 18.3 ± 0.9 (S.E.) cm vertically and 19.8 ± 1.0 cm horizontally. Nest cavities were more spacious internally and on average measured 39.9 ± 2.4 cm by 27.9 ± 1.1 cm. The mean nest depth (N=70) was 125.1 ± 6.6 cm; 74.3 percent of these nests ranged from 51 - 150 cm in depth. Nine percent of the nests had two entrances and 74% (N=52) of the nest cavities had internal ledges or rocky overhangs within them to protect and conceal the eggs. Incubation Clutches of 2 - 6 eggs (mean 3.6 ± 0.2) were laid in late May or early June. The fre¬ quency of 2-egg nests was 4, 3-egg nests was 19, 4-egg nests was 31, 5-egg nests was two and 6-egg nests was one. Egg-laying was asynchronous and eggs were generally laid at two day intervals. Eggs were incubated exclusively by the female. Most females began incubation with the laying of the first egg. Incubation may have been irregular until the clutch was completed. The incubation period was 26 - 28 days. Incubating females never left their nests except to be fed by their mates. There was no relationship (polynomial regression, P>0.05) between the time that a female spent off the nest and the stage of incubation. Females spent an average of 62.1 ± 3.9 min¬ utes per day off the nest. Males spent little time in the nest area except when feeding their mates. When males arrived in the nest area, they perched in a nearby pine or shrub and called to their mates. Females exited the nest on average 1 1 .5 + 1.4 minutes after the male arrived to feed them. Females recognized and responded to the calls of their own mates and never left their nests when neighbouring males called. During incubation the male on average fed the female four times per day with food transfers (N=1 70) lasting an av¬ erage of 16.0 + 0.9 minutes for the various pairs. Most feedings occurred in the morn¬ ing and late evening. Generally, males fed the female in nearby pines in the nest area. A female was fed in the nest only if she remained in the nest after repeated, unsuc¬ cessful male visits and calling. During egg-laying, copulation often followed food trans¬ fers. Copulatory behaviour followed the pattern described for other Amazona species (Skeate 1984, Snyder et al. 1987). The male perched beside the female with one foot and rested his other foot on his mate’s back while she swayed her tail back and forth horizontally against his tail, making cloacal contact. Pairs gave loud, territorial high-squawk calls when they returned to the nest. Females then immediately entered the nest and resumed incubation. Males flew off to forage and roost, often accompanied by neighboring males. Males did not roost in the nest or nest area in the evening. They roosted communally in large flocks (50-60 parrots). Two roost sites were located 300 m and 1000 m from the main nesting area. Nestling Period Eggs hatched asynchronously in late June and early July. The pattern of female care remained unchanged from that which was seen during incubation until about a week after hatching (Figure 1). At this time females started to forage with the males, spend¬ ing more time off the nest but returning to feed and brood the chicks. There was no apparent relationship between brood size and the amount of time the female spent off the nest (ANOVA, P=0.63) (Figure 1). As the chicks grew older, the time that the fe¬ male spent off the nest increased until she foraged consistently with her mate during the day and returned to the nest only to feed the chicks (Figure 1). By the fourth week post-hatching, most females no longer roosted in the nest; at this stage, the chicks’ body feathers had erupted from the shafts, particularly on the back, wings and thighs. 676 ACTA XX CONGRESSUS INTERN ATI ON A LIS ORNITHOLOGIG! Males continued to feed the females directly the first week post-hatching an rare y entered the nest to feed the chick$(Tab+e 1). As females decreased daytime roo ing of the chicks, males gradually increasingly entered the nest and fed the c ic s directly (Table 1). In contrast to the female’s barely observable, secretive entry in o a nest, the male’s initial entries were awkward and took longer. Males lingered on t e nest lip, obviously uneasy at entering. 1 1 brood size 1 brood size 2 brood size 3 10 Weeks after Hatch FIGURE 1 - Time spent off the nest as a function of brood size for female Bahama Par¬ rots Amazona leucocephala bahamensis on Abaco Island, Bahamas, 1985-1988. Values shown are means with their standard errors. Samples sizes (N) for brood size one were 4, 8, 4, 2, 4,1; for brood size two : 6, 8, 5, 7, 2, 5, 5, 3, 3 and for brood size three : 5, 4, 4, 4, 4, 4, 4, 4. Once females no longer roosted in the nest, males and females returned together to the nest four to six times per day to feed their chicks. Males and females showed no differences (ANOVA, P=0.53) in the amount of time that they spent in the nest after the female ceased overnight roosting (Figure 2). Visit time decreased for both males and females at the same rate( ANOVA, P= 0.44) as fledging approached (Figure 2). Fledging Chicks fledged asynchronously (usually 24-48 hours apart) in late August and early September, 56-60 days after hatching. Several days before fledging, chicks began to appear at the nest lip when their parents were in the nest area. The chicks’ initial appearances were brief (less than two minutes) but gradually, more of the chick’s body protruded from the nest cavity until the chick fledged. As fledging approached (eight weeks post-hatching), parents spent little time inside the nest and despite the chicks’ vigorous begging, feedings were brief (Figure 2, Table 1). From a tree or a shrub within 3 m of a nest, parents called to the chicks, apparently coaxing them from ACTA XX CONGRESSUS INTERNATIONALE OHNITHOLOGICI 677 TABLE 1 - Nest attendance by male Bahama Parrots Amazona leucocephala bahamensis on Abaco Island, Bahamas, 1985-1988. Nestling period Daily nest visits Total time spent in nest Males - (minutes per day) (week)* (N) Mean SE Mean SE 1 15 0.5 0.2 5.4 2.6 2 20 1.5 0.4 16.7 4.7 3 14 2.6 0.5 20.4 4.9 4 14 4.1 0.2 29.0 3.2 5 12 4.9 0.3 32.3 4.1 6 11 5.5 0.6 28.9 4.2 7 10 4.2 0.6 19.1 4.1 8 10 3.6 0.5 12.4 2.8 9 3 1.7 0.7 4.3 1.9 * Day first chick hatched counted as day 1. 100 -n 90 - 80 H 4 5 6 7 8 9 10 Weeks after Hatch FIGURE 2 - Total time spent in the nest by male and female Bahama Parrots Amazona leucocephala bahamensis on Abaco Island, Bahamas, 1985-1988. Values shown are means with their standard errors. Sample sizes (N) for males and females were 14,12,11, 10, 10, 3. the nest. I observed nine fledgings; 67% occurred in the morning and always when the parents were present. Upon fledging, chicks flew considerable distances (> 300 678 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI m) accompanied by their parents; chicks had left the nest area by sunset. All fledg¬ ing flights observed were silent and, except for begging during food transfers, the young and their parents remained virtually silent in the days following fledging. Nesting Success An extensive analysis of nesting success data is presented in Gnam (1991), but a brief review of these results follows here. During the years 1985-1988, the mean clutch size at hatching was 3.5 ± 0.1 but the mean number of parrot chicks hatched per nest was 1 .88 ± 0.1 . Forty-six percent of all egg-laying pairs fledged young. Suc¬ cessful pairs which'hatched at least one chick, fledged a mean of 1 .79 ±0.16 chicks. Causes of nest failure were (in decending order of importance): predation by feral cats Felis catus, land crabs Cardisoma guanhumi, snakes, and rats Rattus rattus ; chick deaths from unknown causes; abandoned eggs; flooding of nest cavities; and human disturbance (poaching of chicks). DISCUSSION The general pattern of nesting behaviour in Bahama Parrots follows that seen in other species of Amazona (Snyder et al. 1987, Silva 1989). Some exceptions to this pat¬ tern are noteworthy and likely related to this species’ ecology. Although egg-laying in most Caribbean Amazona species occurs in late winter or early spring (Feb. - Mar.) (Snyder et al. 1987), the Abaco population of the Bahama Parrot lays its eggs in late spring. Its late breeding season coincides with the peak abundance and availability of its food sources during nesting (Gnam 1991). Caribbean pine produces immature(unripe) pine cones during June through August and poisonwood Metopium toxiferum and wild guava Tetrazygia bicolor fruit at this time. These food items ac¬ counted for 74% of the observed diet during the nesting period (Gnam 1991). Although the incubation pattern of Puerto Rican and Bahama Parrot females is simi¬ lar (females spend less than 100 minutes per day off the nest to be fed by their mates), the female patterns differ once nestlings appear (Snyder et al. 1987). During the fourth week post-hatching, female Bahama Parrots no longer roost in the nest in the evening, unlike female Puerto Rican Parrots which roost in the nest until chicks fledge. White-tailed Black Cockatoo Calyptorhynchus funereus females cease roosting in the nest about 2-3 weeks before chicks fledge (Saunders 1982). In the Bahama Parrot, roosting in the nest with chicks may be related to the feather and thermoregulatory development of the chicks, rather than to nest defense from preda¬ tion. Abaco has been devoid of mammalian predators until the relatively recent intro¬ duction of cats and rats by man. On at least two occasions, I observed a female who had ceased overnight nest roosting return to her nest and brood chicks during a heavy rainfall. In the Puerto Rican Parrot, females with larger sized broods reach a plateau of time off the nest earlier in the nestling period than do females with small broods (Snyder et al. 1987). Brood size seems not to affect the time that female Bahama Parrots spend off the nest. My sample size may be too small to detect differences; alterna¬ tively, presently unknown ecological factors may be responsible. Subterranean nest¬ ing may provide a more stable thermal environment, so that females with even small¬ sized broods can safely leave chicks. Food sources may be limiting, patchily distrib- ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 679 uted and/or require considerable handling time and therefore, regardless of brood size, females must forage after the first week post-hatching to feed themselves and their chick(s). The parental care pattern of Bahama Parrots in which females are responsible for the incubation of eggs and brooding of nestlings and both sexes feed the nestlings, is similar to that seen in large psittacines, such as the Puerto Rican Parrot (Snyder et al. 1987) and White-tailed Black Cockatoo (Saunders 1982). Although male and fe¬ male Bahama Parrots visited the nest together to feed the chicks and spent compa¬ rable amounts of time in the nest, they may still differ in their interactions with the nestlings. We could not observe behaviour within a nest. In large broods, parents may feed the chicks differentially as do parakeets (Arrowood & Flint in press). Despite considerable parental care and time expenditure, Bahama Parrots fledge few offspring and reproductive success is low when compared with other Caribbean Amazona species (Gnam 1991). Eighty-two percent of all egg-laying Plispaniolan Parrots Amazona ventralis fledge 2.5 chicks while 69% of all egg-laying Puerto Rican Parrots fledge 1.8 chicks (Snyder et al. 1987). Various factors affect the reproductive potential and output of this island species (Gnam 1991); but its ground-nesting habit has put the species at risk from nest predation by introduced mammals. Nesting be¬ haviour patterns which were selected for in stable environments may be disadvanta¬ geous in a rapidly changing environment. ACKNOWLEDGEMENTS This research represents a portion of a dissertation submitted in partial fulfillment of the requirements for a Ph.D. degree at the Graduate School of the City University of New York. The Bahamas government through the Ministry of Agriculture, Trade and Industry gave me permission to study the Bahama Parrot. Financial assistance is gratefully acknowledged from AAZK Chapter - Audubon Zoo; the American Federation of Aviculture; the Frank M. Chapman Memorial Fund, Ameri¬ can Museum of Natural FHistory ; Friends of the Abaco Parrot; ICBP-Pan American Section; ICBP-U.S. Section; the James Bond Research Fund-Amazona Society; J.McOmie and the Utah Avicultural Society; Mr. Fables Wildlife Conservation Fund; the Roger Tory Peterson Institute; the United States Fish and Wildlife Service, Office of International Affairs; Wildlife Conservation International, New York Zoological So¬ ciety; Wildlife Preservation Trust International; the Women's Research and Develop¬ ment Fund, the City University of New York; the World Nature Association, and the Zoologische Gesellschaft fur Arten und Populationsschutz. During the course of this study, I was supported by an award from the Leopold Schepp Foundation, the Mina Rees Dissertation Fellowship (CUNY) and by a National Institute of Mental Health Animal Behavior Training Grant (MH 1 5341 07A1 ) to Flunter College’s Biopsychology Program. The American Ornithologists’ Union and the Frank M. Chapman Memorial Fund, American Museum of Natural History kindly provided me with travel awards to attend the XXth IOC in New Zealand to present this paper. 680 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI I thank P. Arrowood, E. Bucher, D. Saunders, R. Rockwell, and J. Wiley for their con structive comments on the manuscript. LITERATURE CITED ARROWOOD, P. C., FLINT, E. N. In press. Patterns of parental care in parakeets. Proceedings, ICBP Parrot Specialist Group Meeting, Brazil, 1988. ATTRILL, R. 1981. The status and conservation of the Bahamas Amazon Amazona leucocephala bahamensis. Pp. 81-87 in Pasquier, R. F. (Ed.). Conservation of New World Parrots. ICBP Technical Publication Number 1. Washington, D.C., Smithsonian Institution Press. BOND, J. 1956. Check-list of birds of the West Indies. 4th edition. Lancaster, Pennsylvania, Wickersham Printing Company. CARRAWAY, C., CARRAWAY, P. 1979. The Bahamian Parrot: Amazona leucocephala bahamensis. Avicultural Magazine 85: 18-23. FORSHAW, J. M. 1989. Parrots of the World. 3rd Edition. Melbourne, Lansdowne Editions. GNAM, R. 1991. Breeding biology of the Bahama Parrot Amazonal eucocephala bahamensis. PhD. dis¬ sertation. New York, City University of New York. MERTON, D.V., MORRIS, R.B., ATKINSON, I.A.E. 1984. Lek behaviour in a parrot: the Kakapo, Strigops habroptilus of New Zealand. Ibis 126: 277-283. ROWLEY, I. 1980. Parent-offspring recognition in a Cockatoo, the Galah, Cacatua roseicapilla. Aus¬ tralian Journal of Zoology 28: 445-456. SAS INSTITUTE. 1985. Statistical Analysis Systems. 5th edition. Cary, North Carolina, SAS Institute Inc. SAUNDERS, D. A. 1982. The breeding behaviour and biology of the Short-billed form of the White¬ tailed Black Cockatoo Calyptorhynchus funereus. Ibis 124: 422-455. SAUNDERS, D. A. 1988. Patagial tags : Do benefits outweigh risks to the animal? Australian Wildlife Research 15: 566-569. SILVA, T. 1989. A monograph of endangered parrots. Pickering, Ontario, Silvio Mattacchione & Com¬ pany. SKEATE, S. T. 1984. Courtship and reproductive behavior of captive White-fronted Amazon Parrots Amazona albifrons. Bird Behaviour 5: 103-109. SNYDER, N. F. R., KING, W. B., KEPLER, C. B. 1982. Biology and conservation of the Bahama Par¬ rot. Living Bird 19: 91-1 1 4. SNYDER, N. F. R., WILEY, J. W., KEPLER, C. B. 1987. The parrots of Luquillo: natural history and conservation of the Puerto Ricarr Parrot. Los Angeles, Western Foundation of Vertebrate Zoology. WYNDHAM, E. 1981. Breeding and mortality of Budgerigars Melopsittacus undulatus. Emu 81: 240- 243. ACTA XX CONGRESSUS INTERNATIONAL^ ORNITHOLOGICI 681 SOCIAL BEHAVIOUR AND POPULATION DYNAMICS OF THE MONK PARAKEET E. H. BUCHER, , L. F. MARTIN, M. B. MARTELLA and J. L. NAVARRO Centro de Zoologia Aplicada, Universidad de Cordoba, Casilla de Correos 122, Cordoba 5000, Argentina. ABSTRACT. A marked population of Monk Parakeets Myiopsitta monachus was studied for eight years in two localities in Cordoba, Argentina. We made observations on behaviour and measured population density, natality, mortality, emigration and immigration. Monk Parakeets are resident year-round, with limited movements during winter. They nest solitarily or in colonies. The Monk Parakeet is unique among parrots in that it builds large enclosed stick nests, which are often integrated in a compound nest that may include several isolated chambers. Parakeets continue to roost within their nests through¬ out the whole year. Single nests often develop into large compound nests in which different breeding pairs occupy separate nesting chambers. Members of a compound nest may also include nonbreeding individuals occupying separate chambers. These parakeets do not defend an all-purpose territory, defense being restricted to the nests. Juveniles continue to roost in the parental nest during autumn and winter but leave permanently before the following breeding season. Dispersal distance from birth to breeding place is reduced, averaging 1.2 km. We observed a few cases of allofeeding; helping in nest building, maintenance, and vigilance were observed in all birds occupying a compound nest. Breeding was delayed until at least two years of age. Annual survival of first year birds was estimated to be 61%, whereas adult survival was 81%. Our results suggest that the Monk Parakeet’s breeding system and population dynamics show several characteristics of a communal breeder, a breeding sys¬ tem not previously recorded in parrots. Keywords: Monk Parakeet, parrots, Chaco, Argentina, population dynamics, communal breeding, birds as agricultural pests. INTRODUCTION The Monk Parakeet Myiopsitta monachus occurs in Argentina, Bolivia, Brazil, and Uruguay, where it is considered an agricultural pest (Bucher & Bedano 1976). It is also trapped for the pet trade, around 19,000 birds being imported annually into the USA alone (Nilsson 1990). Despite intensive killing from control campaigns and trap¬ ping, the bird is still abundant throughout its range. Moreover, it has become estab¬ lished in other countries where it escaped or was released accidentally, including Puerto Rico and the continental USA (Bull 1973, Neidermyer & Hickey 1977). The species is unique among parrots in that it builds its own nest of sticks. Nests are often integrated into a compound nest that may include several isolated chambers, each one occupied by a different pair (Forshaw 1978), resembling those of the So¬ ciable Weaver Philetairus socius of Africa (Collias & Collias 1977). As indicated by Brown (1987), compound nests or lodges have been traditionally omitted from most discussions of avian helping and communal breeding systems but nevertheless de¬ serve attention given that collaboration in nest building represents an important form of helping (see below). Intraspecific helping has been found in the Sociable Weaver (Collias & Collias 1977), but not in wild populations of the Monk Parakeet, although S. Emlen (in press) noticed the existence of helpers in captive colonies. 682 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI In this report we provide a broad outline of the basic features of the social organiza tion, breeding biology, and population dynamics of the Monk Parakeet. We discuss the possible existence of communal breeding behaviour as well as the adaptive forces that may have favoured its evolutionary development in the Monk Parakeet. METHODS Research was conducted from 1982 to 1989 in two different study sites in the prov¬ ince of Cordoba, Argentina. The first is located at San Antonio Ranch (31 25 S, 62 59’W), 8 km east of the town of Arroyito; the second at Los Leones Ranch (31° 05 S, 64° 1 1 ’ W), 1 0 km south of Jesus Maria city. Both areas are under a semi-arid climate regime, rainfall being 870 mm at Jesus Maria and 700 mm at Arroyito. The rainfall pattern is very seasonal, with around 80% of the precipitation concentrated during summer (October-March). The original landscape was parkland with patches of xerophytic woodland, which has been largely cleared for agriculture and cattle-rais¬ ing. Observations on general behaviour were made at San Antonio Ranch from an obser¬ vation tower located near nests on both native and introduced trees (mostly Eucalyp¬ tus trees). Birds were trapped from the nests at night using a large net able to em¬ brace an entire compound nest (Martella et al. 1987). The parakeets were kept in small cages during the rest of the night, marked with aluminium rings the following morning and then released. We used anodized colour aluminium rings to identify in¬ dividual birds from a distance, but it proved to be extremely difficult because the Para¬ keet’s legs are very short and covered with feathers. Coloured plastic rings or wing tags also proved unusable because they were rapidly destroyed by the birds, which usually resulted in serious self-inflicted injuries. Some of the birds were also dyed with Rodamine B to help long distance identification. Population dynamics studies were carried out at Los Leones Ranch from 1983 to 1989 in a study area of 604 ha of native vegetation. Total population counts were obtained through trapping birds from the nests at night in March (post-reproductive season), June (winter), and September (pre-reproductive season), complemented with direct observations during daytime. All captured birds were banded with numbered alu¬ minium rings. Recapture of banded birds from the study area and from within a 200 m wide strip surrounding the study area allowed us to estimate survival, emigration, and immigration. During the breeding season all nests below a height of 6 m (the upper limit of our accessibility via a portable ladder) were inspected every 10 days on average, and nestlings were banded with numbered rings at about 20 days old. These observations provided data on productivity, reproductive success, and dispersal. We were unable to differentiate sexes from external characteristics, neither could we age adult birds However, fledglings could be distinguished from adults with reasonable accuracy during the first three months of life because they do not undergo an autumn moult (Navarro et al. unpublished data). ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 683 RESULTS Habitat requirements The Monk Parakeet inhabits semi-arid savanna woodlands and thickets with xerophytic vegetation, particularly the Chaco scrub and woodland, where it prefers open areas with scattered trees. Availability of tall trees or poles is essential for nest¬ ing. Monk Parakeets prefer to locate their nests on the highest available perches, from 3 m to 25 m or more in Eucalyptus trees. Lack of nesting habitat may be limiting, at least in some areas within the parrot’s range, as shown by the Monk Parakeets ten¬ dency to expand into areas with low vegetation whenever introduced tall trees or elec¬ tricity poles become available (Bucher & Martin 1987). Consistent with this tendency, the species has expanded considerably and become abundant in agricultural areas of the open Pampas of Argentina, Uruguay and Brazil, following the widespread in¬ troduction of Eucalyptus trees (Forshaw 1978, Ridgely 1980, Bucher & Martin unpubl. data). Food and foraging Monk parakeets feed both in trees and on the ground. Seeds are the staple food, particularly from small grasses (Bucher et al. unpubl. report) Thistles are very impor¬ tant during the breeding season. The birds also eat fruits and flowers, as well as ani¬ mal food in some cases (mostly insect larvae) (Forshaw 1978, Bucher et al. unpubl. data). Monk Parakeets also feed on grain crops (particularly corn, sunflower, and sorghum) as well as cultivated fruit trees like peaches and pears (Bucher & Bedano 1976). Social system Spacing and feeding behaviour. Monk Parakeets may nest solitarily or in colonies. Colonies include single compound nests, in which within one large structure each pair has its own chamber with an entrance tunnel, or an aggregation of single and com¬ pound nests usually on a single tree or on a few closely-located trees. The parakeets do not defend an all-purpose territory, territorial defense being restricted to the nest colonies. Monk Parakeets feed in flocks of up to several hundred individuals in win¬ ter, flocks being restricted to one to four individuals during the breeding season. Feed¬ ing areas change throughout the year. During the non-breeding season, the birds stay in the general feeding area, roosting in any nests available in the area. In these situ¬ ations up to three or four individuals may roost in each chamber (Martin & Bucher unpubl. data). The basic social unit. The basic reproductive unit is a breeding pair occupying a nest¬ ing chamber, either in a solitary nest or in a compound nest. However, members of a compound nest or a nesting colony may include non-breeding individuals occupy¬ ing separate chambers (usually in pairs or trios). All members of a compound nest, including non-breeding adults and yearlings, participate in bringing material to the nest and building and maintenance activities (Martella & Bucher in press). Social interactions among group members are varied and occur constantly through¬ out the day. The species is very noisy, with calls uttered almost constantly. At least 9 different calls have been identified that elicited specific responses from other indi¬ viduals when played back (Martella & Bucher 1990). 684 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI Breeding behaviour. Nests are made with thorny twigs. Monk Parakeets spen an impressive amount of time and energy in carrying and manipulating material for nes building and maintenance. Nest building and maintenance is particularly intense in spring before the start of the breeding season and in early autumn (Figure 1) (Marte a & Bucher in press). Compound nests originate and grow as other birds build their nests attached to existing chambers. Sometimes newcomers have to overcome strong aggressive challenges from the original nest’s occupants. Providing that adequate support is available, nests may grow to enormous proportions. For example, a geo¬ desic tower in Cordoba supported three massive compound nests, which included more than 200 chambers. FIGURE 1 - Annual variations in the number of sticks being brought to the nest by Monk Parakeets (data from Martella and Bucher in press.). Average nest size in our study area in Jesus Maria was small: 73% of the studied nests had only one chamber, 19% had two, 6% three, and 2% four (n= 123) (Navarro & Bucher unpubl. data). The high predominance of single-chamber nests may have been the consequence of previous nest destruction by the ranch owners. Lack of adequate support may also play an important role in limiting nest size. Strong winds and storms cause the falling of many nests. Larger nests appear to be more likely to fall than smaller nests. In old nests parasites become very abundant (particularly the hematophagous cimicid bug Psitticimex uritui), which may pose a disadvantage for the long-term occupation of large nests. Monk Parakeets add green vegetable material to the lining of the nests during the breeding season, a behaviour that may help to control parasites (Bucher 1988). The parakeets also show low nest site fidelity which may be related to nest in- ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 685 testation: an average of 47% of the population changed nests between years, mov¬ ing a mean distance of 500 m (Figure 2) (Martin & Bucher unpubl. data). Breeding starts in October. Most of first eggs are laid during the second half of Oc¬ tober and the first half of November. Non-breeding pairs usually occupy isolated nests or chambers in a compound nest. Nestling growth rate is intermediate between that of open-nesting and cavity-nesting birds (per-day growth rate constant K= 0.24) (Navarro & Bucher 1990). Fledglings start leaving the nest about 40 days after hatching. On several occasions we observed that fledglings from neighbouring nests concentrated in one nest (creche) where they were fed by adults. FIGURE 2 - Average distance moved by adult Monk Parakeets between censuses in our study area ( data from Martin and Bucher unpubl. data). After leaving the nest, fledglings remained with their parents for several months, but they always left the parental nest before the start of the following breeding season (Figure 3). Dispersal of juveniles from their nest of origin to their first breeding place was recorded for four juveniles. Dispersal distance was 300 m, 1 250 m, 1 400 m, and 2000 m (average 1230 m) (Martin & Bucher unpubl. data). However, this average may be an underestimation, given that our sampling area was limited and we did not have band recoveries besides our own. Helping at the nest, i.e. allofeeding siblings or other juveniles, was not observed as a regular event. However, we noticed a few inciden¬ tal cases of allofeeding: older fledglings fed their younger siblings, a non-breeding trio fed fledglings from a neighbouring nest, and a breeding bird fed a begging juvenile 686 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI from a neighbouring nest. More sibling to sibling allofeeding could have taken place earlier inside the nest but could not be observed (Martella & Bucher unpublished data). Basic demography The Monk Parakeet population in our study area grew constantly during the first years and showed a tendency to stabilize in the last two years (Figure 4). This trend may have occurred because, until the initiation of our study, Monk Parakeets had been controlled by the ranch owners mostly by nest destruction. In any given year of our study period only a proportion of the adults bred, varying between 37% and 60% of the total. The first clutch of the season ranged from one to 11 eggs, averaging 6.0 eggs, whereas replacement clutches had on average 5.0 eggs. Eggs were laid with an av¬ erage laying interval between eggs of 2.1 days. The incubation period averaged 24 days. Replacement clutches after nesting failure were laid by 20% of the pairs, but only 7% of the successful pairs attempted to renest. On average 45% of the cham¬ bers fledged young, and annual productivity per breeding pair averaged 1.38, vary¬ ing from 0.46 to 2.30. In total, 23% of the eggs laid fledged young. We did not find significant difference in pair productivity between colonies of different size, although productivity per pair was slightly higher in four-chamber colonies (1 .58) than in those with three (1 .33), two (1 .33) or one (1 .47) chambers; these differences were not sig¬ nificant (unpubl. data). However, our sample size of large nests was too small (only three four-chamber colonies) to allow us to raise definitive conclusions from our data. FIGURE 3 - Proportion of Monk Parakeet's fledglings remaining at the natal nest chambers at different ages in the study area at San Antonio Ranch. Notice that the following breed¬ ing season starts just after they become nine months old. In the few cases in which para¬ keets remained in their natal nest after that age their parents had already left the nest and bred elsewhere (see text) (data from Martin and Bucher unpubl.). ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 687 FIGURE 4 - Population of Monk Parakeets in the 604 ha study area at San Antonio Ranch, Jesus Maria, during the study period. Annual survival of first year birds was estimated at 61%, whereas in adults it was 81% (Martin & Bucher unpubl. data). These figures do not separate mortality from emigra¬ tion, but are useful for comparisons with other communal breeding species for which similar figures have been provided (see Stacey & Koenig 1990). DISCUSSION Communal breeding in birds is usually characterized by reduced dispersal, delayed breeding, and helping (Brown 1987). In our study area the Monk Parakeet showed delayed breeding, some degree of reduced dispersal, but only incidental direct helping by allofeeding. However, helping in nest building and maintenance, as well as vigi¬ lance of the compound nest, was a constant behavioural feature, although we did not find indications of productivity difference by nest colony size. Incidental cases of help¬ ing by allofeeding suggest a substrate for helping upon which selection could act should helping increase in adaptive value (Brown 1987). Further evidence supporting the existence of a potential for helping in Monk Parakeets comes from the observa¬ tions made by S. Emlen (in press) on Monk Parakeets (of an unknown subspecies) breeding in an aviary, where some non-breeding individuals fed nestlings. 688 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI Delayed breeding maturity seems to be common in parrots and has been reported in the Puerto Rican Amazon Amazona vittata (Snyder et al. 1987), the White-tailed Black Cockatoo Calyptorhynchus funereus (Saunders 1982), and the Galah Cacatua roseicapilla (Rowley 1983), none of which are communal breeders. However, in the Green-rumped Parrotlet Forpus passerinus, a smaller sized species, 9% of the males and 56% of the females breed in first year of life (Beissinger and Bucher in press). This may suggest a direct correlation between size and delayed breeding in parrots. Reduced dispersal has not been described in the published literature on other neotropical parrots, although current research on the Green-rumped Parrotlet in Ven¬ ezuela has shown that most first nestings occur within 1 km of the natal nestbox (S. Beissinger, pers. comm.). Dispersal distance is over 20 km in the White-tailed Black Cockatoo (Saunders 1 982) and the Galah (Rowley 1 983), well over the average 1 .2 km found in Monk Parakeets. Possible constraints on young Monk Parakeets from dispersing and becoming independent breeders may include the saturation of suitable nesting habitat and the energy costs of building large stick nests. Habitat saturation does not seem to provide a general explanation for reduced disper¬ sal in Monk Parakeets. In our study area nesting habitat was not limiting and the population continued to grow during the first part of our research, although it may be in part the result of past destruction of nests. However, habitat saturation may be an important factor in regions where nesting sites are limiting. Even if breeding is not prevented by habitat saturation, reduced dispersal in the Monk Parakeet may be related to the high energetic cost of building and maintaining bulky nests. Because adding a nesting chamber to a compound nest requires less building effort and material than building a new nest, birds may benefit from decreasing en¬ ergy and time expenditure by nesting in an already existing nest. Nesting in colonies may provide an opportunity for inexperienced birds to learn nestbuilding skills. Moreo¬ ver, it is likely that birds could also benefit from communal guarding against preda¬ tors, or the “selfish herd” (Hinde 1961) effect provided by members of the colony. However, continuous growth of a nest beyond a critical size may increase the risk of the whole structure falling. Although we only found incidental helping in our study, it is possible that helping in natural conditions might be favoured under different circumstances, e.g. when disper¬ sal is more difficult than was the case in our study areas (Emlen & Vehrencamp 1983). A greater difficulty in dispersing may occur in areas where nesting habitat becomes limiting as populations reach saturation levels, or where nesting habitat is very scarce, making dispersal highly risky. Finally, the possibility that different populations in each of the three different subspecies of Monk Parakeets may show differences in helping behaviour cannot be ruled out. In conclusion, our results suggest that the Monk Parakeet breeding system shows some characteristics of a communal breeder, including reduced dispersal, delayed breeding, and signs of incipient helping (Brown 1987). However, given that there is little information on other South American psittacines, we cannot be certain whether this set of characteristics is unique to the Monk Parakeet or, on the contrary, is wide¬ spread among neotropical parrots. Further elucidation of the causes and conse¬ quences of the Monk Parakeet social system would require more detailed research ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 689 comparing populations under different ecological conditions and saturation levels, as well as comparative work with other neotropical members of the family. ACKNOWLEDGEMENTS We are grateful to A. Gomez Duran and the Romanutti family for granting permission for the research to be carried out in their properties. We also thank P. Arrowood, S. Beissinger, and D. Saunders for their valuable comments on the original manuscript. Research was funded by the Consejo Nacional de Investigaciones Cientificas y Tecnicas de Argentina (CONICET) and the Consejo de Investigaciones Cientificas de Cordoba (CONICOR). LITERATURE CITED BEISSINGER, S.R., BUCHER E.H. In press. Sustainable harvesting of parrots for conservation, in Beissinger, S.R., Snyder, N.R.F. (Eds). New world parrots in crisis: solutions from conservation biol¬ ogy. Washington, District of Columbia, Smithsonian Institution Press. BROWN, J.L. 1987. Helping and communal breeding in birds. Princeton, Princeton University Press. BUCHER, E.H. 1988. Do birds use biological control against nest parasites?. Parasitology Today 4:1- 3. BUCHER, E.H., BEDANO, P. 1976. Bird damage problems in Argentina. International Studies on Spar¬ rows 9: 3-1 6. Poland. BUCHER, E.H., MARTIN, L.F. 1987. Los nidos de cotorras ( Myiopsitta monachus) como causa de problemas en lineas de transmision electrica. Vida Silvestre Neotropical 1: 50-51. BULL, J. 1973. Exotic birds in the New York city area. Wilson Bulletin 85: 501-505. COLLIAS, N.E., COLLIAS, E.C. 1977. Weaverbird nest aggression and evolution of the compound nest. Auk 94:50-64. EMLEN, S.T. In press. Observations on a captive colony of Quaker Parakeets. The Watchbird. EMLEN, S.T., VEHRENCAMP, S.L. 1983. Cooperative breeding strategies among birds. Pp. 93-120 in Brush, A.H., Clark, G.A. Jr. (Eds). Perspectives in Ornithology. Cambridge, Cambridge University Press. FORSHAW, J. 1978. Parrots of the world. Melbourne, Landsdowne Editions. HINDE, R.A.. 1961. Behavior. Pp. 373-441 in Marshall, A.J. (Ed.). Biology and comparative physiol¬ ogy of birds. Volume 2. New York, Academic Press. MARTELLA. M.B., BUCHER, E.H. 1990. Vocalizations of the Monk Parakeet. Bird Behaviour 8:46-55. MARTELLA. M.B., BUCHER, E.H. In press. Estructura del nido y comportamiento de nidificacion de la cotorra Myiopsitta monachus. Boletln de la Sociedad Zoologica del Uruguay. MARTELLA, M.B., NAVARRO J.L. BUCHER E.H. 1987. Metodo para la captura de cotorras (Myiopsitta monachus) en sus nidos. Vida Silvestre Tropical 1:52-53. NAVARRO, J.L. BUCHER, E.H. 1990. Growth of nestling Monk Parakeets in a wild population. The Wilson Bulletin 102:520-525. NEIDERMYER, L.J., HICKEY, J.J. 1977. The Monk Parakeet in the United States, 1970-75. American Birds 31 :237-278. NILSSON. G. 1990. Importation of birds into the United States. 1986-1988. Washington, Animal Wel¬ fare Institute. RIDGELY, R.S. 1980. The current distribution and status of mainland neotropical parrots. Pp. 233-84 in Pasquier, R. (Ed.), Conservation of New World Parrots. International Council for Bird Preservation. Technical Publication No. 1. Smithsonian Institution Press. ROWLEY, 1.1983. Mortality and dispersal of juvenile Galahs, Cacatua roseicapilla, in the Western Australia Wheatbelt. Australian Wildlife Research 10:329-342. SAUNDERS, D. A. 1982. The breeding behaviour and biology of the Short-billed form of the White¬ tailed Black Cockatoo Calyptorhynchus funereus. Ibis 124:422-455. SNYDER N.F.R., WILEY, J.W., KEPLER, C.B. 1987. The parrots of Luquillo: natural history and con¬ servation of the Puerto Rican Parrot. Los Angeles, Western Foundation of Vertebrate Zoology. STACEY P B., KOENIG, W.D. (Eds). 1990. Cooperative breeding in birds. Cambridge, Cambridge University Press. 690 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI ANNUAL VARIATION IN PRODUCTIVITY OF NORTH ISLAND KAKA ON KAPITI ISLAND, NEW ZEALAND RON J. MOORHOUSE School of Biological Sciences, Victoria University of Wellington, P.O. Box 600, Wellington, New Zealand ABSTRACT. Productivity in North Island Kaka Nestor meridionalis septentrionalis on Kapiti Island, varied markedly between two breeding seasons. In 1988, 26 nests were recorded producing an aver¬ age of 0.88 fledglings per pair, while in 1989 only nine nests were recorded producing 0.22 fledglings per pair. Starvation of nestlings was not observed in 1988, but was the primary cause of nest failure in 1989. As Kapiti Island lacks most of the introduced competitors now widespread on the main islands of New Zealand, this suggests that nesting in the Kaka is predominantly confined to years in which food is relatively abundant. Productivity in the Kaka may be more variable, and in the long-term, lower, than that of most other parrots due to its more variable, temperate environment. Keywords: Nestor meridionalis, productivity, annual variation, food limitation, predation. INTRODUCTION The North Island Kaka Nestor meridionalis septentrionalis [nomenclature follows Kinsky (1970)] is a large (length = 430 mm; weight [maximum range] = 390 - 555 g, n = 20), omnivorous forest parrot endemic to New Zealand. Once widespread, the range of the Kaka has declined since European colonization (Buller 1888, Bull et al. 1985). The North Island Kaka is now considerably more abundant on a few offshore islands than on the North Island mainland (Oliver 1974, Bull et al. 1985). These is¬ lands are notable for their relative lack of the introduced mammalian predators and competitors now widespread on the New Zealand mainland (Atkinson & Bell 1973). While Kaka populations on the North Island mainland are predominantly confined to large remnants of mature native forest (Oliver 1974, Bull et al. 1985), the off-shore islands on which the bird remains common are relatively small, and the vegetation of one has been heavily modified by human-induced fires (Fuller 1985). In view of this, it would seem that introduced predators and competitors, rather than forest size or age, have been the primary factors in the decline of the Kaka on the North Island mainland. In order to understand how North Island Kaka might be affected by introduced com¬ petitors and predators, it is necessary to obtain information on the birds’ productitivity In the only available study of Kaka productivity, in South Island Beech Nothofagus sp forest, Beggs and Wilson (1987, in press) found that only two of 31 radio-tagged birds attempted to breed in six years. The majority of nesting in this population appears to have been confined to one of the six years it was under study (J. R. Beggs pers comm.). Such infrequent nesting suggests a productivity much lower than that re¬ corded in other similarly sized parrots (Rowley 1980; Saunders 1982,1986- Smith & Saunders 1986; Gnam 1991). Infrequent nesting within populations is also known to occur in another New Zealand parrot, the Kakapo Strigops habroptilus (Best & Powlesland 1985. Moorhouse & ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 691 Powlesland in press) and in another New Zealand endemic, the New Zealand Pigeon Hemiphaga novaeseelandiae; (M. N. Clout pers. comm.). Other parrots (Rowley 1980, Gnam 1990, 1991) have occasional years when relatively few individuals nest. The available data on these three New Zealand species, however, indicate that most nest¬ ing occurs at infrequent intervals. Introduced wasps t/espu/a vulgaris and V. germanica and the Australian brush-tailed possum Trichosurus vulpecula , an arboreal herbivore now widespread on the New Zealand mainland, probably had adverse effects on the productivity of the South Is¬ land Kaka population studied by Beggs and Wilson (1987, in press). Wasps compete with Kaka for Beech honey-dew, the sugar rich substance excreted by the scale in¬ sect Ultracoelostoma assimile (Beggs & Wilson, in press) and nectar. Possums are likely to compete with Kaka for fruit and nest cavities. Examination of Kaka produc¬ tivity on an offshore island where the only introduced mammalian competitors are rats, and where wasps are less abundant, could provide a valuable comparison for such research on the mainland. This study sought to determine the factors influencing the breeding productivity of the Kaka on Kapiti Island. STUDY AREA AND METHODS Kapiti Island (2000 ha), a nature reserve administered by the New Zealand Depart¬ ment of Conservation, lies 50 km north of Wellington and 5.5 km off the North Island west coast. Virtually deforested by fire early last century, Kapiti Island is now predomi¬ nantly covered by a variety of serai indigenous forest (Fuller 1985). Except for the Polynesian rat Rattus exulans and Norway rat R. norvegicus, the island is free of feral introduced mammals. The Australian brush-tailed possum, introduced to Kapiti Island in the 1890s, was finally eradicated in 1986. The introduced common and German wasps, competitors of Kaka in South Island beech forest (Beggs & Wilson, in press), are present on Kapiti Island but at much lower densities (pers. obs.). All nest sites found were in tree-cavities, most within three metres of the ground and a high proportion actually at ground level (Table 1). While low (< 3 m), and particu¬ larly ground level nests appear common on Kapiti Island, the high number of such sites found probably also reflects the greater ease with which I was able to detect them. Most (14/20) nest sites within three meters of the ground were detected when hens flushed from the nest chamber, or by examining cavities in large (> 30 cm di¬ ameter at breast height) trees. Two advanced nests were located by the distinctive smell of nestlings’ faeces. In contrast the five nests located above three meters height were found by the relatively time consuming process of observing courting birds. Once found, I checked nests weekly using a bright light and an extendible angle mirror to determine final clutch size, hatching and fledging success. Climbing equipment was used to examine high nests. Because some nests were discovered long after hatch¬ ing had occurred, or were destroyed by predators prior to hatching, I was not able to determine final clutch or brood size for all nests. Estimates of average clutch and brood size were, therefore, derived from the largest available subsample of the total found Four nests in sites previously used within the same season were excluded from estimates of productivity because I could not deterniine whether these were nests of the previously observed pair or a new pair. 692 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI TABLE 1 - Height of Kaka nest sites found on Kapiti Island in 1988 and 1989. Mahoe = Melicytus ramiflorus, Kohekohe = Dysoxylum spectabile, Pukatea = Laurelia novazelandiae, Kamahi = Weinmannia racemosa, Rata = Metrosideros robusta, Puka = Griselinia lucida. height range (m) Tree species Total Mahoe Kohekohe Pukatea Kamahi Rata Puka Dead tree <0-0 7 3 0 0 0 0 0 10 0-1 1 1 0 0 1 0 0 3 1-3 2 3 0 0 0 1 1 7 3-5 0 1 0 1 0 0 0 2 5-7 0 0 0 0 0 0 0 0 7-9 0 0 1 0 0 0 0 1 9-11 0 1 0 0 0 0 0 1 11-13 0 0 1 0 0 0 0 1 Total 10 9 2 1 1 1 1 25 RESULTS Kaka pairs on Kapiti Island laid an average of 3.71 eggs (95% confidence interval = 3.35-4.07, n = 24, range = 1-5), hatched an average brood of 2.09 chicks (95% con¬ fidence interval = 1 .64 - 2.54, n = 23, range = 0-4), and, if their nest did not fail com¬ pletely, fledged 1.69 young (95% confidence interval = 1.18 - 2.21, n = 13, range = 1- 3) (pooled data from 1988 and 1989). The low brood size relative to the number of eggs laid appeared primarily due to poor hatching success. Hatching success in 1988 was 58% (35 chicks from 60 eggs in 16 nests) and 52% (13 chicks from 25 eggs in 7 nests) in 1989, giving an overall value for both years of 56% (48 chicks from 85 eggs in 23 nests). Both 100% and 0% hatching success were rare, occuring in only two and three clutches respectively. Examination of 16 unhatched eggs revealed 13 to be infertile and three to contain dead embryos. In the 1988 breeding season I located twenty-three active Kaka nest sites, three of which were reused in the same season giving a total of 26 actual nests (Table 2 & 3). Despite being more experienced in detecting nests, and making a greater search ef¬ fort (36 days searching for nests in 1989 against 24 days in 1988), in the 1989 sea¬ son I only found eight active sites, one of which was reused within the same season (Table 2 & 3). All but two of the nest sites found in the 1989 season had been used the previous year. Nine of the 17 1988 nest sites not used in 1989 had been success¬ ful while three of the six 1988 sites reused in 1989 had been unsuccessful in 1988. Pairs that did nest in the 1989 season produced an estimated 0.25 fledged young per pair (95% confidence interval = 0.14 - 0.64, n = 8) as opposed to 0.87 fledged young per pair (95% confidence interval = 0.64 - 1.10, n = 23) in the 1988 season (Table 2). Considering pairs which fledged young only, the only two successful nests in the 1989 season produced only one fledgling each while successful pairs in the 1988 season produced an average of 1 .82 fledglings per pair (95% confidence interval = 1 .23 - 2.4, n = 1 1 ) (Table 2). ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 693 TABLE 2 - Productivity of Kaka on Kapiti Island in 1988 and 1989 No. active nest sites No. young fledged No. fledglings per pair No. fledglings per successful pair 1988 23 20 0.87 1.82 1989 8 2 0.25 1.00 Total 31 22 0.71 1.69 TABLE 3 - Fate of Kaka nests on Kapiti Island in 1 988 and 1 989 Total Preyed on by Rat Morepork Weka Abandoned Infertile Flooded Starved Fledged 26 6 3 1 2 1 1 0 12 1988 % 23 12 4 8 4 4 0 46 9 2 0 0 0 2 0 3 2 1989 % 22 0 0 0 22 0 33 22 35 8 3 1 2 3 1 3 14 Total % 23 9 3 6 9 3 9 40 Considering all nests found, 46% were successful in the 1988 season and only 22% in the 1989 season (Table 3). In 1988, rat predation of nestlings was the single great¬ est cause of nest failure accounting for the loss of 23% of nests found (Table 3). In the 1989 season, however, starvation of nestlings was the primary cause of nest fail¬ ure, accounting for the loss of 33% of nests (Table 3). Starvation was the presumed cause of death when chicks were found dead without sign of injury. Such chicks were always in nests containing more than one young and in all cases the younger nest¬ lings died first. The two successful nests in the 1989 season each had only one chick hatch. Overall, in both seasons, only 40% of nests succeeded (Table 3). Rat predation was the predominant cause of nest failure accounting for the loss of 23% of nests found (Table 3). Rat predation was identified using the experimental observations of Moors (1978) who found that rats typically ate eggs and nestlings in situ, leaving remains in the nest. Where eggs and nestlings disappeared without trace, the Morepork Ninox novaeseelandiae , a native owl, or the Weka Gallirallus australis, an endemic rail, were assumed to be the predator. Seven of the eight nests preyed on by rats had an open¬ ing to the nest chamber within 20 cm of the ground, while the other was in a leaning tree which afforded easy access. Only 6% (1/17) of nests in sites over one metre above ground were preyed on by rats. In contrast, 50% (9/18) of nests in sites within one metre of the ground were preyed on by rats. 694 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI DISCUSSION The productivity of successful pairs of North Island Kaka on Kapiti Island over two years is consistent with the available data on other large parrots in terms of small clutch size and the small number of young fledged. The low hatching success of Kaka on Kapiti Island (56%), however, appears more unusual. Smith and Saunders (1986) found 75-100% hatching success in four species of Australian parrots but recorded a comparable 65% hatching success in the Red-tailed Black Cockatoo Cacatua magnificus. Gnam (1991), however, also recorded a hatching success of 56% in the Bahama parrot Amazona leucocephala bahamensis on Great Abaco Island. Gnam (1991) found all eggs examined to be fertile. However, 81% (13/16) of the unhatched Kaka eggs I examined were infertile and only 19% (3/16) contained dead embryos. The difference in the number of active nest sites found in 1988 and 1989 suggests that long-term productivity in Kaka could be lower than in most other parrots. While the infrequent nesting of South Island Kaka observed by Beggs and Wilson (1987, in press), could reflect competition from introduced wasps and possums, productivity in the 1989 season on Kapiti Island appeared low in the absence of possums, and prob¬ ably with no significant increase in competition from wasps. It is possible that I found fewer nests in 1989 because some pairs which had failed nests the previous year renested the following year in sites more difficult to detect. However, this seems un¬ likely as over half the nest sites not used in 1989 had been successful the previous year. The fact that captive Kaka nest, and even renest, every year (M. Sibley pers. comm.) indicates that Kaka can breed annually given adequate food resources. In view of this, the far fewer nests found in 1989 suggests that food was less abundant for the birds in that year. While I have no comparative data on food availability for Kaka in 1988 and 1989, the high incidence of starvation of nestlings in 1989 supports this food limitation hypoth¬ esis. The only two pairs which successfully fledged young in 1989 each had only one chick, whereas successful pairs in 1988 were able to fledge an average of 1.82 chicks. Fruiting in many species of forest trees, including those with fruits taken by Kaka, was more abundant and prolonged in the 1988 than 1989 season (pers. obs.). New Zealand Pigeons, which are predominantly frugivorous in the breeding season (M. N. Clout pers. comm.), also displayed markedly greater productivity on Kapiti Is¬ land in 1988 than 1989 (pers. obs.). Clutch and brood size in many temperate birds often varies in response to annual variation in food availability (Cody 1971). In the Kakapo (Best & Powlesland 1985; Moorhouse & Powlesland in press) and New Zealand Pigeon (M. N. Clout pers comm.), most individuals only nest in relatively infrequent years that are presumably more favourable for breeding. My data, and that of Beggs and Wilson (1987, in press) suggest that the Kaka is similar. The likely longevity of these species, long period of dependency of young, and the former absence of mammalian nest predators in New Zealand, may all have favoured not nesting in years in which conditions are less suit¬ able, rather than reducing clutch or brood size. In effect, these birds could be said to reduce their clutch size to zero in less favourable years. If productivity in Kaka is limited by the availability of plant foods even in the absence of significant competion from introduced species, then the latter are likely to have a ACTA XX CONGRESSUS INTERNATIONAL^ ORNITHOLOGICI 695 significant negative effect on the birds. Kaka may only be able to breed in the pres¬ ence of competitors in habitats where suitable plant foods become periodically su¬ perabundant. Kaka in South Island beech forest appear to depend on the infrequent mast seeding of red beech Nothofagus fusca for successful breeding (J. R. Beggs pers. comm.). As years of high productivity in Kaka could, even in the absence of competitors, be infrequent, Kaka populations may be particularly vulnerable to intro¬ duced predators. On Kapiti Island, rat predation was predominantly confined to nests within a metre of the ground. This suggests that the predominantly ground-foraging Norway rat was the main predator of nestlings rather than the more arboreal Polynesian rat (Atkinson 1985). As I was probably biased toward detecting such low nests, I have probably overestimated the impact of rat predation on Kaka productivity on the Island. None¬ theless, as 50% of nests within one metre of the ground were destroyed by rats, the latter clearly have a significant effect on the productivity of the nest sites accessible to them. In a mainland population, all nests would be vulnerable to arboreal preda¬ tors such as ship rats R. rattus and stoats Mustela erminea, irrespective of height. Stoats are capable of killing female Kaka on the nest (Beggs & Wilson, in press), and so could have a far more immediate effect on productivity than rats. In conclusion, nesting in Kaka on Kapiti Island appears limited by annual variation in plant food availability even in the absence of most introduced competitors present on the mainland. This suggests that the latter may have a significant negative effect on Kaka productivity. Introduced competitors may have already restricted Kaka to habi¬ tats where important plant foods become superabundant with sufficient frequency. Such populations are, however, likely to nest infrequently. High levels of nest preda¬ tion could thus significantly reduce the long-term productivity of such populations. Thus, while introduced competitors may have reduced the range of habitats in which the Kaka can breed successfully, mammalian nest predators may now pose the most immediate threat to the continued survival of the species. Supplementary feeding of Kaka populations may be an effective management technique to increase their pro¬ ductivity. AKNOWLEDGEMENTS I thank the New Zealand Department of Conservation for granting me permission to work on Kaka on Kapiti Island. Special thanks are extended to the ranger on Kapiti Island, Peter Daniel, Linda Daniel, Shona Pengelly and Marion Daniel for their sup¬ port and hospitality. I am grateful to the many people who assisted me in the field, in particular, I thank my regular co-inhabitant of Kapiti Island, Tim Lovegrove. My super¬ visor, Charles Daugherty, provided advice and encouragement, and Rod Hitchmough, Graeme Elliot, Ralph Powlesland and Ben Bell provided productive discussion on Kaka productivity. Finally, I thank Denis Saunders and Patricia Arrowood for their constructive criticism of this manuscript. This work was funded by the New Zealand Department of Conservation, the Native Forests Restoration Trust, the New Zealand Royal Forest and Bird Protection Society, the Nga Manu Trust, the Ornithological Society of New Zealand and the Internal Grants Committee of Victoria University of Wellington. 696 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI LITERATURE CITED ATKINSON, I. A. E. 1985. The spread of commensal species of Rattus to oceanic islands and their effects on island avifaunas. Pp. 35 - 81 in Moors, P. J. (Ed.). Conservation of island birds. 1C ec nical publication No. 3. . ATKINSON, I. A. E., BELL, B. D. 1973. Offshore and outlying islands. Pp. 373-392 in Wiliams, G. H. (Ed.), The natural history of New Zealand. Wellington, Reed. BEGGS, J. R., WILSON, P. R. 1987. Energetics of South Island Kaka {Nestor meridionalis meridionalis) feeding on the larvae of kanuka longhorn beetles (Ochrocydus huttoni). New Zealand Journal of Ecol¬ ogy 10: 143 - 147. BEGGS, J. R., WILSON, P. R. In press. The Kaka (Nestor meridionalis), a New Zealand parrot endan¬ gered by introduced wasps and mammals. Biological Conservation. BEST, H. A., POWLESLAND, R. G. 1985. Kakapo. Wellington, John McEndoe and New Zealand Wild¬ life Service. BULL, P. C., GAZE, P. D., ROBERTSON, C. J. R. 1985. Atlas of bird distribution in New Zealand. Wellington, Ornithological Society of New Zealand. BULLER, W. L. 1888. A history of the birds of New Zealand, 2nd edn. London, the author. CODY, M. L. 1971. Ecological aspects of reproduction. In Farner, D. S., King, J. R. (Eds). Avian biol¬ ogy. Vol. 1 : Pp. 461 -512. New York, Academic Press. FULLER, S. A. 1985. Kapiti Island vegetation: Report on vegetation survey of Kapiti Island 1984 - 85. Department of Lands and Survey, Wellington District Office. GNAM, R. S. 1990. Conservation of the Bahamas Parrot. American Birds 44: 32 - 36. GNAM, R. S. 1991 .Breeding biology of the Bahama Parrot (Amazona leucocephala bahamensis). Unpub. Ph.D. thesis. City University of New York. KINSKY, F. C. 1970. Annotafed checklist of the birds of New Zealand. Wellington, Reed. MOORHOUSE, R. J., POWLESLAND, R. G. In press. Aspects of the ecology of Kakapo (Strigops habroptilus) liberated on Little Barrier Island (Hauturu), New Zealand. Biological Conservation. MOORS, P. J. 1978. Methods for studying predators and their effects on forest birds. Pp 47 - 57 in Dingwall, P. R., Atkinson, I. A. E., Hay, C. (Eds). The ecology and control of rodents in New Zealand nature reserves. Wellington: Department of Lands and Survey Information Series. OLIVER, W. R. B. 1955. New Zealand birds. Wellington, Reed. ROWLEY, I. 1980. Parent offspring recognition in a cockatoo, the Galah, Cacatua roseicapilla. Aus¬ tralian Journal of Zoology 28: 445-456. SAUNDERS, D. A. 1982. The breeding behaviour and biology of the short-billed form of the White¬ tailed Black Cockatoo (Calyptorhynchus funereus). Ibis 124: 422-455. SAUNDERS, D. A. 1986. Breeding season, nesting success and nestling growth in Carnaby’s Cocka¬ too, (Calyptorhychus funereus latirostris), over 16 years at Coomallo Creek and a method for assessing the viability of populations in other areas. Australian Wildlife Research 13: 261-273. SMITH, G. T., SAUNDERS, D. A. 1986. Clutch size and productivity in three sympatric species of Cockatoo (Psittaciformes) in the south-west of western Australia. Australian Wildlife Research 13: 275- 285. ACTA XX CONGRESSUS INTERNATIONAL^ ORNITHOLOGICI 697 CONCLUDING REMARKS: SYMPOSIUM ON THE ECOLOGY AND SOCIAL BEHAVIOUR OF PARROTS AND PARAKEETS PATRICIA C. ARROWOOD' and DENIS A. SAUNDERS2 1 Department of Biology, Dept. 3AF, New Mexico State University, Las Cruces, New Mexico 88003, USA 2 CSIRO Division of Wildlife and Ecology, Western Australian Laboratory, Locked Bag No. 4, PO, Midland, WA 6056, Australia After a long history of little scientific research, progress is being made in research on the members of this interesting group of birds. In this symposium, Saunders showed how the sympatric Carnaby’s and Red-tailed Black Cockatoos have responded differ¬ ently to the same changing environment and examined the factors responsible for those different inter-specific responses. Even though the Red-tailed Black Cockatoo population is increasing in the wheatbelt of Western Australia, Saunders found that its reproductive rate is very low for a species laying only one egg per breeding sea¬ son. A low reproductive rate may not have posed problems for a long-lived species in a stable environment, but in the rapidly-changing environment of Western Australia, the long-term prospects for such a species are grim, particularly so if no conserva¬ tion areas are set aside for such nomadic species. Since most psittacines live in flocks outside the breeding season, Arrowood’s inter¬ est was in analyzing the patterns of interaction within flocks. Using captive flocks of the Neotropical Canary-winged Parakeet, she found patterns of social interaction that, if confirmed in other psittacines, would be different from that of any other bird group. In the Canary-winged Parakeet, affiliative interactions within the pair are egalitarian, with reciprocal allopreening and allofeeding. Affiliative interactions do not extend outside the pair relationship (except to offspring), and males do not seek sexual ac¬ tivities with nonmate females. Aggressive acts by each sex are distributed equally to same-sexed and oppositely-sexed individuals. The only significant unilateral tendency was for males to have higher rates of aggressive acts to males than females did to males. It would be interesting to compare the patterns of interaction found in monomorphic and monochromatic species with those of dimorphic or dichromatic species to see if monomorphic species are more egalitarian in their behaviour. Two of the symposium papers (Gnam, Moorhouse) dealt with the basic biology of large parrots living on small islands. Since both species are declining, such basic in¬ formation is essential for the preparation of preservation and management strategies. The survival of species on small islands is usually tenuous; in the face of human ma¬ nipulation of the environment, the chances for continued survival plummet. The Bahama Parrot (Gnam) relies on three plant species for 74% of the food fed to chicks. Similarly, the Kaka (Moorhouse) uses hard-to-obtain insect larvae and unpredictably- fruiting and flowering plants for the food fed to chicks. On such small islands the birds have few other choices. On top of a precarious situation because of food supplies, chicks are lost to introduced mammals. Cats prey on chicks in subterranean nests 698 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI used by the Bahama Parrot and Norway rats take eggs and chicks in near-ground nests of the Kaka. Ground nesting implies an adaptation to the absence of suitable arboreal nest cavities, a situation uncommon in psittacines. Nesting success in the subterranean-nesting Bahama Parrot was lower than that of the arboreal cavity-nest¬ ing congeneric Hispaniolan Parrot and Puerto Rican Parrot. In the Kaka, however, nesting success on the island may be greater than that on the mainland. If predators could be removed and the islands maintained predator free, they might provide a salvation for the parrots. Bucher et al. presented data on many interesting aspects of the Monk Parakeet’s social life and population biology. The Monk Parakeet is the only psittacine to fabri¬ cate its own nest which it makes of sticks and twigs. This species may be in the in¬ cipient stages of the evolution of a communal breeding system, a system not recorded in any other psittacine. The stick nests that are constructed may be occupied by a single pair or by multiple pairs; non-breeding individuals may also occupy nest cham¬ bers. All occupants of a compound nest participate in its construction, maintenance and defense. Older siblings and nonbreeding and breeding neighbours were observed allofeeding young. The Monk Parakeets had a large clutch size, but low productivity per breeding pair. There was some suggestion that productivity might be higher the greater the number of chambers per compound nest. The highly-visible nest of these parakeets facilitates their destruction by the agricultural community which considers them a pest. On the other hand, population expansion is occurring into areas where recently-introduced tall eucalyptus trees provide attractive substrates for the large compound nests. There are many questions still unanswered about psittacine behaviour and ecology. No one has yet explored whether vocal mimicry plays a role in the natural life of those species with that incredible ability. The life-long, year-round, highly exclusive mo¬ nogamy of many psittacines represents an extreme case among the many avian vari¬ ations on the monogamy theme, but it has received little theoretical or empirical at¬ tention. Since most psittacines are not territorial, many interesting questions exist about the patterns of their flock movements relative to the patterns of abundance of food sources and competition with other species. The study of these distinctive birds will allow us to test the robustness of current principles of avian ecology and behav¬ iour. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 699 SYMPOSIUM 9 BIRD FLIGHT Conveners G. E. GOSLOW and I. . HUMMEL 700 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI SYMPOSIUM 9 Contents INTRODUCTORY REMARKS: BIRD FLIGHT D. HUMMEL and G. E. GOSLOW JR . 701 WAKE STRUCTURE AND FORCE GENERATION IN AVIAN FLAPPING LIGHT JEREMY M. V. RAYNER . 702 NEUROMUSCULAR ORGANIZATION FOR FLIGHT: ISSUES FOR STUDY G. E. GOSLOW, JR . 716 CORE TEMPERATURE RELATIONS OF PIGEONS DURING PROLONGED WIND TUNNEL FLIGHT W. NACHTIGALL and K.-D. HIRTH . 722 ON THE AERODYNAMICS OF THE TAIL IN BIRDS D. HUMMEL . 730 TIME PATTERN OF WING MOTION DURING CRUISING FLIGHT AND ITS IMPORTANCE FOR FLIGHT ENERGETICS H. OEHME . 737 CONCLUDING REMARKS: BIRD FLIGHT D. HUMMEL and G. E. GOSLOW, JR . 748 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 701 INTRODUCTORY REMARKS: BIRD FLIGHT D. HUMMEL1 and G.E. GOSLOW JR2 1 Institute fur Stromungsmechanik, Technische Universitat Braunschweig, Bienroder Weg 3, D-3300 Braunschweig, Germany 2 Section for Population Biology, Morphology and Genetics, Brown University, Box G-BMC, Providence, Rhode Island 02912, USA Bird flight as a mode of locomotion has held us in fascination for centuries. Studies by biologists and engineers have given us some explanation concering the reasons for wingtip slotting, wing kinematics, and adaptive wing shapes, but many problems remain. At the 17th Congress (1978) in Berlin, Germany, as well as at the 18th Con¬ gress (1982) in Moscow, USSR, Professor Oehme from East Germany and Profes¬ sor Nachtigall from West Germany were the convenors for sessions on bird flight. At the last 19th Congress (1986) in Ottawa, Canada, unfortunately no session on our topic took place. Therefore the Scientific Programme Committee of the present 20th Congress (1990) in Christchurch, New Zealand, voted for the organization of a sym¬ posium on “Bird Flight”. We were charged by the Scientific Programme Committee of the Congress to bring to the non-flight specialist a collection of papers that in addition to providing a review, would highlight current issues in bird flight and stimulate students for further study. We believe the selection of papers herein accomplish this charge. 702 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI WAKE STRUCTURE AND FORCE GENERATION IN AVIAN FLAPPING FLIGHT JEREMY M. V. RAYNER Department of Zoology, University of Bristol, Woodland Road, Bristol BS8 1UG, UK ABSTRACT. The mechanics of flapping flight have moulded the flight adaptations of birds, and are subject to a wide range of aerodynamic constraints. Flow visualization experiments and high-speed cinephotography have identified two ‘gaits’, distinguished by the upstroke and the variability of force generation (visualized in the wake vortices) typical of steady flight. In the vortex ring gait - used in slow flight and by birds with low aspect ratio wings - the upstroke is inactive; in the continuous vortex wake gait - in longer-winged birds in fast flight - the wake is a pair of undulating trailing vortices, and the wing deforms only slightly during the upstroke. The vortex ring gait demands intense activity of the supracoracoideus, but in the continuous wake gait this muscle may be unimportant. Theoretical models based on these experiments can predict a range of quantities relevant to flight, including mechanical flight power, and wing root bending moments, and these provide methods of testing the models. Keywords: Birds, bats, aerodynamics, flapping flight, flow visualization, vortices, wingbeat kinemat¬ ics, gait, flight power, efficiency. INTRODUCTION Flapping flight is a demanding adaptation. It demands much, not only from the wings and skeleton to meet aerodynamic stresses and from the physiological system to pro¬ vide sufficient energy flows, but also from the scientist who attempts to study the mechanics of the movement of birds. Until recently much of what we know about flap¬ ping flight has been based on conventional aerodynamic theories, and on analogies and parallels with fixed-wing aircraft. While this approach has been of enormous im¬ portance, and retains fundamental significance for our understanding of flight mechan¬ ics and the aerodynamic design of birds, it is inevitably limited in application to natural flight: the wings of aircraft are not flapped. In aircraft, force generation is steady and continuous, while in flapping flight lift and drag forces vary in strength and direction during the wingbeat. In aircraft horizontal (thrust) and vertical (lift) forces are gener¬ ated by essentially independent mechanisms - aerofoil lift from the wings, and thrust from jet or propeller engines - but birds have only aerofoil action for both lift and thrust. At this point the analogy with aircraft practice breaks down. AERODYNAMICS Aerofoil action and force generation The wings of aircraft and of flying animals act as aerofoils. When held at the appro¬ priate angle to the incident airflow, the tapered, asymmetric profile induces vortices bound onto the wing and trailing from the wingtips (trailing vortices are visible as the vapour trail behind high-flying aircraft); this vortex system forces air downwards be¬ hind the wing, and the wing experiences the reaction of this momentum flow as lift Lift acts at right angles to the direction of movement of the wing, and is proportional to the speed of the wing and the strength or circulation of the bound vortex. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 703 Lift is the largest aerodynamic force, but there are also three components of drag, reflecting friction and surface pressure on the body (parasite drag) and wings (pro¬ file drag), and the energy cost of generating the wake vortices (induced drag); drag is largely horizontal, and retards movement through the air. Thrust from flapping wings Animals must use aerofoil lift both as weight support and as thrust to overcome drag. With wings outstretched lift is fully capable of supporting weight, but there is no thrust, and the animal must be a glider, losing height as energy is expended against drag. Flapping is the mechanism by which a bird is able to sustain level flight using lift alone. It is to provide a horizontal thrust to balance drag that animals must flap their wings. Flapping flight must be modelled as a means of configuring the wingbeat so that on average lift provides both a mean thrust and weight support. The primary movement of the wing is dorso-ventral flapping. As the wing moves downwards it also moves forwards relative to the air, and lift is inclined forwards, acting both as weight support and thrust; this is the main part of the wingbeat, and generates the bulk of the aero¬ dynamic force. Symmetric flapping with upstroke in the same configuration as downstroke gives no mean thrust, since an upstroke with an aerodynamically active wing - while supporting weight - would generate a negative thrust, cancelling that from the downstroke. Down- and upstrokes must be asymmetric, with less lift in the upstroke either by a smaller wing planform or weaker bound vortex. The extent to which up- and downstrokes are asymmetric depends on the magnitude of the required thrust; this is a non-trivial problem, since thrust must balance drag, and drag depends in turn on lift and on wingbeat geometry and kinematics. Vortex action in flapping flight Whenever an aerofoil generates lift it sheds vortices, which in strength and location are symptomatic of the action of the wing. A trailing vortex is shed from the wingtips, equal in strength to the vortex bound on the wing. A transverse vortex is shed along the trailing edge whenever the bound circulation changes strength. Without the vor¬ tex wake there could be no momentum flow away from the bird, and hence no lift. Generation of the vortices represents a significant energy cost (induced drag); some authors claim that the bird could avoid this cost by eliminating all wake vortices, but then there would be no lift, and flight would be impossible. The vortices should be generated as efficiently as possible: the induced drag depends critically on vortex structure; wakes comprising closed circular loops or near-linear elements are most efficient in transporting momentum for minimum energy. I shall describe below that these are the only structures observed in birds in steady flight. Visualization of the wake vortices is therefore a powerful means of clarifying force generation mechanisms in flight. The wake is a transient, three-dimensional structure, and experimental determination of its structure has been difficult. Initial observations of wake vortices in slow-flying pigeons (Magnan et al. 1938) and small fringillid passerines (Kokshaysky 1979) revealed a series of circular vortex rings, each formed by a single downstroke. This implies that all aerodynamic force was generated dur¬ ing the downstroke, and that the upstroke was mechanically passive, as predicted by my initial theoretical model of flapping flight (Rayner 1979, 1980); subsequent experi¬ ments and observations have, however, revealed that the vortex ring is not the only wake observed in birds. 704 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI FIGURE 1 — The wake of a Pigeon Columba livia in slow flight at approx. 3 m/s, show¬ ing a photograph of a single vortex ring, and a reconstruction of the air movements and position of vortex cores. (From Spedding 1982.) Our flow visualization technique developed consists of seeding the air with neutrally buoyant soap bubbles filled with a helium-air mixture, which follow any movements of the air; we obtain an instantaneous 3-D view of the wake by photographing in stereo, with a series of flash guns (Figures 1-4). The bubble traces are longer and more curved when the bubbles are close to the vortex cores, and hence the position and strength of the vortices are determined. Unfortunately the technique reveals little of the airflow around the wings or of the formation of the trailing vortex cores close be¬ hind the wings. First results confirmed Kokshaysky’s vortex ring wake in Pigeon Columba livia (Fig¬ ure 1) and Jackdaw Corvus monedula in very slow flight (Spedding 1986, Spedding ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 705 FIGURE 2 - The wake of a Kestrel Falco tirmunculus in normal flight at approx. 7 m/s. (From Spedding 1982.) et al. 1984, see also Rayner 1986, 1988a). Transverse vortices are shed as the wing begins to generate lift at the start of the downstroke, and as it ceases lifting at the end of the downstroke. These join with the trailing vortices to form a toroidal vortex ring. Measurement of the vortex size and strength showed that the rings carried only two- thirds of the momentum needed to support the bird’s weight for one wingbeat (Spedding et al. 1984). This suggested, disturbingly, that some unconventional un¬ steady mechanism was generating significant aerodynamic force, yet theoretically all momentum should be visible in the wake. Subsequent measurements of body accel¬ eration have shown the Pigeon followed a parabolic flight path, decelerating at ap¬ proximately 0.3 g (J.M.V. Rayner & A.L.R. Thomas, in prep.). 706 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI The momentum discrepancy arose because the rings were smaller than expecte , rather than weaker, and limits to the bound circulation on the wings may present the bird from balancing all of its weight in very slow flight. A very different wake was observed in the European Kestrel Falco tinnunculus. In gliding the wake is a pair of straight line vortices similar to those generated by an aircraft wing (Spedding 1987a). In cruising flight (Figure 2) the upstroke is now aero- dynamically active; no transverse vorticity is visible at any phase, and the vortices undulate following the path of the wingtips (Spedding 1987a, b, see also Rayner 1988a). Vortex strength is constant along the wake (as would be expected with no transverse vorticity), and is sufficient to support the weight: the bird controls wing pitch so that bound circulation remains constant, and ensures a net thrust by sweeping the wingtip back at the wrist during the upstroke so that the effective wingspan - the spac¬ ing between the vortex cores - is reduced (Rayner 1986, 1988a). To investigate whether these two wake patterns were representative we have under¬ taken a survey of the wake patterns in a range of bird and bat species (Rayner, Jones and Thomas 1986, Rayner 1987, 1988a, Rayner & Thomas, in prep.); all fall into one of the same two patterns . In slow flight in all species studied (Pigeon, Jackdaw, Tawny Owl (Figure 3), Barn Owl, Cockatiel, Starling, Zebra Finch, Canary, Blackcap, etc.), and in cruise in slow-flying, broad winged species and in those using bounding flight (fringillid passeriforms, Budgerigar, Little Owl, Quail, Long-eared Bat, Pipistrelle Bat) the wake is vortex rings. In fast flight longer-winged species (Kestrel, Pigeon, Cockatiel, Noctule Bat, Dog-faced Fruit Bat, Swift, Swallow) adopt the continuous vortex wake. Transverse vortices are only observed with the vortex ring wake, and the ‘ladder’ wake of Pennycuick (1988) apparently does not exist in vertebrates (although it may be present in insects (Rayner 1986)). The absence of transverse vortices is not surprising, since the interaction of transverse vortices with the vortex on the wing can dramatically increase induced drag (Rayner 1986). We have limited information about wake transitions in accelerating flight in a number of species including the Noctule Bat (Rayner et al. 1987), the Cockatiel and Meyer’s Conure Policephalus meyeri (Figure 4). In transitions there is a single upstroke with circulation reduced compared to the subsequent downstroke. In each case the ani¬ mal appears to have a critical speed at which gait is changed: the transition is rapid, within a single wingbeat, and at the same time wingbeat kinematics alter as amplitude and frequency both reduce. In decelerating flight the situation appears to be different, birds maintaining the continuous vortex wake down to low speeds during rapid decel¬ eration just prior to landing; there may be advantages in stability and control from maintaining a lifting upstroke. Wingbeat kinematics and gait Patterns of wingbeat kinematics in flying birds have been studied by many authors (see Rayner 1988a), mainly by high-speed cinematography. The patterns observed are complex and varied, but can be simplified when kinematics are used as diagnostic of the aerodynamic function of the wings. Our flow visualizations support classifica¬ tion of wingbeat into clearly defined categories or gaits , associated with flight speed and with wing morphology, and distinguished by the function of the upstroke (see above, also Scholey 1983, Rayner 1988a). I use the term gait by analogy with terres¬ trial locomotion to refer to the pattern of movements of the limbs at different speeds ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 707 and in different types of flight; unlike the terrestrial parallel, transitions between flap¬ ping gaits may be gradual, and need not show the catastrophic discontinuities found in running tetrapods (Alexander 1989). In flapping flight there are strict constraints on the aerodynamic action of the wing during the wingbeat, and hence on gait. At any speed aerodynamic factors demand a particular set of kinematics to ensure that mean lift, weight and drag are in equilib¬ rium and that mechanical energy is minimum. The gait must be consistent with the mechanical properties and physiology of the muscles, because vertebrate muscles contract most efficiently only over a relatively narrow range of contraction strains (re¬ lated to wingbeat amplitude) and strain rates (related to frequency); adequate safety factors must also be maintained in the structures of the pectoral girdle and the wing. We do not know how strict these constraints on gait can be, and they may be suffi¬ ciently inflexible for birds to be seriously constrained in selection of gait. The vortex ring gait The Pigeon (Figure 5a) shows typical wing movements of slow flight. In the upstroke the wing is strongly flexed, and the wingtip is brought close to the body to minimize profile drag and inertia. This phase is aerodynamically inactive (Rayner 1979), gen¬ erates no lift, and the wake is formed of vortex rings generated solely during the downstroke, when the wing remains almost flat and fully stretched. As the upstroke has little aerodynamic significance, the detailed geometry of wing deformation is un¬ important, and it is not analysed in detail. FIGURE 3 - The wake of a Tawny Owl Strix aluco in steady slow flight at approximately 2 5 m/s. Two vortex rings are visible. In each ring the stopping vortex is clearly defined, but the starting vortex is more diffuse; this may reflect a delay in vortex strength develop¬ ment at the start of the downstroke. (Photograph A.L.R. Thomas.) 708 ACTA XX CONGRESSUS INTERNATIONAL^ ORNITHOLOGICI FIGURE 4 - The transitional wake in accelerated flight of Meyer’s Conure Policephalus meyeri. To the left of the image the wake is a vortex ring; this is followed by a weak trans¬ verse vortex and a lifting upstroke, followed by another transverse vortex before the con¬ tinuous vortex wake begins with a downstroke. The bird is in the middle of the following upstroke. (Photograph A.L.R. Thomas.) FIGURE 5a - Wingbeat kinematics. Pigeon Columba livia in very slow flight. (From Guidi 1939.) ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 709 FIGURE 5b - Cruising flight in Black-browed Albatross Diomedea melanophris. (From Scholey 1983.) 710 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI FIGURE 5c - Cruising flight in Ruppell's Griffon Vulture Gyps ruppellii. (From Scholey 1982.) ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 711 This vortex ring gait with an inactive and highly flexed upstroke is typical of slow flight in many birds, and is similar to the gait generally adopted in hovering (except by hum¬ mingbirds). It is also used by many small passeriforms (especially in bounding flight), and is widespread in galliforms, rails and other species with short, rounded wings, regardless of flight speed. The continuous wake gait In longer-winged birds, it becomes possible at higher speeds to use the upstroke to generate lift (Figure 5b). This gait is associated with the continuous vortex wake, and is characteristic of many birds in cruising or fast flight, particularly species with high aspect ratio and pointed wings, including pigeons, falcons, ducks, gulls, petrels (Scholey 1983). As in all gaits the wing is outstretched and planar during the downstroke; during the upstroke it remains nearly flat, but is flexed at the carpal joint and the wingtip is swept back so that wingspan is reduced (Scholey 1983, Rayner 1986, 1988a). Birds flying with this gait can be identified by three factors: the stroke plane is perpendicular to the axis of flight, the upstroke and downstroke are equal in duration, and when seen laterally the track of the wingtip is an ellipse. This gait is simple and efficient, and anatomically straightforward: a mathematical model is dis¬ cussed below. I consider it of considerable phylogenetic importance, as it is probably the primitive gait used by the first birds to evolve flight (Rayner 1988b); this is sup¬ ported by its close relationship with the line vortex wake in gliding. A slightly different gait is found in birds with lower aspect ratio, most commonly in species with low wing loading and square wings, and frequently with separated pri¬ mary feathers (Figure 5c), in particular owls, raptors, herons, cranes, storks and some Pelecaniformes. Again, the wing is fully extended in the downstroke, but wingbeat amplitude is greater, the upstroke is shorter than the downstroke, and the stroke plane is tilted. In the upstroke the arm wing is flexed, and the track of the elbow is highly elliptic. In owls the vortex wake associated with this gait is vortex rings (Figure 3), and the upstroke is passive. I am uncertain whether this is the case in larger birds using this gait, although analysis of film of both Old- and New-World vultures suggests a significant downward deceleration during the upstroke corresponding to the absence of weight support during that phase. Gait selection Most birds change gait as speed increases, with a tendency in longer-winged species to adopt a lifting upstroke so that force equilibrium can be maintained although the components of drag vary independently. The conditions under which gait should change are not yet fully understood, and a number of important questions must be raised: (1) Is it right to assume that birds can select speed freely, or are some species de¬ signed for a narrow range of speeds suited to their ecology? I have shown that pi¬ geons have difficulty flying at low speeds even with the vortex ring gait, and may not be able fully to support their weight, and this slow speed gait may be confined to ac¬ celerated flights. (2) When is it better to use the upstroke to help support the weight, despite the negative thrust, and when should it be passive and all lift be concentrated on the downstroke? (3) At what speed should the gait change, and is this speed spe¬ cific to any bird? (4) Does the selection of gait depend on flight morphology? 712 ACTA XX CONGRESSUS INTERNATIONAL^ ORNITHOLOGICI In terms of mechanical and aerodynamic efficiency the optimum gait is e ous vortex wake. This gait can only work if the bird can maintain constan circu with a shortened wing during the upstroke; this is not possible with short wings ien high wingbeat frequency) or in slow flight because local angles of inci ence are o great. In such circumstances the best alternative is to dispense with force genera ion from the upstroke, and instead to flex the wing to minimize profile drag and inertia^ The result is the vortex ring gait. Thus gait is determined by both morphology and flight speed, and this is in agreement with our observations from flow visualization and high-speed photography. ESTIMATION OF FLAPPING FLIGHT ENERGETICS Energy is a major demand for flying birds, and its quantification is essential if we are to understand the importance of flight in a bird’s biology. There are two approaches to this problem: power in flight can be measured directly through gas exchange or water metabolism, or theoretical models may be derived from which mechanical power (and hence indirectly metabolic power) may be estimated (e.g. Rayner 1986, 1988a, 1990). Theoretical models of flapping flight In the last century many theories were aimed at the construction of ornithopters, and their conspicuous lack of success resulted from misunderstanding of the mechanics of aerofoil action, and from concentration on the need to generate weight support - rather than thrust - by flapping. The first reasonably realistic model of flapping flight was that of Gnosspelius (1925), based on steady-state ‘blade element’ theory; this was the first to demonstrate that mechanical power in birds - as in aircraft - follows a U-shaped curve against flight speed. Like all later blade-element and related models (e.g. Pennycuick 1968) it suffers from the deficiency that aerofoil characteristics in flapping flight vary during the wingbeat, and probably differ substantially from those measured in gliding. A pragmatic model developed by Pennycuick (1969, 1975), and later reformulated by Tucker (1973) and Greenewalt (1975), is loosely based on the steady lifting line theory, central to fixed aerofoil aerodynamics. This incorporates the effects of bound and wake vorticity, but models lift for a fixed wing only and therefore neglects the energy cost of generating thrust, which is the central feature of flapping flight. Although they give little insight into flapping flight aerodynamics, these models are simple to use, and since weight is the dominant force on a flying bird, give rea¬ sonable estimates of total power in flight. I have argued above that the vortex wake is central to force generation in flight. As we now know the wake structure from flow visualization, it is feasible to develop re¬ alistic aerodynamic models of flapping flight, from which various mechanical quanti¬ ties associated with flight may be estimated. A passive upstroke leads to a vortex ring wake, and I developed a time-averaged unsteady lifting-line model to predict total mechanical power (Rayner 1979, 1980). This model predicted that to minimize power, a bird should adjust wingbeat kinematics with flight speed, and that the upstroke should become active at intermediate speeds (but modelling that situation was beyond the capacity of the vortex ring theory). The discovery that in faster flight circulation is constant - and therefore the absence of the complications of transverse vorticity - is the foundation of a model of the continuous wake gait (Rayner 1986, 1987, 1988a. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 713 1990) which is also based on unsteady lifting line theory. Full account is taken of wingbeat kinematics, which are predicted under conditions of force equilibrium and energy minimization. TOTAL induced profile parasite (b) a S § 0£ .s •o a -Q down- up - st roke 10 m/s i | FIGURE 6 - Estimated flight performance of the Kestrel Falco tinnunculus with active upstroke and constant circulation, and continuous vortex wake. For full details see Rayner (1988a). (a) Variation of components of mechanical power with flight speed, (b) Time course of total wingroot roll moments at different flight speeds, including components due to lift, induced drag, profile drag (very small), wing mass and wing inertia; no net work is done against inertia. Peak moment invariably occurs in the latter part of the upstroke. At higher speeds the moment is always positive, indicating that no muscular effort is required to elevate the wing. At lower speeds the moment is negative in the early phases of the downstroke, and the supracoracoideus should become active to elevate the wing. A typical curve of mechanical power against velocity is shown in Figure 6a. The rise in induced power at high speeds is caused by the change in wingbeat kinematics 714 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI needed to maintain thrust as parasite and profile drags increase, this ea ure wa nored in previous models (e.g. Pennycuick 1968). Predicted circulation an mema ic agree closely with those measured in several animals. Kinematics should change wi flight speed, with frequency and amplitude falling as speed increases. At low spee s it is advantageous for the upstroke to be aerodynamically inactive, the anima mus change gait, and the vortex-ring model is more realistic. Estimation of wingroot roll moments at the humeral joint (Figure 6b) generates a number of predictions which may be tested experimentally by methods of Dial et a . (1988). The roll moments are those exerted by the wing on the (rigid) body, and are proportional to the force exerted by the pectoral- muscles (pectoralis and supracoracoideus) as they depress and elevate the wing. A positive moment corre sponds to a downwards force from the pectoralis, and a negative moment indicates that the supracoracoideus is generating force. Muscle force is not in phase with lift because the dominant force for much of the wing stroke is inertia. At medium arid high flight speeds, the moment never falls below zero, and aerodynamic lift is more than sufficient to elevate the wing: the supracoracoideus is not needed and the pectoralis should be active throughout the wingbeat (although the number, and perhaps the type, of active muscle fibres probably varies). At lower speeds the moment is nega¬ tive during the latter part of the downstroke, and the abductor muscles should begin activity in the mid-downstroke rather than simply contracting during the upstroke. At all speeds peak depressor activity occurs during the final half of the upstroke. Ten¬ tative confirmation of these predictions comes from electrophysiological recordings of muscle activity in birds and bats flying at different speeds (Dial et al. 1988; see also Rayner 1986, 1987). Metabolic power and flight efficiency However accurate they may be as descriptions of flight mechanics, all mechanical power models have a serious limitation. They estimate the rate of increase of kinetic energy in to the surrounding air. This quantity may be equated with power output from the flight muscles, but its relation to metabolic energy in flight is tenuous; total energy consumption is presumably the greatest influence on the bird’s decision making. Tucker (1973), Greenewalt (1975) and Pennycuick (1975) all assumed - with no ex¬ perimental justification - that mechanical and metabolic powers are directly propor¬ tional, and that muscle efficiency is constant (of the order of 20-25%). Efficiency is sensitive to a wide range of factors, and comparison of mechanical estimate and metabolic measurement suggests it should increase sharply as mass to the power 1/ 4. as in terrestrial mammals, and for most smaller birds is appreciably lower than the empirical value of 0.25 (Rayner 1990). Moreover, wing root forces and wingbeat kin¬ ematics vary, and efficiency is unlikely to be independent of speed and gait. Far too little is currently known of the thermal physiology of bird flight muscles to explore this with any confidence, and existing estimates of efficiency may be inaccurate. These limitations probably provide at least a partial explanation for the widely reported dis¬ crepancy between predicted and measured metabolic power consumption in some birds (Rayner 1986, 1990). There is currently no reason to doubt either metabolic measurements or the more firmly grounded mechanical predictions of power, but energy and performance criteria predicted on mechanical grounds may have only in¬ direct significance for the true energy demands experienced by the animal (Rayner 1990). The reconciliation of these disparate physical quantities remains the major challenge for the future study of avian flight performance. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 715 ACKNOWLEDGEMENTS These studies have been funded by a Research Fellowship from the Royal Society of London and by grants from the Natural Environment Research Council and the Science and Engineering Research Council. I am grateful to Adrian Thomas for the use of photographs, and Patsy Hughes for commenting on the manuscript. LITERATURE CITED ALEXANDER, R.McN. 1989. Optimization and gaits in the locomotion of vertebrates. Physiological Reviews 69: 1199-1227 . DIAL, K.P., KAPLAN, S.R., GOSLOW, G.E., JR, JENKINS, F.A., JR 1988. A functional analysis of the primary upstroke and downstroke muscles in the domestic pigeon (Columba livia). Journal of Experi¬ mental Biology 134: 1-16. GNOSSPELIUS, O.F. 1925. Notes [to ‘The flight of birds’ by Fullerton, J.D.]. Journal of the Royal Aero¬ nautical Society 29: 543-547 & 648-649. GREENEWALT, C.H. 1975. The flight of birds. Transactions of the American Philosophical Society, vol. 65, part 4. GUIDI, G. 1939. La battuta alare del piccione. Aerotecnica, Roma 19: 121-132. KOKSHAYSKY, N.V. 1979. Tracing the wake of a flying bird. Nature 279: 146-148. MAGNAN, A., PERRILLIAT-BOTONET, C., GIRERD, H. 1938. Essais d’enregistrements cinematographiques simultanees dans trois directions perpendiculaires deux a deux a I’ecoulement de I’air autour d’un oiseau en vol. Comptes rendues hebdomaires des Seances de I’Academie des Sci¬ ences, Paris 206: 462-464. PENNYCUICK, C.J. 1968. Power requirements for horizontal flight in the pigeon Columba livia. Jour¬ nal of Experimental Biology 49: 527-555. PENNYCUICK, C.J. 1969. The mechanics of bird migration. Ibis 111: 525-556. PENNYCUICK, C.J. 1975. Mechanics of flight. Pp. 1-75 in Farner, D.S., King J.R. (Eds). Avian biol¬ ogy, vol. 5. New York, London, Academic Press. PENNYCUICK, C.J. 1988. On the reconstruction of pterosaurs and their manner of flight, with notes on vortex wakes. Biological Reviews 63: 209-231. RAYNER, J.M.V. 1979. A new approach to animal flight mechanics. Journal of Experimental Biology 80: 17-54. RAYNER, J.M.V. 1980. Vorticity and animal flight. Pp. 177-199 in Elder, H.Y., Trueman, E.R. (Eds). Aspects of animal movement. Seminar Series of the Society for Experimental Biology 5. Cambridge University Press. RAYNER, J.M.V. 1986. Vertebrate flapping flight mechanics and aerodynamics, and the evolution of flight in bats. Pp. 27-74 in Nachtigall, W. (Ed.). Bat flight - Fledermausflug. Biona Report 5. Stuttgart, Gustav Fischer. RAYNER, J.M.V. 1987. The mechanics of flapping flight in bats. Pp. 23-42 in Fenton, M.B., Racey, P.A., Rayner J.M.V. (Eds). Recent advances in the study of bats. Cambridge University Press. RAYNER, J.M.V. 1988a. Form and function in avian flight. Current Ornithology 5: 1-77. RAYNER, J.M.V. 1988b. The evolution of vertebrate flight. Biological Journal of the Linnean Society 34: 269-287. RAYNER, J.M.V. 1990. The mechanics of flight and bird migration performance. Pp. 283-299 in Gwinner, E. (Ed.). Bird migration. Heidelberg, Springer Verlag. RAYNER, J.M.V., JONES, G., THOMAS, A. 1986. Vortex flow visualizations reveal change of upstroke function with flight speed in microchiropteran bats. Nature 321: 162-164. SCHOLEY, K.D. 1983. Developments in vertebrate flight. Unpublished PhD. Thesis, University of Bris¬ tol. SPEDDING, G.R. 1986. The wake of a jackdaw (Corvus monedula) in slow flight. Journal of Experi¬ mental Biology 125: 287-307. SPEDDING, G.R. 1987a. The wake of a kestrel (Falco tinnunculus) in gliding flight. Journal of Experi¬ mental Biology 127: 45-57. SPEDDING, G.R. 1987b. The wake of a kestrel (Falco tinnunculus) in flapping flight. Journal of Experi¬ mental Biology 127: 59-78. SPEDDING, G.R., RAYNER, J.M.V., PENNYCUICK, C.J. 1984. Momentum and energy in the wake of a pigeon (Columba livia) in slow flight. Journal of Experimental Biology 1 1 1 : 81 -1 02. TUCKER, V.A. 1973. Bird metabolism during flight: evaluation of a theory. Journal of Experimental Biology 58: 689-709. 716 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI NEUROMUSCULAR ORGANIZATION FOR FLIGHT: ISSUES FOR STUDY G. E. GOSLOW, JR. Section for Population Biology. Morphology and Genetics, Brown University, Box G-BMC Providence, Rhode Island 02912, USA ABSTRACT. The pectoralis, the major depressor muscle of the wing, is structurally and funofonally complex but its neuromuscular organization for flight remains relatively unexplored. We have begun a series of investigations at the motor unit level in two species with contrasting ig s yes, mestic Pigeon Columba livia and the European Starling Sturnus vulgaris. Thirty motor units from the pectoralis of anaesthetised Domestic Pigeons have been isolated. Of these units, 29 were relative y small and one relatively large. When tested for fatigue resistance, 75% of the units were resistant and 25% fatigable. These data are consistent with the hypothesis that within the pectoralis of Domestic Pigeons, two populations of motor units exist which are functionally partitioned for function. Keywords: Flight control, motor units, recruitment. INTRODUCTION Several years ago while reviewing some high speed films of raptors attacking and striking their prey, I recall my wonder at a Goshawk’s ability to alter so quickly and with such grace its entire body orientation in space within a single wingbeat. This split- second precision requires a complex interplay of the sensory and motor components of the flight apparatus. Along with several colleagues and students, I am involved in a series of studies designed to address the neuromuscular organisation of the mus¬ cles responsible for wing movements in two species of birds, the Domestic Pigeon Columba livia and the European Starling Sturnus vulgaris. These two species differ in their flight characteristics which may help us to delineate features specific to each as well as those that are general for flight. This is an exciting time for studies of ani¬ mal flight. Technological advances are enabling scientists to test a number of hypotheses regarding muscle function and efficiency which bear not only on the evo¬ lution of different flight styles and morphologies, but even the evolution of flight in fossil forms (Rayner 1989, Ruben 1990). We employ a number of experimental tech¬ niques to better understand wing kinematics during flight as well as the mechanics of the muscles which underlie movements seen. This summary will focus on some of our ongoing studies of the organization of the motor units for use in the primary depres¬ sor muscle, the pectoralis. Recent studies reveal that the pectoralis of birds is surprisingly complex. Studies of anatomically complex muscles producing jaw and limb movements in reptiles and mammals illustrate a correlation between peripheral neuromuscular organisation and function (Herring et al. 1979, Weijs & Dantuma 1981, English & Weeks 1987). Indi¬ vidual branches of muscle nerves may innervate discrete groupings of muscle fibres within a gross muscle to provide a neuromuscular compartment which, in turn, may be organised to perform a specific behavioural task. Although aspects of this concept are controversial (Windhorst et al. 1990), it provides a useful framework from which ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 717 to begin this paper and for discussing the general functional organisation of the pectoralis. GENERAL ORGANIZATION OF THE PECTORALIS In many birds (including Pigeons and European Starlings) the major part of the pectoralis (pars pectoralis) is divided by a sheet of connective tissue into two heads, the sternobrachialis (SB) and thoracobrachialis (TB) (Simic & Andrejevic 1963, Vanden Berge 1979). Although both heads insert onto the deltopectoral crest of the humerus, they possess separate origins and fibre orientations and apparently are innervated by distinct nerve branches from the brachial plexus (Kaplan & Goslow 1989). Stimulation of these two nerve branches in Pigeons reveals that the SB is capable of humeral depression as well as protraction and the TB of depression as well as retraction (Dial et al. 1988). Anatomical tracer studies of the neurons innervating the SB and TB of Pigeons also illustrates an organisation within the spinal cord sug¬ gestive of a functional differentiation in their use (Sokoloff et al. 1989). As might be anticipated, electromyograms (EMGs) of the SB and TB from free-flying Pigeons (Dial et al. 1987, 1988) and Starlings (Dial et al. 1991) support a hypothesis that the SB and TB are used deferentially during different flight modes. A functional partitioning of the pectoralis has broad implications for the neural con¬ trol of flight and serves as a springboard for subsequent studies. There are data for both Pigeons and Starlings which suggest that within the SB and TB, further functional specialization exists at the motor unit level. A brief review of motor unit organisation and use is in order. MOTOR UNIT ORGANIZATION AND FUNCTION A motor unit, the smallest functional unit of neuromuscular organisation, consists of a single neural component, the motoneuron, and the collection of muscle fibres it in¬ nervates, the muscle unit. Within the ventral horn of the gray matter of the spinal cord, the cell bodies of motoneurons that send axons to a particular muscle are clustered in motor pools. Within a muscle s pool, or across pools for groups of muscles, motoneurons are thought to be selectively activated to complete a locomotor task (for review, Burke 1981, Stuart & Enoka 1983, English 1985). Following the all-important series of papers by Elwood Henneman in the mid-1960’s, studies of the anatomical and physiological properties of isolated motor units have substantially enhanced our understanding of the neuromuscular basis for locomotion. The formulation of an or¬ der of motoneuron recruitment based upon the size principle (Henneman & Mendell 1981, Binder & Mendell 1990) has provided a conceptual framework for motor unit organisation which is testable and may now be extended to birds (but see Zajac 1990). The SB and TB of the Pigeon are comprised of two distinct populations of muscle fi¬ bre types (Figure 1): large fibres (78.2 pm diameter), which appear to have little aero¬ bic capacity (based on histochemistry) and small fibres (33.5 pm diameter), which appear to be richly aerobic (George & Naik 1960, Talesara & Goldspink 1978, Kaplan & Goslow 1989). The electromyographic studies of Dial et al. (1987, 1988) suggest 718 ACTA XX CONGRESSUS INTERNATIONAL^ ORNITHOLOGICI that the large, anaerobic fibres in the Pigeon are used preferentially for takeoff and landing but not for level flight whereas the small, aerobic fibres are used for all modes of flight. The histochemical organisation of the pectoralis of the European Star ing differs in several respects from that of the Pigeon. In contrast to the bimodal popu lation of fibre size and aerobic capacity found in the pectoralis of the Pigeon, this muscle of European Starlings consists entirely of small, aerobic fibres (Rosser & George, 1986). It should be noted, however, that based on differences in aerobic histochemistry, these authors recognise two types of fibres in the Starling pectoralis. FIGURE 1 - Large and small fibres in the pectoralis of Columba livia. Photomicrograph is a frozen section (20 |um) from the SB, stained for myofibrillar ATPase after acid preincubation (pH 4.35). Two populations of fibres are evident. FB - fascicle boundary Scale bar equals 200 |um. After Dial et al. (1987). ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 719 Among flying animals, it seems likely that selection must act to optimize overall me¬ chanical and chemical performance of the muscles that drive the wing while maintain¬ ing plasticity in the locomotor system. Fundamental constraints of contractile protein design dictate the existence of a narrow range of contraction speeds for maximum efficiency (Goldspink 1977, 1981); thus, flight muscles in birds might be designed to operate at a fixed level of power output which allows for predictions of fiber recruit¬ ment. Several studies have addressed fibre recruitment in bird flight muscle based on empirical data or theoretical considerations. Gulls Larus argentatus, like Pigeons, possess two distinctly different sized muscle fibres in the pectoralis. In their study of gulls flying and gliding in a wind tunnel, Goldspink et al. (1978) noted amplitude changes in the EMGs and speculated about motor unit recruitment order, but as in the studies of Dial et al. (1987, 1988) their interpretation was limited by a lack of knowl¬ edge concerning motor unit organisation and contractile properties. In a thoughtful paper, Rayner (1985) draws conclusions about the recruitment of muscle fibres in two groups of birds which characteristically fly with intermittent pe¬ riods of ballistic or gliding flight. He notes that in some of these species (European Starlings included), the pectoralis is comprised of a relatively homogeneous popula¬ tion of aerobic fibres and that wingbeat frequency and amplitude remain relatively constant over a range of flight speeds. Rayner deduced that muscle contraction dy¬ namics remain near optimum and, accordingly, all (or virtually all) aerobic fibres within the pectoralis might be expected to be recruited with each downstroke. Clearly, some knowledge of the neuroanatomical and neurophysiological properties of the motor units which comprise the pectoralis of Pigeons and Starlings is needed in order to better evaluate these fundamental questions regarding motor unit use. MOTOR UNIT STUDIES IN COLUMBA LIVIA Welsford et al. (1991) have initiated studies of isolated motor units from the pectoralis muscle of Domestic Pigeons to determine the functional correlation to the two types of muscle fibres noted as well as to gain some understanding of their contractile prop¬ erties as related to flight. Motoneurons were isolated intracellularly or extracellularly in adult, anaesthetised birds. Of the 30 units isolated from the SB, 29 units were small and generated 0.22% or less of the maximum tetanic tension of the entire SB. One exceptionally large unit developed a peak tension of 0.56% whole muscle tension (Figure 2). When tested for fatigue resistance, 75% of the units were considered re¬ sistant and 25% fatigable (Figure 3). The expectation (based on studies in mammals) is that during takeoff and landing, when airspeed is low and high power output necessary, populations of units capable of producing large forces will be recruited. It is anticipated that within the Pigeon pectoralis, two distinct populations of motor units will be seen relative to size and fatigability. These limited data support such a hypothesis. The units comprising the Starling pectoralis, in contrast, might be expected to fall within a continuum for these two parameters. An investigation of the units in this species remains to be done. Clearly larger samples of units are needed before conclusions can be formulated with confidence. We believe, however, that studies such as these will yield data for our understanding of not only the neural control of flight, but of the evolution of this re¬ markable mode of locomotion. 720 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI FIGURE 2 - Distribution of maximal tetanic tensions of SB muscle units expressed as a percentage of SB tetanic force. Mean = 0.11% (n = 30). After Welsford et al. (1991). FIGURE 3 - Fatigability of SB muscle units during four minutes of stimulation (500 msec of each second) at 60 Hz, expressed as percent of initial muscle unit force (n = 9 from two experiments). After Welsford et al. (1991). ACKNOWLEDGMENTS This work was supported by National Science Foundation Grant DCB-87-1 8727. ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 721 LITERATURE CITED BINDER, M.D., MENDELL, L.M. 1990. The segmental motor system. New York, Oxford. BURKE, R.E. 1981. Motor units: anatomy, physiology, and functional organisation. Pp. 345-421 in Brooks, V.B. (Ed.). Handbook of physiology - the nervous system II. Bethesda, MD, American Physi¬ ological Society. DIAL, K.P., KAPLAN, S.R., GOSLOW, G.E., JR., JENKINS, F.A., JR., 1987. The structure and neu¬ ral control of the pectoralis in pigeons: Implications for flight mechanics. Anatomical Record 218. 284- 287. DIAL, K.P., KAPLAN, S.R., GOSLOW, G.E., JR., JENKINS, F.A., JR. 1988. A functional analysis of the primary upstroke and downstroke muscles in the Domestic Pigeon ( Columbia livia) during flight. Journal of Experimental Biology 134: 1-16. DIAL, K.P., GOSLOW, G.E., JR., JENKINS, F.A., JR. 1991. The functional anatomy of the shoulder in the European Starling ( Sturnus vulgaris). Journal of Morphology 207: 327-344. ENGLISH, A.W. 1985. Limbs vs. jaws: Can they be compared? American Zoologist 25: 351-363. ENGLISH, A.W., WEEKS, O.l. 1987. An anatomical and functional analysis of cat biceps femoris and semitendinosus muscles. Journal of Morphology 191: 161-175. GEORGE, J.C., NAIK, R.M. 1960. Some observations on the distribution of blood capillaries in the pigeon breast muscle. Auk 77: 224-227. GOLDSPINK, G. 1977. Mechanics and energetics of muscle in animals of different sizes, with particular reference to the muscle fibre composition of vertebrate muscle. Pp. 37-55 in Pedley, T.J. (Ed.). Scale effects in animal locomotion. New York, Academic press. GOLDSPINK, G. 1981. The use of muscles during flying, swimming and running from the point of view of energy saving. Symposium of the Zoological Society of London 48: 219-338. GOLDSPINK, G., MILLS, C., SCHMIDT-NIELSEN, K. 1978. Electrical activity of the pectoral muscles during gliding and flapping flight in the Herring Gull. Experimentia 34: 862-865. HENNEMAN, E., MENDELL, L.M. 1981. Functional organisation of motoneuron pool and its inputs. Pp. 423-507 in Brooks, V.B. (Ed.). Handbook of Physiology - the Nervous System II. Bethesda, MD, Ameri¬ can Physiological Society. HERRING, S.W., GRIMM, A.F., GRIMM, B.R. 1979. Functional heterogeneity in a multipinnate mus¬ cle. American Journal of Anatomy 154: 563-576. KAPLAN, S.R., GOSLOW, G.E., JR. 1989. Neuromuscular organisation of the pectoralis (pars thoracicus) of the pigeon ( Columbia livia): implications for motor control. Anatomical Record 224: 426- 430. RAYNER, J. M. V. 1 985. Bounding and undulating flight in birds. Journal of Theoretical Biology 117: RAYNER, J.M.V. 1989. Mechanics and physiology of flight in fossil vertebrates. Transactions of the Royal Society of Edinburgh: Earth Sciences 80: 311-320. ROSSER, B.W.C., GEORGE, J.C. 1986. The avian pectoralis: histochemical characterisation and dis¬ tribution of muscle fiber types. Canadian Journal of Zoology 64: 1 1 74-1 1 85. RUBEN, J. 1991. Reptilian physiology and the flight capacity of Archaeopteryx. Evolution 45:1-17. SIMIC V., ANDREJEVIC, V. 1963 Morphologie und Topographie der Brustmuskeln bei den Hausphasianiden und der Taube. Gegenbaurs Morphological Jahrbuk 194: 546-560. SOKOLOFF, A., DEACON, T., GOSLOW, G. E., JR. 1989. Musculotopic innervation of the primary flight muscles, the pectoralis (pars thoracicus) and supracoracoideus of the pigeon ( Columbia livia): a WGA-HRP study. Anatomical Record 225: 35 - 40 . STUART D G ENOKA, R. 1983. 1983. Motoneurons, motor units, and the size principle. Pp. 471- SI 8 in Rosenberg, R. N. (Ed.). Clinical neurosciences - neurobiology. New York, Churchill Livingston. TALESARA G L GOLDSPINK, G. 1978. A combined histochemical and biochemical study of myofibrillar ATPase in pectoral, I ©g , and cardiac muscle of several species of bird. Histochemistry Journal 10: 695-710. VANDEN BERGE, J. C. 1979. Myologia. Pp. 199-219 in Baumel, J. J., King, A. S., Lucas, A. M., Rroa7ilp J E Evans H E (Eds). Nomina anatomica avium. New York, Academic Press. WE LS FORD I G„ MEYERS, R. A., WILSON, D. S., SATTERLIE, R. A. GOSLOW. <3. E„ JR. 1991. Neuromuscular organisation for “wing” control in a mollusc ( Clione limecina) and a bird (Columba livia): Parallels in design. American Zoologist (In press). WEIJS W A DANTUMA, R. 1981. Functional anatomy of the masticatory apparatus in the rabbit lOrvctolaous cuniculus Netherlands Journal of Zoology 31: 99-147. WINDHORST, U., HAMM. T. M., STUART, D. G. 1989. On the function of muscle and reflex partition¬ ing Behavioural and Brain Sciences 12: 629-681. ., T. „ . 7AIA0 F E 1990 coupling of recruitment order to the force produced by motor units: The size prin¬ ciple hypothesis" revisited. Pp. 96-111. in Binder, M. D„ Mendell, L. M. (Eds). The segmental motor system. New York, Oxford. 722 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI CORE TEMPERATURE RELATIONS OF PIGEONS DURING PROLONGED WIND TUNNEL FLIGHT W. NACHTIGALL and K.-D. HIRTH Arbeitsgruppe Nachtigall, Zoologisches Institut der Universitat des Saarlandes, D-6600 Saarbrucken, Germany ABSTRACT. Using thermistors, core temperature Tc was measured in pigeons Columba livia, (breed “Grippler”) during rest and during flight in a wind tunnel. Mean Tc at rest was 39.8 ± 0.7 C and was independent of ambient temperature Ta (10-30°C). In the first minutes of flight, Tc increased to 1.5- 3.0°C above resting level and remained at this higher level. This hyperthermia increased with Ta (v = const.). It was more or less constant in the low Ta range (10.6-1 3. 9°C) at flight speeds v ranging from 10 to 18 ms 1 and normal body mass, but increased with v and elevated body mass in the high Ta range (23.7-28.8°C). Flight behaviour with and without instrumentation was essentially the same. Hyperthermia during flight was lower in our well trained long flying (more than 3 hrs in the wind tun¬ nel) pigeons than in short flying birds used by other authors, but present in all flights analyzed. The meaning of hyperthermia is discussed. Keywords: Pigeon, Columba livia, temperature, core temperature, wind tunnel flight, flight behaviour. INTRODUCTION The metabolic rate of a bird during flight is at least ten times higher than during rest¬ ing. Taking into consideration that approximately 25% of the metabolic rate is required for muscle activity, the remaining 75% which is transformed into heat must be dis¬ posed of should the bird not become overheated during long flights. According to results obtained in wind tunnel tests (Rothe, Biesel & Nachtigall 1987) the relative metabolic rate of a “Grippler” Pigeon Columba livia of average mass (mean body mass 0.33 kg) was calculated to be 68 W kg1 for free flight in nature. This gives an absolute metabolic rate of approximately 24 W from which 6 W are required for flight performance and 18 W must be discharged as heat. Several areas on the body, wings and extremities are available for discharging sur¬ plus heat (Figure 2a). Heat dissipation is favoured by especially high body tempera¬ tures which may be higher than under resting conditions (hyperthermia). The ques¬ tions arose, which values of hyperthermia are used by the flying animals and what are the internal and external limiting conditions for hyperthermal constancy during long distance flights (which in turn indicate unproblematic heat loss). METHODS Wind tunnel and animals A special wind tunnel was developed and built for these tests (Figure la) and a se ries of well trained pigeons belonging to the “Grippler” race (Figure 1b) was used Details are to be found in Rothe and Nachtigall (1987). ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 723 FIGURE 1 - Pigeon flight in a wind tunnel. (A) Large wind tunnel used by the Arbeitsgruppe Nachtigall at the Institute of Zoology, University of Saarland, and built for the analysis of bird flight. (B) Pigeon flying in the measuring area of the wind tunnel. 724 ACTA XX CONGRESSUS INTERNATIONAL^ ORNITHOLOGICI Start Landing FIGURE 2 - Temperature relationship in flying Grippler pigeons. (A) Heat loss areas (dotted areas) and measurement of core temperature Tc. (B) Changes in core temperature T and ambient temperature Ta at rest and at different speeds v during a long flight. Speed was altered step-wise. High ambient temperature. (C) like B, lower ambient temperature. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 725 FIGURES 2 - Continued. Start Landing Temperature measurements In order to measure the core temperature of the body a Siemens M 85 thermistor (± 0.1°C) was placed deep in the colon. An extremely fine, flexible conducting wire led from the sealed cloaca to a central registering unit which also measured the tempera¬ ture of other areas of the body (not discussed here; see Hirth, Biesel & Nachtigall 1987) in short intervals via a multiplexer. The extra mass of 4 g coming from the measuring system was only 1% of the body mass and did not cause any recognisable anomalies in behaviour. The whole wind tunnel room could be heated or cooled to change the environmental temperatures. Experimental procedure Before beginning with an experiment, the test bird was weighed, the thermistor put in place and the bird was then set on a perch within the test area of the wind tunnel under low light intensity. The core temperature of the bird sitting without wind, sitting with the wind (speed 10-12 ms'1) turned on and flying at different wind speeds was measured until a constant temperature was obtained in each situation. The experi¬ ment usually lasted 30 minutes, during which time the flight speed was altered sev¬ eral times. Immediately after flight was stopped, the resting temperature value was measured again with and without wind in reverse order. RESULTS Flight behaviour The pigeons demonstrated typical flight behaviour, i.e. beating phase punctuated with 726 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI Temperature range ; Temperature range*. 10.6- 13, 9 °C ▲ u o a> L_ D O k_ a> CL E £ <1) L_ o U 42- 41 - 40- 39 -■ o> o CO I o CN CO i , v =14 ms'1 B £ u z a> o >o co i o CO CO M Z O) o CO CO • o o CO Body mass 23,7" 28,8 °C FIGURE 3 - Average core temperature Tc of pigeons during wind tunnel flight. The core temperature is independent of the flight speed (A) and body mass (B) at low ambient tem¬ peratures. At high ambient temperatures, the mean core temperature rises with flight speed (C) and body mass (D). Marks: standard deviation. gliding phases at wind speeds of 10-14 ms 1 and continuous beating at speeds above 14 ms V When ambient temperatures were high, birds with extra high body mass and/ or at especially high flight speeds showed an increasing tendency towards flying ir¬ regularly and attempting to land. On such occasions the birds were seen to pant and spread their wings after landing. Sometimes signs of thermoregulatory behaviour (e.g. beak opening, lowering the legs which are usually tucked under their feathers during ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 727 flight (Biesel & Nachtigall 1987)) were observed. The flight times at high ambient tem¬ peratures (24 - 29.1°C) were normally shorter (23.1 min ± 8.7 min; n = 11) than at lower ambient temperatures (1 0.6 - 1 6.5°C; 57.6 min ± 1 1 .2 min; n = 22). As soon as the upper speed limits at which the bird could still carry out stationary thermoregulation were known, care was taken to remain below these speeds when metabolic rate measurements were being carried out. Time functions of core measurements Figures 2b, 2c show the time functions of the core temperature T and ambient tem¬ perature (wind temperature) Ta from two flights at high and low Ta. Tc remained more or less constant around 40°C during the pre-flight resting period, independent of Ta. After flying had started, Tc increased by a few degrees (1.5 to 3.0°C) and reached a new stationary value after 6 to 10 minutes of flight. An increase or decrease in flight speed resulted in Tc increasing or decreasing accordingly until a new constant value was obtained. After flight, T decreased more or less exponentially to the value be¬ fore flight began. Core temperature during rest The average core temperature Tc during rest was 39.8°C ± 0.7°C (n = 48). A statis¬ tically significant (P=0.05) correlation between T and Ta was not found in the Ta range studied. Thus the core temperature remained independent of the wind temperature and was not influenced by the cooling effect of wind flow. Core temperature during flight T increased during flight. At an optimal flight speed of 12 m s1 Tc lay between 39.7 and 43.3°C (minimum and maximum values measured). During flight Tc increased significantly (P = 0.05) and linearly with Ta within the temperature range studied. Thus a higher hyperthermia (difference between flight Tc and resting Tc) occurs when the ambient temperature rises, i.e. around 1 .4°C at Ta = 1 0°C and around 2.3°C at Ta = 30°C. Within the speed range of 1 0 < v < 1 8 m s*1 tested, the mean core temperature was independent of flight speed v as long as the ambient temperature was low (10.6 - 1 3.9°C). At higher ambient temperatures (23.7 - 29.1°C) Tc increased with v signifi¬ cantly (P = 0.05) and linearly : Tc (°C) = 0.22 v (ms1) + 38.8. Furthermore, flight val¬ ues of T were higher in animals which had a heavier body mass due to feeding be¬ fore a flight experiment, but this effect was only noticable at higher ambient tempera¬ tures (Figure 3 a-d). DISCUSSION Core temperatures compared with measurements by other authors The resting core temperature values of 39.8 ± 0.7°C of our pigeons were similar to those measured by Bernstein (1974), but clearly lower than values obtained by Hart and Roy (1967), Aulie (1961) and Butler et al. (1977). Compared to these data the values of 43.3 to 44.5°C measured by Butler et al (1988) appear to be extremely high. These measurements were all obtained from very short flights. They lasted on an average less than 10 minutes and in some case even under 2 minutes because the 728 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI birds refused to fly longer. Their behaviour after landing (panting and wing spreading) was typical for overheating. One may say that these pigeons were unsuitable for wind tunnel flight and not suffi ciently trained and therefore, the measurements do not represent quasi-natura con¬ ditions. On the contrary the core temperature of our pigeons which flew non s op or more than three hours in a wind tunnel (with an optimal speed of 12 ms an norma body mass), remained constant at low ambient temperatures. In our experiments, unphysiologically high Tc values and symptoms of overheating only occurred when ambient temperatures were high, and the additional stress o ig flight speed (Figure 3c) and/or too high body mass (overfed birds or birds with their crops full) was given (Figure 3d). Even then, maximum core temperatures of around 43.5°C were not as high as those obtained by other authors in pigeons (and other birds; not discussed here). Nevertheless, an increase in resting values compared to flight values was found in all test cases; in our case hyperthermia was 1.5 to 2.5°C at 10°C < Ta < 30°C. These values are not excessive, do not reflect unphysiological situations and can be held constant for hours without any apparent difficulty. Meaning of hyperthermia The importance of a constant hyperthermia may lie in the fact that, obviously due to enzymatic activation, a higher muscle temperature increases the maximum perform¬ ance output and possibly the muscle efficiency as well, in man (Torre-Bueno 1976). This would have a great advantage for birds if it were valid for them too. A bird would then be able to fly just as efficiently with less muscle mass and thus fly for greater distances during long distance flight with the energy reserves available. The hyperthermia observed would therefore have more of a working physiological function than of a heat regulatory one, and is nevertheless important for extremely long migra¬ tion flights. The excessive hyperthermia which depasses the physiological, stationary values and only occurs under the above mentioned stress-conditions, can hardly be explained as being other than an emergency mechanism to dispose of excess heat for a short time. ACKNOWLEDGEMENTS We wish to thank DFG for financial support, Dr H.J. Rothe and Dr W. Biesel for sug¬ gestions concerning the measurements and their comments on the analysis and pres¬ entation of the results and Ms W.Pattullo for translating the German version. LITERATURE CITED AULIE, a. 1971. Body temperatures in pigeons and budgerigars during sustained flight. Comp. Biochem. Physiol. 39 A: 173-176. BERNSTEIN, M.J. 1974. Vascular responses and foot temperature in pigeons. Am. J. Physiol. 226: 1350-1355. BIESEL W NACHTIGALL, W. 1987. Pigeon flight in a wind tunnel IV. Thermoregulation and water homeostasis. Journal of Comparative Physiology B 157: 117-128. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 729 BUTLER, P.J., WEST, N.H., JONES, D.R. 1977. Respiratory and cardiovascular responses of the pi¬ geon to sustained level flight in a wind tunnel. Journal of Experimental Biology 71: 7-26. HART, J.S., ROY, O.Z. 1967. Temperature regulation during flight in pigeons. Am. J. Physiol. 213. 131 1- 1316. HIRTH, K.D., BIESEL, W., NACHTIGALL, W. 1987. Pigeon flight in a wind tunnel III. Regulation of body temperature. Journal of Comparative Physiology B157: 111-116. ROTHE, H.J., NACHTIGALL, W. 1987. Pigeon flight in a wind tunnel I. Aspects of wind tunnel design, training methods and flight behaviour of different pigeon races. Journal of Comparative Physiology B 157: 91-98. ROTHE, H.J., BIESEL, W., NACHTIGALL, W. 1987. Pigeon flight in a wind tunnel II. Gas exchange and power requirements. Journal of Comparative Physiology B157: 99-109. TORRE-BUENO, J.R. 1976. Temperature regulation and heat dissipation during flight in birds. Jour¬ nal of Experimental Biology 65: 471-482. WHITTOW, G.C. 1967. Regulation of the temperature. In: Sturkie P.O. (Ed.). Avian physiology. Bailliere, Tindal Cassel, London. 730 ACTA XX CONGRESSUS INTERNATIONAL^ ORNITHOLOGICI ON THE AERODYNAMICS OF THE TAIL IN BIRDS D. HUMMEL Institui fur Stromungsmechanik, Technische Universitat Braunschweig, Bienroder Weg 3, D - 3300 Braunschweig, Germany ABSTRACT. Windtunnel experiments have been carried out on a rectangular wing for a large variety of tail planform shapes and deflection angles as well as for twisted tails in symmetrical and unsymmetrical free stream flow. As in airplanes the tail of a bird acts as a device to maintain longitu¬ dinal and lateral stability and control. Adding a tail to a wing leads to an increase of longitudinal sta¬ bility. Whereas in airplanes lateral stability is performed mainly by means of a vertical fin, in birds the same is achieved by twisting the tail. Concerning control of the longitudinal motion deflections of the tail up and down change lift and pitching moment, and the corresponding effectiveness is governed by size and aspect ratio of the tail. Control of the lateral motion in airplanes is effected by side forces and yawing moments due to the deflection of the rudder of a vertical fin. In birds, however, such forces and moments are again produced by twisting the tail, which leads to a lateral component of the tail force. Keywords: Aerodynamics, control, stability, tail. INTRODUCTION Stability and control of airplanes are maintained by small additional lifting surfaces which are usually positioned downstream of the wing. A horizontal wing stabilizes the longitudinal motion and a flap at this wing acts as an elevator to control the pitching moment. Concerning the lateral motion a vertical fin leads to increased directional stability and a rudder at this fin is used to control the yawing moment, whereas con¬ trol of the rolling moment is mainly achieved by deflections of the ailerons at the main wing. In birds, however, the situation is different. Birds have at first glance no verti¬ cal fin and the small additional horizontal wing is attached to the trailing-edge of the main wing as a tail. It is common understanding since a long time (Stresemann 1934, Ruppel 1975, Oehme 1976b, Nachtigall 1985, Burton 1990) that deflections of the tail up and down act as an elevator to control the longitudinal motion. The effects of length, size and shape of the tail relative to the wing on the control effectiveness are unknown and systematic investigations on this subject are missing. Concerning lateral control twist¬ ing of the tail has been explained as a device to produce a rolling moment (Oehrne 1976a,c, Nachtigall 1985) and other effects have not yet been discussed so far. The important contribution of a tail to the longitudinal and lateral stability has not been con¬ sidered. Therefore the present knowledge on the aerodynamic effects of a tail is at low standard and systematic investigations are urgently needed. METHODS Comprehensive experimental investigations have been carried out in the 1 ,3m low speed windtunnel of Institut fur Stromungsmechanik at Technische Universitat Braunschweig, Germany. A rectangular wing, having an aspect ratio (span/chord) A ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 731 = b/c = 5.0 and a NACA 3412 airfoil, has been equipped with a series of 26 tails. Length, size and lateral spreading have been varied systematically including forked planform shapes. Three-component balance measurements have been carried out for all these wing-tail configurations for plane (untwisted) tails and for different tail deflec¬ tions up and down. For one typical tail planform shape, which will be discussed sub¬ sequently, also a twisted tail has been investigated. In this case six-component bal¬ ance measurements have been performed again for different tail deflections up and down. The Reynolds number was Re = V c/v = 3.7 • 105. Following here the results for only one wing-tail configuration (called tail B) both un¬ twisted and twisted are demonstrated to show the basic contributions of a tail to lon¬ gitudinal and lateral stability and control. The comprehensive results on the effects of tail planform shape on these contributions will be published elsewhere. RESULTS The rectangular wing and the tail of shape B are shown in Figure 1 . The length I and the width 2a = 2f of this tail are equal to the chord length c of the wing (l/c = 1 .0, f/a = 1 .0, a/c = 0.50) and the tail area ST is therefore 20 % of the wing area Sw. Without a deflection up or down the tail is represented by a flat plate which is adjusted tan¬ gential to the centre line of the airfoil NACA 3412 at the trailing-edge of the wing. The diagram in Figure 1 shows the lift coefficient cL = 2L/pV2S (L lift, p density of the air, V free stream velocity, S area of wing or area of wing plus tail) and the pitching mo¬ ment coefficient about the quarter chord point cm = 2M/pV2S c (M pitching moment nose-up positive, 7T reference chord) as functions of the angle of attack a. The results for the wing alone are discussed first. Lift and pitching moment depend linearly on the angle of attack and the departure from this behaviour at high angles of attack is due to flow separations. The pitching moment slope dcm/dcL turns out to be positive. With increasing lift the nose-up pitching moment increases as well and this leads again to an increase of the angle of attack. This means that the wing alone is unstable. By adding the tail B to the wing the lift L as well as the area S increase. Since the lift coefficients of the wing-tail configuration are lower than those for the wing alone one can conclude that the generation of lift by adding a tail to the wing is very ineffective. This is due to the fact that the tail works in a downwash field which is induced by the wing. Concerning the pitching moment the slope dcm/dcL is now negative for the wing- tail configuration. This means that the wing-tail configuration is stable. By adding the tail to the wing longitudinal stability is considerably improved. By deflections of the tail up (e = - 5°) and down (e = + 5°) the lift and pitching moment curves c (a) and c ^ (a) are shifted parallel. This means that the zero-lift angle of at¬ tack a as well as the pitching moment at zero lift cm0 = cm (cL = 0) can be varied by means° of such deflections e. Therefore the tail of a bird acts as an elevator to con¬ trol the longitudinal motion. For a given positon of the centre of gravity an equilibrium state can be achieved by a proper adjustment of the deflection angle e. The pitching moment slope dc /dcL remains constant for different deflection angles e. This means that the longitudinal stability is kept for different tail deflections. 732 ACTA XX CONGRESSUS INTERNATIONAL^ ORNITHOLOGICI A = b/c = 5.0 q/c = 0.5 f/Q= 1.0 l/c = 1.0 0.02 l i 0 FIGURE 1 - Aerodynamic characteristics of wing and untwisted wing-tail configurations symmetrical flow (Tail shape B) ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 733 The wing alone as well as the wing with an untwisted (v = 0°) and with a twisted (v = 30°) tail of shape B have also been tested in unsymmetrical free stream flow. The results are shown in Figure 2. In the diagrams rolling moment coefficient c, = 4R / pV2Sb (R rolling moment, b span), yawing moment coefficient cn = 4N/pV2Sb (N yaw¬ ing moment) and sideforce coefficient c = 2Y/pV2S (Y sideforce) are plotted against the angle of sideslip B for a constant angle of attack a = 9.5°. The results for the wing alone are described first. In symmetrical free stream flow (B = 0°) all coefficients are zero. For a positive angle of sideslip B > 0° the lift on the windward half is larger than that on the leeward half of the wing. This leads to a nega¬ tive rolling moment and the rolling moment derivative is dc/dB < 0. Since the local drag is correlated with the local lift the drag on the windward half is also larger than that on the leeward half of the wing. A positive yawing moment with a derivative dcn/ dB > 0 results which is called stable because it has the tendency to reduce the an¬ gle of sideslip. The wing alone shows small directional stability. Concerning the sideforce the present windtunnel model showed a very small positive sideforce with a derivative dcy/dB > 0 which is due to the local shape of the wing tips. Adding an untwisted (v = 0°) tail of shape B to the wing leads to the same lateral aerodynamic characteristics as for the wing alone. Therefore an untwisted tail has virtually no effect at all on stability and control of the lateral motion. If the tail is twisted (v = 30°) even in symmetrical free stream flow B = 0° a negative rolling moment turns out which has already been described in the literature (Oehme 1976a, Nachtigall 1985). In addition to this, however, a positive sideforce occurs which acts at the tail and therefore also a corresponding negative yawing moment is found. This means that twisting the tail is a measure to produce rolling moments, yawing moments and sideforces simultaneously. The magnitude of these forces and moments depends linearly on the angle of twist of the tail. If the direction of the twist would be altered (e. g. to v = - 30°) the sign of all three lateral derivatives c,, cn and c at zero angle of sideslip B = 0° would change as well. If the twisted tail (v = 30°) is deflected up (e = - 5°) and down (e = + 5°) in symmetrical free stream flow B = 0° the lateral derivatives c,, cn and cy vary considerably. Due to the deflection e the load¬ ing of the tail is changed primarily: for increasing deflection angle e the forces acting on the tail increase. Since the tail is twisted the positive sideforce as well as the cor¬ responding negative yawing moment increase also. The slight variation of the rolling moment due to changes of the deflection angle e is caused by the sideforce acting below the moments reference point. The results of the present investigations for sym¬ metrical free stream flow B = 0° show that a tail twisted in both directions and de¬ flected up and down acts as a device to control the lateral motion. Other effects of twisting the tail may also be taken from Figure 2 if the variations of the lateral derivatives with the angle of sideslip are taken into account. Concerning the rolling moment the c, (B) curves for all twisted tails (e = - 5°, 0°, + 5°) are shifted parallel as compared with the untwisted tail or the wing alone. The stability derivative dc/dB is not changed by twist v and deflection e. For yawing moment and sideforce however the situation is different. For a twisted tail (v = 30°) the sideforce decreases linearly with increasing angle of sideslip. The stability derivative is dcy/dB < 0 and its value is independent of the deflection angle e. This means that the slightly unstable situation for the wing alone and the untwisted tail has changed towards a stable 734 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI -8° 0° 8° 16° p FIGURE 2 - Aerodynamic characteristics of wing, untwisted and twisted wing-tail configu¬ rations in unsymmetrical flow (Tail shape B) ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 735 behaviour of the configuration with a twisted tail. Similar effects can be seen from the yawing moment results cn(B). The reduction of the sideforce leads to additional posi¬ tive yawing moments with increasing angles of sideslip. This means that the stabil¬ ity derivative dcn/dB is increased in comparison with the wing alone and the untwisted tail. In the present case the directional stability is considerably improved by the twisted tail and the amount of stability is the same for all deflection angles e. If the direction of the twist is changed (e. g. to v = - 30°) there is no effect on the directional stabil¬ ity. This is due to the fact that only the variation of the aerodynamic coefficients with the angle of sideslip has to be taken into account. At an angle of sideslip B > 0 a crossflow in -y direction takes place which causes an additional positive lift on the tail for v < 0 and additional negative lift on the tail for v > 0. The reduction of the sideforce with increasing angle of sideslip is caused by the “cross flow drag force” which acts at the tail in -y direction in both cases v > 0 and v < 0. This means that twisting the tail leads to increased lateral stability and this improvement is independent of the di¬ rection of twist. DISCUSSION The present investigations have shown that longitudinal stability is achieved and lat¬ eral stability is considerably improved by adding a twisted tail to a wing. These ben¬ efits occur simply by the presence of the tail with a certain planform shape and an angle of twist for any deflection angle and for any direction of twist. The improvements in longitudinal stability depend on size and planform shape of the tail, whereas those in lateral stability are additionally governed by the amount of twist Ivl. The longitudinal and the lateral stability discussed so far are called “static” stabilities. They relate the final flow status of the configuration to the original one in the case of a disturbance. Between these two states a time-dependent dynamic process takes place which is governed by the aerodynamic derivatives as well as by the inertia moments of the configuration. Aperiodic as well as oscillatory motions do occur and they can be both stable or unstable. This kind of dynamic stability is also very impor¬ tant, but it has not yet been investigated in birds. Deflections of the tail up and down as well as twisting the tail in both directions can be used to control the longitudinal and the lateral motion simultaneously. The main effects of deflections e up and down are changes of the pitching moment to control the longitudinal motion. Sideforces and yawing moments can be achieved by twisting the tail. It is important, however, that the tail carries some loading in order to be ef¬ fective. For a positive tail loading positive twisting leads to a positive sideforce and to a negative yawing moment. An unloaded twisted tail produces no sideforce and no yawing moment. For a negative tail loading which is achieved for very large negative deflection angles e negative twisting leads again to a positive sideforce and to a nega¬ tive yawing moment. Both ways to produce a positive sideforce at the tail lead to dif¬ ferent pitching moment characteristics which influence again the longitudinal control. In birds movements of the wings forwards and rearwards can also be used to achieve longitudinal control. Compared with airplanes stability and control of the longitudinal motion in birds is achieved in the same way by means of a horizontal tail behind the wing which can be deflected up and down. For the lateral motion however the means are different. In airplanes a vertical fin with a rudder is used to achieve lateral stability and control. In 736 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI birds the roie of the fin with rudder is adopted by a tail which is twisted in the proper direction according to the tail loading. Stability and control are strongly dependent on the size and the shape of the tails. The present investigations contain already a large amount of material and results re ate to the longitudinal motion which will be published elsewhere. Concerning the atera motion the windtunnel experiments will be continued to include twisted tails with i ferent planform shapes. CONCLUSIONS Windtunnel experiments have been carried out on a rectangular wing with tails in sym¬ metrical and unsymmetrical free stream flow. Plane and twisted tails with deflections up and down have been investigated. The following results have been found. 1) The presence of a tail in birds increases the stability of the longitudinal motion as in airplanes. 2) The tail in birds acts as an elevator as in airplanes. Its effectiveness is governed by size and planform shape, but details are not discussed in this paper. 3) Twisting a tail in any direction increases the directional stability of the lateral mo¬ tion. This measure is used by birds instead of the vertical fin of airplanes. 4) Twisting a loaded tail leads to a sideforce and to a yawing moment. Their direc¬ tions depend on the combination of loading and twist. This measure is used by birds instead of the rudder on the vertical fin of airplanes. LITERATURE CITED BURTON, R. 1990. Bird flight. New York, Oxford, Sydney. NACHTIGALL, W. 1985. Warum die Vogel fliegen. Hamburg. OEHME, H. 1976a. Die Flugsteuerung des Vogels. I. Uber flugmechanische Grundlagen. Beitr. Vogelkd. 22: 58 - 66. OEHME, H. 1976b. Die Flugsteuerung des Vogels. II. Kurzer Uberblick uber die Entwicklung der Flugsteuerungstheorien. Beitr. Vogelkd. 22: 67 - 72. OEHME. H. 1976c. Die Flugsteuerung des Vogels. III. Flugmanover der Kornweihe (Circus cyaneus). Beitr. Vogelkd. 22: 73 - 82. RUPPELL, G. 1975. Vogelflug. Munchen. STRESEMANN, E. 1934. Aves. Handbuch der Zoologie, Vol. 7, Berlin und Leipzig. ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 737 TIME PATTERN OF WING MOTION DURING CRUISING FLIGHT AND ITS IMPORTANCE FOR FLIGHT ENERGETICS H. OEHME Forschungsstelle fur Wirbeltierforschung, Alfred-Kowalke-StraRe 17, 0-1 136 Berlin, Germany ABSTRACT. Alternation between powered and non-powered flight is frequent among birds. Two main forms occur characterized by the course of the non-powered part: the true gliding and the ballistic bounding flight. Two topics are discussed on the basis of model calculations: the overall energy sav¬ ing and its role with respect to strain of the muscular apparatus. Intermittent flight reduces the cost of transport in comparison with sustained flapping flight of equal horizontal velocity. This saving is pur¬ chased by increasing power expenditure during the flapping periods. Within species-specific speed ranges birds may fly with the same cost of transport by variation of time pattern of intermittent flight. Keywords: Flight energetics, energy saving, intermittent flight. INTRODUCTION The optimal utilization of flight ability is determined by two tendencies: minimizing of the energetic cost of transport and prevention of overstressing of the flight muscles. These tendencies are partially directed against one another and the exercised modes of flying are certainly a compromise between them. Several birds cover large dis¬ tances by sustained flapping flight, e.g. loons, grebes, ducks, geese, waders, and auks. On the other hand we find the dominance of non-powered flight as genuine soaring in larger birds, e.g. albatrosses, petrels, storks, eagles, and vultures. How¬ ever, a considerable number of species show a more or less regular alternation of powered and non-powered legs during cruising. We find this phenomenon in larger and smaller birds of different taxonomic position. Questions of the energy regime of this alternating or intermittent flight are the subject of this paper. The investigations are more abstract for the present but they try to take into consideration morphologi¬ cal and kinematical characteristics of real species. PRINCIPLES OF MODEL CALCULATIONS Flight styles Four flight styles are distinguished and compared (see Figurel). Style 1 (not figured) is the non-accelerated, horizontal, sustained flapping flight which presents itself as the theoretical reference of the styles of intermittent flight. Style 2 is a flight without al¬ teration of height but with changing air-speed. Compared with Style 1 the peculiarity is, at the same average velocity, the increase of thrust during the powered flight pe¬ riod which is simultaneously the period of acceleration. Powered part (TpJ and gliding part (T ) compose the macrocycle (T mcyc). The gliding period is the part of decelera¬ tion and requires continuous increase of the lift coefficient in the wings with decreas¬ ing air-speed. Style 3 (undulating flight) is a flight with alteration of height but with constant air-speed during both periods of the macrocycle. Since the powered part is climbing flight thrust is also increased compared with that of Style 1 . Style 4 (bounding 738 ACTA XX CONGRESSUS INTERNATIONAL^ ORNITHOLOGICI flight) is similar to Style 3 but besides the change of height there is also a c ange o air-speed during the glide period which represents a ballistic curve where the wings are folded and drawn close to the body. A small lift force is generated by the trun in the acceleration phase of the glide period. Forces, power, cost of transport Calculation of power requirements was carried out as described elsewhere (Oehme 1 985a, b, 1986). The mechanical power of the “engine” sensu Musculus pectoralis is computed from kinematical and morphological data (air-speed, angular velocity of the wings, duration of downstroke and beating cycle, wing length, halfspanwise distribu¬ tion of chord length, mass). Mechanical work done during the upstroke is neglected. A further simplification is that a constant lift coefficient (CL) is used over the wing length. Hence the spanwise distribution of circulation is only determined by the kinematical and geometrical data just mentioned. As to Style 2, for the powered pe¬ riod constant acceleration is assumed in order to calculate the necessary thrust force during the beating cycles. The glide period with its decreasing air-speed and increas¬ ing lift coefficient is calculated spotwise for time intervals of 0.002 seconds. The cal¬ culation of the ballistic curve of Style 4 with its changing velocities is performed af¬ ter Vahlen (1942) and Csicsaky (1977) for the same time interval. The mechanical power during the beating cycle (Pcyc) times T (= Tpow/Tmcyc) yields the average power of the macrocycle. That divided by the average horizontal velocity (vhor) makes the FIGURE 1 - Flight styles. T,t time; v air-speed; s horizontal distance; FH,FH instantaneous and mean aerodynamic force perpendicular to flight path; FD,FD instantaneous and mean drag; F h,FTh instantaneous and mean thrust; G weight force; f wing beat frequency; h height; Indices: pow powered period, gl gliding period, cyc beating cycle, mcyc macrocycle. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 739 FIGURE 1 - Continued 740 ACTA XX CONGRESSUS INTERNATIONAL^ ORNITHOLOGICI mechanical cost of transport. To facilitate comparability these quantities are re a e to body mass: specific power of the cycle (Pcyc/m) ar|d specific cost of transpor m), respectively. Bird models, limits, simplifications Four models are used, the masses of which differ to the highest by the factor (see Figure 2). Considerable differences exist with the wing load (mg/(2Aw). The wing ar¬ eas correspond to the outlines given in Figure 2. These are not the natural outlines but they indicate the semispanwise chord length distribution for which the scheme o Oehme & Kitzler (1975) is used. The shaded area of the Models I and II is that which is assumed for lift production curing the upstroke. The natural prototypes are. the female Goshawk Accipiter gentiiis (Model I); the female Common Kestrel Falco tinnunculus (Model II); the male Flouse Sparrow Passer domesticus (Model III), the female Redstart Phoenicurus phoenicurus (Model IV). Masses and principal linear measures are average values based on own measurements and data given in tne lit¬ erature (Glutz von Blotzheim 1 971 ,1 988, Heinzel et al.1 977 , Mebs 1 989, Peterson et al . 1 976 , Pforr & Limbrunner 1980). The frontal area (Ab) is assumed as a circle with the largest body width as the diameter. The coefficients for body drag = parasite drag (C ) and profile drag (Cn ) were obtained in the Models I and II from gliding with a fineness ratio £ = FD/FL =0.1 supposing minimum induced drag (compare I ucker o Parrott 1970, Videler et al.1 988). Three limiting values are introduced to test the admissibility of the calculation of the required aerodynamic forces FH and FTh in the respective styles and velocities. The lift coefficient must not exceed 1.0 which is a more restrictive precaution (compare Biesel et al.1 985, Knappe & Wagner 1985, Nachtigall et al.1 985). The stroke angle must not exceed 120°. Assuming a uniform relative mass of 17% of both Musculi pectorales in the models the mechanical power output per beating cycle is confined to 30 W/kg in I and II with low stroke frequencies and to 42.5 W/kg in III and IV (com¬ pare Weis-Fogh & Alexander 1977). RESULTS AND DISCUSSION Examples are shown in Figures 3 and 4; further information can be found in the ap¬ pendix. The calculations are complex. The duration of the glide period (T ) along with the given velocity determines the duration of the powered period (Tpow) for which, however, only a positive integer of beating cycles is possible. Additionally, in the Styles 3 and 4 there has to be conformity with the climbing angle of the flight path during Tpow. We find that any break of powered flight reduces the cost of transport (CT) and en¬ larges the instantaneous power of the cycle (Pcyc). Energy saving reaches higher amounts in small birds (Style 4) with low velocities than in larger birds which cannot bound. The strength of the reciprocal action between CT and Pcyr is characterized by the quota of time of the powered phase T= Tpow/Tmcyc. There are only small differences between Style 2 and Style 3 so that a combination of both seems possible. Finally there is some evidence that for a lighter bird with lower stroke frequency it is more difficult to go as fast as a heavier one of about the same size and higher frequency. ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 741 cDp=0-03 CDb1~ °'25 CD62~°-3 CLb2~ °*3 FIGURE 2 - Bird models. CDb coefficient of parasite drag; CDp coefficient of profile drag; CLb lift coefficient of body (Style 4); g gravitational acceleration; m mass; Ab frontal area of the body; A area of one wing; numerical indices in Models III and IV: gliding periods 1 and 2; see also Figure 1 . One may hypothesize that fattening in migrants is not only the preparation of the nec¬ essary fuel but also an improvement of flight performance and thereby of the economy of energy consumption provided that the efficiency of the muscles is adapted to the greater weight. 742 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI FIGURE 3 - Flight characteristics of Models I and II. A%h = 100 x hmax/(spow + sgl ), A%v = 1 00 x ( vmai< - vmin) /vmax, T= Tpow/Tmcy0, cp stroke angle; see also Figurel and text. A peculiarity of the results in the Models I and II shall be emphasized. The values of the instantaneous power output (Pcyc/m) as well as those of the cost of transport (CT/m) vary within a considerable range. This is why the magnitude of the resulting aerodynamic force perpendicular to the flight path during the upstroke is not fixed (see Appendix, factor k). The stronger this force the more balanced is the average lifting force of the beating cycle which compensates in the main the weight force. This is achieved when the strength of the bound vortex (circulation) is nearly equal in the wings during downstroke and upstroke. But this means an increase of the force di¬ rected backwards during the upstroke, say enlargement of the average drag, which must be overcome by stronger thrust of the downstroke. The consequence is an in¬ creasing stroke angle and higher power output compared with a beating mode where the lifting force during the upstroke is smaller. In larger birds with low wing beat fre¬ quency the vertical upward force during the upstroke must be logically as high as ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 743 FIGURE 4 - Flight characteristics of Models III and IV. See Figure 3. possible at low velocities. The smaller the bird and the higher, then, as a rule, the frequency the more the generation of the lifting force of the cycle can be concentrated in the downstroke (e.g. Models III and IV with Style 4). Therefore an appropriate stag¬ gering of the circulation between the stroke phases may be advantageous. This may be attached to the average values of the Models I and II. Thus neither the “motor” will be stressed up to its limit of mechanical performance nor will the lift coefficient reach its upper limit. In the Models III and IV, bounding flight (Style 4), one should reckon with deviations from the ascending flight path of the powered phase as a straight line. It may take the form of a stretched lying sigmoid curve without a decisive change of the facts of the case. But never does the powered phase of the macrocycle start in a downwards directed part of the flight path as it is represented by Rayner (1977) and Norberg (1990). This statement is based on cinematographical investigations on the flight of titmice, redstarts, sparrows, and finches. The time ratio Vindicates the degree of energy saving. The magnitude of CT-reduc- tion when compared with the sustained horizontal flapping flight is negatively corre¬ lated with T . It can only decrease at a given air-speed as far as the aerodynamical 744 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI (CL) and the myomechanical (Pcyc) limits are not passed over u function of the sible for optimizing the gross energy economy but also for the subt 0SDhate “flying motor”. Low T means longer breaks for restitution of adenos P P in the myofibrils. On the other hand a minimum Timplies the entire exha physiological performance during the flapping phase of the ^crocy ^ |araer these constraints a compromise will be established which wi ten o T in dependence on the specific constitution of the muscle. Furthermore we set 3 the bird has ranges of velocities and time patterns within which it can real,z® 9 tactics. The postulate possibly fails that energy saving should be only feasibl® '""e velocity of the alternating flight is higher than the theoretical optimum velocity tor minimum cost of transport in sustained flapping flight (Rayner 1977, Norberg 1990). On the contrary the bird may fly with equal cost of transport by variation of speed ana adequate variation of the time pattern. Then, for the real flying behaviour a fixed “maximum range speed” s.str. does not occur. This could correspond to a Siatemen of Wieser (1986) with this tendency whereas the plain relation between mass an mechanical flight power output by the same author yields much too low values. LITERATURE CITED BIESEL W BUTZ H„ NACHTIGALL, W. 1985. Erste Messungen der Flugelgeometrie bei frei gleitfliegenden Haustauben (Columba IMa var.domestica) unter ®enhu,tzu"9wn®p aTmnNA^Mrt Verfahren der Windkanaltechnik und der Stereophotogrammetrie. In Nachtigall.W. (Ed.), BIONA-report 3, p.139-160, Stuttgart-New York, Gustav Fischer Verlag. . ... 07K OQC CSICSAKY, M. 1977. Body-gliding in the Zebra Finch. Fortschritte der Zoologie 24. 275-286. GLUTZ VON BLOTZHEIM, U.N. (Ed.) 1971. Handbuch der Vogel Mitteieuropas, Vol. 4. Frankfurt (Main), Akademische Verlagsgesellschaft. ....... .... . GLUTZ VON BLOTZHEIM, U.N. (Ed.) 1988 Handbuch der Vogel Mitteieuropas, Vol. 11/1. Wiesbaden, HEINZEl'k, FITTER, R., PARSLOW, J. 1977. Pareys Vogelbuch, 2nd edition. Hamburg-Berlin, Paul KNAPPE H^ WAGNER S 1 985. Die aeroelastische Deformation von Vogelflugeln und ihr EinfluB auf die Profilpolaren. In Nachtigall, W. (Ed.), BIONA-report 3, p.25-44, Stuttgart-New York, Gustav Fischer Verlag. MEBS, T. 1989. Greifvogel Europas. Stuttgart, Franckhsche Verlagshandlung. NORBERG U.M. 1990. Vertebrate flight. Berlin, Springer-Verlag. OEHME H 1985a Moglichkeiten und Grenzen der Flugleistungsbestimmung unter Verwendung aerodynamisch begrundeter Rechenmodelle. In Nachtigall, W. (Ed.), BIONA-report 3, p.231-254, Stutt¬ gart New York, Gustav Fischer Verlag. OEHME, H. 1985b. Biophysik und Verhaltensforschung. Aspekte der Untersuchung des Vogelfluges. Milu 6:361 -37 1 . OEHME, H. 1986. Vom Flug des Habichts Acciplter gentilis (L.). Annalen des Naturhistorischen Mu¬ seums Wein, 88/89 B:67-81. OEHME H.. KITZLER, U. 1975. Zur Geometrie des Vogelflugels. Zoologische Jahrbucher, Abteilung fur allgemeine Zoologie und Physiologies der Tiere 79: 402-424. PETERSON, R.. MOUNTFORT, G., HOLLOM, P.A.D. 1976. Die Vogel Europas, 11th edition. Hamburg- Berlin, Paul Parey Verlag. PFORR M., LIMBRUNNER, A. 1980. Ornithologischer Bildatlas der Brutvogel Europas, Vol. 1, Vol. 2. Neudamm, Neumann-Verlag. RAYNER J. 1977- The intermittent flight of birds. In Pedley, T.J. (Ed.), Scale effects in animal loco¬ motion Pp 437-443. London, Academic Press. TUCKER V.A. PARROTT, G.C. 1970. Aerodynamics of gliding flight in a falcon and other birds. Jour¬ nal of experimental Biology 52: 345-367. VAHLEN T 1942. Ballistic Berlin-Leipzig, Walter De Gruyter Verlag. \/mFi ER J J VOSSEBELT, G., GROENEWEGEN, a. 1988. Indoor flight experiements with trained kestrels I II Journal of experimental Biology 134: 173-199. WEIS-FOGH T., ALEXANDER, R. McN. 1977. The sustained power output from striated muscle. In p HI v T J (Ed ) Scale effects in animal locomotion, Pp. 551-525. London, Academic Press. WEISER W 1986 Bioenergetik. Stuttgart-New York, Georg Thieme Verlag. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 745 -o~ 5 ^ o — CO ‘z: cc cd 0) > Q. ^ CD ® ® ® > ^ CD o Q_ CO -- CD o JZ CD -- 13 O — CD co > CO cz o -4— • CD > CD -3> CD -g o — -Q >> 0 o oo o ~D CD C ^ > CD 0 o J£> -Q E Z3 C CD O Q. JD CO CD CD E CO CD JZ o -5 to d) — _Z3 o ■§, CD C X CD S~ c H o § <0 ° Q. X Sh^ : ^ CD © < .i 0 .. ^ > . CJ >4__> CZ 0- CD 2 ^ = ® < g ^ ZZ. 0 ~0 0 p ^ s 0 0^ S 0 CD C >3 ■o $ = O o ^ Edo X ■0 0 ^ 0 = 0 -g -o O E £Z ^ V) « v CL — L 3 w «- x 0 O LL 0 cn 11 O . - 0 _0 JZ ?.E 0 „ CD |5 o o 0s <3 h- o h. -P cr < Q Q Q I Q Q Q Q O O O 1 O O O O 1 2 2 0 JZ c 0 0 E 0 CL CO CM 00 CD 00 00 ~vP CO ■'t CO CM CO CO O" CM CM CM CM CM CM < . . , + + + + + + 00 CO CD CO CD O CD CD Cvi t- t- l o LO LO o 0 JZ c 0 0 E 0 JZ c 0 0 E ^ 00 -I— N CO N O Ol Tf O O LO N 0 0 0 N O N CD N CM t- CO CZ -r- LO N 0 0 CM CD LO 0“ O O O Is- CO CM CD LO 0- o d CO r^ co 00 co LO LO LO ‘R LO LO CO O si d d d CM CM CM CM p> -r_ T— 1 1 T_ T~ > LO CD t — \ CM CD CO v — ) 1 CD CO -t— 1 X 0 <1 0 CD O CO rL O 0 ■ LO LO ~o X LO Is- LO Is- LO c n E d d -1— 1 CM co CM 0 > T_ -r" T— T— T~ ~o > 1 I c 0 H (/) T- CM CM 1 - CM CM 00 0 JZ c 0 0 E c 0 0 c 0 0 CM LO CM CO 00 CM co co LO O d d o 1 d co r- CO CD O O 0- d CD 0 d CD LO CO CO c 0 0 JZ E 0 JZ E 0 JZ E 00 CO CM CO CO Is- T— 06 d CD d CM 1 1 CM 1 CM CM CM 00 0 CO ■y— 00 CM CD CO CO 0 CM CO O CO CO 00 r^- CO CM LO CM CO CO CM T_ "r_ 'r~ T_ CM CD LO CO CD ■M- CM h- 00 00 d d CO d CO CO T— 1 — CO CM CM + + + + + + + + + CO CM ■M" O CM 00 CM 0 1 0 co 00 LO CM 0 00 CD CM CO 0 0 LO CM CO CO ■M- . 03 CO Is- CO CO LO CO co LO CO T - Is- CO co CD CO CM CD LO T— °o CO 10 LO CO 1— Is- CD CD CM LO T— d d cd id d d d LO cd d d 00 CM d CD CM d CD CO CO CM CM CM CM CM CM CM CM CM I I CM CM CM CM CM CM CM CM CM CM CM r^- CD CO r- "sf 1 | O CO Is- LO "M- to O O CM d LO d d d d d 'T cd d 00 d CO 00 00 00 00 1 00 00 LO 00 LO 00 LO CD CO CO o d d CM CD CO N- CO Is- 00 CM 746 ACTA XX CONGRESSUS INTERNATIONAL^ ORNITHOLOGICI — _ — — 1 — Q O MOD.I Q 0 d O d O 1 Q 0 Q 0 h— 1 O 1 -9 0s < a) b) mean 03 -Q mean S S mean S S mean S S' mean 1 1 -a) b) mean s s mear E 00 0 CO CD CD CO \— CD CD O' CO ID 1- CM CM r— 1- T-! 00 CO CD 0 1 1 1 ID CD 1 1 Is- 1 CD CD 1 1 d 1 CD d 1 1 O' 1 1 1 1 1 1 < CM 1 1 0 > 0 CL -0 CD CD g [5 “2 $ O' CM Is- ID CD O CO CD CO LD CD CD O' CD O r- LD CO O CM Is- O CD t- CD O O' CO 0 CD O' O' T- CM 0 CO 1 1 CM 00 O O' CD CM T- CM Is- CD O' CM Is- LD CM .388 O <1 ■t— * T— - T - 'r_ T“ T_ T_ 1 1 C\i T— T— O LD CO CD CO CD LD 00 O; CM O Is- 1 q 1 1 1 CD O' CO CO + + d CO + O CD O' CO + + 00 CO + 00 LD CO O + + •0 + O CM CD CD + + CM Is- + 1 1 1 1 1 'r_ . O' Is- O' + + CM + E 1 1 0 > 1 0 1 Q. Is- CD 00 Is- CO O CD ■»— O CD CM CD CM CD CO O O 00 O' LD O' CM LD O O CM LD Is- CD 00 CD O' CO Is- CM CM CM t- CD CO r- CD 1 1 Is- T- O' Is- CO 0 Is- CM LD CD CO Is- 00 CD CO 0 CO q 00 cm in LD CD CM -r- CM CD CD CM t- CM LD d CM t- CM d CO CM CM CD CM 1 1 ID CD r- CM T- CM 00 CO CM CM CD CM Ci- O CD _ 1 _ CD 00 CM t- O' LD 1 CD LD O' q 06 O' 00 O' O' Is- t- LD O' Is- t- LD 1 - 1- CO •r- CD r- d 1- CO 1 1 Is- CM ^ LD 06 CO 0 Is- CO CM LD CD LD CD CM 0 CM 1 00 CM CO LD t- ID O LD O LD r- O' T-# 1 LD O O d d d d d d d d 1 d O d 0 h- Is- r- CD CD CD Is- - 1 Is- CD CD CD CD LD CD -C -r- d O d d 1 d O' 1 < CO CD 1 CM co 1 O 0 O CO O' 1 O 0 O CO Is- Is- 0 1 1 O 0 O CM ID 1 1 O 0 Z LD 1 1 8 8 00 CD CO 1 LD CO LD 1 LO ID LD O; O' 1 LD ID ; O d d d d 1 CM CM \> 1 > ID CD 1 CM -9 CD O CD cO <1 O O' CD 0 0 1 O CO ID 1 LD > ID Is- ID LD 1 ID Is- « C O d 1— d d 1 CM CM c > 1 1 > ! H C/) CM CM CO CO 1 1 CM d O c 05 03 -O E CD 00 CO 0; CO CD o- cd c\j T- CO cd cm o- -T - T - 00 CM CO CO -r— O' CM + + + 00 CD 00 CM CM CM O CM CD O CO CD CO CM CM in cd CD CM T- D- Is- co d o CM Is- CD O o o LD CM cn CD CO CM ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 747 > > > > 1 1 1 1 1 > > Q Q d Q 0 O I Q o Q Q Q Q Q ' ci Q o o o o o o 1 o O O o o o o 1 O o h- o 2 2 | 2 c c 1 <1 03 CO 1 CD CD 03 JD E 'co jd' E 1 E 1 i— CO CM LD fO O' CM CO CO i — CM O' CD o 1 T- o CD ^ O T— - CD CD oo 1 — d O' d O' O' d 1 + T CO 1 + 1 1 1 CM 1 1 1 i C\J i CM 1 CO 1 1 1 >* o n C\J O CD 00 CD 00 00 CD o I CM CD CM CD ^ _ CM CD 1 <£> CD CD O O' CD O co CD O' CD oo 1 CM CD CO o CD 00 00 o CM h- co un O' CO CO CD CO T- CD | o CD LO LO CO 1 oo CO < ^ ■«- ■«- ■»- d| CM T- T- CM T- T“ 1 — 1 ^ CM y— 1 1 CM CD O' CD y— CD o CO 1 O' CO 1^ CD CD 1 O' O t- CO 1 - O' y— O' O' O | d O' CM CM d I CM 1 - CO T~ 1 - CD CO 1 CM O' 1 — 1 1 CM O' 1 CM O' 1 CM + + + + + + + + + 1 + + + + + 1 + E 1 'o >» o 1 Q_ 1 lo LO O CD LO CD CD CD CO CD | T— D- oo CD CM 00 CO 1 CD 1 — O CD h- CO O O O O' O CO 00 00 O CM 00 r- 1 D- CO un O' O' CD CD CO LO N- CO CM 1 CM T— O' T- T— LO CM 1 o CD CD t- cri CD CD 00 d T— LO LO 1 LO y— d CM d y— d 1 ^ CM CM CM CM CM CM CM CM CM CO | CM CO co CM CM CO CO 1 CO O' 9- 00 CD O' CD CO LO O' CM , 00 h- CM O O O' o 1 r- o d d cri y— d O' d T— d y— CO d d d d 1 ^ d 1— CD o CO LO CD r- CD CD CO CD CO CD o CM CM T“ T_ T-“ 1 1 CO o- CM 1 o O d o o O o ° 1 O o o o O o o 1 ° o h- h- o O co 1^ 1 CD co co o CM 1 o LO 1 — LO CD CO 1 1 — r- CM CD 1 r- h- CD CD O' CM . CD LO CD LO CO 1 CD sz vX> d d T_ d d o 1 T~ d d 1 — d d d 1 - d O' < o o LO CO O' ' _ ^ _ co CM CO CD CD CD O' h- CO 1 T— 00 T— CO o 1 CM CO O' o CM CO d 1 o CO O' o CO O' d 1 ° O' CD CD CD LO co , O' o LO o LO 1 CM CM LO O 00 00 LO 00 LO 00 D" LO CM O' o CM CO CO ' o CM O' o CM O' 00 o CO z 1 CO CD CD O' I N- T — •j— I''- LO CD CD 8 8 T— 'T— 8 1 0 Is- CO 00 CO LO o 00 O' 1 t''- O' O' LO O' O' CM LO O' CO LO O' CO CM 1 LO CO o CM CM o d d d 1 CM CM CM d d d d 1 CM l> -r" T_ 1 'T— 1 T" 1 1 T"~ 1 1 > CD Is- o I oo O' CO LO 00 1 CD •VO CM CD ■»-; 1 O' o O' o CO CD <1 O o o d d CM I o d o o CM 1 o 1 d in LO LO LO LO LO 1 LO LO LO LO LO LO LO 1 1 ^ LO CO c CM CM d d d O I CM CM CM d d d d 1 CM CM > T“ 1 T_— 1 "" 1 >- 1 /3 CO CO T- O' O' O' I O' O' O' O' O' 1 O' 748 ACTA XX CONGRESSUS INTERNATIONAL^ ORNITHOLOGICI CONCLUDING REMARKS: BIRD FLIGHT D. HUMMEL' and G. E. GOSLOW JR2 1 Institute fur Stromungsmechanik, Technische Universitat Braunschweig, Bienroder Weg , Braunschweig, Germany 2 Section for Population Biology, Morphology and Genetics, Brown University, Box Providence, Rhode Island 02912, USA It is clear that through sophisticated modelling and the application of new techno o- gies, a balance of theoretical and empirical approaches are contained within the pa- pers of this symposium. The application of flow visualization techniques allows for a description of the wake vortices created by flying birds which in turn, provides for the recognition of gait patterns (Rayner). Why do certain sized birds adopt one gait or another and how does gait selection relate to the evolution of flapping flight? If more than one gait is used by a species, does this require a “shifting-of-gears from a neuromuscular control standpoint (Goslow)? Nachtigall s studies stimulate us to won der once again if the high altitude migrations of geese relate to the temperature con¬ trol and dissipation of metabolic heat. For years we have wondered why airplanes possess a vertical tail rudder but most birds do not. Hummel’s wind tunnel studies provide us with many answers and more fascinating questions. An issue of long standing interest relates to wing beat frequency and the cost of transport in birds of different locomotor habits (Oehme). Given a theoretical framework from which to work, can we design the appropriate experiments to test the theory? All of these questions are of great interest and significance for the understanding of bird flight. Much work remains to be done. We would like to thank all participants for their contributions. Special thanks to Dr R. Bannash for his presentation of Prof. Dr Oehme’s paper at the Congress. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 749 SYMPOSIUM 10 NEW ASPECTS OF AVIAN MIGRATION SYSTEMS Conveners S. B. TERRILL and P. Z. ANTAS 750 ACTA XX CONGRESSUS INTERNATIONAL^ ORNITHOLOGICI SYMPOSIUM 10 Contents INTRODUCTORY REMARKS: NEW ASPECTS OF AVIAN MIGRATION SYSTEMS SCOTT B. TERRILL . 751 FOOD AVAILABILITY, FAT DEPOSITION, AND MIGRATORY BEHAVIOUR IN SHORT-DISTANCE VERSUS LONG-DISTANCE MIGRANT SYLVIA WARBLERS SCOTT B. TERRILL . 752 ECOPHYSIOLOGICAL AND BEHAVIORAL RESPONSE TO ENERGY DEMAND DURING MIGRATION FRANK R. MOORE . 753 TEMPERATE VERSUS TROPICAL WINTERING IN THE WORLD’S NORTHERNMOST BREEDER, THE KNOT: METABOLIC SCOPE AND RESOURCE LEVELS RESTRICT SUBSPECIFIC OPTIONS THEUNIS PIERSMA, RUDOLF DRENT and POPKO WIERSMA . 761 IS WATER OR ENERGY CRUCIAL FOR TRANS-SAHARA MIGRANTS? H. BIEBACH . 773 PATTERNS OF AVIAN MIGRATION IN LIGHT OF CURRENT GLOBAL GREENHOUSE’ EFFECTS: A CENTRAL EUROPEAN PERSPECTIVE P. BERTHOLD . 780 CLOSING REMARKS: NEW ASPECTS OF AVIAN MIGRATION SYSTEMS FRANK R. MOORE . 787 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 751 INTRODUCTORY REMARKS: NEW ASPECTS OF AVIAN MIGRATION SYSTEMS SCOTT B. TERRILL H.T. Harvey and Associates, 906 Elisabeth Street, P.O. Box 847, Alviso, CA 95002, USA The scientific study of bird migration has diverged into several major lines of research: navigation and orientation, genetics and endogenous rhythms, behavioral ecology, and ecophysiology. All represent areas of current interest and rapid advancement. I think that this symposium will demonstrate not only recent advances in several of these areas, but also how effectively these subdisciplines can be integrated in the advancement of general knowledge of avian migration systems. The papers in this symposium bring into focus ecophysiology, behavioral ecology and population biology of bird migration. Recognition of the importance of these subdisciplines within the broader scope of bird migration is not necessarily new; in fact, over the past fifteen years the case for the importance of ecological and physi¬ ological research in migration has been made repeatedly. What is new are the levels of sophistication of approach to these problems, and the emerging importance of this research in terms of urgent conservation needs. Researchers in North America and Europe are noting rapid, alarming declines in migrant populations. Sound conserva¬ tion plans will require solid, basic research on migrants, not only on the breeding and wintering grounds, but during migration and at stopover sites as well. A fascinating area of migration research involves the interplay between physiology, behavior and environment that results in the overt behaviors exhibited by migrants. Frank Moore will present a paper on stopover biology and behavioral responses to energy demand. Many migrants are faced with meeting drastically changing environ¬ ments along their migratory routes. As Frank will show, some species meet the de¬ mands of a changing, unpredictable environment by becoming plastic in their forag¬ ing behavior and habitat selection during migration. These behavioral adjustments are the result of constant fine-tuning between internal physiological condition and envi¬ ronmental stimuli. Research concerning how migrants cope with unpredictable and changing environments is not only relevant to basic theory in behavioral ecology, but also has broad implications for conservation. A very pressing question that is highly relevant to all aspects of migration research is: what are the costs and benefits of distance migration? Piersma and his colleagues have selected the Red Knot to address this question. It would be difficult to choose a more ideal species for this endeavor. Red Knot populations breed at similar lati¬ tudes, but migrate vastly different distances to population-specific wintering areas. They successfully examine this question using an energetics approach. The strategies migrants use to cross major ecological barriers is another topic of cur¬ rent interest. Flow birds cope with these barriers has been largely the realm of theory. 752 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI Biebach and others have recently begun to examine empirically strategies for crossing the Sahara Desert. He finds that a fascinating mix of physiological, behavioral and meteorological variables interact to determine the success, or failure, of a migrant traversing the Sahara. Peter Berthold’s paper is, as far as I am aware, an entirely novel attempt to address the large-scale effects of global warming on migrant populations. Berthold provides specific, testable predictions concerning the effects of global warming on Palearctic migrants. His predictions do not bode well for temperate breeding migrant populations in general. The effects of global warming, use of pesticides and other environmental toxins, and massive habitat destruction may mean the extinction of some migrant populations within the next few decades. The importance of this research cannot be overstated, and Peter’s population-level approach is a good starting point. FOOD AVAILABILITY, FAT DEPOSITION, AND MIGRATORY BEHAVIOUR IN SHORT-DISTANCE VERSUS LONG-DISTANCE MIGRANT SYLVIA WARBLERS SCOTT B. TERRILL Dept, of Biology, Siena College, Loudonville, NY 12211, USA ABSTRACT. It is well established that the migratory behaviour of many species of Sylvia warblers is strongly determined by endogenous factors. These factors determine, to a large extent, the timing distance and direction of migration. Recent experiments, however, indicate that there is considerable plasticity in the system. Of particular interest are the relative roles of endogenous and exogenous fac¬ tors in producing overt migratory behaviour in a variable environment. This paper reports on experi¬ ments performed on Blackcaps Sylvia atricapilla, a short-distance migrant, and Garden Warblers Sylvia borin, a long-distance migrant. These experiments were designed to test the effects of various levels of food deprivation on fat deposition and migratory behaviour during the autumn and winter months The results indicate substantial differences, as well as close similarities, in ecophysiological and be¬ havioural strategies in these two closely related migrants. The results are discussed with respect to the different environments these warblers encounter during the autumn migration and winter months ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 753 ECOPHYSIOLOGICAL AND BEHAVIORAL RESPONSE TO ENERGY DEMAND DURING MIGRATION FRANK R. MOORE Department of Biological Sciences, University of Southern Mississippi, Hattiesburg, Mississippi 39406-5018, USA ABSTRACT. Acquiring enough food to meet energetic requirements is an important constraint during migration. Yet, passage migrants must cope with increased energy demand coupled with environmental circumstances that decrease the certainty that demand will be met. Migrants should experience selec¬ tive pressure to diversify their foraging behavior (behavioral plasticity), thereby increasing the likelihood that energetic requirements will be satisfied and en route contingencies met. Birds that replenish de¬ pleted reserves rapidly improve their chances of a successful migration. When the rate of gain (g/d) among Wood Thrushes and Veerys was examined in relation to arrival mass following a flight across the Gulf of Mexico, fat-depleted birds gained mass more rapidly than birds carrying unmobilized re¬ serves. Red-eyed Vireos, too, apparently compensated for increased energy demand and were more likely to gain mass than fatter birds. Foraging observations of Red-eyed Vireos were consistent with the hypothesis that lean individuals adjusted their foraging behavior to mediate compensatory weight increase. As energy demand is satisfied, foraging “decisions” can be expected to change. Keywords: Migration, energetics, Gulf of Mexico, foraging behavior, plasticity. INTRODUCTION Behavioral plasticity among long-distance migrants should come as little surprise given the different vegetation structures, wide variations in resource quality and quan¬ tity, and changes in competitive pressures experienced during their annual cycle (see Keast & Morton 1980, Hutto 1985). Morse (1971, 1980, see also Rabenold 1980, Greenberg 1990, Loria & Moore 1990, Martin & Karr 1990) suggested that the con¬ tingencies which arise during migration place a premium upon plasticity, which is defined here as the ability of an organism to alter its behavior in response to changes in environmental conditions. While en route, a passage migrant must forage in unfa¬ miliar habitats to replenish depleted energy stores, resolve conflicting demands of predator avoidance and food acquisition, compete with other migrants and resident birds for limiting resources, respond to unpredictable and sometimes unfavorable weather, and correct for orientation errors. Moreover, favorable en route habitat, where the bird can safely and rapidly accumulate energy reserves, is probably limited, or effectively so because it may not always have the opportunity to select the best habitats (Hutto 1985, Moore & Simons in press). These “problems” are magnified because the certainty with which contingencies will be met decreases while en route (see Alerstam 1978, Buskirk 1980, Sandberg et al. 1988, Moore & Kerlinger 1991). In this paper, I consider the adaptive value of foraging plasticity in the context of stopover biology and suggest that migrants diversify their foraging behavior (sensu Real 1980) under conditions of high energy demand, thereby increasing the likelihood they will satisfy energetic requirements and meet en route contingencies. How well migratory birds respond to the energy demands of migration affects their survival and 754 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI reproductive success. Although it is difficult to measure directly the effect of en route foraging events on survival or reproductive success, rate of fat deposition represents a suitable fitness surrogate to evaluate the functional consequences of foraging behavior during migration. As energy reserves are mobilized during migration, birds that replenish reserves rapidly improve their chances of a successful migration (Alerstam & Lindstrom 1990). When birds arrive along the northern coast of the Gulf of Mexico, for example, some individuals have mobilized their lipid reserves, are essentially fat-free, and run the risk of a negative energy balance, while others have retained sufficient fat to continue migrating the day of their arrival (e.g., Rappole & Warner 1976, Moore & Kerlinger 1987). Even if a lean migrant achieves a positive energy balance, field studies reveal that lean birds often stay longer during stopover than fatter birds, presumably to re¬ plenish depleted energy reserves (Bairlein 1985, Biebach et al. 1986, Moore & Kerlinger 1987, Safriel & Lavee 1988). One consequence of a longer stopover would be delayed arrival on the breeding grounds and lost breeding opportunities (cf. von Haartman 1968, Francis & Cooke 1986, Lavee & Safriel 1989). It is reasonable then to expect the evolution of compensatory mechanisms in migrant populations to meet nutritional demands and to prevent delays in the migratory schedule. If energetically constrained migrants increased their rate of energy acquisition, a favorable energy budget is achieved more quickly, length of stopover decreases, and the speed of migration increases (see Alerstam & Lindstrom 1990). Is there evidence that neotropical landbird migrants make compensatory adjustments in relation to the en¬ ergy demand experienced during migration? If so, are the adjustments mediated by diversification of foraging behavior? METHODS i he data reported here on the biology of Wood Thrushes Hylocichla mustelina , Veerys Catharus fuscescens, and Red-eyed Vireos Vireo olivaceus were collected when birds stopped following spring trans-Gulf migration at a coastal woodland located in Cameron Parish, Louisiana, USA (see Moore & Kerlinger 1987, Loria & Moore 1990). The woodlands and wooded islands along the northern coast of the Gulf of Mexico are important stopover sites for neotropical landbird migrants (Moore et al. 1990, Moore & Simons in press). These habitats provide spring migrants a place to rest and? replen¬ ish reserves following a nonstop, trans-Gulf flight (18-24 h) of at least 1000 km. Mist-nets (12 x 2.6 m with 30 mm mesh) were used to capture migrants during stopo¬ ver in 1986 (10-21 April), 1987 (1-28 April), and 1988 (26 March-14 May). Birds were weighed to the nearest 0.25 g, fat classed (see Helms & Drury 1960), banded with a USFWS aluminum leg band, and their wing chord (unflattened) measured. Recaptured birds were assigned to a fat class without reference to previous records and bodv mass remeasured. I assume that size-adjusted (mass - chord) changes in mass be¬ tween initial capture and recapture represent changes in stored body fat Duration of stopover and rate of mass change were estimated by analyzing recapture data. Stopover length was conservatively calculated by subtracting the first capture date from the last capture date (Moore & Kerlinger 1987). I assume that the absence of recapture is indicative of departure from the stopover site (e.g., Bairlein 1985 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 755 Biebach, et al. 1986, Rappole & Warner, 1976, Safriel & Lavee 1988). The daily rate of mass change was computed by dividing the change in mass between first capture and last capture by the duration of stopover. I standardized the daily rate of mass change by computing the percent change/h for individuals recaptured on the same day (Moore & Kerlinger 1987, Loria & Moore 1990). The %/h rates were used to cor¬ rect masses to 12:00 h for all birds that were recaptured and to control for diurnal variation in mass due to food consumption. The foraging behavior of Red-eyed Vireos in relation to energy demand was quanti¬ fied by taking advantage of naturally occurring variability in the arrival condition of birds following trans-Gulf migration in spring (see Moore & Kerlinger 1987, Loria & Moore 1990). Foraging observations were made independent of mass and energy reserve estimates, i.e., an energetic condition (“lean” vs. “fat”) was assigned to a for¬ aging bird based upon the average condition of birds captured on a particular day after foraging data were gathered (see Loria & Moore 1990). Fat-free mass in this species is reported to be 14.6 g for females and 15.1 g for males according to Connell et al. (1960), and birds classified as lean (fat score = 0) following trans-Gulf flight averaged 14.3 g, regardless of sex. The difference between the average adjusted masses under these two situations translates to approximately 450 km flight-distance under no-wind conditions (Pennycuick 1969). % OCCURRENCE 60 zero ONE TWO THREE FAT SCORE VEER (N*167) Hi WOTH (N=205) FIGURE 1 - Distribution of fat scores for Wood Thrushes (WOTH) and Veerys (VEER) captured along the northern coast of the Gulf of Mexico in spring 1988. Fat scoring after Helms & Drury (1960). See text for relationship between fat score and body mass. 756 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 30 32 34 36 38 40 42 44 46 48 50 ARRIVAL MASS (gm) F1QIJRE 2 - Relationship between rate of mass change and arrival mass of Wood Thrushes (top) and Veerys (bottom). Line is simple linear regression fitted bv least squares analysis. The regression line has a significantly negative slope (P < o 051 'for both species. v ' ' ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 757 RESULTS & DISCUSSION Arrival condition and mass change among thrushes Migrants would be expected to mobilize fat reserves when crossing the Gulf of Mexico in spring, and half of all Veerys and Wood Thrushes captured in 1988 were essen¬ tially fat-free (Figure 1). Yet, over one-third of the birds carried reserves greater than or equal to the amount for fat class 2, which is enough “fuel” to continue migrating well beyond the northern Gulf coast. The difference in body mass of Wood Thrushes and Veerys given fat class 0 and those given fat class 2 is 4.1 g and 3.5 g, respectively. This difference translates approximately to flight distances under still-air conditions of 430 km and 570 km, respectively (Pennycuick 1969), and could represent a substan¬ tial “savings” in time of migration (see Alerstam & Lindstrom 1990). If lean, fat-depleted thrushes compensate for heightened energy demand, they should gain mass more rapidly during stopover than birds carrying unmobilized reserves. When rate of gain (g/d) is regressed against arrival mass (Figure 2), the slope of the best fit line is negative (P < 0.05) for both species. Lean Wood Thrushes and Veerys gain mass more rapidly than fatter contemporaries, though only a fraction of the to¬ tal variation in rate of gain is explained by arrival mass (r2 = 6% and 19%, respec¬ tively). Low coefficients of determination are not surprising given the different vari¬ ables that probably influence rates of mass change during stopover (Rappole & Warner 1976, Mehlum 1983, Bairlein 1985, Biebach et al. 1986, Moore & Kerlinger 1987, Safriei & Lavee 1988). Changes in Foraging Behavior of Red-eyed Vireos When Red-eyed Vireos stopover at Peveto Woods following trans-Gulf migration, some individuals gain mass at a rapid rate, others lose mass, and many birds main¬ tain their arrival mass (Loria & Moore 1990). Some of the variation in rate of mass change is explained by the birds’ energetic condition. For example, fatter birds (fat score > 1) lost more mass on a daily basis than did lean birds (fat score = 0) in 1988. Like Wood Thrushes and Veerys, Red-eyed Vireos apparently compensated for in¬ creased energy demand and were more likely to gain mass than fatter birds (Loria & Moore 1990). Loria and Moore (1990) hypothesized that lean Red-eyed Vireos adjusted their forag¬ ing behavior to compensate for increased energy demand. They found that fat-de¬ pleted birds (1) moved at a higher mean velocity, (2) increased their degree of turn¬ ing following a feeding attempt, (3) broadened their use of microhabitat space, and (4) expanded their feeding repertoire (Figure 3). The behavior of fat-depleted Red¬ eyed Vireos following trans-Gulf migration, notably their increased use of space and expanded repertoire, is indicative of risk taking in relation to increased energy require¬ ments (see Real & Caraco 1986, Moore & Simm 1986), i.e., the birds were respond¬ ing to an anticipated decrease in expected gain. As a consequence of behavioral ad¬ justments, a lean bird increased the likelihood it would gain mass. As energy demands are satisfied and the deficit reduced, foraging “decisions” can be expected to change (see Sibley & McFarland 1976, Houston & McNamara 1982). Uncertainty and Plasticity Among Passage Migrants Two conditions lead to plasticity: (1) Predictable periods of trophic constraint and (2) unpredictable but probable periods of trophic constraint. Both criteria are fulfilled dur- 758 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI ing migration: passage migrants can anticipate increased energy demand, and they experience increased uncertainty that energy demands will be met (e.g . , Rappole & Warner 1976, Graber & Graber 1983, Loria & Moore 1990, Martin & Karr 1990, Moore & Simons in press). The expectation that foraging plasticity is advantageous in rela¬ tion to energetic constraints may be derived from Real’s (1980) consideration of the diversification of behavior in relation to environmental uncertainty. According to Re¬ al’s model, every behavior will have associated with it some form of probability dis¬ tribution around its expected value and resulting fitness. The variance of this distri¬ bution is essentially the uncertainty that any given behavior will actually result in the expected fitness. While en route between breeding and wintering areas, migrants are required to make “decisions” when outcomes are difficult to ascertain. Assuming that increased expected fitness is desirable and increased uncertainty about that fitness is undesirable, an uncertain strategy will be adopted only if the activity is compen¬ sated by an increased expected fitness. Real (1980) shows that under most conditions strategies composed of a variety of behaviors prove to be most advantageous (i.e., organisms try to minimize the uncertainty associated with outcomes), and hypoth¬ esizes that when fitness is uncertain an organism’s best strategy consists of a diver¬ sity of behaviors whose covariances are negative. Expanding feeding repertoire to satisfy energy demand would be most advantageous when the “rewards” associated with different maneuvers negatively covary. MANEUVERS SUBSTRATES FIGURE 3 - Occurrence (%) of foraging maneuvers (left) and substrates (right) in relation to the energetic condition of Red-eyed Vireos following spring trans-Gulf migration. Maneuvers: N = 117 (lean) and 94 (fat). Substrates: N = 89 (lean) and 117 (fat). Drawinq from Loria & Moore (1990). Altering foraging behavior to meet en route contingencies is not without costs, how¬ ever. Certain maneuvers are energetically more expensive to perform than others (e.g., Norberg 1977), may expose the forager to increased risk of predation, or may be difficult to perform because of morphological constraints (e.g., Moermond 1990) Yet, when a bird’s energy deficit is large, altering foraging behavior to reduce the ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 759 deficit becomes worthwhile because a change in the deficit substantially reduces the cost of changing behavior. As energy demand is met and the deficit reduced, how¬ ever, the cost of changing behavior is less offset (Sibley & McFarland 1976). The en route ecology of migratory birds is profitably viewed as a dynamic interaction between possible behavioral actions and a changing internal state. ACKNOWLEDGEMENTS I am especially grateful to Wang Yong and D. Loria, whose research on migrating thrushes and Red-eyed Vireos, respectively, contributed to the development of this paper. I thank the Baton Rouge Audubon Society for permission to study Neotropical bird migrants at the Holleyman Migratory Bird Sanctuary, near Cameron, Louisiana, USA. LITERATURE CITED ALERSTAM, T. 1978. Reoriented bird migration in coastal areas: dispersal to suitable resting grounds? Oikos 30: 405-408. ALERSTAM, T., LINDSTROM, A. 1990. Optimal bird migration: the relative importance of time, energy, and safety. In: Bird Migration: The physiology and ecophysiology. Gwinner, E. (Ed.). Springer, Berlin. Pp. 331-351. BAIRLEIN, F. 1985. Body weights and fat deposition of Palaearctic passerine migrants in the central Sahara. Oecologia 66: 141-146. BIEBACH, H., FRIEDRICH, W., HEINE, G. 1986. Interaction of body mass, fat, foraging and stopover period in trans-Sahara migrating passerine birds. Oecologia 69: 370-379. BUSKIRK, W. H. 1980. Influence of meteorological patterns and trans-Gulf migration on the calendars of latitudinal migrants. In: Migrant Birds in the Neotropics. Keast, A., Morton, E. S. (Eds). Smithsonian Press, Washington, D.C. Pp. 485-491. CONNELL, C., ODUM, E., KALE, H. 1960. Fat-free weights of birds. Auk 77: 1-9. FRANCIS, C. M., COOKE, F. 1986. Differential timing of spring migration in wood warblers (Parulinae). Auk 103: 548-556. GRABER, J. W., GRABER, R. R. 1983. Feeding rates of warblers in spring. Condor 85: 139-150. GREENBERG, R. 1990. Ecological plasticity, neophobia, and resource use in birds. Studies in Avian Biology No. 13: 431-437. HELMS, C. W., DRURY, W. H., JR. 1960. Winter and migratory weight and fat field studies on some North American buntings. Bird-Banding 31: 1-40. HOUSTON, A., McNAMARA, J. 1982. A sequential approach to risk-taking. Animal Behaviour 30: 1260- 1261. HUTTO, R. L. 1985. Habitat selection by nonbreeding, migratory land birds. In: Habitat selection in birds. Cody, M. L. (Ed.). Academic Press, NY. Pp. 455-476. KEAST, A., MORTON, E. S. (Eds). 1980. Migrant Birds in the Neotropics. Smithsonian Inst. Press, Washington, D.C. LAVEE, D., SAFRIEL, U. N.. 1989. The dilemma of cross-desert migrants - stopover or skip a small oasis? Journal of Arid Environment 17: 69-81. LORIA, D. L.. MOORE, F. R. 1990. Energy demands of migration on Red-eyed Vireos, Vireo olivaceus. Behavioral Ecology 1: 24-35. MANGEL, M., CLARK, C.W. 1988. Dynamic modeling in behavioral ecology. Princeton University Press, Princeton, New Jersey. MARTIN, T. E., KARR, J. R. 1990. Behavioral plasticity of foraging maneuvers of migratory warblers: Multiple selection periods for niches? Studies in Avian Biology No. 13: 353-359. MEHLUM, F. 1983. Weight changes in migrating Robins (Erithacus rubecula) during stop-over at the island of Store Faerder, Outer Oslofjord, Norway. Fauna norv. Ser. C, Cinclus 6: 57-61. MOERMOND. T. C. 1990. A functional approach to foraging: Morphology, behavior, and the capacity to exploit. Studies in Avian Biology No. 13: 427-430. MOORE, F. R., SIMM, P. 1986. Risk-sensitive foraging by a migratory warbler (Dendroica coronata). Experientia 42: 1054-1056. 760 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI MOORE, F. R., KERLINGER, P. 1987. Stopover and fat deposition by North American wood-warblers (Parulinae) following spring migration over the Gulf of Mexico. Oecologia 74: 47-54. MOORE, F. R., KERLINGER, P., SIMONS, T. R. 1990. Stopover on a Gulf coast barrier island by spring trans-Gulf migrants. Wilson Bulletin 102: 487-500. MOORE, F. R., SIMONS, T. R. In Press. Habitat suitability and the stopover ecology of Neotropical landbird migrants. In Ecology and conservation of neotropical landbird migrants. Hagan, J. M., Johnston, D. W. (Eds). Smithsonian Institution Press, Washington, D.C. MOORE, F. R., KERLINGER, P. 1991. Nocturnality, long-distance migration, and ecological barriers. Acta XX International Ornithological Congress. MORSE, D. H. 1971. The insectivorous bird as an adaptive strategy. Annual Review of Ecology & Systematics 2: 177-200. MORSE, D. H.. 1980. Behavioral mechanisms in ecology. Harvard University Press, Cambridge, MA. NORBERG, R. A. 1977. An ecological theory of foraging time and energetics and choice of optimal food-searching method. Journal of Animal Ecology 46: 511-529. PENNYCUICK, C. J. 1 969. The mechanics of bird migration. Ibis 111: 525-556. RABENOLD, K. 1980. The Black-throated Green Warbler in Panama: geographic and seasonal com¬ parison of foraging. In: Migrant Birds in the Neotropics. Keast, A., Morton, E. S. (Eds). Smithsonian Institution Press, Washington, D.C., Pp. 297-307. RAPPOLE, J. H., WARNER, D. W. 1976. Relationships between behavior, physiology and weather in avian transients at a migration stopover site. Oecologia 26: 193-212. REAL, L. A. 1980. Fitness, uncertainty, and the role of diversification in evolution and behavior. Ameri¬ can Naturalist 115: 623-638. REAL, L., CARACO, T. 1986. Risk and foraging in stochastic environments. Annual review of ecology and systematics 17: 371-390. ROBINSON, S. K. 1986. Three-speed foraging during the breeding cycle of Yellow-rumped Caciques (Icterinae: Cacicus cela). Ecology 67: 394-405. SAFRIEL, U. N., LAVEE, D. 1988. Weight changes of cross-desert migrants at an oasis - do energetic considerations alone determine the length of stopover? Oecologia 67: 61 1-619. SANDBERG, R., PETTERSSON, J,, ALERSTAM, T. 1988. Why do migrating Robins, Erithacus rubecula, captured at two nearby stop-over sites orient differently? Animal Behavior 36: 865-876. SIBLEY, R. M., McFARLAND, D. j. 1976. On the fitness of behavioral sequences. American Natural¬ ist 110: 601-617. VON HAARTMAN, L. 1968. The evolution of resident versus migratory habit in birds: some considera¬ tions. Ornis Fennica 45: 1-7. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 761 TEMPERATE VERSUS TROPICAL WINTERING IN THE WORLD’S NORTHERNMOST BREEDER, THE KNOT: METABOLIC SCOPE AND RESOURCE LEVELS RESTRICT SUBSPECIFIC OPTIONS THEUNIS PIERSMA12, RUDOLF DRENT1 and POPKO WIERSMA1 Zoological Laboratory, University of Groningen, P.O.Box 14, 9750 AA Haren, The Netherlands Netherlands Institute for Sea Research, P.O.Box 59, 1790 AB Den Burg, Texel, The Netherlands ABSTRACT. The breeding of Knots Calidris canutus is confined to the tundra areas around the Arc¬ tic Ocean. The four recognized subspecies winter at latitudes varying from north temperate Europe and N. America to tropical W. Africa, southernmost S. America and New Zealand. Their arctic breeding leads to high thermostatic requirements and hence to a high level of energy expenditure during the summer season. In the non-breeding season, thermostatic costs remain at the same high summer level in the north temperate wintering subspecies, but are halved in the tropical winterers. The considerable dif¬ ference in overall energy expenditure in winter appears reflected in subspecific Basal Metabolic Rates measured in midwinter, suggesting that metabolic scope and evaporative water loss have been ad¬ justed in the course of evolution to the tropical winter quarters as an adaptive response. Although saving on thermostatic costs, the tropical winterers face energetic bottlenecks during spring migration. Preliminary calculations suggest that wind assistance is essential to balance their migratory budgets. This finding may be more generally applicable to the problem of the origin of migratory pathways. Keywords: Migration, Knot, Calidris canutus, metabolism, thermoregulation, energetics, BMR, subspecific variation, food resources, travel costs, wind effects. INTRODUCTION Frozen and covered under snow for more than nine months of the year, the sparsely vegetated land around the Arctic Ocean becomes teeming with bird life during the short Arctic summer. Most species annually exploiting this vast temporarily produc¬ tive habitat, winter in the far south, and the Knot Calidris canutus is a typical exam¬ ple of such an extreme longdistance migrant. During the non-reproductive season, Knots roam along the edges of all continents, in the few wildernesses with a sufficient area of intertidal flats still extant. At this season they usually occur in large flocks and depend on small benthic shellfish for food. Knots with a circumpolar but discontinuous distribution, only breed on high Arctic tundra, usually north of the Arctic Circle and north of the 10°C July isotherm (Figure 1 left) Four subspecies are currently recognized. Two subspecies occur in W. Europe. They breed in N.E. Canada and Greenland (C.c.islandica) and in the Taymyr Penin¬ sula region in Siberia ( C.c.canutus ) (Figure 1 right), and will be the focus of this con¬ tribution. The islandica subspecies winters in a few large estuaries in N.W. Europe (50-55°N) During their migrations to and from N. Greenland and N.E. Canada the birds stop-over in Iceland and in N. Norway (Davidson & Wilson 1991). The canutus subspecies winters in W.Africa (10-20°N), the majority of birds frequenting two coastal sites the Banc d’Arguin in Mauritania and the Archipelago dos Bijagos in Guinea- Bissau (Smit & Piersma 1989). The Wadden Sea is their main stopover site during both northward and southward migrations (Dick et al. 1976, 1987, Piersma et al. 1991). 762 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI FIGURE 1 - Circumpolar breeding distribution of Knots (adapted from Tomkovich 1991), the map also showing the 10°C isotherm (from Stonehouse 1989). At right the migration patterns of the canutus subspecies breeding in Siberia and wintering in W. Africa, and the islandica subspecies breeding in Greenland and N.E. Canada and wintering in W. Europe are shown. The figures indicate population sizes at the main wintering sites (from Smit & Piersma 1989). A first step in ecologically evaluating the within-species variation in wintering latitude, is logically the analysis of the energetic repercussions associated with such differ¬ ences (Myers et al. 1985, Castro 1988, Summers et al. 1989, Drent & Piersma 1990). Breeding in the high Arctic, a relatively cold and windy region of the world, is likely to entail a considerable cost to uphold endothermy in waders, since they are not par¬ ticularly well insulated (Kersten & Piersma 1987). We expect these “thermostatic costs” to weigh heavily in the energy budgets of such high Arctic breeders. In winter, the birds might save on thermostatic costs by moving south, but would incur hiqher travel costs the further south they go (Drent & Piersma 1990). In this contribution we present an empirical assessment of the thermostatic costs of Knots on their Arctic breeding grounds in N.E. Canada, and explore how much cheaper life is likely to be on the more southerly wintering grounds. A further element in this assessment is the estimation of migratory cost from information on mass changes, and novel is the find ing that travel costs differ greatly depending on the route followed. These new view¬ points define costs and benefits of current latitudinal range differences and allow speculation on how the spectacular differences in wintering area in Knots are likelv to have come about. ” METHODS Fieldwork on wintering Knots was carried out on the Banc d'Arguin in Mauritania ana in the Dutch Wadden Sea, and on Knots during the breeding season in Arctic Canada” At each of these sites the activity patterns, feeding behaviour and presence in diffa ent microhabitats was studied. To approximate the thermostatic costs of live b' d ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 763 the field, we have chosen to use the approach of “heated taxidermic mounts”, devel¬ oped and advocated by G.S.Bakken and coworkers (Bakken 1976, 1980, Bakken et al. 1981, 1985, P.Wiersma & T.Piersma in prep.). Such “copper Knots” mimic a static bird in its environment as well as its thermoregulatory response (i.e. an increasing metabolism). The heated taxidermic mount used in this study is a hollow copper model of a skinned body covered by a fresh, prepared and complete skin of a Knot, which is mounted in a natural standing position (see Figure 2 bottom). A thermistor inside the model is connected to a thermostat which can be set to keep the internal tempera¬ ture of the model at the natural constant of 41 °C, the required heat being generated by a heating wire embedded in the copper model’s wall. In this way the combined effects of air temperature, wind and radiation, as they affect Knots in different microhabitats, can be measured in an integrated way. The power consumption of the heated taxidermic mounts was calibrated under standardized laboratory conditions (either a pitch black plexiglass respiration box without wind, or a darkened small windtunnel with a wind force of 1 m/s) over a range of temperatures to the metabo¬ lism of live Knots determined from 02-consumption as measured under exactly the same conditions. The methodology of the 02-consumption measurements is described by Kersten & Piersma (1987), with the difference that in this study only values from night-long continuous measurements of post-absorptive resting metabolism under constant conditions were used. It was reassuring that calibrations of heated models to live birds under both wind-free and wind conditions (situations of free and forced convection, respectively) led to the same estimated thermostatic costs under these conditions. The Knots used in these experiments were captured in the Dutch Wadden Sea in winter, and had been in captivity under the normal Dutch photoperiodic regime for 1 .5 years, showing the moult and body mass cycles known from the field as nor¬ mal (W.Teunissen & T.Piersma in prep.; see Kersten & Piersma 1987). Thermostatic cost is defined as the metabolic rate of birds resting or sleeping in dif¬ ferent microhabitats, thus including the basal metabolic rate (BMR), which represents the thermostatic cost under thermoneutral conditions. Thermostatic cost is hence a shorthand for the residual component of the daily energy expenditure excluding ac¬ tivity and energy retention (costs of synthesis), and can be estimated by employing heated taxidermic mounts. Simultaneously to the measurement of thermostatic costs in different microhabitats in the field, three standard metereological parameters (dry bulb air temperature, wind speed at heights between 10 and 15 m, global solar radiation) were monitored on a continuous basis at a nearby location. From half-hourly values of both thermostatic cost (=power reading of heated taxidermic mount) and the three meteorological vari¬ ables measured under a wide range of conditions, simple predictive models to esti¬ mate thermostatic costs in different microhabitats from given values of temperature, wind and radiation were derived. In these statistical models (the multiple regressions always explaining more than 95% of the variance in thermostatic costs), the effects of wind and temperature were assumed to be multiplicative, while that of radiation was assumed to be additive (see Bakken 1976, 1980). Estimates of the field metabolic rate of high Arctic breeding waders in relation to weather conditions, were obtained from a sample of 11 incubating Turnstones Arenaria interpres, studied in June-July 1989 on the tundra of Rowley Island, Foxe Basin Canada (R.I.G. Morrison & T.Piersma in prep.). Field metabolic rates were 764 ACTA XX CONGRESSUS INTERNATIONAL^ ORNITHOLOGICI air temperature (°C) FIGURE 2 - Thermostatic costs of Knots (subspecies islandica ) under laboratory conditions (top) and in different microhabitats in the field (bottom). For the latter, environmental con¬ ditions were standardized as for a cloudy midday with a gentle breeze (3 Beaufort): wind speed is 3 m/s and global solar radiation is 400 W/m2. The data in the upper panel refer to night-long continuous 02 measurements of sleeping Knots in black plexiglass boxes in a climatic chamber, the air temperature reflecting the temperature in the box. The field data are extrapolations of measurements with heated taxidermic Knot-mounts, calibrated in the laboratory to live Knots (see text). In both panels the shaded level indicates the minimum or basal metabolic rate (BMR). calculated from the turnover rates of doubly-labeled water (D.O18). General proce¬ dures and methods of analysis followed Masman & Klaassen (1987). Birds were cap¬ tured on the nest, injected with 2.5 ml doubly-labeled water, kept for 2 hrs after which an initial small blood sample was taken, released and recaptured after 1-4 days to ob¬ tain a second blood sample. Simultaneously to these experiments, the thermostatic costs of Knots on the open tundra and in the nest were continuously measured by heated taxidermic mounts. The Turnstone can serve as an acceptable substitute for Knots in view of their taxonomic affinities, equal sizes and body masses, compara¬ ble levels of BMR (ca. 1 W in either species, see below and Kersten & Piersma 1987) and the close resemblance of their breeding and feeding habitats in the Arctic To obtain a cost factor for each km flown during long distance flights, we have fol¬ lowed a purely empirical approach. Estimates of the loss of fat (with an enerqy con tent of 39.3 kJ/g contributing 90% or more to the energy supplied during a lonq dis tance flight, Piersma & Jukema 1990) and hence the cost per km flown could be ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 765 derived for four trajectories (data from Davidson & Evans 1986, Dick et al. 1987, Prokosch 1988, Piersma 1989, Piersma & van Brederode 1990, Lindstrom 1990, Morrison & Davidson 1990, Gudmundsson et al. 1991): W. Africa to the Wadden Sea (0.3 kJ/km), Wadden Sea to Taymyr (0.5 kJ/km), S.E. England to N. Norway (0.8 kJ/ km), and N. Norway to Ellesmere Island (0.7 kJ/km). For each of these flights there is now ample observational evidence to underpin the interpretation that they are nor¬ mally traversed in one flight, and not carried out in small hops. RESULTS The thermostatic cost curve of Knots of the islandica population under defined labo¬ ratory conditions is presented in Figure 2 (top). The BMR of these birds amounted 0.95 W, a value which approximates the value of 1 W predicted by Kersten & Piersma (1987) for waders of their mass. Their conductance was -0.04 W/°C, a value close to the one for the Turnstone as reported by Kersten & Piersma (1987) and the predic¬ tions by Kendeigh et al. (1977) and Aschoff (1981) for non-passerine birds at this body mass in winter and in the resting phase of their diurnal cycle, respectively. The rela¬ tive thermostatic costs of Knots in different microhabitats in the field as measured by the heated taxidermic mounts (Figure 2 bottom), indicate that shelter from the wind (e.g. when roosting between the vegetation on a saltmarsh or when incubating) can considerably reduce thermostatic costs. The effectiveness of our thermostatic cost estimates to help explain the pattern of energy expenditure of free-living birds, was tested by plotting the field metabolic rate of incubating Turnstones (spending approximately 50% of their time on the nest), to the concurrent thermostatic requirement. The close correspondence in slope (Figure 3) suggests that much of the variation in the field metabolic rate can be explained by compensation for prevailing weather variations. Piersma & Morrison (in prep.) dem¬ onstrate that the estimated thermostatic cost factor explains at least 55% of the vari¬ ation in field metabolic rate. By virtually integrating the combined effects of tempera¬ ture, wind and radiation, the copper mount data more closely parallel the measured field metabolism than any of the meteorological data sets. Judging from the data dis¬ played in Figure 3, the cost of activity can be estimated at 1-1.5 W. A similarly high level of total energy expenditure (average of 3.5 BMR) is reported for the Purple Sandpiper Calidris maritima, studied during the breeding season in high Arctic Spitsbergen by Pierce (1989). Flaving shown the utility of the thermostatic cost (as estimated by the heated taxider¬ mic mounts) in explaining the field metabolism of free-living Turnstones, we can now consider the level of thermostatic costs at different locations and seasons. The pre¬ dictive equations on which Figure 2 (bottom) is based, are used to estimate the ther¬ mostatic requirements of Knots living on both tundra and on temperate and tropical mudflats (Figure 4). During the breeding season the cost of living on open tundra is always close to 3 W. If the physiological maximal intake rate estimated by Kirkwood's (1983) allometric equation at 4-5 W, bears relevance to Knots, this leaves little more than 1-1.5 W “energetic leeway” for the costs of activity, i.e. the cost of activity as measured in breeding Turnstones. The Arctic therefore proves to be a costly place to be and breed. 766 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI The Knots of the islandica subspecies which fly southeast to winter in N.W. Europe face an environment where the average cost of thermoregulation is as high as on the breeding grounds (Figure 4 right): the thermostatic costs when living on the mudflats of the Wadden Sea approximates 3 W. Only Knots which go as far south as W. Af¬ rica experience much more congenial weather, and incur thermostatic costs close to 1 .5 W, a saving of 50%. This pattern of decreasing field costs with decreasing latitude has been empirically confirmed in free-living Sanderlings Calidris alba, studied by Castro (1988) by applying the double-labeled water technique to free-living birds win¬ tering in New Jersey, Texas, Panama and Peru. Although it is therefore likely that it is metabolically cheaper for Knots to spend the winter in W. Africa than in W. Europe, such birds have to fly much greater distances (Table 1). At present we are unable to empirically estimate all costs associated with these migrations (e.g. variable synthesis costs and working levels, variable costs of transport), but we can roughly compare the costs made during long distance flights (Table 1). In spring, both subspecies cover the distance between wintering and breed¬ ing area in two single long flights (Figure 1 right), those of canutus being twice as long as those of islandica (Table 1). In spite of this, the Siberian breeders leave the win¬ tering and the spring staging area at equal or even lower body masses than the FIGURE 3 - Field metabolic rate of free-living Turnstones in relation to the concurrent ther¬ mostatic costs as measured by heated taxidermic Knot-mounts. Field metabolic rate was estimated from D2018-turnover measurements in individuals (9 females and 2 males) recap¬ tured at daily or two-daily intervals while halfway through incubation on the tundra of Rowley Island, Foxe Basin, Canada, in early July 1989 (T.Piersma & R.I.G. Morrison in prep.). In the calculation of the average thermostatic cost per experiment, we assumed that the birds spent half of their time on the open ground and the other half in the nest cup (pers.obs.). Field metabolic rate and thermostatic cost are linearly related by the equation- y= 0.94 + 1 .22x, r2=0.55. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 767 TABLE 1 - Travel distances of, and approximate annual average expenditure on long¬ distance flights by, a temperate ( islandica ) and a tropical (canutus) wintering subspe¬ cies of Knot. Cost factors for long-distance flights were calculated from estimated fat losses during specific flight (see methods section). Flight distances were read from a globe and are from Gudmundsson et al. (1991). Trajectory Subspecies islandica canutus W. Africa - Wadden Sea (km) Wadden Sea - Taymyr (km) Wadden Sea - Iceland (km) Iceland - Ellesmere Island (km) 2 100 2 700 4 600 4 800 Total one way (km) Total return (km) 4 800 9 600 9 400 18 800 Empirical travel cost (kJ/km) Annual flight cost (kJ/yr) 0.7-0. 8 6 720-7 680 0.3-0. 5 5 640-9 400 Average cost (W) 0.22-0.25 0.18-0.30 islandica' s before a comparable leg of the journey. Therefore, canutus ends up with much lower travel costs (this is the cost/km and not the cost/time, the flight cost) than islandica. In the particular case of the flight from W. Africa to the Wadden Sea, there are several lines of evidence suggesting that the low cost can be explained by wind assistance obtained by flying at the appropriate altitude (Piersma & Jukema 1990, T.Piersma, P.Prokosch & D.Bredin in prep.). Concerning the journey from the Wadden Sea to Taymyr, Ebbinge (1989) was able to use tailwind as a factor in explaining Si¬ berian Brent Branta bernicla breeding success ( headwinds in the spring migratory period inhibiting success). Based on the information assembled on wind patterns by Lamb (1972) we propose that favourable winds account for the low apparent costs during the flight from the Wadden Sea to Taymyr. For each subspecies of Knot the average cost of living on an annual basis is thus increased by 0.2-0. 3 W, purely on the basis of net transport cost. At 20-30% of the value of BMR this increment certainly certifies as weighing heavily in the annual energy budgets. DISCUSSION Preliminary measurements of BMR in midwinter indicate a correlation between BMR and the inferred winter metabolism (Figure 4) in the two subspecies of Knot examined. The southerly wintering canutus has a BMR a factor 0.90 that of the temperate win¬ tering islandica (0.1 0 rt nf thp n i itch- M au ritan ian oroiect Banc d’Arauin 1985-1986. WlWO-report 25/RIN re- L. (Eds). Report of the Dutch-Mauritanian project port 89/6, Texel. . . , PIERSMA T VAN BREDERODE, N.E. 1990. The estimation of fat reserves in coastal waders before their departure from northwest Africa in spring. Ardea 78. 221-236. PIERSMA T., PROKOSCH, P., BREDIN, D. 1991. The migration system of Siberian Knots Calidris c canutus Wader Study Group Bulletin, Supplement in press. PROKOSCH P. 1988. Das Schleswig-Holsteinische Wattenmeer als Fruhjahrs-Aufenthaltsgebiet arktischer Watvogel-Populationen am Beispiel von Kiebitzregenpfeifer ( Pluvialis squatarola L. 1758), K tt (Calidris canutus, L. 1758) und Pfuhlschnepfe ( Limosa lapponica, L. 1758). Corax 12: 273-442. 772 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI ROOT, T. 1988. Energy constraints on avian distributions and abundances. Ecology 69: 330-339. ROOT . T. 1 989. Energy constraints on avian distributions: a reply to Castro. Ecology 70: 1 1 83-1 1 85. SMIT, C.J., PIERSMA, T. 1989. Numbers, mid-winter distribution and migration of wader populations using the East Atlantic Flyway. Pp. 24-63 in Boyd, H., Pirot, J.-Y. (Eds). Flyways and reserve networks for water birds. IWRB Spec. Publ.9, Slimbridge. STONEHOUSE, B. 1989. Polar ecology. Blackie, Glasgow. SUMMERS, R.W., UNDERHILL, L.G., CLINNING, C.F., NICOLL, M. 1989. Populations, migrations, biometrics and moult of the Turnstone Arenaria i. interpres on the East Atlantic coastline, with special reference to the Siberian population. Ardea 77: 145-168. TOMKOVICH, P.S. 1991. An analysis of the geographical variation in Knots Calidris canutus. Wader Study Group Bulletin, Supplement in press. WOLFF , W.J., SMIT, C.J. 1990. The Banc d’Arguin, Mauritania, as an environment for coastal birds. Ardea 78: 1 7-38. ZWARTS, L., BLOMERT, A.-M., HUPKES, R. 1990. Increase of feeding time in waders preparing for spring migration from the Banc d'Arguin, Mauritania. Ardea 78: 237-256. ZWARTS , L., ENS, B.J., KERSTEN, M., PIERSMA, T. 1990. Moult, mass and flight range of waders ready to take off for long-distance migrations. Ardea 78: 339-364. ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 773 IS WATER OR ENERGY CRUCIAL FOR TRANS-SAHARA MIGRANTS? H. BIEBACH Max-Planck-lnstitut fur Verhaltensphysiologie, Vogelwarte Andechs, 8138 Andechs, Germany ABSTRACT. For small Passerines crossing the Sahara, two environmental factors, tailwinds and low air temperatures, have been discussed to be critical for the energy and water budget. Fat reserves and water content of unsuccessful, dying birds during migration in the Libyan desert in Egypt were ana¬ lysed. Whereas the reason for the failure of Swallows Hirundo rustica was not obvious, Willow War¬ blers Phylloscopus trochilus, like other small Passerines, ran out of fat, which is the main energy substrate during flight. The water budget was still normal. Keywords: Trans-Sahara migrants, fat, energy, water budget, dehydration. INTRODUCTION Migrants that fly long distances between their breeding and their wintering sites of¬ ten have to cross ecological barriers such as water, glaciers, mountains or deserts. The number and quality of stopover sites and conditions during flight probably present a greater problem than distance per se to migrants. Each particular ecological barrier poses special problems to migrants depending on the species’ characteristics of for¬ aging ecology, flight capabilities and general physiology. In this paper, I investigate critical factors during migration across the Sahara desert and suggest possible reasons for the failure of some birds undertaking the crossing. Most European birds that winter in tropical Africa have to cross the Sahara with only a minimal probability of finding food or water. Two potential strategies of crossing have been proposed: (I) A non-stop flight over the entire breadth of the desert, a dis¬ tance of at least 1800 km in autumn (Moreau 1961,1972) and (2) an intermittent mi¬ gration with flight during the night and stopover during the day (Biebach et al. 1 986). There is no convincing evidence for preferred flights along particular routes such as aiong the Nile, or along a row of oases in N-S direction in Algeria, or along the West- Atlantic coast (Biebach 1990). Based on energy and water budget calculations, two conditions during flight have been suggested to be necessary for successful crossing: a) First, energy reserves before crossing the Sahara in autumn are, for most birds, sufficient to cross without refuelling oniy if they experience tailwinds of at least 8 m/s (Biebach 1990). b . Second, water balance can oniy be maintained if the air temperature during flight js below 10 °C. Otherwise, the birds would have to regulate their temperatures by evaporation and they would quickly reach critical values of dehydration (Biesel & Nachtigall 1987, Hudson & Bernstein 1981, Torre-Bueno 1978, Tucker 1968, 1974). 774 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI These conditions of tailwind and low air temperature cannot be met simultaneously during all days during the autumn migration period. If appropriate tailwinds can be found only at an air layer with temperatures higher than 10°C (normally below 2000 m above ground) the birds cannot balance their water budget. On the other hand, if they choose to fly at an air layer with temperature below 10°C they might confront headwinds and encounter energetic problems. Under these conditions, it has been suggested that birds stop flying and rest until dark, when acceptable temperatures are at lower altitudes where favorable tailwinds prevail. Unfortunately, we currently lack direct measurements of wind and temperature profiles, integrated with studies of stopover behavior of migrants in the Sahara. Based on the hypothesis concerning nonstop versus intermittent migration, we predict that birds stopping over in the desert would have a balanced water budget and sufficient energy reserves to continue the crossing because resting is interpreted as behavior to escape detrimental flight con¬ ditions. And, in fact, birds stopping over in the desert generally have the expected normal water budget and energy reserves (Biebach 1988, 1990). On the other hand, many observers reported exhausted, dying or dead migrants in the Sahara. In some cases, it has been obvious that the birds could not escape long-lasting sand storms, but in most cases the reasons for their failure are unknown. Investigation of such birds could shed some light on critical factors during migration. Yapp (1956, 1962) and Fogden (1972) have previously discussed water as a potential limiting factor on mi¬ gratory range on purely a theoretical basis. METHODS I investigated samples of birds found in a small oasis “Sadat Farm” in the Libyan desert in Egypt and in a mountain ridge called “Sposserberg” during autumn migra¬ tion. Both sites were about 300 km south of the Mediterranean coast and about 200 km west of the Nile (for a detailed description of location and the habitat see Biebach et al. in prep). Two groups of mist-netted birds were investigated for their water and fat content: two samples of 18 “normal” Willow Warblers Phylloscopus trochilus and 10 “normal” Lesser Whitethroats Sylvia curruca. “Normal” means the birds have been caught with mistnets out of a pool of birds with the normal escape behavior. The sec¬ ond group of birds investigated were obviously close to death because they were sit¬ ting on the ground with fluffed feathers and were caught by hand. Eight Swallows Hirundo rustica and two Willow Warblers were analysed. Birds were collected prima¬ rily in the morning, shortly after arrival at the stopover place. Thus the condition of the dying birds were mainly caused by migratory flight. Immediately after capture the birds were weighed to the nearest 0.1 g and deepfrozen for later analysis. In the laboratory the birds were then dried at 60°C to constant mass. The difference between fresh mass and dry mass was the water content. Fat was extracted with petroleum-ether in a Soxhlet apparatus. The extracted fat was directly weighted and the remaining material dried to constant mass was the lean dry mass Water and fat were expressed as percent of lean dry mass. Several problems arise for the interpretation of the water budget. First, we want to know at what point of dehydration birds can no longer maintain basic physiological functions like temperature regulation or heart rate, and second, how one can predict how close a dehydrated bird is to this point. The water content of hydrated birds is ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 775 regulated within a narrow range (Odum et al. 1964); when expressed as % water of total body mass it is close to 66 %. Homeostatic water balance is maintained irrespec¬ tive of fat storage, thus the water content is expressed as % water of the fat-free dry mass or lean dry mass. This index is normally at about 220 for hydrated birds and birds with values below 200 are regarded as dehydrated (Fogden 1972). Desert birds and probably small passerine migrants can tolerate dehydration down to about a water index of 150 (Skadhauge 1974,1981, Fogden 1972). RESULTS Two groups of 18 Willow Warblers and 10 Lesser Whitethroats, regarded as repre¬ sentative samples of birds normally stopping in the Libyan desert, had a mean water index of 209±22 (SD) and 237±26 (SD) respectively and both indices were very similar to a sample of 808 birds from nine species that had been killed during autumn migra¬ tion at a television tower in North America with values between 204 and 225 (Child & Marshall 1970). One bird from the Willow Warbler sample had a water index of 149, WATERBALANCE dying birds □ normal birds FIGURE 1 - Water content (% H20 of lean dry bodymass) of migrants from three species, Hirundo rustica, Phylloscopus trochilus and Sylvia curruca sampled in autumn in the Libyan desert in Egypt and from nine species (means) killed at a television tower in Florida. North America on autumn migration. Dark squares are values from dying birds, light squares from birds with normal behavior. Birds with values above 200 are normally hydrated. Wide dashed zone is tolerance to dehydration, narrow dashed zone is lethal. 776 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI FATRESERVES FIGURE 2 - Fat reserves (% fat of lean dry bodymass) of migrants from three species sampled in the Libyan desert in Egypt. Dark squares are values from dying birds, light squares represent birds with normal behavior. far below 200 - the normal value of hydrated birds. This finding indicates that among normally resting birds, a few might be dehydrated. Two Willow Warblers that were classified as close to death were normally hydrated with values of 21 1 and 215. It is concluded that dehydration was not the reason for their condition. The water indices of the dying Swallows seem to fall into two groups: one group found in 1983 were all below an index of 200 and therefore more or less dehydrated whereas the Swallows found in 1982 had very high water indices well above 200 (Figure 1). With respect to the fat reserves, the normal Willow Warblers and Lesser Whitethroats had moderate to extensive fat reserves with mean fat indices of 94±32 (SD) (n=18) and 1 02±20 (SD) (n=10) respectively. The two dying Willow Warblers had no fat left for metabolism. The group of Swallows with the high water indices still had moderate fat reserves with indices between 50 and 61 (Figure 2). The dehydrated group had no fat left for metabolism. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 777 FATRESERVES 120_ 100. C/3 C/3 CCj 1 $ 80. o X $ cd 50 g) and small (<50 g) migrants found dead or exhausted on spring migration in the Western Sahara. See also legend Figure 1. Data from Haas and Beck 1979. DISCUSSION Is fat as an energy source or is homeostasis of body water critical during long dis¬ tance flights over the Sahara? The poor conditions of the Swallows cannot be ex¬ plained simply by either of the two alternatives. They either had low fat and dehydra¬ tion values or fat and dehydration values had not yet reached critical levels, though the birds were close to death. For this group of birds the physiological causes of their condition remain unclear. Swallows might be different from other small passerines, as they normally migrate during the daytime and are much more airborne than most other migrants. So. the role of dehydration as a limiting factor for Swallows remains to be elucidated. More data are necessary, especially from birds towards the end of the desert crossing. The Willow Warblers and a sample of small passerines (bodymass<50 g) during spring migration in the Algerian Sahara (Haas & Beck 1979) showed no severe de¬ hydration (Figure 3). On the other hand birds stopping-over for one day in the desert may lose as much as 20% of their total bodywater by evaporation and it is likely that 778 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI a second day of stopover brings them to critical levels of dehydration (Biebach 1990). Critical values of dehydration (below a water index of 150) have been reached by six exhausted, large birds (bodymass>50 g) on spring migration (Haas & Beck 1979) (Figure 3). Their fat reserves were all robust. The conclusion is that water as a sin¬ gle limiting factor during desert crossing is still a matter of question. Possibly large birds are more susceptible to dehydration than small ones, as suggested already by Haas & Beck (1979). Depleted fat reserves seemed to be the cause of energetic failure in the Willow War¬ blers (this study) and in small migrants analysed by Haas & Beck (1979) (Figure 3). In addition to the fat extraction data reported here there is more information available concerning the fat reserves from three extensive samples of migrants from two sites in the Libyan desert in Egypt (Biebach et al. 1986) and from the Western Sahara (Bairlein 1985). Fat reserves were scored by inspecting the extent of subcutaneous fat depots in live birds (Cherry 1982). A fat score of 1, indicating no visible subcuta¬ neous fat depots, corresponds quite well with low extractable fat. All six birds in the Swallow and Willow Warbler sample with extractable fat / lean dry mass below 30% had a fat score of 1 . This value is taken as an indicator of birds that fail to cross the Sahara due to low fat reserves. Four percent of Lesser Whitethroats, 10 % of Willow Warblers, and 8 % of Yellow Wagtails Motacilla flava resting in an oasis in the Libyan desert had this low fat score. However, not one of the Lesser Whitethroats or Willow Warblers from a stopover site in a stony plain in the Libyan desert had a fat score of 1 (Biebach et al. 1986). In a sample of different passerine species resting in the Western Sahara about 24% of the birds had fat scores of 0 (which corresponds to 1 in the samples from Egypt because of different scaling - Bairlein 1985). One might conclude that a considerable number of migrants run out of lipid fuel during the desert crossing. However, we do not know to what extent the samples represent the majority of migrants. There are indications that birds resting in oases represent individuals with reduced energy reserves (Biebach et al. 1986). Therefore, we are unable to say how regularly the phenomenon of running out of fuel occurs. As indicated earlier the current hypothesis of desert crossing suggests that an inter¬ mittent flight strategy is adopted because of unfavorable flight conditions. This strat¬ egy may occur when no air layer can be found that simultanously satisfies tail wind and low air temperature requirements, and the birds are forced to stop flying. If the birds have to continue flying because no suitable stopover sites can be found we might tentatively conclude that the birds avoid high temperatures and go for headwinds. Under this scenario, they will then finally run out of fuel. These conclu¬ sions certainly need further confirmation and more data addressing the condition of birds at the end of the desert crossing are needed. LITERATURE CITED BAIRLEIN. F. 1985. Body weights and fat deposition of Palearctic passerine migrants in the central Sahara. Oecologia (Berlin) 66: 141-146. BIEBACH, H. 1988. Ecophysiology of resting Willow Warblers (Phylloscopus trochilus) crossinq the Sahara. Proceedings of the 19th International Congress, Ottawa. 2162-2168. y BIEBACH, H. 1990. Strategies of trans-Sahara migrants. Pp. 352-367 in Gwinner, E. (Ed ) Bird migra¬ tion. Berlin, Heidelberg, Springer. ' J 1 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 779 BIEBACH, H., FRIEDRICH, W., HEINE, G. 1986. Interaction of bodymass, fat, foraging and stopover period in trans-Sahara migrating passerine birds. Oecologia (Berlin) 69: 370-379. BIESEL, W., NACHTIGALL, W. 1987. Pigeon flight in a wind tunnel. Journal of Comparative Physiol¬ ogy B 157:117-118. CHILD, G.I., MARSHALL, S.G. 1970. A method of estimating carcass fat and fat-free weights in mi¬ grant birds from water content of specimens. Condor 72: 1 16-1 19. FOGDEN, M.P.L. 1972. Premigratory dehydration in the Reed Warbler Acrocephalus scirpaceus and water as a factor limiting migratory range. Ibis 114: 548-552. HAAS, W., BECK, P. 1979. Zum Fruhjahrszug palaarktischer Vogel liber die westliche Sahara. Jour¬ nal fur Ornithologie 120: 237-246. HUDSON, D.M., BERNSTEIN, M.H. 1981. Temperature regulation and heat balance in flying White¬ necked Ravens, Corvus cryptoleucus. Journal of Experimental Biology 90: 207-281. MOREAU, R.E. 1961. Problems of Mediterranean Sahara migration. Ibis 103: 373-421, 580-623. MOREAU, R.E. 1972. The palearctic-African bird migration systems. Academic Press, London, New York. ODUM, E.P., ROGERS, D.T., HICKS, D.L. 1964. Homeostasis of the nonfat components of migrating birds. Science 143: 1037-1039. SKADHAUGE, E. 1974. Renal concentrating ability in selected West Australian birds. Journal of Ex¬ perimental Biology 61: 269-276. SKADHAUGE, E. 1981. Osmoregulation in birds. Zoophysiology 12. Berlin, Heidelberg, Springer. TORRE-BUENO J.R. 1978. Evaporative cooling and water balance during flight in birds. Journal of Ex¬ perimental Biology 75: 231-236. TUCKER, V.A. 1968. Respiratory exchange and evaporative water loss in the flying Budgerigar. Journal of Experimental Biology 48: 67-87. YAPP, W.B. 1956. Two physiological considerations in bird migration. Wilson Bulletin 68: 312-319. YAPP, W.B. 1962. Some physical limitations on migration. Ibis 104: 86-89. 780 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI PATTERNS OF AVIAN MIGRATION IN LIGHT OF CURRENT GLOBAL GREENHOUSE’ EFFECTS: A CENTRAL EUROPEAN PERSPECTIVE P. BERTHOLD Max-Planck-lnstitut fur Verhaltensphysiologie, Vogelwarte, Schloss Moggingen, D-7760 Radolfzell, Germany ABSTRACT. There has already been a global increase in the mean temperature of about 0.6 °C since 1900, which has caused considerable concern among climatologists that there may be a further rise of between 1.5 and 5 °C over the coming decades. The increase is primarily due to a build-up of car¬ bon dioxide in the atmosphere. A strong anthropogenic ‘greenhouse’ effect of this magnitude could alter the earth’s vegetation and raise sea levels considerably. In addition, it would change bird life in gen¬ eral and migratory systems in particular. More specifically, residents would benefit from a decrease in winter mortality and an increase in reproductive output. Selective breeding experiments suggest that partial migrant populations would rapidly shift to sedentariness, and that short-distance migrants would shorten their migratory distances. Faced, then, with an overall increase in resident competition for resources, long-distance migrants would decline further. Some more detailed prognoses are given. Possible side effects of man-made climatic changes and other factors may cause an even more pro¬ nounced, general decline of avian diversity. Keywords: Greenhouse’ effects, central European bird fauna, migration systems, evolution, popula¬ tion dynamics. INTRODUCTION The present central European bird fauna has developed during the last postglacial period over the past 15000 years. The highest average species diversity and abun¬ dances were probably reached during the 18th century when the human settlement of large areas, and the alteration of closed woodland into highly structured mosaic landscapes, triggered the additional immigration of many open-country species. The decrease during the first half of the 19th century began slowly at first, but later accel¬ erated. The population decline was initially caused by direct human persecution and later by anthropogenic impairment of all available ecosystems (for review, see Berthold 1990a). The present period is characterized by distinct global climatic changes most likely due to man-made alterations of the atmospheric gas composition (next section), and their accompanying ‘greenhouse’ effects, particularly a global warming tendency which has been predicted for the coming decades. Such basic en¬ vironmental changes would certainly influence bird life in general and migration sys¬ tems in particular. Up to now, few specific predictions for the relationships between global climatic changes and future bird life have been made (Berthold 1990a, Elkins 1990). Therefore some general predicted trends will be outlined here for central Eu¬ rope. MEASURABLE AND PREDICTED ANTHROPOGENIC ‘GREENHOUSE’ EFFECTS Naturally ocurring carbon dioxide and water vapor in the atmosphere absorb some of the long-wave radiation from the earth (mainly originating from the sun) which results in a natural ‘greenhouse’ effect where gas instead of glass acts as the absorber. Man- ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 781 made increases in C02, mainly due to the burning of fossil fuels and forest destruc¬ tion, has led (in combination with other gases) to an enhancement of this effect, which has been termed the ‘greenhouse’ effect. The atmospheric C02 concentration has risen by 25% since pre-industrial times, with a rise of 11% alone in the past three decades. A further doubling of the concentration is expected by the middle of the next century. There has also been a concomitant global warming of about 0.6f C since 1900, most likely representing such a ‘greenhouse’ effect. Predictions for a further global rise in temperature, along with doubling of atmospheric C02, range between 1 .5°C and 5°C (Figure 1), with the greatest increase in polar regions. The climatic patterns of 1989 follow this trend. In central Europe, it was the warmest year of this century with an elevation of 1 .7°C above the mean temperature of the past 21 0 years. If the predictions are substantiated, the meiting of polar ice caps and mountain gla¬ ciers would raise sea levels considerably and would lead to an expansion of the oceans. Under this scenario, current sea-shore and estuarine habitats might well dis¬ appear to a large extent. Predictions with respect to precipitation are rather uncertain, but for higher latitudes an increase would be likely. This fits with the prediction of in¬ creased plant productivity. With a doubling of C02 concentration, the promoting effect could range between 25 % and 50 %. Increased primary production may, however, be accompanied by changes in nutrient dynamics, e.g. by a substantial decrease in the nitrogen content of plant material. Rather dramatic changes would then be ex¬ pected in the distribution of vegetation belts. More southern plant communities could rapidly spread north, and coniferous forest could be replaced by deciduous woodland over large areas. There would, however, be a risk of considerable plant mortality with subsequent loss of the accompanying microfauna and thus dramatic changes in the present ecosystems. The arctic tundra may completely disappear as a habitat due to the extension of shrubby plants from the south and to inundation from northern seas. Further predictions concern a general increase in extreme climatic events like more irregular occurrence of wet and dry periods, increased probability of heatwaves, and storms. These events may exert negative effects on sensitive parts of ecosystems, and storms may increasingly destroy forests. For more details see, e.g., Anonymous (1990), Dobson et al. (1989), Boer et ai. (1990), Elkins (1990), Rocznik (1990). GENERAL ASPECTS OF EXPECTED CHANGES IN BSRDLIFE A general warming, especially during autumn, spring and winter, and an increase in primary production should primarily favour resident species. Reduced winter mortal¬ ity and earlier breeding seasons could theoretically lead to major increases, especially in species with r-strategies such as tits, which normally compensate for high winter mortality with high reproductive rates. There are already indications for current popu¬ lation increases in eastern European tit populations (Berthold 1990a). In western populations, however, possible increases may have been suppressed by negative man-made environmental effects like deforestation due to storm damage, the ‘Waldsterben’, and more direct effects of acid rain. In Holland, Great Tit Parus ma- ior breeding data in areas of nutrient-poor soil, has demonstrated an increase of more than 50 % in the percentage of territorial pairs with no eggs or disturbed eggshell for¬ mation since 1983. The inferior quality of eggshells is assumed to be the resuit of an insufficient calcium content in food and plant material due to acid rain. Meanwhile, there are similar reports for central Germany (for review, see Berthoid ^990a). Thus, a general mass increase in residents may well be prevented in a number of species by counteracting negative environmental factors. 782 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI P 0 cn c 03 sz (J 0 D 03 w_ 0 CL E 0 03 -Q O O I860 1900 1940 1980 2020 2060 2100 Year FIGURE 1 - Global temperature change in °C 1860-1984 and predicted temperature changes according to three models (A-C, from Dobson et al. 1989). Other groups which should be generally favoured by predicted environmental changes are partial and middle to short distance migrants which winter south to the Mediter¬ ranean. Selective breeding experiments with partially migratory Blackcaps Sylvia atricapilla have recently shown that under strong selection pressure, genetically based sedentariness can be expected within four to six generations (Figure 2). Current se¬ lective breeding experiments with Blackcaps indicate that there is also a fairly high heritaoility with respect to the amount of migratory activity within populations which controls the distance of migration. Warmer autumn, winter and spring periods should thus (1) rapidly shift the proportions of obligate partially migratory species towards sedentariness, and (2) individuals of middle to short distance migrants should continu¬ ously be selected for shorter migration distances. After the mild winters of the past few years, local ornithological reports from central European areas have recorded numer¬ ous observations of partial and short-distance migrants wintering within the breedina area. y The obvious losers in light of the predicted environmental changes should in General be long-distance migrants. These species have already shown more marked pooula’ tion declines than any other group of European birds. These decreases are however due to many factors. A major problem is, undoubtedly, the Sahel drouqht fe n Berthold 1990b). Multifactorial analyses of the British Trust for Ornithology (O’Conno 1981) and direct studies of competitive behaviour (for review, see Berthold 1990 1 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 783 have clearly demonstrated that the successful settlement of late arriving, long-dis¬ tance migrants on the breeding grounds depends on the density of resident species and short and middle distance migrants which have already established territories. Substantial increases in both residents and short to middle distance migrants should therefore lead to decreases in long-distance migrant populations. These declines may even result in extinction of those groups for which large-scale habitat loss is likely (e.g., in coastal and tundra areas). Thus, the predicted environmental changes will most likely substantially reduce long-distance migration and promote sedentariness and short-distance migration. The only alternative for long-distance migrants, the shift to short-distance migratory behaviour by microevolutionary processes, is probably too slow an alteration to occur in time. SELECTIVE BREEDING FIGURE 2 - Results of a two-way selective breeding experiment with partially migratory Blackcaps from southern France up to the F6-generation (nonmigrants) and to the F3-gen- eration (migrants). Numbers: individuals bred in each generation, broken lines: mathemati¬ cal functions that best fit to the selection response (from Berthold et al. 1990). SOME SPECULATIONS ABOUT INDIVIDUAL SPECIES In general, in all three groups of birds, i.e. residents, short to middle distance migrants and long-distance migrants, generalists will have a far better chance of coping with the predicted environmental changes than specialists. For example, among residents, specialists for coniferous forests such as the Coal Tit Parus ater are already declin¬ ing and will presumably be further affected whereas generalists such as the Green¬ finch Chloris chloris or Tawny Owl Strix aluco will most likely increase. Further, with respect to deforestation, species which inhabit shrubbery will be more successful than forest specialists. 784 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI TABLE 1 - Four categories of central European bird species with predicted different population trends. Partial and short- to middle-distance migrants A B with high chances for population increase with lower chances for population increase Phalacrocorax carbo Ardea cine re a Buteo buteo Fulica atra Larus r id i bund us Alcedo atthis Motaciila alba Troglodytes troglodytes Prunella modularis Sylvia atricapilla Phylloscopus collybita Phoenicurus ochruros Erithacus rubecula Turd us pilaris Turd us merula Emberiza schoeniclus Serin us serin us Carduelis carduelis Carduelis cannabina Grus grus Vanellus vanellus Columba palumbus A laud a arvensis A nth us prate n sis Turd us viscivorus Turd us pnilomelos Emberiza citrinella Sturnus vulgaris Long-distance migrants C D being endangered to decline strongly or to die out over the next decades with good chances for long-term survival Ixobrychus minutus Ciconia ciconia Anas querquedula Jynx torquilla Lamus senator Acrocephalus palustris Acrocephalus schoenobaenus Sylvia curruca Sylvia communis Saxicola rubetra Phoenicurus phoenicurus Milvus migrans Hirundo rustica Delichon urbica Apus apus ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 785 In Table 1, four categories of species are listed with respect to population trends. Category A summarizes central European partial and short to middle distance mi¬ grants which presently show either increasing or stable populations (e.g. Berthold et al. 1986). Under the predicted environmental changes, these species could readily shift to more-or-less complete sedentariness which would strongly enhance their populations. Most of these species are, in many respects, generalists and should also be able to cope with changing ecosystems. The few specialists such as Phalacrocorax and Alcedo would benefit from human activities. Category B summarizes potentially favoured species where additional circumstances are more limiting. In Grus grus and Columba palumbus, for instance, the destruction of Mediterranean oak woodland as the basis for successful wintering may exert in¬ creasing negative effects (e.g. Gatter et al. 1990). In the other species, habitat loss has either already led to strong population declines like Vanellus vanelius and Anttius pratensis which are already on the German list of endangered species, or may, in combination with other factors, hinder potential population increase. Endangered long-distance migrants are listed under category C. These are species that have already declined substantially and, under the scenario of habitat changes and increased competition, might even disappear. In Category D, finally, four migratory species with presumably good chances for long¬ term survival are listed. All of them are able to feed continuously during migration anc thus may overcome impaired conditions during migration in the future. Among tnese species, the Swallow and the Swift would have practically no competitors, even in a considerably altered central European bird fauna. For all other central European bird species not listed in the table, predictions are, in my opinion, much more difficult to make at this time. OUTLOOK I have to emphasize that the hypothesized changes in central European birdlife dur¬ ing the coming decades due to anthropogenic ‘greenhouse’ effects can only be made with a high degree of uncertainty. Additional changes in ecosystems not considered here may strongly affect various species. For example, the current dramatic increase in the amount of forest fires in the Mediterranean region may destroy wintering grounds for many short and middle distance migrants before these species are able to develop into more sedentary populations. If desertification of the borderland of the Sahara desert reaches the Mediterranean areas, as some climatologists predict, it is uncertain how many long-distance migrants would continue to be capable of cross¬ ing such an expansive, dry area to winter south of it in Africa. Additionally, wintering bv short-distance migrants in the Mediterranean might well become improbable. Tak¬ ing into account all the possibilities, a continued overall reduction of species diversity of central European birdlife is likely. Further anthropogenic climatic changes appear currently highly likely with long-distance migrants suffering most from major negative effects of these changes. 786 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI ACKNOWLEDGEMENT This research is supported by the Deutsche Forschungsgemeinschaft. LITERATURE CITED ANONYMOUS 1990. Coastal concerns. IUCN Bull. 21 (2): 26-29. BERTHOLD, P. 1990a. Die Vogelwelt Mitteleuropas: Entstehung der Diversitat, gegenwartige Veranderungen und Aspekte der kunftigen Entwicklung. Verh. Dt. Zool. Ges 83 (in press). BERTHOLD, P. 1990b. Vogelzug. Darmstadt, Wissenschaftliche Buchgesellschaft (in press). BERTHOLD, P., FLIEGE, G., QUERNER, U., WINKLER, H. 1986. Die Bestandsentwicklung von Kleinvogeln in Mitteleuropa: Analyse von Fangzahlen. Journal fur Ornithology 127: 397-437. BERTHOLD, P., MOHR, G., QUERNER, U. 1990. Steuerung und potentielie Evolutionsgeschwindigkeit des obligaten Teilzieherverhaltens: Ergebnisse eines Zweiweg-Selektionsexperiments mit der Monchsgrasmucke (Sylvia atricapilla). Journal fur Ornithology 131: 33-45. BOER, M. M., KOSTER, E. A., LUNDBERG, H. 1990. Greenhouse impact in Fennoscandia - Prelimi¬ nary findings of a European workshop on the effects of climatic change. Ambio 19: 2-10. DOBSON, A., JOLLY, A., RUBENSTEIN, D. 1989. The greenhouse effect and biological diversity. TREE 4: 64-68. ELKINS, N. 1990. What does global warming mean for our birds? Scottish Bird News 18: 4-5. MATTER, W., GARDNER, R., PENSKI, K. 1990. Abnahme ziehender Ringeltauben Columba palumbus in Suddeutschland. Vogelwelt 111: 111-116. O’CONNOR, R. J. 1981. Comparisons between migrant and nonmigrant birds in Britain. Pp. 167-195 in Aidley D. J. (Ed.). Animal migration. Cambridge, New York, Melbourne, Cambridge University Press. ROCZNIK, K. 1990. Das meteorologische Jahr 1989 in Mitteleuropa. Naturwiss. Rundschau 43- 205- 209. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 787 CLOSING REMARKS: NEW ASPECTS OF AVIAN MIGRATION SYSTEMS FRANK R. MOORE Department of Biological Sciences, University of Southern Mississippi, Hattiesburg, MS 39406-5018, USA Just over a decade ago, S. A. Gauthreaux (1979) lamented justifiably the lack of breadth in our study of bird migration and expressed particular concern that proper emphasis was not being placed on the ecology and evolution of migration. The pa¬ pers presented in this symposium as well as other invited and contributed papers to the XX International Ornithological Congress suggest that “doldrums” is no longer an apt metaphor for research on bird migration. Besides recognizing the progress made in the study of bird migration, I wish to emphasize two, not unrelated, issues raised during this symposium. Not long ago, M. Pienkowski and P. Evans (1985) observed that “...the remarkable lengths of regular migratory journeys by many birds are now so well established that some biologist seem to discount their costs.” Contributions to this symposium focused attention on the factors that contribute to the costs of migration and on the adapta¬ tions that have evolved to offset these costs. While en route, a passage migrant must forage in unfamiliar habitats to replenish depleted energy stores, resolve conflicting demands of predator avoidance and food acquisition, compete with other migrants and resident birds for limiting resources, including food and water, respond to unpre¬ dictable and sometimes unfavorable weather, and correct for orientation errors. Moreover, favorable habitat, where a migrant can safely and rapidly meet physiologi¬ cal demands, is probably limited, or effectively so because it may not always have the opportunity to select the best habitat. Second, conservation efforts on behalf of intercontinental bird migrants depends on knowing when, where, and how migrant populations are regulated. Answers to those questions require a comprehensive interdisciplinary program of research — one that recognizes the relationship between behavior and population biology. Achievement of this disciplinary “synthesis” may not be easy. Behavioral ecologists are usually con¬ cerned with the way the behavior of animals may have evolved, with little reference to the consequences such behavior may have for population dynamics. Conversely, population biologists usually focus on the demographic consequences of environmen¬ tal or biological changes, with little reference to behavioral mechanisms that may underlie changes in birth, death and migration rates. They deal with fundamental de¬ mographic processes of births, deaths, immigration and emigration and factors affect¬ ing them, and the challenge is to expose the crucial factors underlying the patterns or distribution and abundance of a population over a time-scale of several genera¬ tions Although there is an enormous literature on migration and dispersal, there are very few studies explicitly relating these phenomena to the magnitude and stability of the populations in question. One obvious exception is the work on temperate-tropical bird migrants which seeks to comprehend the population dynamics of particular 788 ACTA XX CONGRESSUS INTERNAIONALIS ORNITHOLOGICI species in terms of measurable parameters characterizing the migratory and repro¬ ductive behavior of individuals. Palearctic-African and Nearctic-Neotropical bird mi¬ gration systems represent prime examples of a situation where a phenomenological description of the way populations interact in a spatially heterogeneous environment can, on the one hand, be grounded on an understanding of the behavior of individu¬ als and, on the other hand, lead to insights about population dynamics and commu¬ nity structure. The relevance to the conservation of our intercontinental migrants is obvious. LITERATURE CITED GAUTHREAUX, S. A. 1979. Priorities in bird migration studies. Auk 96: 813-815. PIENSKOWSKI, M. W., EVANS, P. R. 1985. The role of migration in the population dynamics of birds. Pp 331-352 in Sibley, R., Smith, R. (Eds) Behavioural Ecology. Blackwell, Oxford. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 789 SYMPOSIUM 11 ECOLOGICAL AND EVOLUTIONARY CONSEQUENCES OF BODY SIZE Conveners B. A. MAURER and J. F. PIATT 790 ACTA XX CONGRESSUS INTER NATION ALIS ORNITHOLOGICI SYMPOSIUM 11 Contents INTRODUCTORY REMARKS: ECOLOGICAL AND EVOLUTIONARY CONSEQUENCES OF BODY SIZE JOHN F. PIATT . 791 BODY SIZE EFFECTS ON FIELD ENERGY REQUIREMENTS OF BIRDS: WHAT DETERMINES THEIR FIELD METABOLIC RATES? KENNETH A. NAGY and BRYAN S. OBST . 793 SCALING AVIAN ECOLOGY WILLIAM A. CALDER and CYNTHIA CAREY . 800 BODY SIZE AND FORAGING BEHAVIOUR IN BIRDS R. I. GOUDIE and J. F. PIATT . 81 1 POSITIVE CORRELATION BETWEEN RANGE SIZE AND BODY SIZE: A POSSIBLE MECHANISM TERRY ROOT . 817 EXTINCTION RATE, BODY SIZE, AND AVIFAUNAL DIVERSITY BRIAN A. MAURER, HUGH A. FORD, and EDUARDO H. RAPOPORT . 826 CONCLUDING REMARKS: BIRDS, BODY SIZE, AND EVOLUTION BRIAN A. MAURER . . ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 791 INTRODUCTORY REMARKS: ECOLOGICAL AND EVOLUTIONARY CONSEQUENCES OF BODY SIZE JOHN F. PIATT Alaska Fish and Wildlife Research Center, U.S. Fish and Wildlife Service, 101 1 East Tudor Road, Anchorage, Alaska 99503, USA Most ornithologists inevitably give some consideration to the consequences and im¬ plications of body size. Variation in body size within and among species has provided a general framework for interpreting ecological adaptations (e.g., Bergmann’s rule), evolution (e.g., Cope’s law), competition and species diversity. During the last few decades, the study of allometry has yielded a quantitative basis for interpreting rela¬ tionships between body size and other biological parameters. The strength and ubi¬ quity of allometric trends have allowed us to go beyond the explanation of patterns we observe to the prediction of characteristics we cannot otherwise measure. The study of allometry can be broken down into five main components: structure, growth, physiology, behaviour and ecology. The allometry of structure, growth and physiology has been well studied. The scaling of individual body parts in relation to the overall size of birds is often constrained by structural principles. Body parts grow at predictable rates for a given body size, and different species grow at rates that are scaled to their size. The physiological parameters of metabolic rate and turnover of materials are scaled to body size. In each of these three disciplines, it has been found that body size often accounts for 90-99% of the variation in measured parameters — presumably because structure, growth, and metabolism are governed largely by the laws of physics and chemistry. In many cases this scaling is linear. For example, large birds need proportionately as much intestine as small birds to process food. In many instances, however, this scaling is non-linear because of the “economies of scale . The most familiar example of this is the scaling of metabolic rates to M0 75 (Kleiber’s 3/4 Rule). Because small birds lose more heat on a gram-per-gram basis than large birds, they must maintain a higher metabolic rate than large birds. The last two components of allometry — behaviour and ecology — have been stud¬ ied less and are the focus of this symposium. We observe that certain behaviours, for example foraging time and distance, sleep, prey selection, sociality, social dominance, etc are scaled with body size in many taxa. In allometric ecology, we consider the constraints that body size imposes on mass and energy flow, reproduction, and popu¬ lation dynamics. Through consideration of allometric constraints, we hope to gain in¬ sight on evolution and natural selection. Such studies are especially challenging be¬ cause many of the behavioural and ecological parameters that we can measure are often poorly correlated with body size for a variety of reasons. For example, behav¬ iour is quite variable within and between species and often difficult to measure in a standardized fashion. Lumping of different groups may obscure trends that exist within taxa or coexisting guilds. Nonetheless, allometry offers a quantitative and well-estab¬ lished starting point from which we can proceed to understand avian ecology. Once 792 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI we account for ecological variablity owing to body size, we can perhaps assess the degree to which other factors determine patterns of bird abundance and distribution. This understanding may be useful in a world with rapidly changing environments. In the following series of papers, we consider some of the ecological and evolution¬ ary consequences of body size in birds. Kenneth Nagy and Bryan Obst begin by ex¬ amining how field metabolic rates of birds vary with body size and consider what may account for the residual variation in metabolism after the effects of size are removed. William Calder and Cynthia Carey discuss the use of allometry in scaling avian ecol¬ ogy and consider how allometry may be applied to practical management and conser¬ vation issues. Ian Goudie and John Piatt examine the relationship between body size and foraging time budgets in birds and consider how body size constrains behavioural flexibility in variable or extreme environments. Terry Root examines the role of body size in limiting species distribution by considering the implications of body size for competiton, adaptation to extreme environments, habitat use, and prey selection. Fi¬ nally, Brian Maurer, Eduardo Rapoport and Hugh Ford compare the distribution and diversity of Australian and North American avifauna, and consider the role of body size in maintaining diversity and influencing time to extinction for various taxa. ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 793 BODY SIZE EFFECTS ON FIELD ENERGY REQUIREMENTS OF BIRDS: WHAT DETERMINES THEIR FIELD METABOLIC RATES? KENNETH A. NAGY and BRYAN S. OBST* Department of Biology, University of California, Los Angeles, California 90024, USA * Died 9 August 1991 ABSTRACT. Field metabolic rates (FMRs) of birds, measured with doubly labeled water, range from 27 kj/day (0.31 W) in hummingbirds to over 4730 kj/day (54.8 W) in Giant Petrels, a range of 175-fold. Most of this variation is due to differences in body mass. However, after recalculating FMR values in units of kj g0 64 d 1 to account for mass differences, a 6-fold variation in FMR still remains. What ac¬ counts for this residual variation? High FMRs occur in passeriform and charadriiform birds, carnivo¬ rous birds, seabirds (carnivorous birds that forage widely) and birds that live in meadows or marshes. However, these phylogenetic and ecological categories alone do not satisfactorily explain the large residual variation in FMR, partly because of small sample sizes available to date, but also because variation due to season, gender, daily behavior pattern, ambient temperature, social organization, flight mode and duration, food availability and predator pressure is real but is not accounted for in the above categories. INTRODUCTION The fitness of a bird may be determined in large measure by how successfully it ob¬ tains the food it needs for its own growth and maintenance and for production of vi¬ able offspring. Feeding rate is very difficult to measure directly in the field. However, because a large fraction of food energy is metabolized to heat by birds, measure¬ ments of field metabolic rate (FMR), which are readily obtained via the doubly labeled water (DLW) method, can serve as the basis for estimates of food requirements of free-living birds (Nagy 1989). Thus, DLW measurements of FMR can be valuable in evaluating the impact of an environment on the birds living there, as well as the im¬ pact of the birds on their environment, in terms of food resources consumed by the birds. It is unlikely that the field energetics of all bird species under all circumstances will be directly measured by biologists. Thus, it is important to derive ways to predict the food and energy requirements of birds, and to evaluate the accuracy and reliability of such predictions. Many factors are known to influence the energy and food require¬ ments of free-living birds, but body mass is thought to have the largest effect. In this paper, we examine the importance of body size (mass) in explaining the variation in observed FMRs of birds, and then we use analysis of residuals, after correction for body size effects, to try to account for the variation that still remains. MATERIALS AND METHODS The data set we used consisted of all the published studies of FMR in birds, as meas¬ ured with the doubly labeled water method, known to us at the time this analysis was done (21 January 1989). We included data for 25 species from an earlier review (Nagy 794 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 1987) along with new data for an additional 15 species (Table 1). We restricted our analysis to properties of entire species, such as mean body mass, diet and habitat, and did not consider within-species variables (such as gender) in this analysis. When separate FMR and body mass values were reported for different seasons, age groups, study plots, etc. within a species, we used the mean of these values to represent the species as a whole. TABLE 1 - Summary of doubly labeled water measurements of birds published since the review by Nagy (1987). Species Body mass, g FMR, kJ/day Reference Coal Tit Parus ater 9.5 47.4 Moreno et al. 1 988 Crested Tit Parus cristatus 11.1 40.6 Moreno et al. 1 988 Willow Tit Parus montanus 11.3 41.2 Moreno et al. 1 988 Tree Swallow Tachycineata bicolor Leach’s Storm-petrel 22.4 118 Williams 1988 Oceanodroma leucorhoa 44.8 87.1 Ricklefs et al. 1 986 Dipper Cinclus cinclus 63.1 220 Bryant & Tatner 1 988 Bryant et al. 1985 Pied Kingfisher Ceryle rudis 76 210 Reyer & Westerterp 1985 Least Auklet Aethia pusilla South Georgia Diving Petrel 83.5 358 Roby & Ricklefs 1 986 Pelecanoides georgicus Common Diving Petrel 109 464 Roby & Ricklefs 1 986 Pelecanoides urinatrix 137 557 Roby & Ricklefs 1986 Eurasian Kestrel Falco tinnunculus 220 343 Masman et al. 1 988 Black Guillemot Cepphus grylle 420 640 Roby & Ricklefs 1 986 Thick-billed Murre Uria lomvia 834 1475 Roby & Ricklefs 1 986 Little Penguin Eudyptula minor Grey-headed Albatross 1089 997 Costa et al. 1986 Diomedea chrysostoma 3706 2401 Costa & Prince 1 987 We used linear least-squares regression analysis for characterizing allometric relation¬ ships for FMR. Confidence intervals (95% Cl) were used to evaluate (1) differences between predicted and measured FMRs, (2) differences between allometric relation¬ ships, and (3) differences due to taxonomic, dietary and habitat characteristics of groups of species. The latter characteristics cannot be expressed as numerical val¬ ues, making multiple regression procedures inappropriate for our purposes. BODY SIZE (MASS) FMRs of birds studied to date range from about 27 kJ/day for hummingbirds to over 4730 kJ/day for Giant Petrels (Nagy 1987; Table 1), a range of 175-fold. Differences in body mass (as log Mb) account for about 91% of the variation in log FMR (r2 = o 907 for the log10-log10 regression of FMR upon body mass for the 50 data points on 25 species available in 1987). The present availability of data points for 15 additional species affords us an opportunity to address three questions. First, how well does the equation: kJ/day = 10.9 g0 64- based on the 1987 data, predict the FMRs of the re¬ cently-studied species? Second, does addition of the new results to the allometric relationship for birds yield a different equation than that for 25 species? Third does the addition of 15 more data points “explain” more variation (i.e. does r2 increase)? ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 795 TABLE 2 - Relative field metabolic rates (FMRs, adjusted for body mass effects) of 40 species of birds (from Nagy 1987 and Table 1). Common names from Clements (1978). Common name L°g,o mass FMR, % pred.a ord.b diet0 hab. Anna’s Hummingbird 0.65 94 AP N CS Grey-breasted Silvereye 0.98 91 PA F EF Coal Tit 0.98 103 PA r CF Crested Tit 1.05 80 PA r CF Willow Tit 1.05 80 PA r CF Pacific Swallow 1.15 129 PA r TF Sand Martin 1.16 137 PA r TM Savannah Sparrow 1.27 112 PA 0 SM House Martin 1.28 111 PA r TM Swallow 1.31 138 PA r TM Tree Swallow 1.35 148 PA r TM Phainopepla 1.36 98 PA 0 D Blue-throated Bee-eater 1.53 72 CO r TF Wilson’s Storm-petrel 1.63 99 PR c* M Leach’s Storm-petrel 1.65 71 PR c* M Mockingbird 1.68 94 PA 0 DF Purple Martin 1.69 124 PA r DF Dipper 1.80 143 PA r TM Pied Kingfisher 1.88 118 CO c* TF Starling 1.90 151 PA 0 DF Least Auklet 1.92 207 CH c* M South Georgia Diving Petrel 2.04 211 PR c* M Common Diving Petrel 2.14 220 PR c* M Gambel’s Quail 2.16 34 GA 0 D Sooty Tern 2.27 78 CH c* M Sand Partridge 2.28 47 GA 0 D Brown Noddy 2.29 111 CH c* M Eurasian Kestrel 2.34 100 FA c** TM Wedge-tailed Shearwater 2.58 125 PR c* M Black-legged Kittiwake 2.59 185 CH c* M Chukar 2.60 52 GA 0 D Black Guillemot 2.62 123 CH c* M Thick-billed Murre 2.92 183 CH c* M Little Penguin 3.04 104 SP c* M Laysan Albatross 3.49 97 PR c* M Jackass Penguin 3.50 103 SP c* M Grey-headed Albatross 3.57 115 PR c* M Adelie Penguin 3.59 186 SP c* M Giant Petrel 3.61 200 PR c* M Wandering Albatross 3.92 93 PR c* M a Field metabolic rate, as percent of that predicted from body mass (in g) from the equation: FMR (in kJ/day) = 10.9 g067 (Nagy 1987). b Order- AP = Apodiformes, PA = Passeriformes, CO = Coracnformes, PR = Procellaruformes, CH = Charadriiformes, GA = Galliformes, FA = Falconiformes, SP = Sphenisciformes. c Diet- N = nectar. F = fruit, I = insect, O = omnivore, C = carnivore, * = wide-foraging predator, ** = sit-and-wait predator. d Habitat' CS = chaparral scrub, EF = eucalypt forest, CF = coniferous forest, TF = tropical forest, TM = temperate meadow, SM = salt marsh, D = desert, M = marine, DF = deciduous forest. 796 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGIC! To test the predictive capability of the 1987 equation, we solved it using the mean body mass for each new species, and then calculated the 95% Cl of the prediction (which is much larger than the 95% Cl of the regression, due to the increased uncer¬ tainty expected for new data points; see Dunn and Clark 1974, and Figure 5 and Dis¬ cussion in Nagy 1987 for more details). Comparison of predicted with measured FMR values revealed that all 15 new measurements fell within the 95% CIs of the predic¬ tions. The new measurements averaged 133% of predicted FMR values (range 71 to 220% of predicted). Thus, the predictive capability of the 1987 equation is good. 30 CD cti o rH r— H O rO Cti +-> CD £ XI i— i CD • rH Ptn 250 • from Nagy (1987) ▲ new data A* 200 - • XJ ▲ • CD ■+-> O • rH 150 ▲ • XI • • ▲ CD • • ▲ •A Qh • m A 100 _ £ _ , A m *+— 1 o • • • m - m • fc* % • A • 50 • • • 0 - 1_ i j_ t • i ■ i 1 10 100 1000 10000 Body mass, g 25 20 15 10 0 CD o’ I 50 4* a FIGURE 1 - Mass-independent field metabolic rates of 40 species of birds. Horizontal axis is a logarithmic scale. Closed circles represent data taken from Nagy (1987), and filled triangles represent new data published subsequently (Tablel). Does the addition of new data change the allometric relationship for FMR in birds? The 1987 equation included multiple data points for some species, in order to incor¬ porate as much variation in the regression as actually occurs in different cohorts or between seasons within species. To simplify comparison of old and new data, we re¬ calculated the old regression using only species means for body mass and FMR (n = 25). The resulting equation, kJ/day = 10.1 g0 66, does not differ significantly from the 1987 equation, judging by the observations that the intercept (10.1, 95% Cl = 6 5 to 15.5) falls within the 95% C! of the 1987 intercept (10,9, 95% Cl = 8.1 to 14. 7) and the slope (0.66, 95% Cl = 0.58 to 0.74) falls within the 95% Cl of the 1 987 slope (0 64 95% Cl = 0.58 to 0.70). The addition of data points for 15 new species yields the equation kJ/day = 10.4 g 0 67 (N = 40, 95% C! of intercept = 7.5 to 14.6, 95% Cl of slope = 0.60 to 0.73). The intercept and slope of this equation fall within the 95% C!s ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 797 of both of the above equations. Thus, the new data support the descriptive capabil¬ ity of the 1987 equation. Does addition of the new data improve the regression equation’s “explanatory' prop- erty (higher r2 value)? The r2 values are 0.907 (1987 equation, 50 points for 25 spe¬ cies), 0.923 (1987 data, 25 points for 25 species), and 0.919 (40 points for 40 spe¬ cies). Assuming that an r2 value reflects the amount of variation in y explained by variation in x, then all three regressions indicate that variation in log body mass ac¬ counts for 91 to 92% of the variation in log field metabolic rate among species of birds. The addition of data for 15 new species did not substantially increase r2, indi¬ cating that the 1987 equation accurately reflects variation in FMR among bird species. RESIDUAL VARIATION Much variation still remains in FMRs of birds after accounting for body mass influ¬ ences. We recalculated FMR values by dividing by g0 64 to “correct” for size effects. If differences in mass completely accounted for differences between species, all val¬ ues for kJ g 0 64 day1 would be 10.9. In fact, mass-independent FMRs range from 34 to 220% of this value (Figure 1). Thus, the 175-fold variation in whole-animal FMR is largely explained by differences in body mass, but a six-fold variation remains unex¬ plained. We examined this residual variation for trends associated with taxonomic group (order), diet and habitat, in an attempt to improve our accuracy in predicting FMRs of birds. We tested for a difference between the mean mass-independent FMR for a given group and the predicted value of 10.9 kJ g 0 64 d 1 by determining whether the 95% Cl for a group included the predicted value, and by means of a t-test. Results are shown as percent of the predicted value. Phylogeny Passeriform birds (n = 14 spp.) had mass-corrected FMRs averaging 116% (95% Ci = 102 - 130%) of the predicted value. This difference is just barely significant statis¬ tically (P = 0.05). Similarly, the 7 species of charadriiform birds studied also had sig¬ nificantly higher than expected FMRs, with a mean of 145% (95% 01 = 101 - 189%). However, the FMRs of procellariiform birds (8 spp.) averaged 138% (95% Cl = 87 - 189%) of predicted, which is not a statistically significant difference. With a very small sample size (n = 3), galliform FMRs averaged only 44% of predicted (95% Ci = 21 - 67%), the lowest of any taxon. Diet Insectivorous birds (115% of predicted FMR, 95% Cl = 96 — 134%, 11 species) and omnivorous birds (84%, 95% Cl = 45 - 123%, n = 7) did not differ significantly from predicted, but carnivorous birds had higher FMRs (136% of predicted, 95% Cl = 113 159%, n = 20). Among the insectivores and carnivores, those we judged to be widely-foraging species had high FMRs (132% of predicted, 95% Ci = 115 - 149%, n = 28), whereas the three “sit-and-wait” (ambush predator) species had FMRs typical of other birds (97%, 95% Cl = 39 - 155%). 798 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI Habitat Forest-dwelling species did not have unusual FMRs (mean = 103% of predicted, 95% Cl = 87 - 120%, n = 11), but meadow- and marsh-dwelling species had relatively high FMRs (127%, 95% Cl = 112 - 144%, n = 7). Terrestrial nondesert birds (forest, meadow and marsh habitats taken together) had moderately high FMRs (1 13%, 95% Cl = 101 - 125%, n = 18). Desert birds tended to have low FMRs (but not significantly lower than average: 58% of predicted, 95% Cl = 14 to 102%, n = 4), whereas seabirds had high FMRs (140%, 95% Cl = 115 - 165%, n = 18). DISCUSSION Body mass clearly has the largest influence on the field metabolic rate of birds. Once size effects are accounted for, however, much variation (6-fold) still remains. Although the above analyses suggest that phylogeny and ecology may have moderate (ca. 3- fold) effects on FMR, they do not satisfactorily account for the large residual variation in mass-corrected FMR. Several reasons are probably involved. First, current sam¬ ple sizes within categories are perhaps too small to demonstrate important influences. Moreover, strong biases can emerge with small samples, e.g. four of seven species from meadow habitats are swallows. Second, these broad categories incorporate measurements that include variation due to season, gender or behavior. The few stud¬ ies that address these factors within a single species indicate that interseasonal vari¬ ations in FMR range from 1 .2- to 1 .5-fold, that FMRs at various phases of the breed¬ ing season may differ up to 2.5-fold, and that differences due to gender may be 1.3- fold. Third, categorizations of diet (e.g. “omnivore”) and habitat (e.g. “the sea”) are very broad and may be useful only insofar as they accurately describe the species’ ecology at the time of measurement. Finally, ecological and phylogenetic categories often overlap, such that the independ¬ ent effects of each cannot easily be sorted out: nearly all carnivorous birds measured are seabirds; the passerines studied to date are nearly all insectivores or insect-eating omnivores. Other factors worth investigating include ambient temperature, social or¬ ganization, flight mode and duration, food availability, and predator pressure. At present, it appears that no single factor accounts for all, or even much, of the re¬ sidual variation in FMR that is not accounted for by body mass. Several of'these fac¬ tors are probably involved in determining a bird’s FMR. For example, the four species having the highest mass-corrected FMRs (the Least Auklet, the South Georgia Div¬ ing Petrel, the Common Diving Petrel, and the Southern Giant Petrel) are all carnivo¬ rous seabirds that live at high latitudes and spend much time flying and/or have high wing loadings. The three species with the lowest mass-corrected FMRs (Gambel’s Quail, Sand Partridge, and Chukar Partridge) all live in warm deserts, fly infrequently are in the order Galliformes, and are omnivorous (but primarily granivorous). It may eventually become possible to identify and quantify the major factors influencing FMR in enough detail to permit accurate prediction of FMR in birds (e.g. within 10-15%) but present information provides predictive equations (Nagy, 1987) that are only ac¬ curate to between -50% and +100% (95% Cl of predicted kJ/day value). ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 799 ACKNOWLEDGEMENTS Support for this review was provided by US Department of Energy Contract DE-AC03- 76-SF0001 2. LITERATURE CITED BRYANT, D.M., HAILS, C.J., PRYS-JONES, R. 1985. Energy expenditure by free-living Dippers ( Cinclus cinclus) in winter. The Condor 87: 177-186. BRYANT, D.M., TATNER, P. 1988. Energetics of the annual cycle of Dippers Cinclus cinclus. Ibis 130: 17-38. CLEMENTS, J.F. 1978. Birds of the world: a check list. New York, Two Continents Publishing Group. COSTA, D.P., DANN, P., DISHER, W. 1986. Energy requirements of free ranging Little Penguin, Eudyptula minor. Comparative Biochemistry and Physiology 85A: 135-138. COSTA, D.P., PRINCE, P.A. 1987. Foraging energetics of Grey-headed Albatrosses Diomedea chrysostoma at Bird Island, South Georgia. Ibis 129: 149-158. DUNN, O.J., CLARK, V.A. 1974. Applied statistics: analysis of variance and regression. New York, John Wiley and Sons. MASMAN, D., DAAN, S., BELDHUIS, H.J.A. 1988. Ecological energetics of the Kestrel: daily energy expenditure throughout the year based on time-energy budget, food intake and doubly labeled water methods. Ardea 76: 76-81. MORENO, J., CARLSON, A., ALATALO, R.V. 1988. Winter energetics of coniferous forest tits Paridae in the north: the implications of body size. Functional Ecology 2: 163-170. NAGY, K.A. 1987. Field metabolic rate and food requirement scaling in mammals and birds. Ecologi¬ cal Monographs 57: 111-128. NAGY, K.A. 1989. Doubly labeled water studies of vertebrate physiological ecology, Pp. 270-287 in Stable isotopes in ecological research. Rundel, P.W., Ehleringer, J.R. Nagy, K.A. (Eds). New York, Springer-Verlag. REYER, H-U., WESTERTERP, K. 1985. Parental energy expenditure: a proximate cause of helper recruitment in the Pied Kingfisher (Ceryle rudis). Behavioral Ecology and Sociobiology 17: 363-369. RICKLEFS, R.E., ROBY, D.D., WILLIAMS, J.B.. 1986. Daily energy expenditure by adult Leach’s Storm-petrels during the nesting cycle. Physiological Zoology 59: 649-660. ROBY, D.D., RICKLEFS, R.E. 1986. Energy expenditure in adult Least Auklets and Diving Petrels during the chick rearing period. Physiological Zoology 59: 661-678. WILLIAMS, J.B. 1988. Field metabolism of Tree Swallows during the breeding season. Auk 105: 706- 714. 800 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI SCALING AVIAN ECOLOGY WILLIAM A. CALDER1 and CYNTHIA CAREY 2 1 Department of Ecology and Evolutionary Biology, University of Arizona, Tucson, Arizona 85721 , USA 2 Department of Environmental, Population and Organismic Biology, University of Colorado, Boulder, Colorado 80309-0334, USA ABSTRACT. Interspecific scaling patterns derived from laboratory data have facilitated analysis of many physiological and morphological variables of both embryonic and adult birds. Some of these variables (e.g. metabolic rates, turnover times) have ecological parallels (e.g. home range, longevity). Correlations for scaling relationships calculated from organismic data collected in the laboratory are generally much higher than are ecological ones; ecological relationships are complicated by a number of uncontrollable variables like spatio-temporal variation in resources, size-dependent species diver¬ sity, and scaling biases in research techniques. Nevertheless, it is theoretically possible and practically desirable to use scaling methodology for predicting ecological requirements, for analyzing resource use and ecological time scales and for conservation purposes. INTRODUCTION Body size is an excellent predictor of many physiological and morphological charac¬ teristics of embryonic and adult birds (Paynter, et al. 1974, Rahn & Paganelli 1981, Calder 1984, Rahn et al. 1985, Walsberg 1987, Nagy 1987, Vleck & Vleck 1987). Allometric equations formulated from data derived from hummingbirds to ostriches have fostered considerable research by identifying general patterns and highlighting those species which do not fit the trends. The use of scaling on the ecological level has received much less study, but scaling is potentially useful for avian ecoloqy in many ways: a) The most obvious use is simply to render the data manageable and to obtain em¬ pirical summaries. In a less than perfect situation, scaling allows pattern analy¬ sis of spotty but available data. b) A significant correlation arising from scaling is a generalization which might sug¬ gest sources of variation after factoring out body size. c) Predictions from scaling equations are first approximations for estimating ecologi¬ cal requirements, for designing experiments with species for which actual meas¬ urements are now missing, or for determining what reserve size is needed to pre¬ vent a species from extinction. d) Comparisons of allometric generalizations derived separately may suggest i) the ways in which internal function is adapted to environmental factors and commu¬ nity structure, and ii) how physiological capacity has responded to natural selec¬ tion in the face of biotic and abiotic realities. Notably and intentionally missing from the possible applications listed above is estab¬ lishment of causality and theory. An existing bird is an outcome, an adaptive suite of form, function, and iife history, for which we cannot deduce anything about the genetic basis of scaling of body proportions and life history variables. Ornithology has been ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 801 advanced considerably by documentation of patterns generated by least squares re¬ gressions (LSR) of log-transformed data on log body mass. The past decade focused on procedures: how we should measure size, which data we should include, which model of regression we should use. These debates provided valuable insight about statistics, taxonomic effects, and interpretations (Harvey 1982, Harvey & Mace 1982, Clutton-Brock & Harvey 1984, Smith 1984a,b, Bennett & Harvey 1987, Read & Harvey 1989. See LaBarbera 1989 for a provocative review). Prudence and caution dictate the need to refine statistical approaches. However, considerable effort may produce refinements which exceed that warranted, given the coarse nature of the original data, which tend to exclude valid but unpaired data, and which inhibit speculation and hy¬ pothesis formulation. Awareness of limitations to application should not preclude pro¬ ductive use of biological scaling techniques. In this paper we explore aspects of the extent to which scaling might be fruitfully ap¬ plied at the population or community level, which has considerably more variables than can invade the controlled environment of a laboratory. METHODS Body mass (m, in g) is the measure of body size in the scaling equation, Y = a mb. We used one or an average value for each measured variable Y per species (n = number of species). The exponential scalings (b) were derived from log-transformed data. Least-squares regression (LSR, Model I) analysis yields a straight line with slope b, the proper form for predictive applications (Sokal & Rohlf 1981 , pp. 547-549). LSR should be used with caution unless body mass is assumed to be an essentially error- free, independent variable. LSR and reduced major axis (RMA, Model II) scaling lines are identical when the correlation coefficient r is 1 . As the correlation weakens, the two diverge. The RMA slope b is “the ratio of the standard deviations of points meas¬ ured on the y axis and on the x-axis”, sy/sx (Harvey & Mace 1982), while the LSR b is the product r(s /sx) (Snedecor & Cochran 1967). Therefore the RMA scaling expo¬ nent can be determined from the LSR as RMA b = LSR b/r. RMA “assumes that error variance is the same proportion of the total variance on each axis, which may be closer to reality . . .” (Clutton-Brock & Harvey 1984) and has the least bias (Rayner 1985). However, “. . . both estimation of functional relationships and prediction are carried out best by means of simple linear regression, but when both variables are subject to error the appropriate method depends on the nature of the data... The subject of Model II regression is one which research and controversy are continuing and definitive recommendations are difficult to make. Much will depend on the intentions of the investigator.” (Sokal & Rohlf 1981, pp. 547-549; see also Seim & Saether 1983). If preservation of biotic diversity is the concern and specific data are lacking, predictions [for which Model I or “least squares regression” (LSR) is best] may offer a first approximation. Rising & Somers (1989) decided body mass was inferior to lengths or principle com¬ ponent analysis for expressing body size, on the basis of larger coefficients of vari¬ ation (CV) for mass than for any of several linear measurements from a singie spe¬ cies’ body form. However, mass, like volume, is the product of three linear dimen¬ sions, so its CV would be expected to be compounded to about three times the 802 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI TABLE 1 - Correlation Strength and Level of Regression. FUNCTION SCALING R FUNCTION SCALING R MAMMALS Body surface area 0.67a 0.99 Home range area 1.02b 0.93 Gut contents 1.05cH 0.99 Fieldmet. rate 0.81 d 0.98 Metabolic rate 0.76b 0.98 Population density -0.786 -0.80 Max heart rate -0.19' -0.99 Intrinsic rate of BIRDS (ADULTS) increase -0.269 -0.69 Body area 0.67h 0.998 Home range area 1.17' 0.793 Resting metabolic Field metabolic Rate 0.67' 0.99 Rate 0.64d 0.95 Feather density -0.27k 0.905 Population density -0.491 0.42 Lifespan in captivity 0.1 9b 0.70 Lifespan in nature 0.20bS 0.78 BIRDS (EGGS) 0.1 4>G 0.52 Conductance to water vapour 0.81 m 0.97 Waterloss 0.75p 0.93 Shell thickness 0.46m 0.97 Incubation time 0.22m 0.73 Pre-pipping 02 consumption 0.71° 0.97 Data in the right and left columns were collected in the field and laboratory, respectively Sources : a Peters 1983; b Calder 1984; c Demment 1983; d Nagy 1987; e Damuth 1987 ' Baudinette 1978; a Henneman 1983; h Walsberg & King 1978; 1 Calder 1990b; i Bennett & Harvey 1987- k Weltv 1982, Pettingill 1985; 1 Juanes 1986; m Ar &Rahn 1978; n Ar et al. 1974- 0 Hoyt & Rahn 1980- p Ar l Rahn 1980. ’ ’ Qualifiers: H = herbivores; G = 215 longest-lived by genus; s = 152 longest-lived species. average CV for a linear measurement. In fact, the CV for mass was less than such a value (Strauss & Calder MS). Lengths, judging from the diversity often represented in allometric regressions (e.g. duck to darter, Anhinga), would not likely predict meta¬ bolic requirements as well as masses do. Examples of tightness of correlation, comparing r values from captive and field meas¬ urements, appear in Table 1. Damuth (1981,1987) reported that population densities of 467 species of mammals were proportional to M 0 78, with r = -0.80. The scalinq was similar in magnitude but opposite in sign from basal metabolism (M0 76, r = 0 98) meas¬ ured in laboratories. For birds, the difference in correlations for density and metabo lism is more dramatic: r = -0.42 and 0.99, respectively, but such results are due to slope differences, (higher b giving higher r) and to scatter in the data. Body surface ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 803 areas of mammals and birds (M0 67) correlate more tightly with body mass (r = 0.99) than is the case for land surface areas which they occupy (mammals: r = 0.87, birds, r = 0.79). Thus confounding variables could have a differential effect on RMA scalings, although this is not clear for many cases, perhaps due, in part, to effects of sample size and the fact that r is influenced by regression slope (r is equal to b divided by the complex denominator Sy/Sx). The residuals remind us that regression cannot tell it all - environmental adaptation and history have their influences. The correlation coefficient (r) for scaling might be expected to decrease as ever more variables come into play as the scaling technique is employed from egg and bird to avian ecology. For example, there are natural spatial and annual variations in produc¬ tivity with consequences for resource availability and carrying capacity, niche speciali¬ zation, size- dependent species diversity, anthropogenic disruption, errors in meas¬ urement and calculation, seasonal hormonal states, and local adaptations. Better correlations are often obtained by regressing life history variables on each other directly, bypassing body size (Partridge & Harvey 1988, Read & Harvey 1989, Harvey, this volume). However, in some situations or types of study, body mass scal¬ ing is still the most practical approach for dealing with: 1) Evolution and consequences of size (e.g. Bergmann’s and Cope’s rules, commu¬ nity and fossil assemblages) and first approximations to variables for species lack¬ ing actual measurements. 2) If not of primary interest, size is factored out via ratios of observed values to size- predicted values from general allometry, and 3) Limited data sets, too few species with data for both variables of interest: regres¬ sions can be run with species hypothetical A, B, C, D, and E (distributed over a good range in size) on mass for variable 1, and on species D-J on mass for vari¬ able 2. A ratio of scalings (Stahl 1962) in two or more regressions gives a first approximation to the relationship (see example of sound and area, below). This is, of course, compoundedly crude because of variances around both slopes (LaBarbera 1989), but central tendencies provide something more for thought than mere failure to find patterns through rigid adherence to ideal criteria which cannot be met with data presently available. SCALING OF RESOURCE USE Home areas Space use and claims are important ecological characteristics of birds. Schoener (1968) found that territory scaled as the 1.09 power and home range as the 1.16 power of body mass. Adding nine territory sizes of other species from the literature does not change the basic scaling: area, ha = 0.033 m1 17 (1) (m = mass, range 3.2 to 4500 g; n = 86; SE of b = ±0.098; r=0.793; F = 143.7, P<0.0001 Calder 1990a). 804 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI Reserve size and population size Multiplying the area scaling by some number for minimum viable population size for species survival (e.g. the “500” of Frankel & Soule 1981) might be one “back of an envelope” approach to preliminary estimation of minimum reserve size: 500 x 0.033 m1 17 (7) This predicts a minimum of 244 ha for a hypothetical 10 g species, 3610 ha for a 100 g species, and 53,393 ha for a 1 kg species, estimates perhaps no more tenuous than the number 500 itself. Song power output Birds vocalize to claim space and resources, to warn intruders, and to attract mates. Singing conserves energy by reducing the need for flight and direct confrontation. Sound output power should have implications for ecology, behavioral energetics, and community structure. Direct correlation would be preferable, but with too few species with data on both home or territorial areas and sound outputs, separate body size regressions were necessary. The relationship of milliwatts of sound output power (corrected to 1 m distance) to body mass (range 6 to 10800g) was: power output, mW = 0.042 m1 14 (2) (SE of b = + 0.084; F =185.5, P < 0.001.). Body size accounts for 86.1% of the vari¬ ability in sound output (r = 0.928; Calder 1990a). The mass scalings for area (b = 1.17) and sound output (b = 1.14) are statistically indistinguishable (P > 0.25; F-test), so mass can be factored out by combinina eaua- tions (1 ) and (2): sound (mW)/ area (ha) = 1.28 nr0 03 ( 3) Thus vocal capacity and spatial requirements appear to have evolved together, with a size independent of body size, resulting in a ratio of 1.3 mW of sound output at 1 m distance per ha. The scaling of equation (2) can be used with the inverse square rule to calculate the radial distance from a song perch at which the intensity would fall to a minimal audi¬ bility, say 30 dB (20 dB is a whisper). Under “ideal” conditions (no attenuation or sig¬ nal degradation due to wind, turbulence, or vegetation, in an environment completely free of sounds other than birdsong) the census taker with perfect hearing would hear a hypothetical 1 kg bird within a 4.2 km radius, while a hypothetical 10 g bird would have to be within 304 m to be heard. Thus, the area from which a census taker could hear all the less vociferous 10 g birds would be only 2 per cent of the area from which all 1 kg birds could be heard. The Breeding Bird Survey was developed “to detect population trends within species from year to year” (Bystrak 1981). However, “species differ in how easy they are to see or hear, so the effective area sampled from a point or a line will differ between ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 805 species. Each species is therefore measured on a separate scale, and the counts of different species may not be added together or used in species-diversity calculations . . (Dawson 1981). Caution should be exercised in estimating population densities from transect censuses that depend heavily on calls and song. Population densities Much time and energy must be invested to obtain accurate population density data. Damuth (1981) combined allometries of herbivorous mammalian densities (D) and basal metabolic rates (E) to obtain a relationship for population energy requirements. Since the exponents were of opposite sign but numerically identical (D = k M 75; E = k M+0 75), the scalings canceled, suggesting that community net primary productiv¬ ity (NPP) was apportioned independent of consumer size. This calculation overlooks probable biases of studying species that are more abundant in local distribution (Brown & Mauer 1986, 1987, Lawton 1989) and the fact that species diversity is in¬ versely related to body size (more small species than medium-sized ones, and more mediums than large (Van Valen 1973, May 1978, Calder, 1984 p. 298, Brown & Mauer 1986, Lawton 1989). Scaling of D in birds is a much shallower inverse relationship with weaker correlations than in mammals (Peters & Wassenberg 1983, Juanes 1986, Brown & Maurer 1986). The product of these Ds and basal metabolic scaling yielded positive exponents (b > 0), indicating that large birds consumed more of NPP than small species. These calculations overlook the additive effects when the environment is utilized at a finer grain by a greater number of species in smaller body size ranges (Maurer & Brown 1988, Lawton 1989). For example, in the compilation of North Ameri¬ can bird weights (Dunning 1984), more than half (57%) are under lOOg, a weight that is 8% of that of the largest passerine and under 1% of the largest non-passerine. Lawton (1989) argued “. . . it is impossible to say what the real relationship between population density and body mass is; or rather what the relationships are . . .”. He distinguished between upper and lower bounds in and to scatter plots of log density vs. log body size, and pointed out that conclusions about per capita use of resources were based on upper bounds, the scaling of which would be valid only if the lower bounds were parallel. Another matter of concern is that density and diversity tend to fall off from the inverse trend at the small end of the size range within a class or smaller grouping (Dial & Marzluff 1988, Maurer & Brown 1988). We feel it is premature to abandon examination of density/size relationships before adequate true densities are available. It seems desirable that we explore further and improve knowledge of such bounds with awareness of errors that could come from improper interspecific use of linear transect counts pointed out above. Since spatial scales (environmental grain) are perceived in a size-dependent way, ecological den¬ sity is based on area of intense use. The reciprocal of home range or territory has the dimensions of D [1 /(area/bird) = (birds/area)], and thus better approximates ecologi¬ cal density. It is dimensionally correct for birds that are dispersed in space (not flocked or colonial, seasonally or by nature). While this overlooks the existence of “floaters” or birds who have not succeeded in staking territorial claims, so do linear transects that utilize only the conspicuous songs, posting, and activities associated with own¬ ership; 0 = 1/ 0.033 m1 17 = 30.3 nr1-17 (4) 806 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI The field metabolic rates (FMRs) of birds (Nagy 1987) are scaled: E = 10.9 m 0 64 ( 5) This is similar to an earlier derivation by Walsberg (1983): E = 13.1m061 (6) The products, DE, scale as m 0 53 to'0 56. Smaller, but still negative exponents are found when passerine and non-passerine birds are analyzed separately (passerine -0.35; nonpasserine -0.08). When analysis is limited by phylogeny, the smaller range in bird sizes used in the analysis reduces the strength of correlation. Hence it may be that small birds take more, not less than the large, of the NPP, par¬ ticularly if the size-dependency of species diversity is taken into account. Avian com¬ munity partitioning of NPP deserves further study, but a better data base of true popu¬ lation densities is probably a prerequisite for success in this. ECOLOGICAL TIME SCALES A life history is comprised of many time periods. They scale with the following mass exponents: from heartbeat to heartbeat (0.23), incubation (0.20), fledging time (0.18 to 0.20), sexual maturity (0.23), life expectancy (0.46), and maximum lifespan (0.14 to 0.20) (Calder 1984, 1990b). We will examine scaling relationships which bear on two major emphases of ecology: foraging and reproduction. Digestion time Karasov et al. (1986) showed that periodic cessation of foraging activity by energy-maximizing hummingbirds could be due to limitations in rate of digestive processing. Determination of gut passage or retention times is complicated by differ¬ ences in passage of solids varying in digestibility, laxitivity, and fluid components of the diet. Owing to paucity of measurements these distinctions are ignored to make a first approximation of the scaling of gut transit time from intake to first appearance of stained material in the faeces (Figure 1): “gut passage time”, min. = 23.7 m031 (8) [n = 36 species, 3.5 to 38,000 g; r = 0.751; F = 43.90, P < 0.005; SE of b = 0.046; data from Gasaway et al. (1975), Walsberg (1975), Sturkie (1976), Herd & Dawson (1984), Herrera (1984), Karasov et al. (1986)]. This scaling exponent is statistically indistinguishable from the scaling ratio for gut size/metabolic rate in birds, and from the 0.28 to 0.30 derived for ruminant and nonruminant mammals (Demment 1983, Demment & Van Soest, 1985). Such scaling is consistent with the apparent lower size limit for microbial fermentation of crude fibers (a minimum of 2 h for the 425 g ptar¬ migan Lagopus mutus), indicative of size constraints on food habits and feeding behavior. Rapid digestibility, not energy density (J/g wet), is the primary considera¬ tion in the diet of very small birds; grass is about as rich (J/g) as the average floral nectar. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 807 Longevity Maximum longevity (tmax) scales with a somewhat smaller body mass exponent than the 1/4 scaling of metabolic turnover and other physiological and life-history scalings (Calder 1984). Lindstedt & Calder (1976) suggested that avian tmax data could have been biased by the fact that small birds with shorter lifespans had been more exten¬ sively sampled in the few decades of banding and recaptures than large birds. Klimkiewicz & Futcher (1989) extracted t records of 498 species of wild banded birds from the computer files of the US Fish and Wildlife Service Bird Banding Labo¬ ratory, through 1987. Limited sample sizes contributing the oldest records probably understate considerably the ages actually attained for each species. Typically less than 10% of banded birds are subsequently recaptured, collected, or found dead, except for colonial-nesting marine birds, and hunted species of fowl, waterfowl, and doves. Intense banding/recapture efforts are necessary to obtain longevity records which actually reflect life’s limits. Attempting to eliminate the understudied, Calder (1990b) used only the maximum longevity within each of 215 genera represented. Mass from Dunning (1984) accounted for 27% of variance: tmax, years = 5.45 m014 (9) GUT PASSAGE TIMES OF BIRDS LOG BODY MASS (GRAMS) FIGURE 1 - Times for passage of ingested food through bird digestive tracts. N = nectar- feeder, S = seed-eater, O = frugivore, X = other non-frugivore. 808 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI (r = 0.518, P <0.001, SE of slope 0.016). This predicts shorter tmax for all birds over 33g in body mass, compared to Lindstedt & Calder’s (1976) equation for wild birds. Total numbers of bandings from which the records could have been sampled, as re¬ ported for 192 of these species by Clapp et al. (1982,1983) and Klimkiewicz et al. (1983, 1987) were regressed as log number of bandings per species vs. log body mass to check for any systematic bias towards smaller birds, more likely to be caught by back-yard banders. The correlation between banding frequency and bird size was weak and statistically insignificant (r = 0.014). Similarly, tmax did not correlate signifi¬ cantly with number of birds banded (r = 0.038). Hence, we have no explanation for why these tmax scaling exponents are consistently less than 1/4. The addition of over two decades’ worth of captive mammal and wild bird longevity records has not shown a significant tmax increase, suggesting that we may be approaching a point of dimin¬ ishing returns in derivation of t scaling. LITERATURE CITED AR, A., PAGANELLI, C.V., REEVES, R.B., GREENE, D.G., RAHN, H. 1974. The avian egg: water vapor conductance, shell thickness, and functional pore area. Condor 76: 153-158. AR, A., RAHN, H. 1978. Interdependence of gas conductance, incubation length, and weight of the avian egg. Pp. 227-236 in Piiper, J. (Ed.) Respiratory function in birds, adult and embryonic. Berlin, Springer-Verlag. AR, A., RAHN, H. 1980. Water in the avian egg: overall budget of incubation. American Zoologist 20: 373-384. BAUDINETTE, R. U. 1978. Scaling of heart rate during locomotion. Journal of Comparative Physiol¬ ogy 127: 337-342. BENNETT, P.M., HARVEY, P.H. 1987. Active and resting metabolism in birds: allometry, phylogeny and ecology. Journal of Zoology, London 213: 327- 363. BROWN, J.H., MAURER, B.A. 1986. Body size, ecological dominance and Cope’s rule. Nature 324: 248-250. BROWN. J.H., MAURER, B.A. 1987. Evolution of species assemoiages: effects of energetic constraints and species dynamics on the diversification of the North American avifauna. American Naturalist 130: 1-17. BYSTRAK, D. 1981. The North American breeding bird survey. Studies in Avian Biology 6: 34-41. CALDER, W.A. 1984. Size, function, and life history. Cambridge, USA Harvard University Press. CALDER, W.A. 1990a. The scaling of sound output and territory size: are they matched?. Ecology 71 : 1810-1822. CALDER, W.A. 1990b. Avian longevity and aging. Pp. 185-204 in Harrison, D.E. (Ed.) Genetic effects on aging II. Caldwell, N.J., Telford Press. CLAPP, R.B., KLIMKIEWICZ, M.K., KENNARD, J.H. 1982. Longevity records of North American biras: Gaviidae through Alcidae. Journal of Field Ornithology 53: 81-124. CLAPP, R.B., KLIMKIEWICZ, M.K., FUTCHER, A.G. 1983. Longevity records of North American birds: Columbidae through Paridae. Journal of Field Ornithology 54: 123-137. CLUTTON-BROCK, T.H., HARVEY, P. H. 1984. Comparative approaches to investigating adaptation. Pp. 7-29 in Krebs, J.R., Davies, N.B. (Eds), Behavioural ecology: An evolutionary approach. Oxford, Blackwell. DAMUTH, J. 1981. Population density and body size in mammals. Nature 290: 699-700. DAMUTH, J. 1987. Interspecific allometry of population density in mammals and other animals: the in¬ dependence of body mass and population energy-use. Biological Journal of the Linnean Society 31: 193-246. DAWSON, D.G. 1981. Counting birds for a relative measure (index) of density, pp. 12-16 in Ralph, C.J Scott, J. M. (Eds) Estimating numbers of terrestrial birds. Studies in Avian Biology 6. DEMMENT, M.W. 1983. Feeding ecology and the evolution of body size of baboons. African Journal of Ecology 21: 219-233. DEMMENT, M.W., VAN SOEST, P.J. 1985. A nutritional explanation for body size patterns of ruminants and non-ruminant herbivores. American Naturalist 125: 641-672. ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 809 DIAL, K.P., MARZLUGG, J.M. 1988. Are the smallest organisms the most diverse? Ecology 69: 1620- 1624. DUNNING, J.B. Jr. 1984. Body weights of 686 species of North American birds. Western Bird Band¬ ing Association Monogr. 1. FRANKEL, O.H., SOULE, M.E. 1981. Conservation and evolution. Cambridge, Cambridge University Press. GASAWAY, W.C., HOLLEMAN, D.F., WHITE, R.G. 1975. Flow of digestion in the intestine and cecum of the rock ptarmigan. Condor 77: 467-474. HARVEY, P.H.1982. On rethinking allometry. Journal of Theoretical Biology, 95:37-41. HARVEY, P.H., GODFRAY, H.C.J. 1987. How species divide resources. American Naturalist 129: 318- 320. HARVEY, P.H., MACE, G.M. 1982. Comparisons between taxa and adaptive trends: problems of meth¬ odology. Pp. 343-361 in King’s College Sociobiology Group (Eds). Current problems in sociobiology. Cambridge, Cambridge Univ. Press. HAUCK, M.A., GAUTHIER, J.A., STRAUSS, R.E. 1990. Allometric scaling in the earliest fossil bird, Archaeopteryx lithographica. Science in press. HENNEMAN, W.W. 111.1983. Relationship among body mass, metabolic rate and the intrinsic rate of natural increase in mammals. Oecologia, Berlin 56: 104-108. HERD, R.M., DAWSON, T.J. 1984. Fiber digestion in the Emu, Dromgius novaehollandiae, a large bird with a simple gut and high rates of passage. Physiological Zoology 57: 70-84. HERRERA, C.M. 1984. Adaptation to frugivory of Mediterranean avian seed dispersers. Ecology 65: 609-617. HOYT, D., RAHN, H. 1980. Respiration of avian embryos — a comparative analysis. Respiration Physi¬ ology 39: 255-264. JUANES, F. 1986. Population density and body size in birds. American Naturalist 128: 921-929. KARASOV, W.H., PHAN, D., DIAMOND, J.M., CARPENTER, F.L. 1986. Food passage and intestinal nutrient absorption in hummingbirds. Auk 103: 453-64. KLIMKIEWICZ, M.K., CLAPP, R.B., FUTCHER, A.G. 1983. Longevity records of North American birds: Remizidae through Parulinae. Journal of Field Ornithology 54: 287-294. KLIMKIEWICZ, M.K., FUTCHER, A.G. 1987. Longevity records of North American birds: Coerebinae through Estrildidae. Journal of Field Ornithology 58: 318-333. KLIMKIEWICZ, M.K., FUTCHER, A.G. 1989. Longevity records of North American Birds Supplement I. Journal of Field Ornithology 60: 469-494. LABARBARA. M. 1989. Analyzing body size as a factor in ecology and evolution. Annual Review of Ecology and Systematics 20: 97-117. LAWTON, J.H. 1989. What is the relationship between population density and body size in animals? Oikos 55: 429-434. LINDSTEDT, S.L., CALDER, W.A. 1976. Body size and longevity in birds. Condor 78: 91-94. MAUER, R.A., BROWN, J.H. 1988. Distribution of energy use and biomass among species of North American terrestrial birds. Ecology 69: 1923-1932. MAY, R.M. 1975. Patterns of species abundance and diversity. Pp. 81-120 in Cody. M L., Diamond, J. M. (Eds). Cambridge, Harvard University Press. NAGY. K.A. 1987. Field metabolic rate and food requirement scaling in birds and mammals. Ecologi¬ cal Monographs 57: 111-128. PARTRIDGE, L., HARVEY, P.H. 1988. The ecological context of life history evolution. Science 241: 1449-1454. PAYNTER, R.A. Jr. 1974. Avian energetics. Cambridge, Nuttall Ornithological Club Publication 15. PETERS, R.H. 1983. The ecological implications of body size. Cambridge, Cambridge University Press. PETERS R.H., WASSENBERG, K. 1983. The effect of body size on animal abundance. Oecologia 60: 89-96. PETTINGILL, O.S. 1985. Ornithology in laboratory and field, 5th ed. Orlando, Academic. QUIRING, D.P. 1950. Functional anatomy of the vertebrates. New York, McGraw-Hi!i. RAHN, H., PAGANELLI, C.V. 1981. Gas exchange in avian eggs: Publications in gas exchange, physi¬ cal properties, and dimensions of bird eggs. Buffalo, New York, SUNY Department of Physiology. RAHN, H., WHITTOW, G.C., PAGANELLI, C.V. 1985. Gas exchange of avian eggs- Vol. 11. Buffalo, New York,’ SUNY Dept, of Physiology. RAYNER J.M.V. 1985. Linear relations in biomechanics: the statistics of scaling functions. Journal of Zoology, (London). A 206: 415-439. READ A F., HARVEY, P.H. 1989. Life history differences among the eutherian radiations. Journal of Zoology (London) 219: 329-353. RISING J D., SOMERS, K.M. The measurement of overall body size in birds. Auk 106: 666-674 . SCHOENER, T.W. 1968. Sizes of feeding territories among birds. Ecology 49: 123-141. 810 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI SEIM, E., SAETHER, B.-E. 1983. On rethinking allometry: which regression model to use? Journal of Theoretical Biology 104: 161-168. SMITH, R.J. 1984a. Determination of relative size: the “criterion of subtraction” problem in allometry. Journal of Theorical Biology 108:131-42. SMITH, R.J. 1984b. Allometric scaling in comparative biology: problems of concept and method. Ameri¬ can Journal of Physiology 246 (Regulatory, Integrative, and Comparative Physiology 15): R152-R160. SNEDECOR, G.W., COCHRAN, W.G. 1967. Statistical methods. Ames: Iowa State University Press, p.177. SOKAL, R.R., ROHLF, F.J. 1981. Biometry: The principles and practice of statistics in biological re¬ search, 2nd Ed. New York, W.H. Freeman. STAHL, W.R.1 962. Similarity and dimensional biology. Science 137: 205-212. STRAUSS, R., CALDER, W.A. MS. Dimensional errors in the measurement of overall body size STURKIE, P.D. 1976. Avian physiology, 3rd ed. New York, Springer-Verlag. VAN VALEN, L. 1973. Body size and numbers of animals. Evolutioh 27: 27-35. VLECK, C.M., VLECK, D. 1987. Metabolism and energetics of avian embryos. Journal of Experimen¬ tal Zoology Supplement 1: 111-125. WALSBERG, G.E. 1975. Digestive adaptations of Phainopepla nitens associated with the eating of mistletoe berries. Condor 77: 169-174. WALSBERG, G.E. 1983. Avian ecological energetics. Pp. 161-220 in Farner, D.S., King, J.R. (Eds). Avian biology, Volume 7. Orlando, Academic Press. WALSBERG, G.E., KING, J.R. 1978. The relationship of the external surface area of birds to skin sur¬ face area and body mass. Journal of Experimental Biology 76:185-189. WELTY, J.C. 1982. The life of birds, 3rd ed. Philadelphia, Saunders. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 811 BODY SIZE AND FORAGING BEHAVIOUR IN BIRDS R. I.GOUDIE1 and J. F. PIATT 2 ’ Canadian Wildlife Service, P.O. Box 9158, St. John’s, Newfoundland A1 A 2X9, Canada 2U.S. Fish and Wildlife Service, Alaska Fish and Wildlife Research Center, 1011 E. Tudor Road, Anchorage, Alaska 99503, USA AB8TRACT. Foraging time in birds is an inverse function of body size. All else being equal, smaller species must spend more time feeding because of higher metabolic demands on a gram per gram basis (Kleiber’s rule). This can limit the flexibility of small species to adjust activity budgets to fluctua¬ tions in food abundance or environmental conditions. As well, greater energy reserves of larger spe¬ cies can promote a strategy of “deferred foraging” to await better environmental conditions. Selection for large body size would appear to be an evolutionary force in harsh (cold) environments (Bergmann’s rule). Keywords: Body size, time budgets, allometry, feeding behaviour, Kleiber’s rule, Bergmann’s rule. INTRODUCTION The relationship between metabolic rate and body size in birds follows Kleiber’s 3/4 rule (aM0 75, where M=mass; Kleiber 1961, Calder 1984). This well-known allometric relationship results from the thermodynamic consequences of body size. On a gram per gram basis, small animals must maintain a higher metabolism than large animals to compensate for greater heat loss to the environment (Peters 1983). Empirical data reveal that ingestion rates for birds and other taxa are similarly scaled with body size (aM0 63-M084, Peters 1983, Calder 1984) — presumably because metabolic rate largely determines food requirements. The volume of the digestive system of birds (gut ca¬ pacity), however, is a linear function (aM10) of body size (Quiring 1950, Calder 1984). Thus, energy demand (daily consumption) is greater on a gram per gram basis for small birds than for large ones — but the maximum amount of food able to be con¬ sumed on any foraging trip is scaled linearly with body size. From these considerations, one can calculate (Calder 1984) that daily consumption divided by gut capacity equals the frequency of foraging trips required to obtain food _ which should be approximately proportional to M 0 25. However, predators vary considerably in their modes of foraging and in the size and quality of prey consumed. Empirical data on “killing rates” (ingestion rates divided by prey size) indicate that fre¬ quency of foraging in homeotherms is proportional to M 0 49 (Peters 1983). In any case, the data indicate that small birds must spend more time foraging than large birds. This constraint has important ecological consequences for birds, especially in harsh envi¬ ronments (Goudie & Ankney 1986). FIELD STUDIES OF FORAGING ACTIVITY BUDGETS There have been relatively few studies of feeding time budgets in multi-species feed¬ ing guilds (Calder 1974). In his classic study of tits ( Parus spp. and Aegithalos 812 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI cauclatus) and the Goldcrest Regulus regulus, Gibb (1954) found that the proportion of time spent feeding by each species was inversely related to body size for these 5g to 18g passerine insectivores. This trend was most conspicuous in fall and winter, when birds were most active and food was least abundant (Figure 1). The largest species (Great Tit Parus major ) spent between 59% and 81% of its time foraging in the months of September to March, whereas the tiny Goldcrest was foraging 88-100% of the time it was observed in those months. Blue Tits P. caeruleus and Marsh Tits P. palustris are both about the same size (ca. 10g), but Marsh Tits consistently spent more time foraging than Blue Tits. This departure from the general body size trend may be explained by differences in foraging behavior between the two species (see aiso below). Blue Tits concentrated their foraging in relatively few productive habitats, whereas Marsh Tits were more widely dispersed in a greater variety of foraging habi¬ tats. Despite this added component of variation, body size still explained 92% of the variation between species in winter feeding activity and time spent feeding was pro¬ portional to M 0 25 in this feeding guild (Figure 2). WINTER FEEDING ACTIVITY OF TITS FIGURE 1 - Activity time budgets of tits in winter. Compiled from Table 3 in Gibb (1954) for the months of September to March. Species codes: GR TIT - Great Tit Parus major, BL TIT - Biue Tit P. caeruleus ; CO TIT - Coal Tit P. ater, MA TIT - Marsh Tit P. palustris ; LT TIT - Long-tailed Tit Aegithalos caudatus ; GCrest-Goldcrest Regulus regulus. Pearson (1968) studied the feeding ecology of eight coexisting seabirds at the Fame Islands, Scotland. Species ranged 17-fold in size from Arctic Terns Sterna paradisaea (104 g) to Shags Phalacrocorax aristotelis (1785 g) and included surface-feeders (terns, gulls) and divers (aicids, shags). Dietary overlap between species was high and most fed on small forage fishes (77-100% of diet) ranging in length from 25 to 150 mm. For each species, Pearson calculated the percent of daylight hours spent in fish¬ ing activities that would be required for adults feeding one chick — or for those spe¬ cies with multi-egg clutches (terns, gulls, shag) — several chicks. Regression of those data (Figure 3) indicate that feeding frequency is proportional to M 081-M 0-86 in this feeding guild and that body size accounts for 81-91% of the variation in feeding fre¬ quency. Shags required as little as 4-8% of daylight hours to feed chicks, whereas ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 813 Arctic Terns had to forage over 100% of the day to rear two chicks. The scaling con¬ stant is considerably lower in this feeding guild than predicted or observed elsewhere (e.g., M 0 25-M 049). This is possibly because time spent feeding included time spent by adults foraging for themselves, foraging for chicks, and for delivery of food back to their chicks; so that whatever time advantage the larger species gained was multiplied several-fold. Also, foraging time may have been underestimated for some of the larger species (e.g., Common Murre Uria aalge, Burger & Piatt 1990). 0.7 0.8 0.9 1.0 1.1 1.2 1.3 LOG OF BODY MASS FIGURE 2 - Regression of (log) time spent feeding by tits versus (log) body size (g). Data from Gibb (1954). See Figure 1. Finally, Goudie & Ankney (1986) studied the winter feeding ecology of four coexist¬ ing species of sea-ducks in Newfoundland. Body sizes ranged about three-fold from Harlequin Duck Histrionicus histrionicus (610 g) to Common Eider Somateria mollissima (1790 g). Differences in quality of food and choice of foraging habitat prob¬ ably influenced feeding rates (e.g., Harlequin Ducks tended to dive in shallower wa¬ ter and on prey with higher energy density), but body size still accounted for 53% of the variation in time spent feeding — which was proportional to M 0 28. Whereas the empirical and theoretical evidence suggest that body size accounts for most of inter-specific variation in time spent feeding, foraging behaviour (style) prob- abjy accounts for much of the residual variation (see also Nagy & Obst; this sympo¬ sium) The importance of behaviour in foraging energetics and activity budgets has ^een documented best in lizards — which range in behaviour from those that “sit-and- wait” in isolation to opportunistically consume dispersed prey, to those that “widely- search” in large groups for dense prey aggregations (Huey & Pianka 1981). Daily pneray expenditures in widely-searching species appear to be about 1 .3-1 .6 times eater than in sit-and-wait predators, but the extra effort pays off because gross food intake may be twice that achieved by sit-and-wait predators (Huey & Pianka 1981). 814 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI O < OC O ./y SEABIRDS REARING 1-3 CHICKS ’y = 3.8299 - 0.85909X RA2 = 0.912 LL y = 3.4758 - 0.81 070x RA2 = 0.810 O O'' O O 0 2.0 2.5 3.0 3.5 LOG OF ADULT BODY MASS FIGURE 3 - Regression of (log) time spent feeding by seabirds versus (log) body size (g) for seabirds rearing 1 chick (squares) or seabirds rearing 1-3 chicks (triangles). Data from Pearson (1968). See text for details. Furthermore, widely-searching species may spend two to three-fold less time forag¬ ing than similarly sized sit-and-wait species because high density prey aggregations, when found, can be rapidly exploited (Nagy et al. 1984). Although less well docu¬ mented, it appears that within most feeding guilds of birds there are behavioural equivalents to the “sit-and-wait” and “widely-searching” types of species found in liz¬ ard communities. Within foraging guilds whose members exhibit similar morphology and foraging styles, however-, it appears that body size is the predominant factor in¬ fluencing feeding frequency (Goudie & Ankney 1986, Piatt 1987). ECOLOGICAL IMPLICATIONS In harsh environments, small species have little flexibility to adjust their activity budg¬ ets because most of the day is spent seeking food [e.g. Nilsson 1970 for Common Goldeneye Bucephala clangula and Oldsquaw Clangula hyemalis ; Pearson 1968 for terns ( Sterna spp.); Goudie & Ankney 1986 for Harlequin Duck and Oldsquaw], Fur¬ thermore, starvation is a more imminent threat to relatively small species. Survival time for birds in thermally neutral environments is proportional to M0 26, and survival time for passerines at 0°C scales to M0 47 (Peters 1983). Thus, large birds take lonoer to starve than small birds, and small birds are more adversely affected by cold tem¬ peratures than large birds. For example, Common Eiders could survive 40% to 50% longer without food than Harlequin Ducks (Goudie & Ankney 1986). King Penguin chicks Aptenodytes patagonica (ca. 5-7 kg) can tolerate fasts of four to six months ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 815 and a 70% decrease in body mass (Cherel et al. 1987). In contrast, Dippers Cinclus cinclus are relatively small (50-70 g), and fat reserves allow only four to six hours survival in winter without food should their open-water habitat freeze over (Lehikoinen & Hakala 1988). Golden-crowned Kinglet lipid and food reserves can support individu¬ als for less than a day in winter (Blem & Pagels 1984). Hummingbirds must spend most of the day actively foraging just to exist, and added stress from rain or cold tem¬ peratures may cause them to enter a nocturnal torpor in order to conserve energy reserves overnight and compensate for lost food intake during the day (Hainsworth et al. 1977, Calder 1974). Between-sex differences in energy reserves have also been demonstrated for species exhibiting sexual size dimorphism — which would confer some advantage in winter to the larger sex [e.g., Common Goldeneye (Nilsson 1970); Capercaillie Tetrao urogallus, (Gjerde & Wegge 1987)]. The differential ability to deal with increasing energy stress may result in differing foraging strategies for large and small birds. Nilsson (1970) and Goudie & Ankney (1986) demonstrated that the Common Eiders decreased feeding intensity with de¬ creasing ambient temperatures, a strategy also noted for the Svalbard Rock Ptarmi¬ gan Lagopus mutus hyperboreus (Stokken et al. 1986) and Snow Geese Anser caerulescens (Frederick & Klass 1982). Thus, large species may defer foraging un¬ der extreme conditions to avoid wasting energy — a strategy that works as long as energy reserves permit. In contrast, smaller species may have to increase feeding intensity with decreasing temperatures [e.g. Nilsson (1970) for Goldeneye, Jorde et al. (1984) for Mallards Anas platyrhyncos], or when stressed for food [e.g. White- crowned Sparrows Zonotrichia leucophys (Ketterson & King 1977)]. Within limits, the response of a bird to increased energetic demands is directly affected by the quality and quantity of food available. For example, Black Ducks Anas rubripes may increase or decrease feeding in winter depending on food quality (Brodsky & Weatherhead 1985). The greater flexibility of time budgets in large species may also help breeding birds buffer against seasonal fluctuations in food supply. Burger & Piatt (1990) studied the activity budgets of Common Murres rearing chicks at a colony in Newfoundland over four consecutive breeding seasons. Local prey abundance varied 10-fold within sea¬ sons and between years, but this was not reflected in chick meal delivery rates or overall breeding success. Instead, adult Murres compensated for fluctuations in prey abundance by adjusting their foraging effort. Common Murres are the largest mem¬ ber of the alcid family of seabirds, and therefore probably have the greatest ability to buffer against fluctuations in food abundance (Piatt 1987). CONCLUSIONS Large birds have greater flexibility than small birds to adjust time budgets under harsh environmental conditions or during periods of food stress. On the other hand, they must extract an overall greater biomass of food from their environment. Small birds are forced by energetic demands to forage more frequently, and to search for higher quality prey — but they can subsist on a lower overall density of food. Therefore, small birds are more vulnerable to extreme environmental conditions, but may be able to exploit a greater variety of habitats. These factors may explain some patterns of spe¬ cies distribution, and geographic variation in body size within species. Within feeding guilds small species may winter in areas with warmer ambient temperatures. Within 816 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI species, large body size may be selected for in cold, northern environments (Bergmann 1847). LITERATURE CITED BERGMANN, C. 1847. Uber die verhaltnisse der Warmese ekonomie der Thiere zu ihrer Grosse (con¬ cerning the relationship of heat conservation of animals to their size). Gottinger Studien 3(1): 595-708. BLEM, C.R., PAGELS, J.F. 1984. Mid-winter lipid reserves of the Golden-crowned Kinglet. Condor 86: 491-492. BRODSKY, L.M., WEATHERHEAD, P.J. 1985. Variability in behavioral response of wintering black ducks to increased energy demands. Canadian Journal of Zoology 63: 1657-1662. BURGER, A.E., PIATT, J.F. 1990. Flexible time budgets in breeding common murres: buffers against variable prey abundance. Studies in Avian Biology No. 14. CALDER, W.A. 1974. Consequences of body size for avian energetics. Pages 86 - 151 in Paynter, R.A. (Ed.). Avian energetics. Nuttall Ornithological Club 15, Cambridge, Mass. CALDER, W.A. 1984. Size, function and life history. Harvard University Press, Cambridge, Mass. CHEREL, Y., STAHL, J.C., LEMAHO, Y. 1987. Ecology and physiology of fasting in King Penguin chicks. Auk 104: 254-262. FREDERICK, R.B., KLASS, E.E. 1982. Resource use and behavior of migratory snow geese. Journal of Wildlife Management 46(3): 601-614. GIBB, J. 1954. Feeding ecology of tits, with notes of treecreeper and goldcrest. Ibis 96: 513-543. GJERDE, I., WEGGE, P. 1987. Activity patterns of Capercaillie, Tetrao urogallus, during winter. Holarctic Ecology 10: 286-293. GOUDIE, R.I., ANKNEY, C.D. 1986. Body size, activity budgets, and diets of seaducks wintering in Newfoundland. Ecology 67: 1475-1482. HAINSWORTH, F.R., COLLINS, B.G., WOLF, L.L. 1977. The function of torpor in hummingbirds. Physi¬ ological Zoology 50: 215-222. HUEY, R.B., PIANKA, E.R. 1981. Ecological consequences of foraging mode. Ecology 62: 991-999. JORDE, D.G., KRAPU, G.L., CRAWFORD, R.D., HAY, M.A. 1984. Effects of weather on habitat se¬ lection and behavior of Mallards wintering in Nebraska. Condor 86: 258-265. KETTERSON, E.D., KING, J.R. 1977. Metabolic and behavioral responses to fasting in the White- crowned Sparrow (Zonotrichia leucophys gambelu). Physiological Zoology 50: 1 15-129. KLEIBER, M. 1961. The fire of life: an introduction to animal eneraetics. Wiley New York New York USA. LEHIKOINEN, E., HAKALA, J. 1988. Variation in weight of migratory Dippers (Cinclus cinclus) in their Finnish winter quarters. Bird Study 35: 101-108. NAGY, K.A., HUEY, R.B., BENNETT, A.F. 1984. Field energetics and foraging mode of Kalahari Lacertid lizards. Ecology 65: 588-596. NILSSON, L. 1970. Food-seeking activity of south Swedish diving ducks in the non-breedinq season Oikos 21: 145-154. PEARSON, T.H. 1968. The feeding biology of seabird species breeding on the Fame Islands. North¬ umberland. Journal of Animal Ecology 37: 521-551 PETERS, R.H. 1983. The ecological implications of body size. Cambridge University Press Cambridge PIATT, J.F. Behavioural ecology of Common Murre and Atlantic Puffin predation on capelin- implica¬ tions for population biology. Unpubl. Ph.D. thesis, Memorial University of Newfoundland St John’<; Newfoundland. QUIRING, D.P. 1950. Functional anatomy of the vertebrates. McGraw-Hill, New York. STOKKEN, K.A., MORTENSEN, A., BLIX, A.S. 1986. Food intake, feeding rythm, and body mass regu¬ lation in Svalbard rock ptarmigan. American Journal of Physiology 251 : R264-R267 y ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 817 POSITIVE CORRELATION BETWEEN RANGE SIZE AND BODY SIZE: A POSSIBLE MECHANISM TERRY ROOT School of Natural Resources, 430 East University, University of Michigan, Ann Arbor, Ml 48109-1115, USA ABSTRACT. Empirical studies have shown that range size often correlates positively with body size (e.g. Brown & Maurer 1989). Factors influencing this relationship include: (1) increased environmen¬ tal variability with bigger ranges, and (2) increased ability of larger species to maintain homeostasis under various conditions. In a select group of wintering North American passerines, larger birds were found to extend their ranges farther north than smaller species. The fat available at dawn for one of these species, Northern Cardinal Cardinalis cardinalis, was sufficient to maintain a metabolic rate of 2.5 times basal for 1 1 . 1 ±1 .8 hours in Tennessee and Indiana, but only for 4.0±2.1 hours in Michigan. Therefore, at the northern range limit, Cardinals were close to their maximal ability to maintain ther¬ mal homeostasis. This suggests that the positive correlation between range and body size is partly due to direct relationships between body size and homeostatic variability, this variability and environmen¬ tal variability, and the latter variability and range size. Keywords: Homeostasis, environmental variability, geographic variability, thermogenesis, birds, North¬ ern Cardinal, Cardinalis cardinalis. INTRODUCTION Many studies have empirically shown a positive correlation between range size and body size within specific taxa (e.g. Averill 1933, Van Valen 1973, Reaka 1980, Brown. 1981, Brown & Maurer 1989). Mechanisms causing these patterns are not known. However, larger bodied animals, in general, have an ability to maintain homeostasis over a wider array of conditions than smaller bodied ones, and larger geographic ranges of species generally encompass more environmental variability than smaller ranges (Figure 1). Species with small body sizes, in contrast, will not be able to main¬ tain homeostasis over a diversity of environmental conditions, resulting in local extirpations in more extreme environments. Consequently, their ranges will be re¬ duced in size. Large bodied animals can easily maintain homeostasis over a small range, but the energy needed to maintain their large body sizes would require that food within the small ranges be of high quality and quantity. Such situations are rare, and when they do occur species’ abundances will generally be low, making extinction probable (Rabinowitz 1981). Therefore, based on homeostatic ability and environmen¬ tal variability the expected pattern is one of direct association between range size and body size (Figure 1). If homeostatic ability and environmental variability are indeed defining the pattern, then many environmental factors could be structuring such a pattern. Ambient tem¬ perature is one that could be easily observed (Root 1988b); range edges would be limited by maximum thermogenic ability. To investigate this possibility I compared different energy-storage and energy-utiliz¬ ing variables across the winter range of the Northern Cardinal Cardinalis cardinalis 818 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGIC! E N V V A I R High R I 0 A Extirpation BS : LARGE RS : LARGE N B M I BS: small E L Low Extinction N I RS: small T T A Y L Low High HOMEOSTATIC VARIABILITY FIGURE 1 - Matrix of possible outcomes for body size (BS) and range size (RS) in envi¬ ronments with low and high variability, and species with low and high homeostatic variabil¬ ity. to determine if energy availability restricts its northern range boundary. Energetic con¬ straints could be manifest in several ways. First, the latitudinal increase in depot fat may not be sufficient to compensate for the latitudinal increase in energetic demands of thermogenesis. This could be due to a limited abundance of food, or not enough foraging time due to a latitudinal decrease in day length. Second, depot fat cannot be metabolized quickly enough to fuel the necessary thermogenesis. In this study I examined geographic variation in metabolic rates, fat content, and fat- metabolizing enzymes of four populations of Cardinals across their winter range. The goal of this study was to determine if any of these physiological properties constrained this species from expanding its range into colder habitats. OBSERVED PATTERN In a previous study (Root 1988a), I found that metabolic rates at northern range boundaries (NBMR; kJ*[bi rd*day] 1 ) of fourteen passerines were positively associated with body mass (M measured in g) when both variables were logarithmically trans¬ formed. The resulting regression equation was: log(NBMR) = log(4.06) + 0.92 log(M), (1) which has r2=0.87. A reduced major axis analysis of log(NBMR) and ambient tempera¬ ture resulted in the following equation: H log(NBMR) = 1.54 - 0.031 T, (2) where T is ambient temperature (average minimum January temperature in °C ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 819 r2=0.7l). By combining equations (1) and (2), the following predicted relationship be¬ tween body mass and ambient temperature was obtained: T = 30.05 - 29.68 log(M). This equation was based on fourteen species, whose northern range boundaries were associated with a particular isotherm and for which winter physiological data were available. Values for thirty-six other species that had northern range boundaries as¬ sociated with particular isotherms but for which no physiological data were available, also fit this line (Root 1988b; Figure 2). Consequently, for these select species, larger bodied birds have ranges extending farther north than smaller species. This suggests that the general pattern of range size correlating positively with body size may be partly due to larger species possessing the capability to maintain thermal homeostasis in a wider variety of habitats. MATERIAL AND METHODS I examined populations of Cardinals along a 1250 km transect that spanned the lon¬ gitudinal length of its winter range in eastern North America. Individuals were captured at four transect locations from December 15 to February 28 in 1987/1988, 1988/1989, and 1989/1990 (hereafter referred to as 1988, 1989, and 1990, respectively). These sites included Ann Arbor, Michigan (Ml; 42° 3’N, 83° 36’W), Bloomington, Indiana (IN; 39° 12’N, 86° 30’W), Smyrna, Tennessee (TN; 36° 0’N, 86° 24'W), and Eufaula, Ala¬ bama (AL; 31° 42’N, 85° 6’W). Birds were collected during the two-hour period before sunset. Metabolic tests A flow-through respirometry system was used to measure oxygen consumption in 1990. Within an hour of dusk, individuals were weighed and placed in an 1800 ml pressure-sealed metal chamber with air input and output ports. The chambers were painted flat black. Birds stood on a wire-mesh platform at the bottom of each can. Chambers were immersed in a bath of water and ethylene glycol. The bath tempera¬ ture was preset to either 25°, 23°, -3°, or -8°C. Air dried with Drierite was supplied to each chamber at a regulated flow of 900 ml*min T The birds were maintained undis¬ turbed for at least four hours before their metabolic rates were monitored. Output air from the chamber was dried again with Drierite and the C02 removed with Ascarite before the oxygen content was measured using a portable paramagnetic oxygen analyzer. Standard metabolism values were determined by finding the minimum 10- minute average of metabolic values calculated using methods described in Withers (1977). Conductance (C; J*[g*h*°C] ’) was determined using C = H / (Tb - T ), where H is metabolic rate in J*(g*h)\ Tb is body temperature in °C and Ta is ambient temperature in °C. Only metabolic rates recorded at -3° and -8°C, which are well be¬ low thermal neutrality, were used to calculate C. 820 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI Body temperature, which was recorded at the end of trials, was determined within a minute of initial disturbance of the bird by inserting a thermocouple approximately 1 cm into the cloaca of the subject. The body temperature of every individual was not recorded; I ran more than one subject each evening and the removal of the first po¬ tentially disturbed the others. Only males were used and metabolic tests were run on the day of capture. Subjects were weighed to the nearest 0.01 g before the test. Individuals were sacrificed after removal from their chamber. They were weighed to the nearest 0.01 g. Stored Fat Fat was extracted from these sacrificed birds. Their guts were assumed empty be¬ cause they had sat in metabolic chambers for a minimum of 5.5 hours. Each specimen was lyophilized to a constant weight. The carcasses, with feathers intact, were then cut into pieces 1 cm3 or smaller. Neutral lipids were extracted us¬ ing petroleum ether in a Soxhlet extraction apparatus. When complete, the residual ether was evaporated by heating at 80°C for 0.5 hr. After cooling, the lean, dry mass was determined. The difference between total-body dry and total-body lean, dry mass provided mass of the neutral lipids to the nearest 0.01 g. Enzyme Assays The activities of two different fat-metabolizing enzymes in the flight muscles were examined: citrate synthase (CS) and betahydroxyacyl-CoA dehydrogenase (HOAD). In 1988 birds collected for these assays were humanely sacrificed and within five min¬ utes of death carcasses were frozen in liquid nitrogen. Birds collected in 1989 were humanely sacrificed, the left pectoralis muscle removed, placed in a Nunc tube, and frozen in liquid nitrogen within 30 minutes of death. The 30 minute lag time presum¬ ably had no influence on the enzyme activity (Root and O’Connor unpublished data on House Sparrows Passer domesticus). While being kept on ice in the laboratory, approximately 0.25 g of pectoralis muscle was finely minced. This was then added to 10 volumes of homogenizing buffer con¬ taining 100 mM phosphate and 2 mM EDTA at a pH of 7.3 at 5°C. Homogenization occurred in a hand-held, glass-glass tissue grinder that was maintained on ice. Homogenates were sonicated three times for 15 s at 45 s intervals. The assay protocols from Marsh (1981) were followed and duplicate assays were run. Maximum activity values for each bird were averaged at each location. RESULTS Body Temperature A regression analysis showed that body temperature was not dependent on ambient temperature (r2-0.047), and consequently a mean value of 39.26 ± 0.1 8°C (n=13) was used. Metabolism Student’s t-tests for small samples revealed no significant differences in the mean conductance values, and thus, these data were combined (2.22 ± 0.04). This value ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 821 was used as the slope of a line relating metabolic rate (H) with ambient temperatures (T) below the lower critical temperature. The principles of Newtonian cooling, which apply due to the independence of body and ambient temperature, require that this line pass through the point at which metabolic rate is zero and ambient temperature equals body temperature. The resulting equation is: H = 87.16 - 2.22 T. Metabolic rates above 23°C at each location were averaged to calculate BMR values. Student’s t-test for small samples revealed no significant difference in mean values- from IN and TN (P=0.38). The mean of these two states (60.04 ± 1.36) was found to differ significantly from that in Ml (P=0.04) and approached significance with the AL mean (P=0.07). The difference between Ml and AL was highly significant (P=0.006). Stored Fat Fat content of birds at dusk was estimated from individuals used in metabolic tests. During these tests they were held a known length of time in a metabolic chamber at a constant temperature. The estimated amount of fat used during these tests (F meas¬ ured in g) was derived by the following equation: F = M * K * t, where M is standard metabolism in J*(bird*h) \ K is a conversion factor (37.7kJ=1g of fat; see Blem & Shelor 1986), and t is time in the chamber. To approximate the amount of fat available at dusk, F can be added to the amount of fat extracted from the individual carcasses (Table 1). This provided only a rough estimate of absolute fat content, because birds use digestive heat in the early evening to supplement thermogenesis (King 1972), and the minimum 10-minute average was the assumed metabolic rate for the entire trial. These estimates, however, allow comparisons among locations, because methods were constant across all locations. Available Energy from Stored Fat Due to the geographic scale of this study, the amount of depot fat available at dusk is not important for comparative purposes because influencing factors, such as night length and ambient temperature, vary latitudinally (King & Farner 1966, Dawson et al. 1983, Nolan & Ketterson 1983). Consequently, I converted these data to the number of hours the estimated amount of fat can sustain a specific metabolic rate (Table 1). The metabolic rate I used was 2.5xBMR, which is the estimated rate at the northern boundary of select species’ ranges (Root 1988b). The total number of hours this metabolic rate could be sustained was determined at each location by calculating the grams of fat needed per hour to fuel a metabolic rate of 2.5xBMR (J*[bird*h] '). Divid¬ ing this into the estimated available fat provided the number of hours the amount of fat would sustain that metabolic rate. Night length, which varied by location, was sub¬ tracted from this vaiue to obtain the number of hours individuals can survive after dawn without feeding, still assuming a metabolic rate of 2. 5xBMR (Figure 3). In AL where individuals would rarely if ever have to maintain a metabolic rate as high as 2 5xBMR the available fat at dawn was enough to allow them to survive for an av- eraqe of 1 0 ± 1 1 hour after dawn without feeding. The average number of hours in TN was shorter than in IN, but not significantly so (P=0.20). Consequently, these data were combined resulting in an average of 11.1 ± 1.8 hours. This mean was signifi- cantly higher than either AL (P-0.0001) and Ml (P=0.02; Figure 3). 822 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI TABLE 1 - Means, standard errors, and sample sizes for various parameters at dif¬ ferent locations. Parameter Ml IN TN AL Conductance* 2.28±0.07 2.22±0.07 2.28±0.08 2.1210.08 (n) (3) (7) (9) (9) BMR+ 64.87±1 .86 60.39±0.70 59.6913.22 56.9711.13 (n) (5) (2) (2) (4) Estimate Fat# 3.87±0.55 4.72±0.67 4.4510.26 2.421C.18 (n) (8) (9) (9) (13) Hours++ 4.0±2.1 1 2.9±3.9 9.711.2 1.011.1 (n) (7) (9) (11) (13) Citrate Synthase” — 1 47±1 0 146112 128110 (n) (4) (4) (4) HOAD” — 1 5.9±1 .5 16.111.4 13.412.4 (n) (4) (4) (4) + in J*(g*h*°C)'1 * in J*(g*h) 1 # in g. Text explains the derivation. ++ Number of hours after dawn an individual sustaining a metabolic rate of 2.5xBMR could survive. " in micro-moles (g*min) 1 Emzyme Assays The activities of both CS and HOAD, and the ratio of HOAD to CS were not signifi¬ cantly different among the states (Table 1). They had a grand mean of 140 ± 6 and 15.2 ± 1.0, with ranges of 106 to 178 and 8.6 to 20.0, respectively. DISCUSSION A pattern of larger birds extending their ranges farther North into areas with colder ambient temperatures is seen in passerines that winter in North America and that have northern range boundaries associated with particular isotherms (Figure 2). Un¬ derstanding the mechanisms influencing this pattern will help us understand the com¬ monly found empirical pattern of larger-bodied species generally having larger ranqes (e.g. Averill 1933, Van Valen 1973, Reaka 1980, Brown 1981, Brown & Maurer 1989). Due to colder temperatures and longer nights, Cardinals in the more northerly portions of their ranges must either decrease their heat loss due to conductance, or be able to sustain higher metabolic rates for a longer period of time than do more southerly conspecifics. Comparisons of metabolic responses from four locations indicates that heat loss due to conductance does not vary geographically (Table 1). Therefore, the insulative abilities, i.e. heat loss, of individuals are apparently constant from the south¬ ern to the northern extremes of it range. ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 823 _og(Mass) FIGURE 2 - The line defined by T = 30.5 - 29.68 log(M) (see derivation in “OBSERVED PATTERN” of text) with the values for 50 North American wintering passerines that had northern range boundaries associated with ambient temperature. T is measured in °C and M in g. Depot fat, which fuels thermogenesis (Carey et al. 1978), could be limited in two ways. First, enough fat cannot be obtained from available food. Second, the enzymes needed to convert the fat to usable energy are not available in sufficient concentra¬ tions to facilitate conversion rapidly enough. The enzyme assays of CS and HOAD provided no indication that enzymes were lim¬ ited anywhere across their range. The activities of these enzymes per g of muscle did not vary geographically (Table 1). I have found evidence that the amount of depot fat may indeed be constraining car¬ dinals from extending their ranges farther north. Estimates of the amount of fat avail¬ able at dusk provides a means of examining energy availability among locations. The meaningful comparison is not of the amount of fat but rather the length of time it can provide sustainable energy for thermogenesis at each of the locations. Night length, ambient temperature and BMR vary with location (Table 1). The number of hours after dawn that a metabolic rate of 2.5xBMR can be sustained is significantly shorter for in¬ dividuals in the North than in the middle portion of the cardinal’s range (Ml=4.0 ± 2.1 , IN/TN=1 1 1 ±1.8; P=0.02; Figure 3). Therefore, the amount of depot fat in northern individuals provides very little surplus energy to ensure survival without actively for- 824 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI aging throughout most day-light hours. Cardinals in the middle of their range are bet¬ ter buffered against inclement weather, the fuel will maintain thermogenesis longer without feeding. Hence, the amount of depot fat seems to be constraining the range of the Northern Cardinal. FIGURE 3 - The number of hours past dawn that the average amount of depot fat in North¬ ern Cardinals can sustain a metabolic rate of 2.5xBMR in AL, TN, IN, and Ml. The aver¬ age values are indicated with a bar. Similar studies need to be done on species of different body sizes and range sizes Finding results similar to the ones in this paper would suggest that homeostatic and environmental variabilities are indeed driving the positive correlation between body size and range size within given taxa. This would be due to the direct relationship be¬ tween body size and homeostatic variability, this variability and environmental vari¬ ability, and the latter variability and rangesize. LITERATURE CITED AVERILL, C. K. 1933. Geographical distribution in relation to number of eggs. Condor 25-93-97 BLEM, C. R., SHELOR, M.H. 1986. Multiple regression analyses of midwinter fatteninq of the WhitP throated Sparrow. Canadian Journal of Zoology 64:2405-241 1 . BROWN, J. H. 1981. Two decades of homage to Santa Rosalia: toward a general theory of diversity American Zoologist 21 :877-888. ' ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 825 BROWN, J. H., MAURER, B.A. 1987. Evolution of species assemblages: effects of energetic con¬ straints and species dynamics on the diversification of the North American avifauna. American Natu¬ ralist 130:1-17. CAREY, C, DAWSON, W.R., MAXWELL, L.C., FAULKNER, J.A. 1978. Seasonal acclimatization to temperature in cardueline finches. II. Changes in body composition and mass in relation to season and acute cold stress. Journal of Comparative Physiology 1 25B : 101-1 13. DAWSON, W. R., MARSH, R.L., BUTTEMER, W.A., CAREY, C. 1983. Seasonal and geographic vari¬ ation of cold resistance in House Finches Carpodacus mexicanus. Physiological Zoology 56:353-369. KING, J. R. 1972. Adaptive periodic fat storage by birds. Proceedings of 15th International Ornithologi¬ cal Congress: 200-217. KING, J. R., FARNER, D.S. 1966. The adaptive role of winter fatting in the White-crowned Sparrow with comments on its regulation. American Naturalist 100:403-418. MARSH, R. L. 1981. Catabolic enzyme activities in relation to premigratory fattening and muscle hyper¬ trophy in the Gray Catbird (Dumetella carolinensis). Journal of Comparative Physiology 141 B:41 7-423. NOLAN, V., JR., KETTERSON, E.D. 1983. An analysis of body mass, wing length, and visible fat de¬ posits of Dark-eyed Juncos wintering at different latitudes. Wilson Bulletin 95:603-620. RABINOWITZ, D. 1981. Seven forms of rarity. In Synge, H. (Ed.). The biological aspect of rare plant conservation. Wiley, Chichester. REAKA, M. L. 1980. Geographic range, life history patterns and body size in a guild of coral-dwelling mantis shrimps. Evolution 34:1019-1030. ROOT, T. 1988a. Environmental factors associated with avian distributional boundaries. Journal of Bio¬ geography 15:489-505. ROOT, T. 1988b. Energy constraints on avian distribution and abundance. Ecology 62:330-339. VAN VALEN, L. 1973. Body size and number of plants and animals. Evolution 27:27-35. WITHERS, P. C. 1977. Measurements of V02, VC02, and evaporative water loss with a flow-through mask. Journal of Applied Physiology 42:120-123. 826 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI EXTINCTION RATE, BODY SIZE, AND AVIFAUNAL DIVERSITY BRIAN A. MAURER1, HUGH A. FORD2, EDUARDO H. RAPOPORT3 1 Department of Zoology, Brigham Young University, Provo, UT 84602, USA 2 Zoology Department, University of New England, Armidale, NSW 2351, Australia 3 Universidad Nacional del Comahue, C. 1336, 8400 Bariloche, Argentina ABSTRACT. Many factors determine the diversity of an avifauna. Although it is not possible to measure all of these factors for all birds in an avifauna, three variables in particular seem to provide a summary of the effects of these many different factors. We examined the statistical distribution of Australian terrestrial birds within a three dimensional space formed by log transformed values of their average abundance, average body mass, and geographic range size. We found that Australian birds had very similar statistical distributions to those documented for North American terrestrial birds. We used data on North American birds to estimate extinction times for various taxa, and found that body size per se was not closely related to estimated extinction time, but that estimated extinction rates were signifi¬ cantly different for four different passerine taxa that differ in their ecological flexibility. We conclude that body size, inasmuch as it affects the complex of adaptations possessed by a taxon, may influence rates of extinction. However, we stress that taxa that show a greater degree of ecological flexibility than other taxa also have lower rates of extinction. INTRODUCTION The diversity of an avifauna is determined by many processes occurring on many scales. At one extreme, the success with which individual birds are able to meet their life history requirements determines the stability and rates of change of populations, and hence, the likelihood of extinction for each species. At the other extreme, the history of geological and climatic events on continents through geological time deter¬ mines patterns of geographical isolation and availability of resources for continental populations, and hence, rates of speciation and extinction. Measuring all of these factors for all species in a continental avifauna is impossible. Hence, we are left with the task of finding a small set of variables that are sufficient to describe the major ecological and evolutionary characteristics that affect the persistence of species if we are to develop an empirically based theory for the maintenance of species diversity. Three variables seem particularly suitable for this task (Brown & Maurer 1987). As has been discussed in this symposium, body mass is related to a number of factors affect¬ ing the ecological properties of species such as metabolic rate, life histories, and for¬ aging behavior. In addition, geographic range size is related to the breadth of ecologi¬ cal conditions that a species can tolerate and average population density is a measure of how successful a species is in turning resources into birds. Hence, examining the statistical distribution of a large number of species in an avifauna among these three variables should provide an indication of the kinds of processes affecting the persist¬ ence of species, and hence how species diversity is regulated (Brown & Maurer 1987 1989). Each species in an avifauna can be represented by a point in a three dimensional space defined by body mass, average density, and geographic range size (Brown & ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 827 Maurer 1987). For any species, if we multiply its values for these variables together, we get the volume of a rectangular box in the space, which is an estimate of the to¬ tal biomass of the species (Rapoport 1982). The distribution of species of North American terrestrial birds was confined to limited regions of this space (Brown & Maurer 1987). Some of the under-represented combinations of variables were attrib¬ uted to the higher likelihood of extinction of some species that occupy certain regions of the space (Brown & Maurer 1987). The implication is that certain “strategies” or combinations of density, body mass, and geographic range size are more likely to result in persistence than are others. Species with similar total biomasses may have very different likelihoods of extinction, depending on the particular combinations of the three component variables they possess. Our purpose in this paper is to compare the distribution of species of Australian landbirds among these variables with the results Brown & Maurer (1987) obtained for North American landbirds. If the avifaunas of these two continents show similar pat¬ terns, this would suggest that the processes determining these patterns were similar. Furthermore, we examine the relationship between body size and rate of extinction in North American terrestrial birds in order to examine hypotheses regarding the role of body size variation among species in regulating continental diversity. METHODS Data on geographic range size of Australian landbirds were taken from Blakers et al. (1984). We used the number of 1 degree latitude-longitude blocks occupied by a spe¬ cies as a measure of the size of its geographic range. A crude measure of average abundance was calculated by dividing the number of records of a species by the number of blocks it occurred in to give an estimate of the average number of records per block occupied (Ford 1990). This measure has many potential shortcomings as a measure of average abundance, which we acknowledge, but there are no published data from comprehensive surveys attempting to estimate abundance for Australian birds comparable to those for North American birds. Therefore, the results obtained using this measure should be considered tentative. Estimates of body masses for most Australian birds were obtained from data collected for the Australian Bird Band¬ ing Scheme and from the literature. To estimate extinction times for North American terrestrial birds, we obtained crude estimates of total population densities as follows. Data from the Breeding Bird Sur¬ vey (BBS) conducted by the U.S. and Canadian Fish and Wildlife Services (Robbins et al. 1986) were used to calculate the average number of birds counted per BBS route for each of about 350 species. Since a BBS route is composed of 50 stops, and at each stop a circle of radius 0.4 km is censused, a total of about 25 km2 are sur¬ veyed by each route. Flence, a crude estimate of average density for a species is the average number of individuals seen per BBS route divided by 25 km2. This measure of average density was multiplied by the area of a species’ geographic range obtained from planimetery of range maps in Robbins et al. (1983) to give an estimate of the total population sizes of birds. There are many sources of potential bias in these data, but like the Australian data, they are the only data that exist to answer the questions pertinent to the present investigation. We view our findings as tentative until better data have been accumulated. 828 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI Using estimated total population sizes for the North American birds, we calculated extinction times using Goodman’s (1987a,b) model for average extinction time. His model expresses average time to extinction as a function of total population size, the rate of population change (r), and the variance in the rate of population change (V). Estimates of r were calculated by obtaining data on clutch sizes for North American birds from C. Harrison (1978) and H. Harrison (1979). Death rates were assumed to average around 50% (Henny 1972, Saether 1989), although this figure is probably too low. Rate of population change was calculated as the difference between birth rate (i.e., clutch size, accounting for multiple broods per year) and death rate. Since these data were measured per year, the resulting extinction times were given in years. The variance in rate of population change was calculated as Vr = 5r using a value inter¬ mediate between Belovsky’ s (1987) extremes of Vr = 2r and Vr = 7r. These extinc¬ tion times must be considered as preliminary estimates because of the crude nature of the data, but they provide an initial indication of how extinction likelihood varies among North American birds. Extinction rates of four North American passerine taxa were estimated using the fol¬ lowing procedure. An extinction curve for each taxa was constructed by plotting ex¬ tinction time of a species against the number of species with longer extinction times. FIGURE 1 - Log body mass distribution of 426 species of Australian landbirds. L0G(B0DY MASS (g)) ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 829 The tangent of this curve at a given time divided by the number of species at that time is an estimate of the per-species extinction rate. The taxa used were Tyrannidae, Vireonidae, Parulinae, and Icterinae. These taxa represent a range of different eco¬ logical strategies from the small-bodied insectivorous parulines and vireonids, to the larger-bodied insectivorous tyrannids, to the even larger, ecologically diverse icterines. Analysis of variance was used to test for significant differences in mean log- transformed per-species extinction rates among the four taxa. Tukey’ s method for comparison among means (Sokal and Rohlf 1981) was used to test for differences among pairs of taxa. RESULTS Distribution of density, body mass, and geographic range size among Austral¬ ian terrestrial birds The distribution of body masses for Australian birds was positively skewed on a loga¬ rithmic scale (Figure 1 ). This result is similar to the results that Maurer & Brown (1 988) found for North American birds, and Maurer et al. (1991) found for mammals from several continents (see also Brown & Maurer 1989). Ford (1990) plotted log abun¬ dance against log geographic range size for Australian terrestrial birds and found a LOG(BODY MASS (g)) FIGURE 2 - Relationship between log body mass and log geographic range size for Aus¬ tralian landbirds. LOG ABUNDANCE 830 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI pattern similar to that found by Brown and Maurer (1987) for North American terres¬ trial birds, with relatively few species with small geographic ranges achieving high densities. The distribution of Australian species with log geographic range size and log body mass indicated that the minimum geographic range size increases with in¬ creasing body mass (Figure 2). This pattern is nearly identical to Brown & Maurer’s (1987) results for North American birds and Brown’s (1981) results for North Ameri¬ can mammals. The plot of log abundance against log body mass differed in some details to Brown & Maurer’s (1987) results for North American birds. Where Brown & Maurer (1987) found that the maximum density of species declined for species both less than and greater than lOOg, the Australian data indicated that maximum abun¬ dance declined only for species greater than lOOg (Figure 3). Extinction in North American birds A plot of estimated average extinction times against body mass for North American terrestrial birds indicated that there was no straightforward relationship between body mass and extinction time (Figure 4), although the minimum time to extinction in¬ creased with body size above lOOg. 2.5 0.0- 1 - r - » - » - 1 - » - ' - ' - 1 - » - » - ' - -| - 1 - 1 - 1 - , - , - -r 0 12 3 4 LOG(BODY MASS (g)) FIGURE 3 - Relationship between log average abundance and log body mass for Austral ian landbirds. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 831 Extinction curves for four taxa of North American passerines were significantly differ¬ ent (Figure 5). The rate of extinction was clearly greater for the more speciose wood warblers (Parulinae) and flycatchers (Tyrannidae). Of the equally speciose vireos (Vireonidae) and blackbirds (Icterinae), the blackbirds had a lower rate of extinction. Average log-transformed per-species extinction rates were significantly different among the four taxa examined (F = 4.12; df = 3, 83; P = 0.009). The blackbirds had the lowest per-species extinction rates, and the flycatchers and wood warblers had significantly higher rates (Table 1). TABLE 1 - Average extinction rates of four North American passerine taxa. Taxon Number of Species Average Per-species Extinction Rate1 (/1000y) Standard Deviation Icterinae 13 0.093a 0.189 Vireonidae 8 0.1 03ab 0.184 Parulinae 43 0.1 45b 0.265 Tyrranidae 23 0.281 b 0.593 1 Means with the same letter are not significantly different at P = 0.05, Tukey’s test, after log transform¬ ing data. t 0 ~r 3 lOG(BODY MASS (g)) T 4 FIGURE 4 - Relationship between log average extinction time and log body mass for North American landbirds. 832 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI TABLE 2 - Characteristics of distributions of log body mass among all Australian spe¬ cies of landbirds and those which are considered rare or endangered. Statistic All species Rare and endangered species Number of species 426 30 Average log body mass 1.67 2.04 Median log body mass 1.55 1.76 Modal log body mass 1.04 1.74 Standard deviation log body mass 0.64 0.64 Interquartile range log body mass 0.93 1.08 EXTINCTION CURVES FOR 4 BIRD FAMILIES FIGURE 5 - Extinction curves for the North American passerine taxa of wood warblers (squares), flycatchers, (closed circles), blackbirds (open circles) and vireos (triangles). The slopes of the lines are all significantly different (P < 0.05). DISCUSSION The similar relationships among body mass, geographic range size, and abundance obtained by Brown & Maurer (1987) for North American birds and the present study for Australian birds suggests that the patterns documented here for avifauna are quite general. Similar patterns have been obtained recently by a number of workers exam¬ ining taxa other than birds (Brown 1981, Strayer 1986, Damuth 1987, Gaston 1988 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 833 Gaston & Lawton 1988, Morse et al. 1988, Maurer et al. 1991). We suggest that what¬ ever processes are ultimately operating to generate these patterns in North American birds, they must also be operating in Australian birds, and other taxa as well. There were some differences between the Australian and North American results, particu¬ larly in the distribution of species with log body mass and log average abundance. However, given the different methods of estimating average abundance used for the two avifauna, we hesitate to conjecture whether these differences are due to differ¬ ences in methods of data analysis or true biological differences between the avifauna. Brown & Maurer (1987) postulated that the patterns they examined were determined by differential speciation and extinction through geological time. Maurer et al. (1991) further went on to show that the skewed distributions of log body masses seen for many taxa (see also Figure I) implied that directional selection on body size within lin¬ eages was coupled with high rates of extinction in large-bodied taxa and high rates of speciation in small-bodied taxa. Our data for North American birds suggest that extinction likelihood does not increase significantly with body size, but this conclusion must be viewed with caution until better estimates of extinction times are available. Data for Australian birds indicate that the distribution of log body masses among spe¬ cies which are considered rare or endangered is disproporionately made up of spe¬ cies of larger body mass (Table 2). The most intriguing result we obtained was the finding that different taxa have signifi¬ cantly different rates of extinction. The blackbirds are an ecologically diverse and relatively large-bodied group of passerines. They often maintain relatively high den¬ sities when compared with other passerines. Our data suggest that there is a relation¬ ship between the tendency of this group of birds to be ecologically diverse and their low rates of extinction when compared to more ecologically specialized taxa. Although body size is one important characteristic of the blackbirds that set them apart from many other passerine taxa, other ecological characteristics they possess also are important in determining their overall diversity. Since each of the groups compared have different evolutionary histories leading to different sets of ecological adaptations, phylogenetic history may be more important in determining the diversity of a group than body size per se. ACKNOWLEDGEMENT We would like to thank the Australian Bird and Bat Banding Scheme for providing us with weights of many Australian birds. LITERATURE CITED BELOVSKY, G.E. 1987. Extinction models and mammalian persistence. Pp. 35-58 in Soule, M. (Ed.). Viable populations for conservation. Cambridge University Press. BLAKERS, M., DAVIES, S.J.J.F., REILLY, P.N. 1984. The atlas of Australian birds. Melbourne Univer¬ sity Press. BROWN, J.H. 1981. Two decades of homage to Santa Rosalia: toward a general theory of diversity. American Zoologist 21: 877-888. BROWN J H.. MAURER, B.A. 1987. Evolution of species assemblages: effects of energetic constraints and species dynamics on the diversification of the North American terrestrial avifauna. American Natu¬ ralist 130: 1-17. 834 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI BROWN, J.H., MAURER, B.A. 1989. Macroecology: the division of food and space among species on continents. Science 243: 1145-1150. DAMUTH, J. 1987. Interspecific allometry of population density in mammals and other animals: the in¬ dependence of body mass and population energy-use. Biological Journal of the Linnean Society 31. 193-246. FORD, H.A. 1990. Relationships between distribution, abundance and foraging specialization in Aus¬ tralian landbirds. Ornis Scandinavica 21: 133-138. GASTON, K.J. 1988. Patterns in the local and regional dynamics of moth populations. Oikos 53: 49- 57. GASTON, K.J., LAWTON, J.H. 1988. Patterns in body size, population dynamics, and regional distri¬ bution of bracken herbivores. American Naturalist 132: 662-680. GOODMAN, D. 1987a. Consideration of stochastic demography in the design and management of bio¬ logical reserves. Natural Resource Modeling 1: 204-234. GOODMAN, D. 1987b. The demography of chance extinction. Pp. 11-34 in Soule, M.E. (Ed.). Viable populations for conservation. Cambridge University Press. HARRISON, C. 1978. A field guide to the nests, eggs, and nestlings of North American birds. Stephen Greene Press, Brattleboro, Vermont. HARRISON, H.H. 1979. A field guide to western bird’s nests. Houghton Miflin, Boston. HENNY, C.J. 1972. An analysis of the population dynamics of selected avian species. U.S. Fish and Wildlife Service, Wildlife Research Report 1. MAURER, B.A., BROWN, J.H. 1988. Distribution of energy use and biomass among North American terrestrial birds. Ecology 69: 1923-1932. MAURER, B.A., BROWN, J.H., RUSSLER, R.D. 1991. The macro and micro in body size evolution. Evolution, in press. MORSE, D.R., STORK, N.E., LAWTON, J.H. 1988. Species number, species abundance and body length relationships of arboreal beetles in Bornean lowland rain forest trees. Ecological Entomology 13: 25-37. RAPOPORT, E.H. 1982. Areography: geographical strategies of species. Pergamon Press, Oxford. ROBBINS, C.S., BRUUN, B., ZIM, H.S. 1983. The birds of North America. Golden Press, New York. ROBBINS, C.S., BYSTRAK, D., GEISSLER, P.H. 1986. The Breeding Bird Survey: its first fifteen years, 1965-1979. U.S. Fish and Wildlife Service, Resource Publication 157. SAETHER, B.-E. 1989. Survival rates in relation to body weight in European birds. Ornis Scandinavica 20: 13-21. SOKAL, R.R., ROHLF, F.J. 1981. Biometry, 2nd edition. W. H. Freeman, San Francisco. STRAYER, D. 1986. The size structure of a lacustrine zoobenthic community. Oecologia 69: 513-519. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 835 CONCLUDING REMARKS: BIRDS, BODY SIZE, AND EVOLUTION BRIAN A. MAURER Department of Zoology, Brigham Young University, Provo, UT 84602, USA Studies of birds have provided a great deal of theoretical and empirical information regarding ecology and evolution. Lack’s (1947, 1954, 1966) studies exemplify this contribution to biology. Recent interest in the ecological and evolutionary conse¬ quences of body size (Peters 1983, Calder 1984) has stimulated a great deal of ef¬ fort to understand how body size evolves and influences the evolution of other sys¬ tems of adaptations. It is only natural that studies of the ecological and evolutionary consequences of body size in birds should become an important pursuit of ornitholo¬ gists. As has been indicated in this symposium, there are many interesting and important things we have learned from studies of birds and body size. Yet these studies also indicate that there is much more to be learned. The general pattern that seems to emerge from these studies is that body size can explain some of the biology that has the generated patterns observed, but there is sufficient unexplained residual variation after the effects of body mass have been removed to suggest that body size interacts with many other factors to determine the observed ecological characteristics of spe¬ cies. Probably the most important factor that interacts with body size to influence ecologi¬ cal patterns is phylogeny. Phylogeny is the history of evolutionary relationships among species. Species inherit ecological attributes from their ancestors. Body size is no exception to this rule. Species within higher taxa generally tend to be similar in size and have similar ecological attributes. To make a rough evaluation of the relative roles of body size and phylogeny in determining population densities, I took data on popu¬ lation densities of 380 species of North American birds obtained by Brown & Maurer (1987) and performed a random effects ANOVA (Sokal & Rohlf 1981) to examine the effects of body size and taxonomic status of each species. I divided body sizes into nine classes (see Maurer & Brown 1988) and assigned each species to one of these classes. To evaluate the role of phylogeny, I used the family to which each species belonged as an estimate of phylogenetic relationships among species. Body size class and family were used as random main effects, and I estimated the variance components for each of these effects. The family variance component explained 41% of the total variance in population density, while body size class only explained 1%. Clearly, an explanation for variation in population densities among species must in¬ clude information on phylogeny. Nagy and Obst’s paper in this symposium indicates that residual variation among species in field metabolic rates can be partially ex¬ plained by phylogeny. In general, to study the evolution of ecological attributes of species, we need better information on phylogenies than we now have for birds. Brooks and McClennan (1990) show how powerful evolutionary explanations become when phylogenetic information is available. 836 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI The pursuit of energy by birds is a major factor in determining the ecological success, and hence, fitness of individuals. Body size has a major influence on the rate at which energy is needed (Nagy & Obst) and actually obtained (Goudie & Piatt) by birds. This in turn, influences the ability of individuals to persist in the environment, and more importantly, to use a wider variety of environments (Root). The statistical pattern of individual successes across space determines the size of the geographic range, and Root’s intriguing results suggest that body size influences the success of individuals of different species by determining the physiological limitations of individuals to en¬ vironmental extremes. Root’s hypothesized mechanism for limitation by temperature could be extended to any kind of limiting factor. When species with different geographical ranges are compared, body size emerges as a positive correlate of range size because many of the factors that limit individu¬ als in their response to environmental extremes, such as thermoregulation and for¬ aging rates, are directly tied to the size of the individual. It would be interesting if other correlations between ecological attributes of species, such as population density, can be examined using Root’s general approach of searching for factors that limit individu¬ als and asking how the statistical pattern of individual successes is related to body size. In attempting to catalogue biological diversity, it is evident that we cannot describe the life histories of every species on the planet, indeed, we may not be able to do so even for groups as well known as birds. Carey and Calder’s suggestion that body size might provide a way to obtain some estimates of life history requirements of different spe¬ cies may be an important tool for conservation biologists. The picture developed by the presentations in this symposium suggests that body size is intimately involved in the day to day ecological activities of individuals, and conse¬ quently has a significant impact on both individual fitness, and stability of populations. Hence, the evolution of body size might be influenced both by natural selection and by differential speciation and extinction. Natural selection is the consequence of dif¬ ferences in fitness among individuals in a population. If variation in body mass among individuals in populations is heritable (and there is good evidence that it is), then natural selection may operate to change average body sizes within species (Boag & Grant 1981). At the same time, body size influences the rates of life history processes, so that it must also influence the rate of population change. This in turn must ulti¬ mately influence the likelihood of extinction of a species. Hence, differential extinc¬ tion of species of different size may also contribute to the evolution of body size (Brown & Maurer 1986). There is much yet to be learned about the mechanisms in¬ volved in generating correlations between body size and ecologically important vari¬ ables such as metabolic rate, foraging rate, and geographic range size. The answers that we obtain from asking questions about such correlations will tell us much about how evolution occurs. Many of the answers to these questions will come from stud¬ ies of birds. As we learn more, we will gain new tools to apply to the ever increasing threat of global losses of biodiversity. Such tools cannot be developed without pushing the frontier of knowledge farther than it is today. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 837 LITERATURE CITED BOAG, P.T., GRANT, P.R. 1981. Intense natural selection in a population of Darwin’s Finches (Geospizinae). Science 214:82-85 BROOKS, D.R., McCLENNAN, D.A. 1990. Phyloqeny, ecoloqy, and behavior. University of Chicago Press, Chicago. BROWN, J.H., MAURER, B.A. 1986. Body size, ecological dominance, and Cope’s rule. Nature 324:248-250. BROWN, J.H., MAURER, B.A. 1987. Evolution of species assemblages: effects of energetic constraints and species dynamics on the diversification of the North American avifauna. American Naturalist 103:1- 17. CALDER, W.A., III. 1984. Size, function, and life history. Harvard University Press, Cambridge, Mas¬ sachusetts. LACK, D. 1947. Darwin's finches. Cambridge University Press, Cambridge. LACK, D. 1954. The natural regulation of animal numbers. Oxford University Press, Oxford. LACK, D. 1966. Ecological adaptations for breeding in birds. Methuen, London. MAURER, B.A., BROWN, J.H. 1988. Distribution of energy use and biomass among species of North American terrestrial birds. Ecology 69:1923-1932. PETERS, R.H. 1983. The ecological implications of body size. Cambridge University Press, Cambridge. SOKAL, R.R., ROHLF, F.J. 1981. Biometry, 2nd edition. W.H. Freeman, New York. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 839 SYMPOSIUM 12 ECOLOGICAL AND BEHAVIOURAL ADAPTATIONS OF SOUTHERN HEMISPHERE WATERFOWL Conveners D. F. McKINNEY and M. J. WILLIAMS 840 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI SYMPOSIUM 12 Contents INTRODUCTORY REMARKS: ECOLOGICAL AND BEHAVIOURAL ADAPTATIONS OF SOUTHERN HEMISPHERE WATERFOWL MURRAY WILLIAMS . 841 BREEDING ADAPTATIONS OF SOUTHERN HEMISPHERE, ARID ZONE DUCKS S. V. BRIGGS and W. G. LAWLER . 843 MATING SYSTEMS OF TROPICAL AND SOUTHERN HEMISPHERE DABBLING DUCKS L. G. SORENSON . 851 THE BLUE DUCK MATING SYSTEM - ARE RIVER SPECIALISTS ANY DIFFERENT? CLARE J. VELTMAN, SUSAN TRIGGS, MURRAY WILLIAMS, KEVIN J. COLLIER, BRIAN K McNAB, LISA NEWTON, MARIE HASKELL and IAN M. HENDERSON . 860 MALES PARENTAL CARE IN SOUTHERN HEMISPHERE DABBLING DUCKS FRANK McKINNEY . 868 ECOLOGICAL AND BEHAVIOURAL RESPONSES OF AUSTRAL TEAL TO ISLAND LIFE MURRAY WILLIAMS, FRANK McKINNEY and F. I. NORMAN . 876 CONCLUDING REMARKS: ECOLOGICAL AND BEHAVIOURAL ADAPTATIONS OF SOUTHERN HEMISPHERE WATERFOWL FRANK McKINNEY . 885 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 841 INTRODUCTORY REMARKS: ECOLOGICAL AND BEHAVIOURAL ADAPTATIONS OF SOUTHERN HEMISPHERE WATERFOWL MURRAY WILLIAMS Department of Conservation, P.O. Box 10-420, Wellington, New Zealand In late 1987 the Delta Waterfowl and Wetlands Research Station convened, in Win¬ nipeg, Canada, a week-long symposium on the ecology and management of breed¬ ing waterfowl. It was a marvelous talkfeast for swan- goose- and duckophiles from many nations. For me, the highlight was a series of invited review papers which pro¬ vided superb ‘state-of-the-art’ summaries of many aspects of waterfowl biology. Their publication is imminent (Batt in press) and they will do much to stimulate and direct studies of waterfowl ecology and behaviour over the next decade. More recently I have had the opportunity to read a delightful text on waterfowl ecol¬ ogy (Owen & Black 1990), a book that aims to (and indeed will) serve and stimulate the interests of undergraduate and immediate postgraduate students. I suspect it will also prove a useful review document for many long-established waterfowl biologists. But one conspicuous feature of these compilations, especially when viewed from my home range, is their scant reference to southern hemisphere species and their envi¬ ronments. This is not because Delta’s reviewers or Myrfyn Owen and Jeffrey Black have been parochial or less than exhaustive in their literature surveys. It reflects the fact that tropical and southern hemisphere waterfowl are poorly represented in the ornithological literature and have not been the subject of intense or enduring studies like so many of their northern hemisphere counterparts. This is a pity for it is the tropical and southern hemisphere species that best illustrate and emphasise the diversity and adaptability of the Anatidae. The ecology of northern hemisphere waterfowl is dominated by the intense seasonality of continental climates. Migration is a dominant feature of their life history. That marked seasonality has impacted upon the evolution of their social systems and breeding strategies and promoted major physiological adaptations. In contrast, the climates of tropical and southern land masses appear less season¬ ally extreme and migration is essentially unknown. Nomadism is the response to arid areas but, in general, southern hemisphere waterfowl are sedentary animals. As a consequence social systems are more diverse and breeding seasons are more vari¬ able in timing and duration. In bringing this symposium together, my co-convener Frank McKinney and I have sought to illustrate some of this diversity amongst tropical and southern hemisphere waterfowl. We have chosen specific examples from amongst the ducks (Anatini) and concentrated on the general fields of ecology and behaviour. 842 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI Our contributors (and we regret we have none from southern Africa or continental South America) have been asked to highlight behavioural responses to differing eco¬ logical conditions. In the process the emphasis has been narrowed down to a consid¬ eration of breeding strategies and mating systems. Arid zone breeding adaptations (Sue Briggs), breeding behaviour of a sedentary tropi¬ cal resident (Lisa Sorenson), mating system of a river specialist (Clare Veltman), parental care in southern hemisphere Anas (Frank McKinney), and ecological and behavioural responses to island life (Murray Williams) is our lineup but they represent just a few of the topics that warranted consideration in this symposium. While all of these contributions will be presenting new and exciting findings, none should be re¬ garded as the last word on their topics. We hope the papers will both inform and, more importantly, stimulate further interest in those waterfowl which live close to or south of the equator. LITERATURE CITED BATT, B.D.J. (Ed.) (In press). Ecology and management of breeding waterfowl. Minneapolis, Univer¬ sity of Minnesota Press OWEN, M., BLACK, J.M. 1990. Waterfowl ecology. Glasgow. Blackie. ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 843 BREEDING ADAPTATIONS OF SOUTHERN HEMISPHERE, ARID ZONE DUCKS S. V. BRIGGS and W. G. LAWLER National Parks and Wildlife Service (N.S.W.), P. O. Box 1967, Hurstville, NSW 2220, Australia Current address: C /- Division of Wildlife and Ecology, CSIRO, P.O. Box 84, Lyneham, ACT 2602, Australia ABSTRACT. Ducks inhabiting the arid zones of the southern hemisphere have similar clutch sizes to other non-sedentary ducks, and longer incubation and fledging times than northern migratory species. Arid zone ducks appear able to use the resource peaks in their environment to produce clutches of equivalent size to most other ducks. However, their incubation and fledging times do not conform with the very brief (< three months) inundation of some wetlands in arid areas, in spite of their ability to breed on temporarily flooded wetlands in arid country. Keywords: Arid zone, ducks, southern hemisphere, breeding, clutch size. INTRODUCTION Arid and semi-arid lands are characterised by low (< 500 mm per year) and variable rainfall (MacLean 1976, West 1983, Evenari 1985). Short, heavy bursts of rain are typically followed by long, dry periods (Nicholls & Wong 1990). These alternating wet and dry conditions produce intermittently flooded, productive wetlands in arid and semi-arid regions, on which large numbers of waterbirds feed and breed (Brand 1966, Frith 1967, Uys & MacLeod 1967, Siegfried 1970, Geldenhuys 1982, Gentilli & Bekle 1983, Maher & Carpenter 1984, Maher 1988). The specific aims of this review are to determine how clutch sizes, and incubation and fledging times differ between ducks that breed in arid (including semi-arid) regions, and those that breed in non-arid (humid plus sub-humid) regions. It is based on a longer paper which will be published elsewhere (Briggs ms.). The species covered in this review are those in the tribes Anatini and Aythyini (subfamily Anatinae) and in the monospecific genera Stictonetta and Malacorhynchus (Livezey 1986). Common and scientific names follow Johnsgard (1978), except where superceded by Marchant & Higgins (1990). METHODS Ducks were grouped in six categories (Briggs ms.) by the climate of their breeding habitat (arid/semi-arid or humid/sub-humid), their movement patterns (non-arid taxa only), and whether their mid-breeding latitude is in the northern or southern hemi¬ sphere. The categories were thus: arid , southern hemisphere ducks (AS); arid , north¬ ern hemisphere ducks (AN); non-arid, partial migrants of the southern hemisphere (PS); non-arid, regular migrants of the northern hemisphere (RN); non-arid , seden- 844 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI tary , southern hemisphere ducks (SS); and non-arid, sedentary, northern hemisphere ducks (SN) (Appendix). Waterfowl in the AS and AN categories that mostly or always breed in arid or semi-arid environments (highly arid ducks) were further divided from waterfowl that only sometimes breed in arid or semi-arid environments (partly arid ducks) (Appendix). Clutch sizes, and incubation and fledging times of ducks in these categories were compared. Throughout this paper, arid includes both arid and semiarid, humid or non-arid in¬ cludes both humid and subhumid, and ducks and waterfowl refer specifically to the species and sub-species in Appendix. Masses of eggs and female ducks were ob¬ tained from Rohwer (1988). The source data and references are available from the senior author. Statistical tests follow Zar (1984). All means are expressed with ± SE. RESULTS AND DISCUSSION Clutch characteristics The mean clutch size of AS ducks is similar to the mean clutch size of RN and PS migrants, and to that of SN ducks (Figure 1). However, the average clutch sizes of waterfowl in all these groups are significantly larger than the average clutch size of SS waterfowl (ANOVA, F = 5.64, df = 5, 68, Tukey test, q = 6.50, P < 0.001) (Figure 1). Clutch sizes did not differ between highly arid waterfowl, partly arid southern, and partly arid northern waterfowl (ANOVA, F = 1.32, df = 2, 19, P > 0.25). The mean clutch size (both hemispheres combined) of sedentary island ducks is 6.7 ± 0.7 (n=10), and that of sedentary mainland ducks is 7.7 ± 0.5 (n=15). 10 9 0) N CO 8 _C o 3 o 7 6 FIGURE 1 - Mean clutch sizes of non-arid sedentary, southern (SS); non-arid partial mi gratory, southern (PS); non-arid sedentary, northern (SN); arid, southern (AS); and regu Jar migratory, northern (RN) ducks. Vertical bars are standard errors; numbers in the his tograms are sample sizes. Data from Briggs (ms.). ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 845 The average ratios of egg to female mass (Rohwer 1988) in all categories of ducks, except island forms, are similar (arid (AS plus AN) 0.075 ± 0.003, n=20; regular mi¬ grants (RN), 0.072 ± 0.002, n=22; partial migrants (PS), 0.076 ± 0.005, n = 8; sed¬ entary mainland ducks (combined northern and southern hemisphere taxa), 0.077 ± 0.007, n=11). In sedentary island ducks (combined hemispheres) the ratio of egg to female mass is 0.093 ± 0.006 (n=7). Their similar average clutch sizes, and egg to female body mass ratios, indicate that egg production in arid zone ducks is not more constrained by low food resources (or by other environmental factors) before or during laying, than it is in PS or RN ducks. Soils in arid and semi-arid regions generally contain fewer nutrients than soils in hu¬ mid areas, although there is considerable variation between continents (West 1981, Stafford Smith & Morton 1990). However, even wetlands in infertile arid country may be as high in nutrients as wetlands in more fertile, wetter regions (Williams et al. 1970, Briggs et al. 1985, unpubl. data). This is because run-off water carries nutrients into temporarily flooded run-on areas (Stafford Smith & Morton 1990) which form wetlands, and because wetting and drying of soils, and associated plant growth, enhance cy¬ cling of nutrients and organic matter (West 1981, Briggs et al. 1985, references therein). This alternating flooding and drying of wetlands in arid areas enhances pro¬ duction of waterfowl food resources, particularly invertebrates, which allow waterfowl to breed (Frith 1959, Brand 1966, Braithwaite & Frith 1969, Maher & Carpenter 1984). Food supplies for waterfowl increase in wetlands in both arid and non-arid regions following watering after a dry period (Swanson & Meyer 1977, Van der Valk & Davis 1978, Danell & Sjoberg 1982). However, the effects of wetting and drying on produc¬ tivity of waterfowl foods may be more marked in wetlands in arid regions because the greater variation in rainfall (Nicholls & Wong 1990) increases the area of wetlands that dry and reflood in such regions. Breeding in arid zone waterfowl, especially in AS ducks, is related to a combination of wetland flooding and photoperiod, with the effect of the former usually predominat¬ ing (Siegfried 1965, Braithwaite & Frith 1969, Braithwaite 1976 a,b, Halse & Jaensch 1989, Lawler & Briggs 1991). In highly arid ducks, notably in Pink-eared Duck Malacorhynchus membranaceus, Cape Teal Anas capensis and Grey Teal Anas gracilis, laying is entirely dependent on good rains and wetland flooding (Siegfried 1974, Braithwaite 1976 a,b, MacLean 1985;, Halse & Jaensch 1989, Lawler & Briggs 1991). At least in Australia, highly arid ducks rarely breed in years when rainfall is sparse and intermittent wetlands are dry or almost so (Frith 1959, Gentilli & Bekle 1983, Halse & Jaensch 1989, Lawler & Briggs 1991). Clutch sizes of arid zone ducks thus appear to be adapted to the resource peaks of their breeding habitat rather than to its average resource availability. The smaller clutch sizes of SS ducks compared with non-sedentary ducks supports Ashmole’s hypothesis which proposes that avian clutch sizes are related to differ¬ ences in resource levels between breeding and nonbreeding seasons (Ashmole 1963, Ricklefs 1980). Resource levels in the habitats of non-sedentary waterfowl should vary much more than those in the habitats of sedentary taxa, especially ducks on islands. Island waterfowl have lower clutch sizes than mainland taxa Weller (1980). However, clutch sizes of SN ducks are similar to those of non-sedentary ducks, and larger than those of SS ducks. 846 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI Incubation and fledging times Mean incubation time in AS waterfowl is not significantly different from mean incuba¬ tion times in PS, SS and SN ducks (Figure 2). However, average incubation times in these four groups of ducks are significantly longer than the average incubation time in RN migrants (ANOVA, F = 3.64, df = 4, 64, Tukey test, q = 4.42, P < 0.05) (Fig¬ ure 2). Incubation times of highly arid, of partly arid southern, and of partly arid north¬ ern waterfowl do not differ (ANOVA, F = 0.97, df = 2, 19, P > 0.25). Limited data on fledging times indicate that RN ducks fledge significantly sooner than AS or seden¬ tary ducks (ANOVA, F = 7.46, df = 2, 36, Tukey test, q = 6.23, P < 0.01) (Figure 3). Category of Waterfowl FIGURE 2 - Mean incubation periods of non-arid regular migratory, northern (RN); arid, southern (AS); non-arid partial migratory, southern (PS), sedentary, northern (SN); and sedentary southern (SS) ducks. Vertical bars are standard errors; numbers in the histo¬ grams are sample sizes. Data from Briggs (ms.). Durations of incubation and fledging in waterfowl generally decrease as latitude in¬ creases (Lack 1968, Johnsgard 1978). Average incubation and fledging periods in AS and AN species (Figure 2) are as expected for ducks breeding at their latitudes. The relatively long incubation and fledging times in AS and AN ducks, compared with the short times in RN migrants, suggest that wetlands in arid areas are usually inundated long enough for ducklings to hatch and fly (> three months). Not all wetlands in arid ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 847 country persist this long, however (Frith 1959, Geldenhuys 1982, Maher 1988), and arid zone waterfowl usually breed on wetlands that hold water for several months, although they may feed on more ephemeral waters (Maher 1988, Lawler & Briggs 1991). Thus, arid zone waterfowl apparently have not reduced their incubation and fledging times to conform with the brief flood periods of some wetlands in deserts and semideserts (also see Fullagar et al. 1988). Category of Waterfowl FIGURE 3 - Mean fledging periods of non-arid regular migratory, northern (RN); arid, south¬ ern (AS); and combined sedentary, southern (SS) plus sedentary, northern (SN) ducks. Vertical bars are standard errors; numbers in the histograms are sample sizes. Data from Briggs (ms.). CONCLUSIONS In summary, arid zone ducks are able to use the fluctuating food resources of their environment to produce clutches of similar size as other non-sedentary ducks. They cope with their variable environment by laying only when conditions are good, not necessarily every year. Arid zone ducks take longer or as long to fledge their duck¬ lings as most other ducks, in spite of their ability to breed on temporarily flooded wetlands. ACKNOWLEDGEMENTS The comments of Peter Fullagar, Richard Kingsford, Ian Norman, Murray Williams and Frank McKinney substantially improved the manuscript. The Division of Wildlife and Ecology, CSIRO provided library and laboratory facilities. The National Parks and 848 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI Wildlife Service (N.S.W.) funded the study. We are grateful to these people and or¬ ganisations. LITERATURE CITED ASHMOLE, N. P. 1963. The regulation of numbers of tropical oceanic birds. Ibis 103B: 458-473. BRAITHWAITE, L. W. 1976a. Breeding seasons of waterfowl in Australia. Proceedings of the Interna¬ tional Ornithological Congress 16: 236-247. BRAITHWAITE, L. W. 1976b. Environment and timing of reproduction and flightlessness in two spe¬ cies of Australian ducks. Proceedings of the International Ornithological Congress 16: 489-501. BRAITHWAITE, L. W., FRITH, H. J. 1969. Waterfowl in an inland swamp. Ill Breeding. CSIRO Wild¬ life Research 14: 65-109. BRAND, D. J. 1966. Nesting studies of the Cape Shoveller Spatula capensis and the Cape Teal Anas capensis in the Western Cape Province 1957-1959. Ostrich Supplement 6: 217-221. BRIGGS, S. V. ms. Characteristics of arid zone ducks of the southern hemisphere. Submitted to Corella. BRIGGS, S. V., MAHER, M.T., CARPENTER, S. M. 1985. Limnological studies of waterfowl habitat in south-western New South Wales. I. Water chemistry. Australian Journal of Marine and Freshwater Research 36: 59-67. DANELL, K., SJOBERG, K. 1982. Successional patterns of plants, invertebrates and ducks in a man¬ made lake. Journal of Applied Ecology 19: 395-409. EVENARI, M. 1985. The desert environment. Pp. 1-22 in Evari, M., Noy-Meir, I., Goodall, D. W. (Eds). Ecosystems of the World, volume 12A. Hot deserts and arid shrublands, A. Amsterdam, Elsevier. FRITH, H. J. 1959. The ecology of wild ducks in inland New South Wales. IV. Breeding. CSIRO Wildlife Research 4: 1 56-181 . FRITH, H. J. 1967. Waterfowl in Australia. Sydney, Angus and Robertson. FULLAGAR, P. F., DAVEY, C. C., RUSHTON, D. K. 1988. Is it true that Australian ducks are differ¬ ent? Pp. 81-98 in Gilligan, B., Maddock, M., McDonald, K. (Eds). Proceedings of an International Sym¬ posium on Wetlands. Newcastle, Shortland Wetlands Centre. GELDENHUYS, J. N. 1982. Classification of the pans of the western Orange Free State according to vegetation structure, with reference to avifaunal communities. South African Journal of Wildlife Re¬ search 12: 55-62. GENTILLI, J., BEKLE, H. 1983. Modelling a climatically pulsating population; Grey Teal in south-west¬ ern Australia. Journal of Biogeography 10: 75-96. HALSE, S. A., JAENSCH, R. P. 1989. Breeding seasons of waterbirds in south-western Australia - the importance of rainfall. Emu 89: 232-249. JOHNSGARD, P. A. 1978. Ducks, geese and swans of the world. Lincoln, University of Nebraska Press. LACK, D. 1968. Ecological adaptations for breeding in birds. London, Methuen. LAWLER, W., BRIGGS, S. V. 1991. Breeding of Maned Ducks and other waterbirds on ephemeral wetlands in north-western New South Wales. Corella 15: in press. LIVEZEY, B. C. 1986. A phylogenetic analysis of recent Anseriform genera using morphological char¬ acters. Auk 103: 737-754. MacLEAN, G. L. 1976. Arid-zone ornithology in Africa and South America. Proceedings of the Inter¬ national Ornithological Congress 16: 468-480. MacLEAN, G. L. 1985. Roberts’ birds of southern Africa. Cape Town, John Voelcker Bird Book Fund. MAHER, M. 1988. Wetlands and waterbirds in the arid Australian inland - some principles for their conservation. Pp. 285-294 in Gilligan, B., Maddock, M., McDonald, K. (Eds). Proceedings of an Inter¬ national Symposium on Wetlands. Newcastle, Shortland Wetlands Centre. MAHER, M., CARPENTER, S. M. 1984. Benthic studies of waterfowl breeding habitat in south-west¬ ern New South Wales. II. Chironomid populations. Australian Journal of Marine and Freshwater Re¬ search 35: 97-1 1 0. MARCHANT, S., HIGGINS, P. (Eds). 1990. Handbook of Australian, New Zealand and Antarctic birds. Oxford, Oxford University Press. NICHOLLS, N., WONG, K. K. 1990. Dependence of rainfall variability on mean rainfall, latitude, and the southern oscillation. Journal of Climate 3: 163-170. RICKLEFS, R. E. 1980. Geographical variation in clutch size among passerine birds: Ashmole’s hy¬ pothesis. Auk 97: 38-49. ACTA XX CONGRESSUS INTERNATIONAL^ ORNITHOLOGICI 849 ROHWER, F. C. 1988. Inter- and intraspecific relationships between egg size and clutch size in wa¬ terfowl. Auk 105: 161-171 . SIEGFRIED, W. R. 1 965. The Cape Shoveller Anas smithii in southern Africa. Ostrich 36: 1 55-1 98. SIEGFRIED, W. R. 1970. Wildfowl distribution, conservation and research in southern Africa. Wildfowl 21: 89-98. SIEGFRIED, W. R. 1974. Brood care, pair bonds and plumage in southern African Anatini. Wildfowl 25: 33-40. STAFFORD SMITH, D. M., MORTON, S. R. 1990. A framework for the ecology of arid Australia. Jour¬ nal of Arid Environments 18: 255-278. SWANSON, G. A., MEYER, M. I. 1977. Impact of fluctuating water levels on feeding ecology of breed¬ ing Blue-winged Teal. Journal of Wildlife Management 41: 426-433. UYS, C. J., MacLEOD, J. G. R. I. 1967. The birds of the De Hoop Vlei region, Bredasdorp, and the effect of the 1957 inundation over a 10-year period (1957-1966) on the distribution of species, bird numbers and breeding. Ostrich 38: 233-254. Van der VALK, A. G., DAVIS, C. B. 1978. The role of seed banks in the vegetation dynamics of prai¬ rie glacial marshes. Ecology 59: 322-335. WELLER, M. W. 1980. The island waterfowl. Ames, Iowa State University Press. WEST, N. E. 1 981 . Nutrient cycling in desert ecosystems. Pp. 301 -324 in Goodall, D. W., Perry, R. A. (Eds). Arid land ecosystems. Volume 2. Cambridge, Cambridge University Press. WEST, N. E. 1983. Approach. Pp. 1-2 In WEST, N. E. (Ed.). Ecosystems of the world. Volume 5. Tem¬ perate deserts and semi-deserts. Amsterdam, Elsevier. WILLIAMS, W. D., WALKER, K. F., BRAND, G. W. 1970. Chemical composition of some inland sur¬ face waters and lake deposits of New South Wales, Australia. Australian Journal of Marine and Fresh¬ water Research 21: 103-115. ZAR, J. H. 1984. Biostatistical analysis. Second edition. Englewood Cliffs, Prentice-Hall. APPENDIX Mid breeding latitudes of arid and semi-arid zone ducks (AS and AN), non-arid regular (RN) and partial (PS) migratory ducks, and sedentary (SS and SN) ducks. The semi- arid and arid zone ducks are split according to whether they mostly (highly arid) or sometimes (partly arid) breed in such environments. The sedentary waterfowl are split into sedentary, mainland (sedentary) and sedentary, island (island) taxa. Data from Briggs (ms.). 1 Regarded as southern hemisphere breeders. Taxon Latitude Taxon Latitude Highly arid Stictonetta neavosa 30°S A. u. undulata 1 5°S Malacorhynchus membranaceus 30°S A. erythrorhyncha 1 8°S Anas capensis 1 2°S A. smithii 23°S A. gracilis 30°S A. rhynchotis 35°S Partly arid A. sibiiatrix 45°S A. platalea 35°S A. f. flavirostris 40°S Chenonetta jubata 30°S A. platyrhnchos diazi 25°N Marmaronetta angustirostris 37°N A. castanea 35°S Netta rufina 33°N A. superciliosa rogersi 28°S N. erythropthalma brunnea 8°S A. v. versicolor 37°S Aythya australis 28°S A. hottentota 1 7°S Aythya nyroca 40°N 850 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI Taxon Latitude Taxon Latitude Regular migrants Aix sponsa 42°N A. disco rs 48°N A. galericulata 45°N A. cyamopter septentrionalium 40°N Anas penelope 58°N A. clypeata 55°N A. americana 55°N Aythya valisineria 58°N A. falcata 62°N A. americana 50°N A. strepera 45°N A. ferina 45°N A. formosa 60°N A. co Haris 52°N A. crecca 56°N A. baeri 50°N A. p. platyrhynchos 49°N A. fuligula 58°N A. rubripes 47°N A. m aril a 62°N A. a. acuta A. querquedula 58°N 51 °N A. affinis 58°N Partial migrants Nettapus pulchellus 1 5°S A. s. specularioides 45°S N. coromandelianus 1 0° A. georgica spini cauda 27°S N. auritus 1 0° Callonetta leucophyrys 25°S Anas platyrhynchos maculosa A. spec u laris 27°N 46°S Netta peposaca 35°S Sedentary Pteronetta hartlaubi 8°N A. poecilorhyncha 20°N Cairina moschata 5°S A. s. superciliosa 45°S C. scutulata 7°N A. b. bahamensis 1 7°N Anas waigiuensis 5°S A. versicolor puna 1 5°S A. s. sparsa 22°S Amazonetta brasiliensis 10°S A. g. gibberifrons 5°S Rhodonessa caryophyllacea 30°N A. aucklandica chlorotis A. playhynchos fulvigula 41 °S 27°N Aythya novaeseelandiae 41 °S Island Anas gibberifrons albogu laris 1 2°N A. pelewensis 5°S A. a. aucklandica 50°S A. luzonica 1 2°N A. platyrhynchos wyvil liana 20°N A. acuta eatoni 52°S A. p. laysanensis 25°N A. g. georgica 54°S A. melleri 20°S Aythya innotata 20°S ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 851 MATING SYSTEMS OF TROPICAL AND SOUTHERN HEMISPHERE DABBLING DUCKS L. G. SORENSON Bell Museum of Natural History, Dept, of Ecology and Behavioural Biology, University of Minnesota, Minneapolis, Minnesota 55455, USA Current address: Smithsonian Institution, Conservation and Research Center, National Zoological Park, Front Royal, Virginia 22630, USA ABSTRACT. Many aspects of the behaviour of northern hemisphere dabbling ducks (e.g. seasonal mo¬ nogamy, female-only parental care) are probably the result of their migratory lifestyle and short, an¬ nual breeding season. The sedentary lifestyle and extended or irregular breeding seasons of White¬ cheeked Pintails Anas bahamensis bahamensis in the Bahamas may explain several aspects of this species’ mating system that differ from its northern counterparts. During a three year field study, I found that: 1)4- 9% of paired males had two mates each year; 2) both long-term pair bonds and mate changes occurred: 3) females alone cared for ducklings but some males continued to escort and de¬ fend their mates during brood-rearing; and 4) courtship and competition for mates occurred year-round. A review of the breeding ecology of eight tropical/southern hemisphere Anas suggests that variation in movement patterns and in the timing and duration of breeding seasons in these species contributes to greater variation in mating systems and greater intraspecific variability in behaviour. Keywords: White-cheeked Pintail, Anas bahamensis, Anatidae, mating system, polygyny, pair-bond duration, courtship, reproductive strategies, tropics, southern hemisphere. INTRODUCTION The mating system of dabbling ducks ( Anas spp.) breeding in temperate or sub-arc¬ tic regions of the northern hemisphere (NH) varies little among species: all have monogamous pair bonds, highly seasonal courtship and pairing, and female-only parental care. Tropical and southern hemisphere (TSH) dabbling ducks appear to show much greater variation in mating systems: pair bonds may persist for more than one year, courtship may occur throughout the year, polygyny has been documented in several species and some species have biparental care. The factors favouring these characteristics are poorly understood, primarily because few TSH species have been studied intensively. Several recent reviews of mating systems conclude that the basically monogamous mating system of temperate NH ducks is a product of their highly seasonal breeding environment and consequently migratory lifestyle (McKinney 1985, 1986, Rohwer & Anderson 1988, Oring & Sayler in press). New pairs form each year on the wintering grounds and return together to the female’s natal area in spring. After escorting and defending their mate during winter and spring, males desert the female during incu¬ bation and move to a safe area for the wing-moult. Opportunities for polygyny are lim¬ ited because 1) males cannot follow more than one female back to the breeding grounds and 2) synchronous breeding and male-biased sex ratios limit the ability of males to acquire a second mate during the breeding season. Benefits to males of early wing-moult and preparation for fall migration apparently outweigh benefits of remaining with the female and providing parental care and/or 852 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI maintaining the pair bond until the next year. Long-term pair bonds, biparental care and polygyny have not been documented in any of these species. Ecological factors are strikingly different for dabbling ducks breeding in the tropics and southern hemisphere. First, mild climates and variable patterns of rainfall result in more variation in the timing and duration of breeding seasons. In some species, breeding may occur in any month of the year and may continue for many months if suitable wetland conditions persist (Siegfried 1974, Braithwaite 1976a,b, Johnsgard 1978). A second consequence of mild climates is that many TSH species do not un¬ dergo long-distance seasonal migrations; some species or populations are sedentary, while others are nomadic (see Briggs, this symposium). In this paper, I examine the relationship between these basic ecological factors and the variable mating systems of TSH dabbling ducks with emphasis on my own study of the White-cheeked Pintail Anas bahamensis bahamensis in the Bahamas. A more detailed analysis of the White-cheeked Pintail mating system is presented in Sorenson (in review). WHITE-CHEEKED PINTAIL BREEDING SCHEDULES AND MATING SYSTEM The bahamensis subspecies of the White-cheeked Pintail is resident in the West Indies and an adjacent part of South America. Brackish or salt water ponds and mangrove marshes are the preferred habitat. Dur¬ ing a three year field study (1985 - 1987) in the Bahamas, I recorded the breeding activities, social interactions and pair-bond relationships of a color marked population on Paradise Island and other nearby islands and cays northeast of Nassau (Sorenson 1990, in review). The timing and duration of breeding seasons in this population are variable and greatly influenced by the timing and amount of winter and spring rainfall. The earli¬ est nests were initiated in February in 1987 but not until May in 1985. Breeding sea¬ sons also may be extended: nests were initiated over an 86 day interval in 1987. Breeding seasons are even more variable in populations closer to the equator. In Puerto Rico for example, nests have been found in every month of the year (E. Rodriguez, pers. comm.). Males of breeding pairs in my study population were highly territorial throughout the female’s breeding cycle and many males engaged in forced extra-pair copulation at¬ tempts. Also, non-breeding was a regular occurrence: an average of 33.6% of females apparently made no attempt to breed each year. These aspects of the White-cheeked Pintail’s breeding ecology and social system will be addressed elsewhere (Sorenson, in prep.). Polygyny White-cheeked Pintails in the Bahamas usually paired monogamously, but a low level of polygyny (4 - 9 % of paired males) occurred regularly (Table 1). In four of the six trios studied, the male divided his time between two mates, spending more time with one female that was in pre-laying or laying condition and then switching attention to ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 853 the second female when the first was incubating and tending her brood. A fifth trio was formed when a male, while still maintaining a pair bond with his first mate who was raising a brood, courted and paired with another brood female whose mate had been killed. In each of these five cases, aggression between the two females was evident whenever the trio was together. TABLE 1 - Proportion of males pairing polygynously and number of unpaired males for 1985, 1986 and 1987 field seasons (from Sorenson (in review)). Year No. paired males which were polygynous (%)* No. marked unpaired males 1985 1 / 14 (8.3) 9 1986 3 / 34 (8.8) 11 1987 2 / 45 (4.4) 13 * From Sorenson (in review). Data include monogamous pairs and polygynous trios with at least one mate marked. In the sixth trio, all three individuals associated closely with one another with very little aggression between the females. In this case, the two females bred synchronously, initiating nests within one day of each other. Initially, the male of this trio was simul¬ taneously paired to three females. He copulated with all three females and was en¬ gaged in almost constant mate defence as many unpaired males repeatedly tried to approach and court his three mates. After several weeks, one female finally paired with another male. The maintenance of these polygynous pair bonds was remarkable given that the sex ratio of this population was strongly male-biased (1.3 - 1.4 males : 1.0 females) and many males remained unpaired during the breeding season (Ta¬ ble 1). All females were paired and frequently were courted and harassed by both unpaired and paired males. Two factors probably allow the formation of polygynous bonds in this population. First, because most individuals are sedentary, males have an opportunity to form and main¬ tain pair bonds with two females breeding in the same location. This is not possible for NH dabbling ducks because pair formation occurs away from the breeding grounds. Moreover, individuals are able to interact with one another year-round and year after year, allowing the formation of stable dominance relationships. The fact that individuals “know” each other may facilitate female assessment of male quality. Fe¬ males that were not protected by their mates from courtship and harassment by other males spent less time feeding and often failed to initiate a nesting attempt. Polygynous males were extremely aggressive, guarded their mates effectively, and were dominant over other males they encountered. Thus, females might pair polygynously with a known male of high quality rather than monogamously with a poor quality male. Second, extended breeding seasons result in asynchrony in female breeding sched¬ ules and may create opportunities for males to acquire more than one mate (McKinney 1985). A male may escort and guard an early breeding female during her pre-laying and laying period and then switch attention to a second female when the first becomes occupied with incubation and brood-rearing. Although this sequence of events occurred in most of the polygynous relationships I documented, the two 854 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI females in one trio nested synchronously, indicating that breeding asynchrony was not a prerequisite for polygyny. Pair-bond Duration Many pairs remained together for two or more breeding seasons but mate changes were also frequent even when both members of a pair survived to the next season (Sorenson, in review). Of 37 marked pairs studied in 1985 and 1986, both members of 23 pairs were still alive in the following year. Ten of these 23 pairs (43%) remained together in the following breeding season while 13 pairs (57%) divorced. Including data from a brief visit to the study area in 1988, three pairs were known to remain together for three consecutive years. There are several potential advantages for both male and female birds of remaining with the same mate from year to year. Established pairs may obtain better feeding sites or territories, save time and energy by avoiding courtship, and, through famili¬ arity with patterns of individual behaviour, better coordinate activities and movements (Rowley 1983, McKinney in press). Mate retention is most feasible in non-migratory species (Rowley 1983), and the reunion of mates is more likely in sedentary populations of ducks even when pairs split up during brood-rearing and/or the non¬ breeding season. Considering the potential advantages of retaining the same mate, the divorce rate in this population of White-cheeked Pintails seems high. Contrary to results from other studies (e.g. Coulson & Thomas 1983), I found no relationship between breeding success and subsequent mate retention: rates of divorce were similar among pairs that were successful in raising ducklings (7 of 1 1 pairs) and those that were unsuc¬ cessful (6 of 12 pairs, G = 1.37, P > 0.2). Courtship and Pairing Several additional aspects of the White-cheeked Pintail’s mating system differ from that of its NH counterparts (Sorenson, in review). First, at least some birds remained paired year-round. Although males did not provide parental care, many males contin¬ ued to escort and defend their mates for at least part of the brood-rearing period, while other pairs which had separated during brood-rearing reunited after ducklings fledged. Several pairs remained together during the wing-moult, and pairs were also seen during the non-breeding season (September - November). As in NH dabbling ducks, courtship activity and pair formation occurred well in ad¬ vance of the breeding season but, in marked contrast to NH species, some social courtship was observed year-round. In particular, males showed intense interest in and actively courted females tending broods. Courtship was also observed in June and July when breeding activity had ended and some birds were beginning the wing- moult. This suggests that the seemingly low level of mate fidelity among successfully breeding pairs in my study may have been a consequence of continuing social court¬ ship and competition for mates throughout the year. It may be very difficult for males and females to keep a good mate from one year to the next. Siegfried (1974) suggested that if the onset of conditions suitable for breeding is un¬ predictable, birds may benefit from remaining in pairs year-round so that nesting can begin as soon as conditions permit. Although breeding was seasonal in my study ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 855 c/d ctj c: Q) CD JZ CL CO E CD 0 JZ ZJ o 0 TO £Z 0 "aj o Cl O g> ’0 E 0 00 0 "0 o o 0 0 0 o 0 CL 0 0 TO CZ 0 C 0 -4— ' 0 CL •4— » C 0 E 0 > o E cf o 0 0 0 0 CO c TO 0 0 CD CM LU CD < h- a> o QC c o _0 c CD E E CD (1) > 2Z O T3 2 GL 0 .2 o CD a CO C\J o T— v— *T— CO CO T— -T— T- CO CM LO CO CD 'I— y— o c 0 CD 0 TO CO c0 > o c 0) to CO c0 > 0 0 o c: 0 CD C^- o c o c o c 0 CD C/D o3 CO O) c c 0 c TO 0 TO 1— _0 -= TO 55 0 E TO TO o o c 0 o c 0 V) 0 0 CO) 9 0 0 0 0 0 0 0 0 X k_ (D X CQ 0 0 0 L_ 0 0 CO CO c o 0 CO CD 0 0 O 0 0 0 CL 0 CD c .2 o ^ =3 < Q 0 .0 0 0 I- C 0 g cl 9 0 g o 3, 0 : 0 i— 0 TO ^ (D 0 z 8 o ^ 0 CL C/D 0 -= .52 s » i- s c c 0 0 0 0 0 «? CO TO ^ fl CD 0) CD X »- 0 0 E * Q_ * _ TO * 0 0 0 C SZ 0 V E 0 0 .tr -c -C 0 £ s 06 CO 0 o CD ° .0 0 ^ > q> CD 0 0 0 0 >> o3 CO C o 0 0) 0 0 >> 0 0 >* 0 3 0 0 0 0 O O 0 0 o 0 >- o c > >> c > C^- c^- c^- CO) CO) c c E Z E Z »- CJ i- o » 0 2 2 2 2 2 2 2 2 ■= TO 0 0 E TO 0 c TO 0 TO i5 ° c o c 3 m 0 0 0 0 CO) 0 0 X 0 X 0 0 0 0 V— 5 | 0 0 0 0 0 H r~~ o >> O ~ TO 0 0 6 2. 5 3 Movement pattern: S = Sedentary, N = Nomadic, M = Migratory. Mating System: M = Monogamy, P = Polygyny. * Field study with marked birds. ** Studied in captivity Reference numbers: 1 . McKinney et al. 1 978; 2. Siegfried 1974; 3. Siegfried et al. 1976; 4. Stolen & McKinney 1983; 5. Standen 1976; 6. McKinney 1985' 7 McKinney & Brewer 1989; 8. Moulton & Weller 1984; 9. McKinney & Bruggers 1 983; 10. Sorenson 1990, in review; 1 1 . Weller 1 968; 12. Brewer 1990; 13. Braithwaite 1976a, b; 14. Marchant & Higgins 1990. 856 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI population, the start of the breeding season was quite variable and breeding is more irregular on other islands in the archipelago. This unpredictability probably favours year-round courtship and maintenance of pair bonds in the non-breeding season in White-cheeked Pintails. Where breeding seasons are extended, female dabbling ducks may be able to raise more than one brood in a single season (Frith 1959, Braithwaite 1976a,b, Fullagar & Davey, in press). Although not documented during my study, one marked female in this population in 1990 (L. Rutan, pers. comm.) successfully fledged her first brood and then initiated a second nest while still paired to the same male (the fate of the second nesting attempt was unknown). Thus, in certain years, the possibility of rais¬ ing two broods within a 12 month period might favour continued pair-bond mainte¬ nance by males during the brood-rearing period. Males also may court brood females to whom they are not already paired in attempts to establish pair bonds with females of proven breeding ability (as evidenced by the presence of ducklings) for a future breeding attempt, either in that season or the next. Both of these male strategies are facilitated by a sedentary lifestyle. OTHER TROPICAL AND SOUTHERN HEMISPHERE SPECIES Information is available on the breeding ecology and mating system of seven other TSH dabbling ducks (Table 2). Although intensive field studies of marked birds have been carried out only on the African Black Duck A. sparsa, Laysan Teal A. laysanensis, and Grey Teal A. gracilis, preliminary comparisons with findings on the White-cheeked Pintail are possible. Polygyny The White-cheeked Pintail is the only dabbling duck for which polygyny has been recorded regularly in a wild population. Instances of polygyny have been reported in captive Cape Teal A. capensis and Speckled Teal A. flavirostris, however, and the presence of year-round courtship in both of these species suggests further similarity to the White-cheeked Pintail system. In captivity, the formation of polygynous trios in all three species began with a paired male courting a second female while his first mate was incubating. In each case, the polygynous male succeeded in dominating the second female’s mate and breaking up their pair bond. Once paired to the second female, the division of his time between his two mates depended on the stage of their breeding cycles. As suggested for White-cheeked Pintails, the extended and/or irregu¬ lar breeding seasons of Cape Teal and Speckled Teal may provide opportunities for males to obtain two mates. Factors associated with a sedentary lifestyle also may facilitate polygyny in some populations of Speckled Teal. Cape Teal, on the other hand, are highly nomadic. Siegfried (1974) notes, however, that birds usually travel in pairs and small flocks. If these flocks represent fairly stable groups of individuals, similar factors (e.g. stable dominance relationships) could also apply to this species Pair-bond duration Long-term pair bonds have been documented in African Black Duck, Laysan Teal Grey Teal and Cape Teal and they are thought to occur in Silver Teal A. versicolor and Chiloe Wigeon (A. sibilatrix). African Black Duck and Laysan Teal are sedentary seasonal breeders but neither has biparental care, and pair bonds weaken or break ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 857 temporarily during the brood-rearing period (African Black Duck) or incubation (Laysan Teal). The African Black Duck is a highly territorial river specialist: pairs stay together and in residence on the territory almost year-round. Pair bonds can last for several years and only territory holding pairs breed. The importance of cooperation by mates in defense of the territory may favour mate retention in this species (McKinney et al. 1978). Both Cape Teal and Grey Teal have extended and irregular breeding seasons. As in White-cheeked Pintails, it may be advantageous for individuals to remain paired for unpredictable breeding opportunities and for subsequent breeding attempts after the first brood has fledged or is lost (Sorenson, in review). Grey Teal, Cape Teal, Silver Teal, and Chiloe Wigeon all have biparental care and conspicuously strong pair bonds which apparently remain intact through the wing- moult. In sedentary populations or where pairs remain together during the wing-moult and migration, benefits of desertion for males should be lower as compared to NH dabbling ducks, making both biparental care and the maintenance of long-term pair bonds more likely. Alternatively, biparental care in certain TSH species may enhance duckling survival by improving detection of avian predators (McKinney & Brewer 1989; McKinney, this symposium). Regardless of what factors select for biparental care in TSH species, if it occurs, mate retention may be indirectly affected as well. Strong selection for biparental care would reduce the relative costs of maintaining long-term pair bonds and advantages associated with retaining the same mate may be particu¬ larly important in species with biparental care. Selection for either long-term pair bonds or biparental care should indirectly make the other more likely. Biparental care does not occur, however, in all species with long-term pair bonds (e.g. White-cheeked Pintail, African Black Duck, Laysan Teal). Although the factors influencing mate retention in Laysan Teal are unknown, the an¬ nual divorce rate in this species is high (58%, n = 19) as in White-cheeked Pintails. Observations of birds in pairs year-round, biparental care and/or male attendance of brood females in many additional species probably contribute to the long-standing belief that pair bonds in TSH dabbling ducks are permanent (Weller 1968, Kear 1970, Siegfried 1974, Johnsgard 1978). Mate-switching has now been found to be common in two of these species. Future studies of individual species need to critically assess whether long-term pair bonds are the rule or simply one variant. CONCLUSION Limited and/or nomadic movements, unpredictability in the timing of breeding, and extended breeding seasons all result in a much less ordered annual cycle for TSH dabbling ducks than for their NH counterparts and probably contribute to greater vari¬ ation in mating systems. Differences in these basic ecological factors probably explain why polygyny, long-term pair bonds, and year-round courtship and pairing occur in White-cheeked Pintails but not in NH dabbling ducks. Other sedentary species (or species in which the sexes travel together) with extended and irregular breeding sea¬ sons may share many of the features of the White-cheeked Pintail mating system. White-cheeked Pintails also show great variability in individual patterns of behaviour. In particular, pair-bond duration and the length of time pairs remain together during the annual cycle is highly variable. Although long-term pair bonds appear to be the 858 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI predominant pattern in some species (e.g. Chiloe Wigeon, African Black Duck), it is apparent that not all TSH Anas can be characterized as having either long-term or seasonal pair bonds. Similarly, biparental care, male attendance of brood females, year-round courtship, mate-switching, territoriality, and polygyny may occur at vary¬ ing frequencies in different species or populations. Future study of TSH species, par¬ ticularly those showing intraspecific variability in behaviour, should greatly enhance our understanding of how ecological pressures shape the social system. ACKNOWLEDGEMENTS I am very grateful to Frank McKinney and Michael Sorenson for many valuable com¬ ments on the manuscript. I also thank my many field assistants for their dedication and hard work, the Bahamas Ministry of Agriculture and Fisheries for permission to conduct this research and the Bahamas National Trust, Pericles and Christina Maillis, and Michael Lightbourne for logistical support. Financial support was provided by the Dayton Natural History Fund of the Bell Museum of Natural History, the Chapman Memorial Fund of the American Museum of Natural History, Sigma Xi, the Explorer’s Club, a University of Minnesota Doctoral Dissertation Fellowship and grants to F. McKinney from the National Science Foundation (BNS-831 71 87). LITERATURE CITED BRAITHWAITE, L. W. 1976a. Breeding seasons of waterfowl in Australia. Proceedings of the Interna¬ tional Ornithological Congress 16: 235-247. BRAITHWAITE, L. W. 1976b. Environment and timing of reproduction and flightlessness in two spe¬ cies of Australian ducks. Proceedings of the International Ornithological Congress 16: 486-501. BREWER, G. L. 1 990. Parental care behavior of the Chiloe Wigeon (Anas sibilatrix). Unpub. Ph.D. the¬ sis. University of Minnesota. COULSON, J C., THOMAS, C. S. 1983. Mate choice in the Kittiwake Gull. Pp. 361-376 in Bateson, P. (Ed.). Mate choice. Cambridge, Cambridge University Press. FRITH, H. J. 1959. The ecology of wild ducks in inland New South Wales. IV. Breeding. CSIRO Wildlife Research 4: 156-181 . JOHNSGARD, P. A. 1978. Ducks, geese, and swans of the world. Lincoln, University of Nebraska Press. MARCHANT, S., HIGGINS, P. 1990. Handbook of Australian, New Zealand and Antarctic birds. Ox¬ ford, Oxford University Press. (Text on Grey Teal Anas gracilis). MCKINNEY, F., SIEGFRIED, W. R., BALL, I. J., FROST, P. G. 1978. Behavioural specializations for river life in the African Black Duck (Anas sparsa Eyton) Zeitschrift fur Tierpsychologie 48: 349-400. MCKINNEY, F., BRUGGERS, D. J. 1983. Status and breeding behavior of the Bahama pintail and the New Zealand blue duck. Proceedings of the 1983 Jean Delacour/IFCB Symposium on breeding birds in captivity, Pp. 211-221. Hollywood, Calif. MCKINNEY, F. 1985. Primary and secondary male reproductive strategies of dabbling ducks. Pp. 68- 82 in Gowaty, P. A., Mock, D. W. (Eds). Avian monogamy. Ornithological Monographs 37. McKINNEY, F. 1 986. Ecological factors influencing the social systems of migratory dabbling ducks. Pp. 153-171 in Rubenstein, D. I., Wrangham, R. W. (Eds). Ecological aspects of social evolution. Princeton, Princeton University Press. McKINNEY, F., BREWER, G. 1989. Parental attendance and brood care in four Argentine dabbling ducks. Condor 91: 131-138. McKINNEY, F. (In press). Courtship, pair-formation and signal systems of waterfowl, in Batt, B. D. J. (Ed.). Ecology and management of breeding waterfowl. Minneapolis, University of Minnesota Press MOULTON, D. W., WELLER, M. W. 1984. Biology and conservation of the Laysan Duck (Anas laysanensis). Condor 86: 105-117. ORING, L.W., SAYLER, R.D. (In press). The mating systems of waterfowl, in Batt, B. D. J., (Ed.). Ecol¬ ogy and management of breeding waterfowl. Minneapolis, University of Minnesota Press. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 859 ROHWER, F. C., ANDERSON, M. G. 1988. Female-biased philopatry, monogamy, and the timing of pair formation in migratory waterfowl. Current Ornithology Vol. 5, Pp. 187-221. New York, Plenum Press. ROWLEY, I. 1983. Re-mating in birds. Pp. 331-360 in Bateson, P. (Ed.). Mate choice. Cambridge, Cambridge University Press. SIEGFRIED, W. R. 1974. Brood care, pair bonds, and plumage in southern African Anatini. Wildfowl 25: 33-40. SIEGFRIED, W. R., FROST, P. G. H., HEYL, C. W. 1976. Long-standing pairbonds in Cape Teal. Os¬ trich 47: 130-131. SORENSON, L. G. 1990. Breeding behaviour and ecology of a sedentary tropical duck: the White¬ cheeked Pintail (Anas bahamensis bahamensis). Unpub. Ph.D. thesis. University of Minnesota. SORENSON, L. G. (in review). The variable mating system of a sedentary tropical duck: the White-cheeked Pintail (Anas bahamensis bahamensis). STANDEN, P. J. 1976. The social behaviour of the Chilean Teal. Unpub. Ph.D. thesis. University of Leicester, England. STOLEN, P., MCKINNEY, F. 1983. Bigamous behaviour of captive Cape Teal. Wildfowl 34: 10-13. WELLER, M. W. 1968. Notes on some Argentine Anatids. Wilson Bulletin 80: 189-212. 860 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI THE BLUE DUCK MATING SYSTEM - ARE RIVER SPECIALISTS ANY DIFFERENT? CLARE J. VELTMAN1, SUSAN TRIGGS2, MURRAY WILLIAMS2, KEVIN J. COLLIER2, BRIAN K. MCNAB3, LISA NEWTON1, MARIE HASKELL1, and IAN M. HENDERSON1 ! Department of Botany and Zoology, Massey University, Palmerston North, New Zealand 2 Science and Research Directorate, Department of Conservation, Box 10 420, Wellington, New Zealand 3 Department of Zoology, University of Florida, Gainesville, FI 3261 1 , USA ABSTRACT. The reproductive behaviour of male and female Blue Ducks Hymenolaimus malacorhynchos leads to enduring pair bonds and permanent territories. To evaluate hypotheses about how stability of the river habitat shaped the evolution of this mating system, we measured reproduc¬ tive success and paternity in a Blue Duck population on the Manganuiateao River, New Zealand. Breeding pairs fledged 1.2 offspring per year, averaged over 23 pairs and 58 breeding attempts. We related reproductive success to pair tenure, and found that pairs of long duration fledged significantly more young than pairs of short duration. Some long-established males therefore exhibited high repro¬ ductive success. DNA fingerprinting confirmed the observation that extra-pair copulations did not oc¬ cur, so cuckoldry was probably nonexistent in this population. Energy requirements were estimated from the basal metabolic rate and represented a very small fraction of the energetic value of the avail¬ able prey, indicating that Blue Ducks did not defend food resources. Our evidence indicates that male Blue Duck, in common with other waterfowl, defend their mates. Keywords: Blue Duck, Hymenolaimus malacorhynchos, mating system, territoriality, food availability, energetics, evolution, waterfowl. INTRODUCTION Worldwide, six species of waterfowl live permanently on rivers. They are the Torrent Duck Merganetta armata, Salvadori’s Duck Anas waigiuensis, the African Black Duck Anas sparsa, Blue Duck Hymenolaimus malacorhynchos, the Bronze-winged Duck Anas specularis and the Brazilian Merganser Mergus octosetaceus (Johnsgard 1978, Madge & Burn 1988). All six riverine species are distributed in the Southern Hemi¬ sphere. In the Northern Hemisphere, Harlequins Histrionicus histrionicus, Goosanders Mergus merganser and the Chinese Merganser Mergus squamatus breed on rivers but migrate to other habitats in autumn. All of the river specialists studied so far have similar social behaviour. Territories are defended year-round by monogamous adult pairs, and broods are accompanied by the male parent as well as the female except in African Black Duck (Moffett 1970, Kear 1972, 1975, Ball et al. 1978, Williams 1991). Kear (1975) extrapolated from anatomical similarities to suggest that ecological and behavioural adaptations shared by the Southern Hemisphere river ducks originated with a Gondwanaland ancestor, but this phylogenetic hypothesis still lacks supporting biochemical evidence. Brush (1976) was unable to demonstrate shared recent ancestry using feather protein analy¬ sis. Other attempts to understand the evolution of perennial monogamy and territoriality in river ducks have focused on the value of resource defence or mate defence. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 861 McKinney et al. (1978) wrote that exclusive ownership of a stretch of river by African Black Duck pairs evidently served to ensure access to resources like food, nest sites or safe locations for maintenance activities. Such all-purpose territories of ducks in river and shoreline habitats were related to insectivory in particular (McKinney 1985) and the temporal and spatial distribution of food resources in general (McKinney 1986). Oring and Sayler (1989) also related long-term pair-bond behaviour to limited or defendable food resources. The value of mate defence by male waterfowl was stressed by Gauthier (1988), who argued that the territoriality exhibited by river specialists may represent one end of a continuum from home range behaviour by waterfowl in unstable habitats to enduring site attachment in stable habitats. He reasoned that the expected reproductive suc¬ cess of a drake is made up of two additive components: the productivity of his bonded mate and the offspring resulting from forced extra-pair copulations. In stable habitats where females attempt breeding most years, drakes have most to gain from intense mate guarding behaviour and territoriality. Conversely, males breeding in unstable habitats where nest failure rates may be high could reduce reproductive variance by extra-pair copulations. Such a causal relationship between habitat stability and behaviour during the breeding season assumes that drakes defend mates rather than food or other resources (Gauthier 1988). This seems a reasonable assumption because overt mate guarding behaviour is widespread in waterfowl (McKinney 1986). A second assumption of the habitat stability hypothesis is that forced copulation is an alternative mating tactic for paired drakes. This has general acceptance (McKinney et al. 1983, Gauthier 1988, Oring & Sayler 1 989). Whether perennial monogamy and territoriality evolved in riverine species because habitat stability promoted mate and site attachment or because habitat stability offered defendable food supplies for breeding pairs and/or their offspring remains to be re¬ solved. Here, we explore the likelihood of food defence using measurements of the use of space, prey availability and metabolic rate in New Zealand’s Blue Duck. To our knowledge these are the first quantitative data about food availability obtained from a river duck. We also describe defence behaviour and evaluate Gauthier’s two pre¬ dictions that forced copulations will be rare and the realised reproductive success of drakes will be relatively high with low variance. METHODS This appraisal is based on data from a series of completed and ongoing studies of a population of Blue Ducks on the Manganuiateao River in the central North Island of New Zealand. The description of the mating system and measurements of adult sur¬ vival and productivity were obtained by methods described in Williams (1991). The DNA fingerprinting procedure is reported in Triggs et al. (1991). We refer to earlier analyses of territorial behaviour (Eldridge 1986, Veltman & Williams 1990), diurnal activity (Veltman & Williams 1990) and diet (Kear & Burton 1971), and direct readers to those papers for methodological details. The rate of metabolism under standard conditions was determined (by BKMcN) from oxygen consumption in two captive-reared adult Blue Ducks maintained in sealed 862 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI chambers held at constant temperatures from 10° through 35°C. It was assumed that 20.08 kJ were liberated during consumption of 1 litre of oxygen. The dry weight of larval Chironomidae per square metre of rock surface was estimated (by MH) in May 1986 and temporal variations in density of Chironomidae were as¬ sessed (by KJC) at two-monthly intervals throughout 1989. The two riffles sampled in both studies were the most frequently used foraging area in the territory of the respec¬ tive resident pairs. Samples for the measurement of dry weight were collected from an area of 123 cm2 using a rigid plastic tube with a rubber washer on one end to re¬ duce slippage on the rock surface, and a net on one side. Vertical slits on the up¬ stream side allowed water to flow through and wash invertebrates into the net when rock surfaces were scrubbed. A total of 36 samples was collected in this manner in each of the riffles, from the top, front and sides of submerged rocks. The larvae were later measured, and dry weight was estimated using conversion equations in Smock (1980). Temporal variations in the number of Chironomidae were determined by removing from each riffle 10 small (0.41-1.33 m2) boulders on each date (5 in July), and scrub¬ bing the larvae from all surfaces into plastic buckets. Faecal droppings were collected concurrently with invertebrate samples from boulders in 1989 (by LN, IMH and CJV) and Chironomidae were counted in subsamples using mandibles, clypea, and whole heads. RESULTS Territorial behaviour and the food supply Adult pairs of Blue Duck were dispersed at approximately 1 km intervals along the middle section of the Manganuiateao River, and their home ranges were non-overlap¬ ping (Williams 1991). Most aggressive interactions between resident adults and their neighbours or newcomers in their area took the form of same-sex challenges. Intrud¬ ers were usually males, and the male resident responded very aggressively. The bony wing spurs were most pronounced in adult males. Female aggression was mainly di¬ rected to juvenile birds of both sexes (Eldridge 1986). While pairs occupied up to five pool-riffle systems, most of their daily activities were concentrated in only a small part of their range (Veltman & Williams 1990, Williams 1991). Little time in any day was spent in boundary areas. The Blue Ducks vocalised in social contexts and during human disturbance, but did not vocally advertise their residency as do territorial songbirds. Pairs foraged most frequently in riffles, within a few metres of the bank, and certain riffles were more heavily used (Veltman & Williams 1990). While the remainder of the territory represented potential foraging substrate, the ducks made little use of it. Ter¬ ritory limits did not change markedly from year to year but the number of resident territorial pairs doubled over a ten-year period (Williams 1991). Except for incubation by females, the diurnal time activity budget of territorial pairs prior to and during breeding was dominated by inactivity (Veltman & Williams 1990). ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 863 Within these home areas, Blue Ducks caught and consumed a diverse range of aquatic invertebrate larvae, similar to that recorded on other rivers by Kear & Burton (1971). Chironomidae were the most abundant prey in 68 faecal samples, followed by Ephemeroptera. The mean number of chironomid midge larvae per square metre and temporal changes in their relative abundance are shown on Table 1. Food requirements and prey availability in territories The mean basal metabolic rate measured from non-moulting Blue Ducks in the thermoneutral zone was 0.790 cm302g'1h'1. Over a 24 hour period, with the basal metabolic rate elevated three times for an active individual, the average energy con¬ sumption is 818.64 kJ for a 717g Blue Duck. The prey requirement can therefore be estimated, for invertebrates of known energy value. Fourth instar Chironomus zealandicus larvae generate 17593 + 190J per gram of dry weight (mean of three samples Ryan 1982). This species is found in New Zealand lakes rather than rivers but is similar to the Chironomidae consumed by Blue Ducks on the Manganuiateao River. We do not know the digestibility of these prey, but will assume 50% of their calorific value is assimilated by the ducks. To meet energetic outgoings of about 818 kJ per day eating only chironomid larvae, a Blue Duck must consume approximately 818000/(17593 x 0.5) = 93 g dry weight of prey. The mean estimated dry weight of chironomid larvae in May 1986 was 924 mg per 123 cm2 of exposed rock surface. Thus prey were distributed at 75 g m 2 and a Blue Duck needed to glean less than 2 m2 of riverbed per day to meet the total daily energy requirement from this one prey type alone. Chironomid larvae comprised 65-70% of invertebrates on rock surfaces in May 1989 (Table 1). TABLE 1 - Seasonal variation in mean numbers of larval chironomids per square me¬ tre, and in percent abundance of total invertebrates on rocks in two riffles of the Manganuiateao River, 1989. Month Territory One Territory Two Number Percent Number Percent January 375.4 10.6 1955.5 49.0 March 1269.5 22.5 1271.0 23.0 May 9264.7 66.5 10213.3 68.5 July 1486.9 33.5 6780.8 55.7 September 10935.8 87.0 8443.0 75.0 November 390.7 16.0 1677.6 26.8 The mating system The relationship of paired birds in the Manganuiateao River population was “long-term and constant” (Williams 1991). Females nested between August and December each year, while males stationed themselves nearby throughout the 35-day incubation period. Both adults accompanied the brood, which broke up and dispersed from the natal territory after 70 to 80 days. Williams (1991) recorded 16 cases of pair formation and an additional 6 instances of two birds acting as a pair to acquire a territory in an undefended area. Three pairs formed after solitary birds had established their occupancy and then attracted a part¬ ner and one pair formed when a resident male was challenged and ousted by an 864 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI unpaired male. In the remaining 12 cases, pair formation followed death of a partner. Widowed males were seen to displace their male neighbours or bond with formerly unpaired females. Male Blue Ducks have never been observed to attempt a forced copulation in our study population. The absence of cuckoldry amongst Blue Ducks was supported from DNA fingerprinting analysis which confirmed parentage of 14 offspring (from a total of 10 clutches) of four territorial pairs (Triggs et al. 1991). It seems that an exclusive partnership with a female is the principal route to successful reproduction for male Blue Duck in the Manganuiateao River population. Survival and productivity Generally, Blue Ducks are long-lived birds. Males may survive for 7 years or more once they acquire territorial status, but territorial females risk predation while nesting (Williams 1 991 ). Female survival was lower than male survival in five of the nine study years, and overall mean annual survival estimates were 0.91 for males and 0.83 for females. Females may thus have been a limited resource for males. Of all the nests initiated by Blue Ducks during the study, 54% were successful. On average, 1.2 (SD 1.5) offspring per pair were raised to fledging each year from a mean clutch of 6.0 eggs (Williams 1991). Ten percent of nest failures were followed by renesting (Williams 1991). When productivity is analysed in relation to pair bond duration, significantly more offspring were fledged in each breeding attempt by four pairs with a median pair bond duration of 5 years than by 16 pairs with a median du¬ ration of 1 year (Mann-Whitney, W=751.0, P<0.05). In fact, 69% of 64 fledged young raised in the study area were produced by the four long-established Blue Duck pairs (Figure 1). High reproductive success is therefore an attribute of some males in the Blue Duck population, even if productivity is low overall. DISCUSSION Blue Ducks in territories we sampled on the Manganuiateao River did not appear to be defending economic territories, in which defence effort and territory size vary with resource availability. Their foraging activity was limited to a small fraction of the day in a small part of the home range, and prey availability greatly exceeded their calo¬ rific requirements by our admittedly crude estimates. Chironomid larvae represented bread-and-butter prey, less preferred than caddis larvae (our unpub. data) and varying temporally in availability, but numerically abundant. A 1 km territory averaging 10 m in width offers up to 10,000 m2 of planar foraging substrate, so the ballpark figure of 1.24 m2 per adult per day represents a small requirement given the very rapid colo¬ nization rates for Chironomidae (Williams & Flynes 1976). Given also that the number of breeding Blue Duck pairs packed into 9.3km of river habitat doubled in ten years, there is little evidence from our results that these Blue Duck defended their food re¬ sources. Rather, we found behaviour consistent with the assumption that Blue Duck males defended their mates. Males engaged in fierce fighting with other males in the pres¬ ence of females, and such episodes were frequently followed by mate swaps. Males replaced dead partners by breaking the bond of a neighbouring pair, and left the core area of their home ranges if successfully challenged by a newcomer. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 865 Territoriality seemed to isolate breeding pairs, providing a buffer zone and reduced social disturbance of females. Female survival may be lower than male survival, mak¬ ing females a limiting resource for drakes. Pairings Ranked in order of Productivity FIGURE 1 - Offspring production from 20 Blue Duck pairings There was no observational or biochemical evidence of cuckoldry. Thus the realised reproductive success for males depended on breeding with a bonded partner. This is consistent with Gauthier’s (1988) hypothesis, but his prediction of relatively high re¬ productive success and low variance for territorial males is more problematic. We have found few comparative data on reproductive success in territorial waterfowl. In New Zealand the nesting success of Paradise Shelduck Tadorna variegata was 63- 68% and breeding pairs fledged an average of 2. 6-2. 8 ducklings per year (Williams 1979). Similar productivity was measured in Buffleheads Bucephala albeola by Gauthier (1989), who found 65% of nests were successful and pairs raised an aver¬ age of 2.23 offspring. In the latter study a coefficient of variation of about 100% was obtained for reproductive success (G. Gauthier, pers. comm.), compared with 125% for Blue Ducks. In contrast, waterfowl breeding in unstable habitats exhibit very low nesting success. Klett et al. (1988) reported average nest success rates of 6-17% for Mallard Anas platyrhynchos, Gadwall Anas strepera , Bluewinged Teal Anas discors, northern Shoveler Anas clypeata and northern Pintail Anas acuta , so we expect that offspring production per pair in these species will exhibit extremely high coefficients of variation. Greatest reproductive output in the Manganuiateao River population was achieved by a few males in long-term pairings, indicating that pair-bond duration or longevity may be important factors in this territorial system. 866 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI Predation by introduced mammals accounted for three of 1 1 known nest failures and probably also explained the disappearance of six incubating females (Williams 1991). Therefore we doubt that we can legitimately estimate the variance in male mating success under natural conditions. Nevertheless, the partners of paired males made a breeding attempt every year and occasionally re-nested after nest failure. Territo¬ rial males thus had a high expectation of reproduction in any year. Based on our results, territoriality in Blue Ducks probably evolved via mate defence as suggested for other waterfowl living in stable habitats (Gauthier 1988). McKinney et al. (1978) found that male African Black Ducks made “heavy and prolonged invest¬ ment of time and effort in a single female” and Eldridge (1986) described how terri¬ toriality was “inextricably tied to pair formation and pair-bond maintenance” in Blue Ducks. Defense of an area larger than required for reproduction and their everyday needs may even be a handicap for drakes, signalling their vigour to potential mates (A. Zahavi, pers. comm.). We do not rule out the possibility that Blue Duck populations experience severe prey shortages following extraordinary floods (P. Ryan, pers. comm.) but our study found no support for the resource defence hypothesis. ACKNOWLEDGEMENTS We thank Amotz Zahavi, Ed Minot and Charley O’Kelly for very helpful discussions of our results, and Paddy Ryan, Gilles Gauthier, Frank McKinney, Lisa Sorenson and Ed Minot for their comments on an earlier version of the manuscript. LITERATURE CITED BALL, I.J., FROST, P.G.H., SIEGFRIED, W.R., McKINNEY, F. 1978. Territories and local movements of African Black Ducks. Wildfowl 29: 61-79. BRUSH, A.H. 1976. Waterfowl feather proteins: analysis of use in taxonomic studies. Journal of the Zoological Society of London 179: 467-498. ELDRIDGE, J.L. 1986. Territoriality in a river specialist: the Blue Duck. Wildfowl 37: 123-135. GAUTHIER, G. 1988. Territorial behaviour, forced copulations and mixed reproductive strategy in ducks. Wildfowl 39: 102-114. GAUTHIER, G. 1989. The effect of experience and timing on reproductive performance in buffleheads. Auk 106: 568-576. KEAR, J. 1972. The Blue Duck of New Zealand. Living Bird 1 1 : 175-192. KEAR, J. 1975. Salvadori’s Duck of New Guinea. Wildfowl 26: 104-1 1 1 . KEAR, J., BURTON, P.J.K. 1971. The food and feeding apparatus of the Blue Duck Hymenolaimus. Ibis 113: 483-493. KLETT, A.T., SHAFFER, T.L., JOHNSON, D.H. 1988. Duck nest success in the prairie pothole region. Journal of Wildlife Management 52: 431-440. MADGE, S., BURN, H. 1988. Wildfowl. London: Christopher Helm. McKINNEY, F. 1985. Primary and secondary reproductive strategies of dabbling ducks, in Gowaty, P.A., Mock, D.W. (Eds). Avian monogamy. Washington DC: American Ornithologists’ Union. Pp 68-82'. McKINNEY, F. 1986. Ecological factors influencing the social systems of migratory dabbling ducks, in Rubenstein, D.I., Wrangham, R.W. (Eds). Ecological Aspects of Social Evolution, pp 153-171. Princeton: Princeton University Press. McKINNEY, F., DERRICKSON, S.C., MINEAU, P. 1983. Forced copulation in waterfowl. Behaviour 86: 250-294. McKINNEY, F., SIEGFRIED, W.R., BALL, I.J., FROST, P.G.H. 1978. Behavioural specializations for river life in the African Black Duck ( Anas sparsa Eyton). Zeitschrift fur Tierpsychologie 48: 349-400. MOFFETT, G.M. 1970. A study of nesting Torrent Ducks in the Andes. Living Bird 9: 5-27. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 867 ORING, L.W., SAYLER, R.D. 1 991 . The mating systems of waterfowl, in Batt, B.D.J. (Ed.). Ecology and management of breeding waterfowl. To be published. RYAN, P.A. 1982. Energy contents of some New Zealand freshwater animals. New Zealand Journal of Marine and Freshwater Research 16: 283-287. SMOCK, L.A. 1980. Relationships between body size and biomass of aquatic insects. Freshwater Bi¬ ology 10: 375-383. TRIGGS, S., WILLIAMS, M., MARSHALL, S., CHAMBERS, G. 1991. Genetic relationships within a population of Blue Duck ( Hymenolaimus malacorhynchos). Wildfowl 42. In press. VELTMAN, C.J., WILLIAMS, M. 1990. Diurnal use of time and space by breeding Blue Ducks. Wild¬ fowl 41:62-74. WILLIAMS, M. 1979. The social structure, breeding and population dynamics of Paradise Shelduck in the Gisborne-East Coast district. Notornis 26: 213-272. WILLIAMS, M. 1991 . Some social and demographic characteristics of blue duck. Wildfowl. In the press. WILLIAMS, D.D., HYNES, H.B.N. 1976. The recolonization mechanisms of stream benthos. Oikos 27: 265-272. 868 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI MALE PARENTAL CARE IN SOUTHERN HEMISPHERE DABBLING DUCKS frank McKinney Bell Museum of Natural History and Department of Ecology, Evolution and Behavior, University of Minnesota, 10 Church St. S.E., Minneapolis, Minnesota 55455, USA ABSTRACT. All northern hemisphere dabbling ducks (Anatini) have female-only parental care. Females brood their ducklings, protect them from predators, and lead them to feeding and resting sites. Males have been recorded with females and broods in many southern hemisphere species, but male parental care (indicated by consistent presence, vigilant guarding, warning, defense, and response to duckling alarm calls) occurs regularly in only nine species. Male presence with broods is variable in another seven southern hemisphere species and may reflect male interest in the female rather than the duck¬ lings. Ecological factors (e.g. hazardous brood habitats) and social factors (e.g. benefits of desertion versus long-term pair-bonding) appear to influence male brood-attendance patterns in this group. Field studies with a focus on intra-specific variation in male behaviour are needed. Keywords: Southern hemisphere, dabbling ducks, Anatini, Anas, parental care, mate-guarding, pre¬ dation, pair-bond, moult, breeding strategies. INTRODUCTION The general pattern of parental care roles in the family Anatidae has been known for many years (Delacour & Mayr 1945) and has been reviewed in detail by Kear (1970) and Afton & Paulus (in press). Of the eight major tribes, biparental care of the young is the rule in whistling ducks* (Dendrocygnini), swans and geese (Anserini), and shelducks and sheldgeese (Tadornini), while female-only care is almost universal in stiff-tails (Oxyurini), sea ducks (Mergini), and pochards (Aythyini). The remaining two tribes, the perching ducks (Cairinini) and dabbling ducks (Anatini), are of special in¬ terest because they include species showing both biparental and female-only care. The large tribe of dabbling ducks, including 36 species with a world-wide distribution, provides especially good opportunities to identify ecological and social factors that could influence these variable patterns of parental care. Available information for this group is reviewed here in an attempt to develop promising hypotheses and draw at¬ tention to research needs. Three types of parental attendance patterns can be distinguished in dabbling ducks (McKinney 1985, McKinney & Brewer 1989): (a) only the female is present, (b) both male and female are present, (c) male presence is variable. To investigate why males are always present in certain species but not in others, and why male presence is variable in certain species, detailed information is needed on what males are doing while they accompany females with ducklings, what other options males have, and whether males derive fitness benefits from attending broods. * Footnote: Common and scientific names of waterfowl tribes and species follow Johnsgard (1978). ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 869 A major clue to questions about male presence with broods comes from the geo¬ graphic distribution of these patterns. Of the 36 species of Anatini, 16 have male at¬ tendance (regular or variable), and all of these are southern hemisphere species (Ta¬ ble 1). Factors that could promote this phenomenon in the southern hemisphere are the focus of this review. TABLE 1 - Incidence of male presence with broods and male parental care in tropi¬ cal and southern hemisphere dabbling ducks (Anas species) *. Anas spp Male care usual Male presence variable Male usually Male absent behaviour unknown waigiuensis X spars a X Sibil atrix X flavirostris X capensis X bernieri X gibberifrons X castanea X aucklandica X melleri X undulata X poecilorhyncha X luzonica X specu laris X specularioides X georgica X bahamensis X erythrorhyncha X versicolor X hottentota X cyanoptera X platalea X smithii X rhynchotis X * Based on McKinney 1985, more recent information in references cited in text, and unpublished data. BEHAVIOUR OF MALES WITH BROODS Observations on the nine species of Anas in which males are normally present with females and broods (Table 1) have revealed several kinds of male brood-care behav¬ iour (Norman & McKinney 1987, Buitron & Nuechterlein 1989, McKinney & Brewer 1989, Brewer 1990, Marchant & Higgins 1990). Most conspicuously, males spend much time in alert, vigilant postures while the ducklings are feeding and moving about. They also escort isolated ducklings, respond to duckling alarm calls, and behave aggressively toward other water birds near the brood. In response to predators, males have been recorded giving alarm calls and distraction displays, and they may even attack the predators. In exceptional circumstances, if the female is absent, males of several species have been reported raising the ducklings alone. The usual pattern is 870 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI for male and female to collaborate closely in these species with biparental care. Fe¬ males play the major role in maintaining vocal communication with the ducklings, brooding them when they are small, and leading them to feeding and resting sites. In one study of Chestnut Teal A. castanea, broods attended by both parents were larger than those accompanied by a single parent (Norman & McKinney 1987). This sug¬ gests that duckling survival may be enhanced when two parents are present, but re¬ moval experiments are needed to demonstrate this convincingly. In contrast to these biparental species, the behaviour of males in seven Anas species with variable male attendance (Table I) is often more difficult to interpret. In these species, when the male is present with the brood he is usually more interested in the female than the ducklings, but there may be specific differences (see below). FACTORS FAVOURING DESERTION BY THE MALE Migratory northern hemisphere Anas In these species, pair bonds typically break during the incubation phase, when the male deserts his mate and leaves the breeding area (Bellrose 1976, Cramp & Simmons 1977, McKinney 1986). Many of these males are known to move consider¬ able distances to reach traditional moulting places (usually large marshes) where they spend the flightless period during wingmoult. Three potential advantages for males have been suggested: (a) the moulting sites enhance individual survival by providing rich food supplies and better escape cover than the breeding grounds; (b) desertion at an early stage in the breeding season allows males to advance through the post¬ breeding moult, regain nuptial plumage, and build up energy reserves for fall migra¬ tion, winter courtship, and competition for mates; (c) departure of the male reduces competition for food for the female and brood. Although direct tests of these three hypotheses are difficult, indirect evidence tends to provide support only for the first two. When females lose their initial clutches to predators, many make renesting attempts, and the same pair bond often persists through the laying of these clutches. In some species, “pairs” have been seen on the breeding grounds late in the breeding season, and it is speculated that these may be failed nesters. Generally this is an uncommon phenomenon. Southern hemisphere Anas A similar pattern of male desertion appears to be usual in at least some populations of certain southern hemisphere species (e.g. Red Shoveler A. platalea, Australasian Shoveler A. rhynchotis) in which males leave the breeding areas and move to tradi¬ tional moulting lakes (Fjeldsa & Krabbe 1986, M. Williams, pers. comm.). The timing of desertion may vary, however, and in Cape Shoveler A. smithii some males appar¬ ently remain with their mates until early in the brood-rearing phase (Siegfried 1974). In these species the advantages of desertion to males may be the same as in north¬ ern hemisphere migratory species. Information on the occurrence of moult migration is lacking for a number of southern hemisphere species. ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 871 FACTORS FAVOURING MALE PRESENCE WITH BROOD FEMALES A male’s presence with a brood could be primarily a consequence of pair-bond main¬ tenance, and this seems to be the case in some species with variable male attend¬ ance. By preserving his bond with the female, a male can ensure that he will be able to father any subsequent clutches she lays during the same breeding season. Al¬ though males of northern hemisphere Anas rarely stay with the female after the end of incubation, desertion can occur during incubation or brood-rearing in some tropi¬ cal and southern hemisphere species (Sorenson this symposium). Another possibility is that males remain with brood females to maintain bonds for longer than the current breeding season. Various potential advantages of long-term pair bonds have been identified for birds in general (Rowley 1983), and several au¬ thors have suggested that these could apply to southern hemisphere waterfowl. Main¬ tenance of pair bonds could save time when conditions favourable for breeding arrive and/or persist in unpredictable habitats (Siegfried 1974). Alternatively, after a suc¬ cessful breeding effort it could be advantageous for the same compatible individuals to remain paired for subsequent breeding attempts, and compatibility of mates should be especially important in species with biparental care. The presence of an escorting male can be beneficial to females also. In addition to benefits for both partners relating to future breeding attempts (as noted above), male vigilance and guarding can enable females to feed without disruption. This factor is especially relevant for females in the early stages of brood-rearing when they need to recover from the energetic costs of incubation. FACTORS FAVOURING BIPARENTAL CARE Little attention has been given to identifying factors that promote male care in certain Anas species but not in others. Males activities while escorting ducklings suggest three potential benefits for the young: (a) promoting brood cohesion, (b) providing protection against predators, (c) enhancing duckling feeding efficiency. I propose that characteristics of brood habitats and the predation pressure to which ducklings are exposed may impose a greater need for male assistance in certain species. There is suggestive evidence along these lines for six biparental species. River habitats Two Anas species, the African Black Duck A. sparsa and Salvadori’s Duck A. waigiuensis raise their broods on rivers. The latter species occupies “white water" stretches of mountain streams, similar to the habitats preferred by the Blue Duck Hymenolaimus malacorhynchos and Torrent Duck Merganetta armata , and biparental care is strongly developed in all three species. These river habitats are especially hazardous for ducklings because of the high risk of being swept away by the current and becoming separated from their parents. Apart from being exposed to greater danger from attack by predators, lack of parental vigilance may adversely affect the growth of isolated ducklings (Williams 1991), presumably by reducing feeding fre¬ quency and/or efficiency. Female-only brood care (at least during daylight hours) is characteristic of African Black Ducks in the regions where this species has been ob¬ served, but many of these rivers are relatively slow-flowing and perhaps they do not present such extreme hazards for ducklings. 872 ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI Open habitats Three biparental species use brood-rearing habitats that are often very open and devoid of escape cover. Cape Teal A. capensis use open saline lakes and lagoons in southern and eastern Africa. Crested Ducks A. specularioides breed on Patagonian and Andean lakes and seacoasts with rocky shorelines. Chiloe Wigeon A. sibilatrix broods use open wetland habitats where they feed on aquatic plants at the surface, often far from shore. Terrestrial habitats Waterfowl that feed on land by grazing are especially vulnerable to predators because often there is no escape cover and they cannot use diving as a method of escape. Brewer (1990) has shown that Chiloe Wigeon broods frequently leave the safety of wetlands to come out on land to graze on nearby grasslands where avian predators hunt, and that males can effectively contribute to brood defense by remaining vigilant and giving alarm calls. Similar hazards may be faced by some populations of island dwelling Austral teal (eg. A. aucklandica, Williams et al. this symposium) in which males show vigorous defensive behaviour against predators while escorting broods. The importance of biparental care in ducks that specialize in grazing has also been pointed out by Kingsford (1990) in regard to the Australian Wood Duck Chenonetta jubata, a member of the perching duck tribe (Cairinini). Occluded habitats In addition to the hazards associated with their terrestrial life, aucklandica broods characteristically occupy dense vegetation where vision is often impaired and the risk of ducklings becoming lost may be high. Preliminary observations on Silver Teal A. versicolor broods indicate that they also prefer to feed in dense emergent vegetation (McKinney & Brewer 1989). There remain three biparental species for which habitat features that might promote male brood-care have not been identified. These are the Chestnut Teal, Grey Teal A. gibberifrons, and Bronze-winged Duck A. specularis. VARIABLE MALE PRESENCE WITH BROODS The presence of males with females and broods is variable in seven species (Table 1). Field studies on Speckled Teal and Brown Pintail (unmarked birds) indicate that a number of factors probably contribute to these variations (McKinney & Brewer 1989). Male Speckled Teal appear to vary individually in their attendance patterns; some broods consistently have a male present while others have no male present. Similar consistent patterns of male attendance were observed in many Brown Pintail broods, but part-time attendance was shown by some males. Observations on a marked population of White-cheeked Pintails (Sorenson ms in review) revealed great variation in male attendance; some males were consistently present, others were sometimes present, others were absent. In all three species, males show little or no interest in ducklings and, apart from consistent following by some individuals, males show no obvious brood-care behaviour. Attending males do show interest in the brood female, however, and therefore in these species male presence with broods appears to be primarily a consequence of persistence of the pair bond in anticipation of sub¬ sequent breeding attempts. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 873 In these species it is often difficult to decide whether an attending male is the female’s mate. Sorenson (ms in review) has shown that brood-tending female White-cheeked Pintails are preferentially courted by males (including males that are paired to other brood-tending females) and this can lead to the formation of new pair bonds. Court¬ ship of brood females has been noted also in Red Shoveler (McKinney & Brewer 1989). COMPETING MALE REPRODUCTIVE OPTIONS Male dabbling ducks have several reproductive options in addition to seasonal mo¬ nogamy (McKinney 1985) and the possibility of conflicts with male brood-care should be considered. The combination of seasonal monogamy and forced extra-pair copu¬ lation (FEPC) in a mixed reproductive strategy appears to be a common pattern in migratory species of Anas with short, regular breeding seasons in high latitudes. In these species the main period for FEPC activity (i.e. while females are fertile) is usu¬ ally over before broods hatch, and therefore FEPC is not likely to inhibit male brood attendance. Male breeding options are more varied in tropical and southern hemisphere species with extended and/or irregular breeding seasons. In addition to monogamy+FEPC, males of several species also engage in extra-pair courtship, mate-switching, and polygyny (McKinney 1985, Sorenson 1990). These activities entail investments of time and effort in assessment and securing of alternative or additional mates, and they are likely to conflict with male brood attendance. Little is known about the ways in which males resolve such conflicts, but information from White-cheeked Pintails in the Ba¬ hamas is enlightening (Sorenson, ms in review). The occurrence of complex mixed male strategies (including mate-guarding, territory-defense, FEPC, and polygyny), together with a high incidence of non-breeding, evidently produced great individual variation in male behaviour in this population. One consequence was great variation in pair-bond duration; some birds remained paired together over several breeding seasons, while others switched mates. This study provided firm evidence that males may attend broods in order to preserve their pair bonds, and that such behaviour need not entail male parental care. Other males may desert their mates and broods, or associate with them only part of the time while they also spend time courting other breeding females. Similar complex mixed male strategies and variable male presence with broods (without obvious parental care) are present in Speckled Teal (McKinney 1985, McKinney & Brewer 1989, unpublished data). The extent to which males which do contribute parental care are able to engage also in polygyny is not yet clear. The Cape Teal may be a species in which males can combine these activities, because courtship occurs year-round and polygyny has been recorded in captives (Stolen & McKinney 1983). On the other hand, tendencies to maintain long-term pair bonds appear to be very strong in other biparental species (notably Chiloe Wigeon and Silver Teal), and perhaps polygyny is not an option in these species. 874 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI CONCLUSIONS Male brood-care patterns in dabbling ducks fall into three categories: (a) male deserts and provides no care, (b) male actively contributes to protection of ducklings, (c) male sometimes accompanies female and brood but gives little or no care. Benefits to males of deserting and leaving their mates to care for the ducklings alone appear to be especially important in migratory Anas species with short, regular, annual breed¬ ing seasons in high latitudes. This is characteristic of most northern hemisphere spe¬ cies. By deserting, males can move to safe sites for the post-breeding wing-moult and they can begin the wing-moult earlier in preparation for the courtship season. Biparental care occurs in nine southern hemisphere species. Male participation may be favoured in six of these species because they breed in habitats that appear to be hazardous for ducklings. Biparental care and long-term pair-bonding are likely to be closely linked and mutually reinforcing characteristics. Variable male presence with broods is characteristic of seven tropical/southern hemi¬ sphere species and appears to reflect male interest in the female rather than the duck¬ lings. Opportunities for mate-switching and polygyny are probably greater in species with extended and/or irregular breeding seasons, and participation in these activities helps to account for variable male presence with broods. The extent to which male brood-care conflicts with polygyny is not clear. Regular, annual breeding seasons Extended and/or irregular breeding seasons Opportunities for polygyny and mate-switching / Long-term pairbonds Male deserts before ducklings hatch Variable male brood attendance Biparental care Advantages of Habitats hazardous molt migration for ducklings FIGURE 1 - Factors believed to influence the occurrence of male parental care in dabbling ducks. Highest priority for future research is for long-term studies of marked birds. The pres¬ ence of males with broods and the roles they may be playing in brood-care is an in¬ triguing topic for research, calling for detailed examination of the overall social sys¬ tem of each species. Especially valuable insights on male options and priorities should come from species showing intrapopulation and intra-specific (eg. racial) variations in male behaviour. More information is needed on the relative vulnerability of duck¬ lings to predators in different habitats. Field experiments (e.g. removals) are needed to test hypotheses on male parental care. ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI 875 ACKNOWLEDGEMENTS The studies on which this review is based were supported by grants from the National Science Foundation (grants no. BNS-8317187 and BNS-8820065), the Bush Founda¬ tion sabbatical program, the Graduate School, University of Minnesota, and the Na¬ tional Geographic Society. I thank Gwen Brewer, Sue Briggs, Lisa G. Sorenson, Michael D. Sorenson, W. Roy Siegfried, and Murray Williams for helpful discussion and comments. LITERATURE CITED AFTON, A.D., PAULUS, S.L. (In press). Incubation and brood care in waterfowl. In Batt, B.D.J. (Ed.) Ecology and management of breeding waterfowl. Minneapolis, University of Minnesota Press. BELLROSE, F.C. 1976. Ducks, geese and swans of North America. Harrisburg, PA, Stackpole. BREWER, G.L. 1990. Parental care behavior of the Chiloe Wigeon (Anas sibilatrix). Unpub. PhD thesis. University of Minnesota. BUITRON, D., NUECHTERLEIN, G.L. 1989. Male parental care of Patagonian Crested Ducks Anas ( Lophonetta ) specularioides. Wildfowl 40:14-21. CRAMP, S., SIMMONS, K.E.L. (Eds) 1977. Handbook of the birds of Europe the Middle East and North Africa. The birds of the Western Palearctic. Vol. 1. Oxford, Oxford University Press. DELACOUR, J., MAYR, E. 1945. The family Anatidae. Wilson Bulletin 57:3-55. FJELDSA, J., KRABBE, N. 1986. Some range extensions and other unusual records of Andean birds. Bulletin British Ornithologists’ Club 106:115-124. JOHNSGARD, P.A. 1978. Ducks, geese, and swans of the world. Lincoln, University of Nebraska Press. KEAR, J. 1970. The adaptive radiation of parental care in waterfowl. Pp. 357-392 in Crook, J. H. (Ed.). Social behaviour in birds and mammals. London, Academic Press. KINGSFORD, R.T. (1990). Biparental care in the Australian Wood Duck Chenonetta jubata. Wildfowl 41: 83-91. MARCHANT, S., HIGGINS, P. (Eds). 1990. Handbook of Australian, New Zealand and Antarctic birds. Oxford, Oxford University Press. (Text on Anas gracilis Grey Teal). McKINNEY, F. 1985. Primary and secondary male reproductive strategies of dabbling ducks. Pp. 68- 82 in Gowaty, P. A., Mock, D. W. (Eds). Avian monogamy. Ornithological Monographs 37. McKINNEY, F. 1986. Ecological factors influencing the social systems of migratory dabbling ducks. Pp. 153-171 in Rubenstein, D.I., Wrangham, R. W. (Eds). Ecological aspects of social evolution. Princeton, Princeton University Press. McKINNEY, F., BREWER, G. 1989. Parental attendance and brood care in four Argentine dabbling ducks. Condor 91 :1 31 -1 38. NORMAN, F.I., McKINNEY, F. 1987. Clutches, broods, and brood care behaviour in Chestnut Teal. Wildfowl 38:1 17-126. ROWLEY, I. 1983. Re-mating in birds. Pp. 331-360 in Bateson, P. (Ed.). Mate choice. Cambridge. Cambridge University Press. SIEGFRIED, R.W. 1974. Brood care, pair bonds and plumage in southern African Anatini. Wildfowl 25:33-40. SORENSON, L.G. 1990. Breeding behaviour and ecology of a sedentary tropical duck: the White¬ cheeked Pintail (Anas bahamensis bahamensis). Unpub. PhD thesis. University of Minnesota. SORENSON, L. G. (Ms in review). The variable mating system of a sedentary tropical duck: the White¬ cheeked Pintail (Anas bahamensis bahamensis). STOLEN, P., McKINNEY, F. 1983. Bigamous behaviour of captive Cape Teal. Wildfowl 34:10-13. WILLIAMS, M. (1991). Social and demographic characteristics of Blue Duck. Wildfowl 42: in press. 876 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI ECOLOGICAL AND BEHAVIOURAL RESPONSES OF AUSTRAL TEAL TO ISLAND LIFE MURRAY WILLIAMS1, FRANK McKINNEY2 and F. I. NORMAN3 1 Department of Conservation, P.O. Box 10-420, Wellington, New Zealand 2 Bell Museum of Natural History and Department of Ecology, Evolution and Behavior, University of Minnesota, 10 Church St S.E., Minneapolis, Minnesota 55455, USA 3 Arthur Rylah Institute for Environmental Research, Department of Conservation and Environment, P.O. Box 127, Heidelberg, Victoria 3084, Australia ABSTRACT. Responses of Austral teal to island life were investigated by comparing the morphology, habitats and foods, breeding characteristics, social system, and courtship behaviour of Chestnut Teal Anas castanea, New Zealand Brown Teal A. aucklandica chlorotis, Auckland Islands Teal A. a. aucklandica and Campbell Island Teal A. a. nesiotis. By treating them as a sequence from continen¬ tal ( castanea ) to large island ( chlorotis ) to small island ( aucklandica , nesiotis) forms there was: (i) a reduction in body size and a disproportionate decrease in wing length leading to flightlessness; (ii) a shift from a generally omnivorous to a predominantly carnivorous diet with a concomitant increase in metabolic rate; (iii) increased exploitation of terrestrial habitats and a more cursorial habit; (iv) a re¬ duction in clutch size accompanied by increased egg size, longer laying interval, longer incubation pe¬ riod, a larger hatchling and, in aucklandica , faster duckling growth; (v) a change in social organisation from the seasonally social and dispersive to year-round territoriality; (vi) an increase in aggression; and (vii) loss of conspicuous attention-seeking displays (Head-up-tail-up, Down-up) and those occurring in long display sequences and greater use of more subtle displays derived from comfort movements. Keywords: Austral teal, Chestnut Teal, Brown Teal, Auckland Island Teal, Campbell Island Teal, Anas castanea, Anas aucklandica, morphology, ecology, behaviour, insularity. INTRODUCTION David Lack (1970) and Milton Weller (1980) both appraised the ecological character¬ istics of waterfowl (Anatidae) inhabiting small oceanic islands. They noted the ten¬ dency for small islands to have but one species present, for that species to be derived from a nearby extant continental form (usually Anas) with a tolerance of estuarine or marine habitats, for the island form to be smaller and with different body proportions to those of its presumed ancestor, and for it to occupy a wider ecological niche. Both authors noted the tendency for island forms to have shifted from an r- to a K-selected reproductive strategy (by reduction of clutch size and increase of egg size) and from dichromatic to monochromatic plumages. In this paper we give substance to some of these generalisations as they apply to some of Delacour’s (1956) “Austral teal” and we extend the analysis to consider some behavioural changes that appear to be consequences of the island lifestyle. The Austral teal we consider are the Australasian inhabitants comprising the pan-con¬ tinental and nomadic Grey Teal Anas gibberifrons* of Australia, New Zealand and * Nomenclature follows Kinsky (1970). ACTA XX CONGRESSUS INTERNATIONALIS ORNITHOLOGICI 877 New Guinea (but ranging as far south as Macquarie Island and north to the Solomons), the Chestnut Teal A. castanea of southwest and southeast Australia and Tasmania, the Brown Teal A. aucklandica chlorotis, once widespread on New Zea¬ land’s three main islands and Chatham Islands, and two flightless sub-Antarctic forms restricted to the Auckland Islands A. a. aucklandica and Campbell Island A. aucklandica nesiotis. However, in this presentation most of our comparisons involve only the four brown-plumaged forms, castanea, chlorotis, aucklandica and nesiotis. Austral teal appear to be very closely related (Livezey 1990). Charles Daugherty (in prep., pers. comm.), using gel-electrophoresis of blood proteins, determined that castanea and chlorotis were separated at a genetic distance most commonly asso¬ ciated with infra-specific variation in birds (Nei 1978, Barrowclough & Corbin 1978) but chlorotis and the sub-Antarctic teal were further apart; chlorotis and aucklandica could be distinguished at one of the 14 loci examined and, although chlorotis and nesiotis also differed at one locus, this was different from that which separated chlorotis and aucklandica. Given that all 54 chlorotis examined were homogeneous at all loci, and castanea were polymorphic at only one, these results suggest that the sub-Antarctic teal are the result of two quite independent colonisations at different times and, perhaps, from different sources and that flightlessness has evolved inde¬ pendently in both sub-Antarctic forms. This may imply that the sub-Antarctic teal are descended from continental, not New Zealand stock and that chlorotis is the most recent colonist of the New Zealand region. RESULTS Morphometries Livezey (1990) has compared the morphology of the Austral teal from museum study skins and skeletal material. Our field measurements of live birds confirm his findings and show (Figure 1): (i) castanea and chlorotis are of similar size; (ii) the sub-ant¬ arctic teals are substantially smaller than the others and with nesiotis being approxi¬ mately 20% smaller than aucklandica ; (iii) there is greater sexual size dimorphism in the sub-Antarctic teals (26 - 30% cf. 10 -14% for other two); and (iv) the three island forms have reduced wings and the sub-Antarctic teals are flightless. In all four forms there is a linear relationship between bill length and tarsus length and body mass, and the relationship holds also for gibberifrons. But a similar relationship does not hold for wing length (Figure 1); the wing length of chlorotis falls midway between the flighted and non-flighted forms. Wing bones of the now extinct Chatham Island population of chlorotis were shorter than those of New Zealand birds (P. Millener pers. comm.) and although this form was probably not functionally flightless, wing reduction was apparent. Ecological characteristics The principal ecological trends in the move to insular life involve a change in diet consequent upon a change in habitat, and a change in reproductive rate. Habitat and foods. Although all four teal occupy freshwater habitats (Table 1), castanea and chlorotis show a tolerance of brackish and estuarine waters and the animal foods therein (Frith 1977, Norman & Mumford 1982, Norman & Brown 1988, 878 ACTA XX CONGRESSUS INTERNATIONALE ORNITHOLOGICI Weller 1974, Hayes & Williams 1982) and the sub-Antarctic forms have extended their exploitation to the marine foreshore (Weller 1975a, 1980, Williams 1986). In the ab¬ sence of ground predators (and probably competitors - the diminutive but now extinct Euryanas finsci was abundant in New Zealand’s post-Pleistocene wetlands and for¬ ests becoming almost flightless during this period (Worthy 1988)), chlorotis also ex¬ ploited occluded swamplands and swamp forests (Buller 1 888,Guthrie-Smith 1927). It became more dependent on invertebrates and detrital foods and adopted a more cursorial or terrestrial habit in the process, an adaptation also very evident in the two sub-Antarctic forms and for which Livezey (1990) has identified skeletal changes. WING LENGTH 150 I I — ee 'C « o c