POPULATION DYNAMICS OFPopulation Ecology, Life-History, Population Genetics, Behavioural Ecology, Avian Disease Ecology and Conservation. More specifically, it employed techniques

POPULATION DYNAMICS OF

GREAT WHITE PELICANS

IN SOUTHERN AFRICA

�

�

CAUSATIVE FACTORS AND INFLUENCE ON OTHER SEABIRDS

�

�

�

�

�

��

MARTA DE PONTE MACHADO

Lda. Biología (BSc Biology) University of La Laguna (Spain)

MSc Conservation Biology University of Cape Town

Thesis presented for the Degree of

DOCTOR OF PHILOSOPHY

in the DEPARTMENT OF ZOOLOGY

UNIVERSITY OF CAPE TOWN

�

�

�

��

I hereby declare that all of the work presented in this thesis, titled “Population

dynamics of Great White Pelicans in southern Africa– causative factors and influence

on other seabirds”, is my own, except where otherwise stated in the text. This thesis

has not been submitted in whole or in part for a degree at any other university.

Handed in for examination in Cape Town in September 2009.

Marta de Ponte Machado Date of graduation: JUNE 2010

�

�

�

�

�

�

�

To the Wild Nature

iv

��

Table of contents v

Abstract x

Acknowledgements xi

Chapter 1 Introduction 1

Section 1 Population dynamics and ecology

Chapter 2 The role of agricultural waste in the population expansion of Great White Pelicans in the Western Cape, South Africa 29

Chapter 3 Life-history and population ecology of Great White Pelicans in southern Africa 67

Section 2 Predation and impact on prey species

Chapter 4 Pelicans threaten local seabirds: adapting cooperative hunting strategies from water to land 103

Chapter 5 Can seabirds survive pelican predation? Results of a management intervention on the islands of the West Coast National Park, South Africa 123

Section 3 Population genetics

Chapter 6 Incorporating analysis of Great White Pelican population structure and gene flow into managing ecosystem cascades 155

Chapter 7 Development and characterisation of microsatellite loci from the Great White Pelican and widespread application to other members of the Pelecanidae 197

Section 4 Conclusions and recommendations

Chapter 8 Conclusions, future research and management 203 considerations

Appendix A.1 Prevalence of pathogens in Great White Pelicans Pelecanus onocrotalus from the Western Cape, South Africa 217 Appendix A.2 Photographs 225

v

��

Chapter 1: Introduction

Pelicans of the world....................................................................................................... 1

The Great White Pelican................................................................................................. 5

The Western Cape population.................................................................................. 7

Previous studies on Great White Pelicans in Africa.............................................. 9

Conceptual background and objectives....................................................................... 10

Rationale........................................................................................................................... 13

Outline of this thesis....................................................................................................... 14

Special considerations.................................................................................................... 17

References......................................................................................................................... 19

Chapter 2: The role of agricultural waste in the population expansion

of Great White Pelicans in the Western Cape, South Africa

Abstract............................................................................................................................. 29

Introduction..................................................................................................................... 30

Methods............................................................................................................................ 33

Breeding censuses and chick production................................................................ 33

Population surveys.................................................................................................... 34

Counts at the pig farm and evidence for dietary subsidisation.......................... 36

Pelican movements and ringing data..................................................................... 37

Results.............................................................................................................................. 39

Breeding population trends...................................................................................... 39

Population surveys and the role of the pig farm................................................... 40

Movements and dispersal......................................................................................... 42

vi

Discussion........................................................................................................................ 43

The role of immigration in the population expansion.......................................... 43

Intrinsic population growth and the role of the pig farm.................................... 44

Population censuses................................................................................................... 45

Are pelicans in the Western Cape food-limited? .................................................. 46

Some ecological consequences of a subsidised population................................. 48

References........................................................................................................................ 51

Tables and figures........................................................................................................... 57

Appendix 2.1: Breeding estimates and growth rates................................................. 65

Appendix 2.2: Pelican counts at regular sites in the Western Cape......................... 66

Chapter 3: Life-history and population ecology of Great White

Pelicans in southern Africa

Abstract............................................................................................................................. 67

Introduction..................................................................................................................... 68

Methods............................................................................................................................ 70

Breeding phenology and colony location on Dassen Island................................ 70

Ringing scheme........................................................................................................... 71

Survival estimates....................................................................................................... 72

Satellite transmitters................................................................................................... 74

Results............................................................................................................................... 75

Breeding phenology and colony location................................................................ 75

Age of first breeding................................................................................................... 76

Longevity..................................................................................................................... 76

Survival estimates....................................................................................................... 76

Spatial distribution of resightings and mortality................................................... 77

Behavioural observations at the breeding colony.................................................. 79

Discussion......................................................................................................................... 80

References......................................................................................................................... 89

Tables and figures............................................................................................................ 93

vii

Chapter 4: Pelicans threaten local seabirds: adapting cooperative

hunting strategies from water to land

Introduction, main findings and discussion................................................................ 103

Methods summary........................................................................................................... 108

Predation observations............................................................................................. 108

Number of pelicans involved in predation........................................................... 108

Prey species breeding success................................................................................ 109

Pelican diet analysis................................................................................................. 110

Proportion ringed pelicans...................................................................................... 110

References......................................................................................................................... 111

Box 1: Discontinuity of offal increases pelican predation on seabirds.................... 115


Chapter 5: Can seabirds survive pelican predation? Results of a

management intervention on the islands of the West Coast National

Park, South Africa

Abstract............................................................................................................................. 123

Introduction...................................................................................................................... 124

Methods............................................................................................................................ 126

Study sites.................................................................................................................... 126

Management intervention......................................................................................... 128

Training activities....................................................................................................... 129

Pelican monitoring..................................................................................................... 130

Prey species monitoring: pelican-exclusion areas versus control areas............. 130

Results............................................................................................................................... 132

Pelican monitoring...................................................................................................... 132

Prey species monitoring............................................................................................ 132

Discussion......................................................................................................................... 135

References......................................................................................................................... 141


Appendix 5.1: List of management options to reduce pelican predation............... 154

viii

Chapter 6: Incorporating analysis of Great White Pelican population

structure and gene flow into managing ecosystem cascades

Abstract............................................................................................................................. 155

Introduction...................................................................................................................... 156

Methods............................................................................................................................. 158

Sampling and DNA extractions................................................................................ 158

Genotyping and sequencing..................................................................................... 159

Statistical analysis....................................................................................................... 160

Results............................................................................................................................... 163

Mitochondrial DNA................................................................................................... 163

Hardy-Weinberg equilibrium (HWE) and exploratory analysis......................... 164

Population structure................................................................................................... 165

Assignment tests......................................................................................................... 166

Population bottleneck and demographic changes................................................. 167

Discussion......................................................................................................................... 168

Population structure.................................................................................................. 168

Demographic history................................................................................................. 171

Ecological and conservation implications............................................................... 175

References......................................................................................................................... 177


Appendix 6.1: Raw microsatellite scores..................................................................... 189

Appendix 6.2: Allele frequency distributions per loci and sampling locality........ 193

Appendix 6.3: Allele frequency distributions (graphical)......................................... 195

ix

Chapter 7: Development and characterisation of microsatellite loci

from Great White Pelican and widespread application

to other members of the Pelecanidae

Abstract...................................................................................................................... 197

Introduction..................................................................................................................... 197

Methods........................................................................................................................... 198

Results and discussion................................................................................................... 199

References........................................................................................................................ 200


Chapter 8: Conclusions, future research and management

considerations

Main conclusions............................................................................................................. 203

Novel contributions of this study................................................................................. 207

Future research and management considerations..................................................... 210

References........................................................................................................................ 214

Appendix A.1: Prevalence of pathogens in Great White Pelicans

from the Western Cape

Abstract.............................................................................................................................. 217

Introduction...................................................................................................................... 217

Materials and methods................................................................................................... 218

Results and discussion................................................................................................... 219

References......................................................................................................................... 221


Appendix A.2: Photographs

x

��

The exponential growth of the Western Cape Great White Pelican Pelecanus onocrotalus

population during the 20th century was partly caused by the increased availability of

agricultural offal to opportunistic and scavenging bird species. This superabundance had

ecological and behavioural implications, triggering a new feeding behaviour by pelicans on

the islands off the coast of the Western Cape. Pelicans adapted cooperative hunting

techniques used to capture aquatic prey to the land, shifting from a mostly piscivorous diet

to eat large numbers of seabird chicks. The spread of these novel foraging techniques in the

Western Cape suggested cultural transmission of behaviour and caused concern for the

conservation of declining populations of local breeding seabirds. The Western Cape pelican

population was found to be genetically less variable than other southern African breeding

colonies, possibly due to the demographic bottleneck experienced during the early part of

the 20th century and to the low frequency of immigration into this population. However,

pelicans from the Western Cape dispersed and some individuals entered into contact with

pelicans further north, indicating that cooperative seabird-eating behaviour could be

exported to other populations. Also, concern was raised with regards to the close contact

between domestic and wild animals feeding on agricultural offal and the potential risk for

the spread of disease. Although the incidence of disease in the Great White Pelican

population did not indicate an immediate risk, the prevalence of pathogens on this

population indicated that they could constitute a reservoir of zoonoses of ecological and

economic relevance. In addition, with the objective of curbing the impact of pelican

predation on seabird populations, a management intervention was implemented on two

islands of the West Coast National Park and proved successful to reduce predation. This

study amalgamates a diverse set of concepts and methodologies from the fields of

Population Ecology, Life-History, Population Genetics, Behavioural Ecology, Avian Disease

Ecology and Conservation. More specifically, it employed techniques such as mark-

recapture, direct observation of behaviour, molecular analysis of nuclear and mitochondrial

DNA, demographic modelling, microbiological analyses and adaptive management, in

order to investigate the complex ecological interactions among Great White Pelicans, human

landscapes and locally breeding seabird species.

xi

��

chieving a doctoral degree is much more than collecting data and answering more

or less complicated scientific questions. To me, this has been an opportunity to

learn and to grow as a person. It is my wish that this work also contributes to

further our understanding of the wondrous natural world we have the fortune to live in,

and in this way to preserve its beauty and the important processes that support the

diversity of life from the changes that we humans are causing on the planet.

It has been only with the help of many of you that I have been able to accomplish the

work included in this thesis. Thanks to all for providing the most valuable help, advice,

encouragement and support. I could have not done it alone!

To Piérre Nel, who is responsible for the successful achievement of this PhD. We know it

would have not been possible without you. For your constancy, love and support in all

levels: scientific, logistical and emotional. Thanks for being there for me, from waking

up at impossible times to do fieldwork to talking me through the most acute PhD crises,

but specially for allowing me the space to be myself.

To my supervisors, and especially to Les Underhill, who first spoke to me about pelicans

and who has guided me throughout numerous challenges. He always believed on me,

and that is the most valuable gift he has bestowed me with. During this time Les has

been infinitely patient and flexible, listening to my adventures and misfortunes on

islands (and piggeries) on the pursuit of pelicans and data. Thanks to him and to all the

good work he and his team do at the ADU.

Rauri Bowie offered me a life-enriching opportunity by inviting me to join him at the

Museum of Vertebrate Zoology at the University of California at Berkeley, where I had

the fortune to meet people that encouraged me to pursue my scientific career and where

I felt at home thanks to their exemplary generosity. Thanks Rauri for your excellent

guidance and shared enthusiasm about our work. I am also indebted to both of you,

Rauri and Verna, for offering me to stay in your home and for those shared home made

meals and glasses of wine. Peter Ryan contributed with his insight and suggestions to

shape this work and read through meandering versions of the first manuscripts,

managing to improve them in ways I didn’t think possible. Rob Crawford kindly shared

his knowledge and understanding of the Benguela ecosystem and helped me draft the

questions included on these chapters.

A mis padres, Gaspar y Marta, who have always encouraged my academic and

professional wanderings. I owe them my fascination for the natural world and a burning

curiosity to explore the planet. Se que conmigo han aprendido a “esperar lo inesperable”,

por eso y por su apoyo incondicional, ¡Gracias!

A mi abuela, Tita, one of the strongest and most fascinating women I have ever met.

Gracias por esperarme, y por ser capaz de leerme por dentro. Espero haber heredado tu

fortaleza y tus ganas de vivir.

A

xii

Al resto de mi familia por recordarme que soy querida y añorada. A mis amigas y amigos

por la hermosa responsabilidad de contribuir en hacerme ser quien soy hoy, y formar

parte de mi vida, para siempre. A Francis, por ser la fuerza externa que más a

contribuido a mi comunión personal con el mundo natural; por ser un eterno artista, que

admiro desde lo más profundo, y al que deseo que encuentre esa paz interna tan a ratos

elusiva. Thanks for your help doing fieldwork and the beautiful drawings that open

Chapters 1 and 8. A Elena por diseñar esta maravillosa portada y tu infinita amistad.

To Pelucho (Manuel Nogales) and Aurelio Martín for offering me my first glimpse of the

life on these magic places that are islands inhabited by seabirds. Fue en el islote de

Alegranza donde me enganché y desde entonces no he dejado de sentir un inmenso

placer con ese aroma tan característico de las aves marinas.

To Dave Gwynne-Evans, Maya Pfaff, Kathleen Reaugh-Flower and Piérre Nel. Thanks

for being my family away from home. For being an uncontainable source of inspiration

and for showing me some of the beauty and complexity of this country and this

continent. For being true to yourselves and some of the most fascinating people in the

planet. Maya, having the fortune of getting to know you has been one of the most

exciting and beautiful journeys of my life. Thanks for sharing your beautiful spirit and

magic with me.

To my Californian family: Anna Sellas, Sharon DeCray, Steven Long, Ângela Ribeiro,

Verna and Rauri Bowie, Sandra Nieto and David Vieites. Thanks for making space for

me in your homes and your lives; I treasure each one of your friendships and I know

that many of them will become life-long. Anna Sellas and Sampath Loukugalappatti

showed me almost anything one need to know to work in the lab, and many others of the

vibrant community at the MVZ improved my knowledge and skills both in the lab and

during the data analysis.

Besides the advice and ideas provided by my supervisors, Tony Williams and John

Cooper contributed to shape this work from the early days of research design and

exploration, providing literature suggestions and exciting discussions about what would

be the most interesting questions to ask and ideas on how to answer them. Other people

that contributed greatly with insight and observations were Phil Hockey, Res Altwegg,

Jan Hofmeyr, Piérre Nel, Anton Woldfaardt, Rob Simmons and (although we finally met

in 2009, it is never too late!) Alfredo Guillet.

Lara Atkinson and Maya Pfaff were part of the most inspiring PhD working group in

our joint race to the final deadline. I’m so proud of the work of the three of us! Thanks

for helping me to stick to the deadlines, structuring ideas, data analysis and your sharp

comments on the early manuscripts.

Jan Hofmeyr, Ângela Ribeiro, Piérre Nel, David Gwynne-Evans, Rauri Bowie, Peter

Ryan and Les Underhill also commented on early versions of the manuscripts. René

Navarro helped with Figure 2.6 and the ArcView analysis for Figures 3.4 and 3.5. Cathy

Boucher digitized the maps for Figure 3.2. Res Altwegg provided essential insights and

experience for building the survival models in Mark 5.1 and interpreting the results.

Thanks to José Bismark Poveda, Patricia Assunção and all others at the research group

xiii

at the Epidemiology Unit of the Faculty of Veterinary of the University of Las Palmas de

Gran Canaria (Spain) for the microbiological analysis and results.

To my colleagues at the Animal Demography Unit and the FitzPatrick Institute: Ralf

Mullers, Samantha Petersen, Lorien Pichegru, Marius Wheeler, Doug Harebottle, René

Navarro, Res Altwegg, Ross Wanless, Andrea Ángel, Anton Woldfaardt, Odette Curtis,

Anuradha Rao, Steven Low, Katta Ludynia, Yakhat Barshep, Aeshita Mukheriee, Linda

Tsipa, Marja Wren-Sargent , Dieter Oschadleus, Michael Brooks, Oscar Noels, Candice

Lakay, Donella Young, Marienne de Villiers, Nola Parsons, Silvia Mecenero, Steven

Kirkman and Madga Remisiewicz, thanks for all the ideas, discussions, encouragement,

insight, advice and laughter provided. Many especial thanks to Sue Kuyper for being the

soul of the ADU, for always having an ear to listen and the warmest hug ever, able to

heal and comfort all sorts of problems. Thanks for offering your help to all of us and for

your hard work and efficiency. To Andrea Plos for being keen to solve all the problems,

and also for your contagious optimism and friendship over the years.

To Brigitta Tummon for showing me the balance between strength and flexibility, and

about facing life’s challenges with both courage and acceptance. For being a source of

inspiration, joy and beauty. Thanks for helping me to keep healthy, happy and sane

during this time.

Financial support that made possible this study was obtained from the project

LMR/EAF/03/02 of the Benguela Current Large Marine Ecosystem (BCLME)

programme, an NRF grant-holder bursary to Les Underhill in the SeaChange

Programme. The ADU, the EarthWatch Institute and a Gordon Sprigg Bursary also

contributed with funds. Rauri Bowie and Cabildo de Tenerife (Canary Islands, Spain)

provided funding for the genetics work at the University of Stellenbosch and at

University of California at Berkeley.

Marine and Coastal Management (MCM) and CapeNature provided permits to work on

Dassen Island and access to the pelican breeding data. The Nambian Ministry of

Environment and Tourism provided permits to work in Namibia. Sampling permits for

the genetic and microbiological work were granted by CapeNature in the Western Cape

and the Ministry of Environment and Tourism for Namibia. The West Coast National

Park (SANParks) granted permits, provided logistical support and funding for the island

trips. Jointly, we designed and implemented the management intervention that is

presented in Chapter 5.

n addition, over the course of these years in the field and in the lab I have had the

fortune to work with more than 200 people, to whom I am indebted. I will name

them in sections according to what their most important contribution was, although

some of you contributed transversally on many aspects of my work. I have made an

effort to remember all people that helped me with this work, my sincere apologies if by

mistake or a lapsus of memory I leave any of you out.

I

xiv

��

The late Mr and Mrs Fisk and their family kindly allowed us to work on their farm.

George Fisk and his father, Stanley, gave us valuable information on the history of

pelican food subsidisation on their farm from the 1980s. Jan Hofmeyr performed regular

pelican counts and invaluable observations at the monitored pig farm.

Thanks to all that helped ringing and collecting samples at the pig farm: Marius

Wheeler, Doug Harebottle, Silvia Mecenero, Carla Mecenero, Barbara and John

Paterson, Francisco Aguilar, Marta Coll, Eladio García, Samantha Petersen, Maria

Honig, David Swanepoel, Linda Tsipa, Robyn Rorke, Sandra Bouwhuis, Nozuko Lugalo,

Sue Kuyper, Ralf Muller, David Thompson, Tijl van der Winkel, Adam Welz, Vincent

Ward, Pete Mayhew, Lola García López de Hierro, Manfred Waltner and Patrick

Wildjang.

Marine and Coastal Management (MCM), the South African Ringing Unit (SAFRING)

and the Animal Demography Unit (ADU) from the University of Cape Town purchased

the rings. SAFRING granted the ringing permits and the ringing-recapture data. MCM

generously supplied the two satellite transmitters fitted on the pelicans. Jan Hofmeyr,

Pieter van Wyk, Morné Carstens, Holger Kolberg, Keith Wearne and many others

reported resightings of ringed pelicans.

��

Transport to Dassen Island was provided by JC Smith from Portnet and the crew of the

RV Sardinops. Johan Visagie, Leshia Upfold and Bruce Dyer provided logistical support

during the ringing expeditions to Dassen Island. Bruce Dyer (MCM) assisted with

analysis of diet samples. Thanks to all that assisted with capturing, ringing and

collecting samples from pelicans chicks: Les Underhill, John Cooper, Peter Ryan, Johny

Witbooi, Nyameko Pita, Andrea Plos, Silvia Mecenero, Steven Kirkman, Sue Kuyper,

Lucy Kemp, Maria Honig, Irene García Fernández, Aanyiah Omardien, Res Altwegg,

David Grémillet, Pete Mayhew, Linda Tsipa, Shannon Neills, Paul Hanekom, Magda

Remisiewicz, Hannah Thomas, Philippa Schultz and the Zoology Honours students of

2006 and 2009.

��

Holger and Claire Kolberg, Dirk Heindrick, Marc Dürr, Mark Boorman and the late

Keith Wearne provided logistical support during the ringing field trips. Barbara and

John Paterson and Sandra Dantu and Mark Boorman offered us company and a place to

stay. Dr Clinton Hay (Ministry of Fisheries and Marine Resources, Hardap Dam,

Namibia) assisted with analysis of diet samples from Hardap Dam. Mark Boorman

identified the diet samples from the pelican trapped in Walvis Bay.

Ricky Taylor (KZNWildlife), Meyrick and Karin Bowker facilitated the trip to St Lucia

and showed us the wonders of the lake and its surroundings. Bruce Dyer and Rob

Crawford were fantastic travel and field companions. Ricky Taylor, Meyrick Bowker and

Caroline Fox kindly kept me informed of the state of the lake and the fluctuating

numbers of pelicans.

xv

��

This important part of my work would have not been possible without the valuable

conservation-aimed work of the late Norah Kreher from The Bateleurs, who sponsored

the aerial surveys. Pilots Dr Harold Bloch and his friend Graham Smith, members of

The Bateleurs generously contributed with their knowledge, enthusiasm, aeroplane and

time, flying us safely over the waterbodies of the Western Cape counting and

photographing pelicans. Pierre Nel provided continuous support and contributed to the

aerial and ground surveys. Marine and Coastal Management (MCM) and CapeNature

personnel undertook counts of the breeding colony on Dassen Island. About 60

volunteers from local bird clubs, CWAC participants and conservation officials

undertook regular pelican counts at mainland sites. Especial thanks to Ann and Rae

Gordon, Anton Oudendal, Avril Young, Brian Dennis, Brian Herman, Colin and Pam

Codner, Dalton Gibbs and the students at Rondevlei, David Swanepoel, Doug

Harebottle, Eileen Wallace, Elbe Cloete, Engela Duvenage, Felicity and Nick Strange,

Gilly Smith, Gordon Scholtz, Helené Mostert, Jan Hofmeyr, Jannet van der Vyve, Johan

Visagie, John Carter, Joy Kruger, Judy New, Keith Harrison, Keith Spencer, Kevin

Shaw, Koos Retief and staff at Rietvlei Nature Reserve, Lee Burman, Leshia Upfold,

Lorraine from Hermanus, Louise de Roubaix, Manfred Low, Margaret McCall,

Marguerite van der Merwe, Mariana Delport, Marius Wheeler, Marjorie de Kock, Mark

Anderson, Maud Agrealla, Morné Carstens, Otto Schmicht, Penny Palmer, Peter

Chadwick, Peter Vijoen, Pieter Albertyn, Pieter van Wyk, René Navarro, Roelof Jalving,

Sally Blake, Sheilla Walker, Sylvia and Paul Slabbert, Vincent Ward, Willem

Grobbelaar and Yvonne Weiss. Stephany Taylor, Pete Hancock from BirdLife Botswana

coordinated counts in Botswana and Holger Kolberg and Keith Wearne in Namibia.

��

Eddie Papier, Sipho Jivindava, Bruce Dyer, Mwema Musangu, Colin and Pam Codner,

Lorien Pichegru, Ralf Mullers, Azwianewi Makhado, Nicola Okes and Johan Visagie

counted pelicans and seabirds on the islands and provided first hand observations.

Thanks to the field rangers and Honorary Rangers of the West Coast National Park for

their dedication and hard work counting and chasing pelicans and monitoring seabirds

on Malgas and Jutten Island. Colin Codner made us aware of the problem on Schaapen

Island, and besides being a veteran on Malgas and Jutten, took the responsibility of

monitoring pelican numbers on Schaapen Island. Eddie Papier, Pierre Nel and Wallace

Benjamin took us on board of Pelecanus II safely to the islands. Thanks to field rangers

and WCNP personnel Jan Swarts, Jan Peters, Thembile Mpongoza, Andile Manana,

Huzalle Blake, Chris Wilkinson, Karen Engelbrecht, Claude Coraizen, Johannes

Kleinsmidt, Faroeshka Gool, Ashlyn Klein, Jansen Botha, and to all the volunteers,

many of them Honorary Rangers: Colin and Pam Codner, Alec and Janet Celliers, Binks

MacKenzie, Dirk and Karen Havenga, Heidi Duncan, Wendy Wentzel, Sandy

Youngkrantz and Gerard Oosthuizen, Janice Maltby, Dee Cartoulis, Heili and Tony

Templer, Mike and Jenny Lodge, Elbeth, Edelberth and Ebert Le Roux, Monika Forner,

Christo Bredell, Shannon Neills, Reena Cotton, Dagmar and Michael Forcioli, Ben and

Delia Dilley and Annina Vinzens.

xvi

��

Numerous people assisted with the collection of samples in both South Africa and

Namibia (see above). David Allan at the Durban Natural Science Museum allowed us to

sample egg membranes from their egg collection and toe-pads from museum specimens.

Ricky Taylor provided tissue samples from a pelican of the KwaZulu-Natal population.

Mike Harman and Katja Koeppel at the Johannesburg Zoo provided tissue from their

captive pelicans. Mark Brown and the Museum of Vertebrate Zoology at the University

of California at Berkeley provided samples of other pelican species. Microsatellite

enrichment was carried out by Kevin Feldheim in the Pritzker Laboratory for Molecular

Systematics and Evolution and was partially funded by the Grainger Foundation.

Special thanks to Anna Sellas, Sampath Lokugalappatti, Ângela Ribeiro, David Vieites,

Sandra Nieto Román, Ricardo Pereira, Jérôme Fuchs, Krystle L. Chavarría, Julie

Woodruff, Lydia Smith and all my colleagues at the Museum of Vertebrate Zoology and

Behaviour Lunch at UC Berkeley for their insightful discussions and generosity.

��

Thanks to Tumeka Mpongoma, Jose Enrique Ascanio Suárez, Annina Vinzens, Monika

Forner and Candice Lakay for their help at entering data. I borrowed spotting scopes,

photo and video equipment from David Gwynne-Evans, Peter Ryan, Allan Ellis, Callan

Cohen, Phil Hockey, Harold Bloch, EarthWatch Institute, the Zoology Department at

UCT and the West Coast National Park. Tom Peschak, Alberto del Hoyo, Mark

Anderson, Peter Ryan, Jan Hofmeyr, Ralf Mullers, Madga Remisiewicz, Leshia Uphold,

Eladio García, Lucy Kemp, Francisco Aguilar, Marta Coll and Adam Welz provided

images and video footage that I used in talks and other events showcasing this work.

Thanks for all to Sue Kuyper, Linda Tsipa and Marja Wren-Sargent at the ADU; to

Gilly Smith at the Marine Biology Research Centre in the Zoology Department, who

considered me a ‘marine’; to Hayley Battle, Meg Ledeboer, George du Plessis and

Granville Faulmann and Andrea Plos from the Department of Zoology; to Chris Tobler

and Hilary Buchanan from the FitzPatrick Institute; to Anna Hippolito, Steven Long

and others at University of California at Berkeley; and to Margaret Koopman at the

Niven Library at UCT.

inally, thanks to the fascinating Nature of southern Africa: the islands, vleis,

mountains and sea; especially to the pelicans and the seabirds. To all the

wonderful places where this PhD has taken me. To all those weeks and months of

solitude while doing observations from 5am to 8pm on Jutten and Malgas Island. I can

now smile at my tears when the pelicans ate one chick too many, and still worry when I

remember the feeling of being truly alone whenever all pelicans left the island, or at

closing my eyes and only seeing white birds, that I started counting automatically. I will

never forget the pelicans eating ‘spaghetti’ at the pig farm, or the early flights with

Harold’s aircraft and the incredible bird’s eye views of the Western Cape; the trips to the

islands in all kinds of seas: in the fog, through gigantic swells, among whales,

cormorants and little penguin heads. And also the trips along thousands of kilometres in

southern Africa, and further ones in the folds of my mind as a continuation of this

project.

F

�

CHAPTER 1

��

1

��

“Let a squadron of southbound pelicans but feel a lift of prairie breeze (...) and

they sense at once that here is a landing in the geological past, a refuge from

that most relentless of aggressors, the future. With queer antediluvian grunts

they set wing, descending in majestic spirals to the welcoming wastes of a

bygone age”.

Aldo Leopold 1949

�

�

��

Although in this case describing a flock of American White Pelicans Pelecanus

erythrorhynchos, the quotation above could as well relate to the ecologically

equivalent species in the Old World, the Great White Pelican P. onocrotalus.

The eight species of pelicans of the world belong to the genus Pelecanus, which

constitutes a highly conserved group in spite of an ancient phylogenetic origin.

Fossils ascribed to Pelecanus have been dated as far back as the Miocene (ca.

20 mya) and share key anatomical features with extant species (Johnsgard

1993). Pelicans are classified into three major groups: 1) the ground-nesting

large-bodied species, in which American White and Great White Pelicans are

included together with Dalmatian Pelican P. crispus and Australian Pelican P.

conspicillatus; 2) Pink-backed P. rufescens and Spot-billed Pelican P.

philippensis, both tree-nesting species with more solitary habits than the

previous four species; and 3) two marine species, Brown Pelican P. occidentalis

and Peruvian Pelican P. thagus, previously considered a subspecies of the Brown

Pelican but recently accepted as a separate species (American Ornithologists'

Union 2007; Banks et al. 2008). Pelicans have colonised and are distributed

across all continents except Antarctica (Table 1.1).

Ta

ble

1.1

: D

istrib

utio

ns,

pop

ula

tio

n e

stim

ates a

nd

con

serva

tio

n s

ta

tu

s o

f th

e e

igh

t s

pecie

s o

f p

eli

ca

ns o

f th

e w

orld

.

Da

ta

com

pil

ed

from

: (1

) M

arch

an

t &

Hig

gin

s (

19

90

); (

2)

del

Hoyo e

t a

l. (

19

92

); (

3)

Joh

nsga

rd

(1

99

3);

(9

) C

riv

ell

i et

al.

(1

99

8a

); (

10

) S

hie

lds (

20

02

); (

11

) S

imeon

e e

t a

l.

(20

03

); (

12

) K

nop

f et

al.

(2

00

4);

(1

3)

Keit

h (

20

05

); (

14

) N

els

on

(2

00

5);

(1

5)

Wetla

nd

s I

ntern

atio

na

l (2

00

6);

(1

6)

Am

eric

an

Orn

ith

olo

gis

t’ U

nio

n (

20

07

); (

17

) IU

CN

(2

00

9).

See t

ext a

nd

th

e c

ited

gen

era

l refe

ren

ce b

ook

s f

or a

more d

eta

iled

recou

nt o

f th

e p

rim

ary l

itera

tu

re.

Sp

ec

ies

D

istrib

utio

n

Po

pu

latio

n s

ize

(b

re

ed

ing

pa

irs

)1

5

Ma

in b

re

ed

ing

sit

es

M

ov

em

en

ts

P

op

ula

tio

n

Sta

tu

s (

IUC

N) 1

7

Gre

at W

hit

e P

eli

ca

n

Pe

lec

an

us

on

oc

rota

lus

Su

b-S

ah

ara

n A

fric

a

Eu

ra

sia

(so

uth

-e

aste

rn

Eu

ro

pe

,

form

er U

SS

R,

Ch

ina

,

Mo

ng

olia

)9

23

0 0

00

(7

5 0

00

–1

00

00

0)

35

10

0–

62

90

0

(7

50

0–

10

50

0)

Se

ne

ga

l, M

au

rita

nia

Eth

iop

ia,

Ke

nya

, T

an

za

nia

, Z

am

bia

,

Bo

tsw

an

a,

So

uth

Afr

ica

, N

am

ibia

Ro

ma

nia

, G

re

ece

, T

urke

y,

Ira

n,

Pa

kis

tan

?,

Ka

za

kh

sta

n,

Ru

ssia

,

Tu

rkm

en

ia?

, U

zb

ekis

tan

? 9

Su

b-S

ah

aria

n p

op

ula

tio

ns r

esid

en

t o

r

dis

pe

rsiv

e

Eu

ra

sia

n p

op

ula

tio

ns m

igra

tory (

ove

rw

inte

rs in

Su

da

n/E

thio

pia

).

Sta

ble

in

Afr

ica

(L

ea

st

co

nce

rn

)

De

clin

ing

in

th

e

Pa

lea

rctic

Am

eric

an

Wh

ite

Pe

lic

an

P.

ery

thro

rhy

nc

ho

s 1

2,1

5

No

rth

Am

eric

a,

inla

nd

an

d w

est

>1

80

00

0

( 5

4 6

00

)

Britis

h C

olu

mb

ia (

Ca

na

da

);

Ca

lifo

rn

ia,

Mo

nta

na

, N

orth

Da

ko

ta

(U

SA

)

Bo

th r

esid

en

t a

nd

mig

ra

tory in

div

idu

als

in

sa

me

po

pu

latio

ns.

Tw

o m

eta

po

pu

latio

ns,

Ea

st

an

d

We

st.

Dis

pe

rse

/ m

igra

tes t

o c

en

tra

l A

me

ric

a.

Sta

ble

/ in

cre

asin

g

(L

ea

st

co

nce

rn

)

Da

lma

tia

n P

eli

ca

n

P.

cri

sp

us

9

Yu

go

sla

via

, B

lack S

ea

Ch

ina

10

40

0–

13

80

0

(4

00

0–

5 0

00

)

21

–2

2 c

olo

nie

s in

fo

rm

er U

SS

R,

Gre

ece

, T

urke

y,

Ba

lka

ns,

Ch

ina

Dis

pe

rsiv

e in

Eu

ro

pe

, m

igra

tory in

Asia

D

eclin

ing

(V

uln

era

ble

)

Au

stra

lia

n P

eli

ca

n

P.

co

ns

pic

illa

tus

1

Au

str

alia

an

d T

asm

an

ia

10

0 0

00

–

1 0

00

00

0

Off

-sh

ore

isla

nd

s (

Ta

sm

an

ia);

sm

all

sa

nd

y isla

nd

s o

r a

lon

g s

ea

sh

ore

in

lake

s,

sw

ap

s (

we

st

an

d s

ou

th-e

ast)

Re

sid

en

t o

r d

isp

ersiv

e

Sta

ble

(L

ea

st

co

nce

rn

)

Pin

k-b

ac

ke

d P

eli

ca

n

P.

rufe

sc

en

s

Su

btr

op

ica

l a

nd

tro

pic

al A

fric

a,

an

d t

he

Re

d S

ea

50

00

0–

10

0 0

00

S

en

eg

al, E

thio

pia

, S

ou

th A

fric

a

(K

wa

Zu

lu-N

ata

l),

Bo

tsw

an

a,

SW

Ara

bia

Dis

pe

rsiv

e o

r r

ela

ted

to

wa

ter c

on

ditio

ns

Wid

esp

re

ad

an

d

co

mm

on

(L

ea

st

co

nce

rn

)

Sp

ot-b

ille

d P

eli

ca

n

P.

ph

ilip

pe

ns

is

Sri L

an

ka

, S

E I

nd

ia,

Ca

mb

od

ia

po

ssib

ly in

Su

ma

tra

13

00

0–

17

00

0

(ca

. 1

30

0)

Sri L

an

ka

(2

3 c

olo

nie

s) a

nd

SE

In

dia

(4

co

lon

ies)

Po

orly

kn

ow

n

De

clin

ing

(N

ea

r t

hre

ate

ne

d)

Bro

wn

Pe

lic

an

P.

oc

cid

en

tali

s 1

0

(P

.o.o

cc

ide

nta

lis

,

P.o

. c

aro

lin

en

sis

,

P.o

. c

ali

forn

icu

s,

P

.o.

mu

rph

y,

P

.o.

uri

na

tor)

Co

asta

l N

orth

, C

en

tra

l a

nd

So

uth

Am

eric

a:

5 s

ub

sp

ecie

s

(C

arib

be

an

,

tro

pic

al A

tla

ntic,

Pa

cific

Ca

lifo

rn

ia–

Me

xic

o,

Co

lom

bia

–N

Pe

ru

,

Ga

lap

ag

os I

sla

nd

s)

29

5 0

00

(9

5 0

00

)

We

st

Ind

ies,

Ba

ha

ma

s,

Cu

ba

,

Ja

ma

ica

, A

ntille

s,

Co

lum

bia

, A

ru

ba

,

Ve

ne

zu

ela

, T

rin

ida

d;

Ma

ryla

nd

,

Vir

gin

ia,

N./

S.

Ca

ro

lin

a,

Ge

org

ia,

Flo

rid

a,

Ala

ba

ma

, L

ou

isia

na

, T

exa

s,

Me

xic

o,

Be

lize

, H

on

du

ra

s,

Pa

na

ma

;

Ca

lifo

rn

ia;

Ecu

ad

or;

Ga

láp

ag

os

Dis

pe

rse

s c

oa

sta

lly,

so

me

mig

ra

tio

n a

lso

occu

rs.

Me

tap

op

ula

tio

n s

tru

ctu

re

Sta

ble

/ in

cre

asin

g

(L

ea

st

co

nce

rn

)

Pe

ru

via

n P

eli

ca

n

P.

tha

gu

s 1

1,1

6

Co

asta

l P

eru

an

d t

o c

en

tra

l

Ch

ile

40

0 0

00

–6

00

00

0

Pe

ru

; Is

las M

och

a,

Ca

ch

ag

ua

an

d

Pá

jaro

Niñ

o in

no

rth

-ce

ntr

al C

hile

Dis

pe

rse

s a

lon

g t

he

co

ast

du

rin

g t

he

no

n-

bre

ed

ing

se

aso

n.

Sym

pa

tric

with

P.o

. m

urp

hy

fro

m E

cu

ad

or t

o C

hile

.

Re

co

ve

rin

g

(N

ea

r t

hre

ate

ne

d)

��

3

All pelicans share gregarious habits, breeding communally and in some cases

roosting and fishing in groups. They depend on aquatic habitats and most use

their large beaks and extensible pouches to capture fish in shallow waters. The

exception are the Brown and Peruvian Pelicans, which plunge-dive, thus

accessing fish a few metres beneath the surface. Prey items are not exclusively

restricted to fish, and some species also eat crustaceans, amphibians and other

aquatic prey. Some species, e.g. Australian, Brown and Great White Pelicans,

are opportunistic feeders and scavengers, occasionally eating dead organisms or

begging for scraps from fishermen. Their large wing-spans, light skeletons and

the presence of air sacs under the skin enable them to fly long distances despite

being among the heaviest of flying birds. They often use thermals for long-

distance movements and even to commute several hundred kilometres between

nesting and feeding places during the reproductive period (del Hoyo et al. 1992;

Johnsgard 1993).

Traditionally pelicans have been included in the order Pelecaniformes together

with cormorants (Phalacrocoracidae), anhingas (Anhingidae), sulids (Sulidae),

tropicbirds (Phaethontidae) and frigatebirds (Fregatidae) based of shared

morphological and behavioural characteristics (e.g. totipalmate feet, cranial

anatomy, unfeathered gular poach, foraging habits and social behaviour;

Lanham 1947; Cracraft 1985). Since the advances in molecular techniques

involving DNA-DNA hybridisation (Sibley et al. 1988; Sibley & Ahlquist 1990)

the structure of the group has changed profoundly, to include species not

previously considered in the order and implying that the Pelecaniformes do not

constitute a natural taxonomic group (Johnsgard 1993). Recent genetic work

confirms that the group is indeed polyphyletic and pelicans seem to be more

closely related to the Shoebill Balaeniceps rex (Balaenicipitidae) and the

Hamerkop Scopus umbretta (Scopidae), and even to ibises (Threskiornithidae)

and herons (Ardeidae), than to the families that were formerly included in the

Order Pelecaniformes (Ericson et al. 2006; Hackett et al. 2008; Figure 1.1).

��

4

Figure 1.1: Avian molecular phylogenetic relationships including pelicans and

members of the Pelecanidae after (a) Ericson et al. (2006) and (b) Hackett et al.

(2008).

Although early work considered pelicans as a subfamily of a larger group that

included tropicbirds, frigatebirds, anhingas, cormorants and sulids (Elliot 1869),

the first taxonomic classification of the pelicans by Ogilvie-Grant (1989)

comprised a total of nine species, all included in the genus Pelecanus. He

recognised P. thagus as separate from P. occidentalis (then named fuscus) and

considered a species, P. roseus, which has later been considered a synonym of P.

onocrotalus. Posterior classifications by Peters (1931) and Dosrt & Mougin (1979)

introduced minor changes to this scheme, although they split the genus into

three distinct phyletic groups. Consistently, P. occidentalis (including the

Peruvian Pelican P. thagus) has been included in the Leptopelicanus group, and

P. erythrorhynchos in the Cyrtopelicanus group, including the remaining species

into the nominal subgenus Pelecanus. Brown (1982) also consider three pelican

superspecies based on comparative biological traits, i.e. fishing strategies

(communal, solitary or plunge-diving) and breeding behaviour (large ground

colonies or smaller tree-nesting colonies) (Table 1.1).

Among the pelican species of the world, American White and Brown Pelicans

have been widely studied in some of their areas of distribution (see Shields 2002;

Knopf & Evans 2004 for a detailed list of the primary literature). Numerous

studies have been carried out on the threatened Dalmatian Pelican (e.g. Crivelli

& Vizi 1981; Crivelli et al. 1988; 1989; 1991a; 1997a; 1998b; Pyrovetsi &

Papazahariadou 1995; Catsadorakis et al. 1996; Peja et al. 1996; Pyrovetsi 1997;

(a (b)

��

5

Pyrovetsi & Economidis 1998; Catsadorakis & Crivelli 2001), although important

aspects like survival, age of first breeding and philopatry require further

research (Crivelli et al. 1997a). Less well-known are the Australian Pelican (but

Vestjens 1977; Marchant & Higgins 1990) and the Pink-backed Pelican (see

Burke & Brown 1970; Din & Eltringham 1974; Bowker & Downs 2008a; c). The

Spot-billed Pelican is the only species endemic to Asia, and comprises a small

number of breeding colonies in Sri Lanka, India and possibly Cambodia (Table

1.1). Many aspects of its breeding biology and ecology are not known, although

research has been carried out in India and Sri Lanka (Neeklakantan 1949; Gee

1960; Nagulu & Rao 1983; Nagulu 1984; Collar & Andrew 1988). The Peruvian

Pelican is the least studied of the group (Wetmore 1945; Ridgely & Greenfield

2001), having been split from the Brown Pelican in 2007 (Banks et al. 2008).

Due to their comparable life-history traits and usage of similar aquatic habitats,

threats to their populations are similar for all pelicans (Crivelli & Schreiber

1984). The greatest threats to pelican populations are disturbances at breeding

colonies (Johnson & Sloan 1976; Anderson & Keith 1980; Anderson 1988;

Boellstorff et al. 1988; Anderson et al. 1989), habitat destruction and drainage of

wetlands (Crivelli & Vizi 1981; Crivelli & Schreiber 1984; van Halewyin &

Norton 1984; Crivelli et al. 1991a; 1998a; Catsadorakis et al. 1996; King &

Michot 2002; Murphy 2005), competition with man for fish (Anderson et al. 1980;

Anderson et al. 1982; Duffy 1983; Schreiber & Mock 1988; Shmueli et al. 2000a),

illegal shooting (Schreiber & Mock 1988; Crivelli et al. 1991a), collision with

transmission lines, wires and fences (McNeil et al. 1985; Crivelli et al. 1988;

Shmueli et al. 2000b) and pesticide contamination (WB Anderson et al. 1977;

Blus 1982; Crivelli et al. 1989; Gress 1995; DW Anderson et al. 1996).

��

The Great White Pelican breeds in Europe and Africa (Crivelli et al. 1991b; del

Hoyo et al. 1992). The global population was estimated in 2006 to consist of

ca. 200 000 birds (Wetlands International 2006). Two geographically

��

6

independent populations have been identified (Crivelli & Schreiber 1984; Crivelli

1994): the declining Euro-Asiatic population, currently about 7 500-10 500 pairs

(Wetlands International 2006), and the African population which has been

considered to have remained stable at ca. 75 000 pairs since the 1960s (Brown &

Urban 1969; Crivelli et al. 1991a; 1998a; Wetlands International 2006; Dodman

in press).

In Africa, they have been recorded to breed at more than 26 sites (Urban 1984).

Some populations are nomadic and nest opportunistically when conditions are

adequate, sometimes in large numbers (20 000–30 000 pairs). The largest annual

breeding colonies occur in Isle of Arel (Mauritania), Park du Djoudj (Senegal),

Lake Shala (Ethiopia), Lake Elmenteita (Kenya) and Lake St Lucia (South

Africa) (Urban 1984). The western African population was estimated at 60 000

birds in the early 2000s, whereas the pelican population of eastern and southern

Africa was estimated at 150 000 birds, 50 000 pairs in eastern Africa and 6 000

in southern Africa (Dodman in press). Because of the difficulty of counting at all

sites simultaneously, the ability of pelicans to move large distances and the

absence of counts at some important pelican sites due to their remoteness (e.g.

Lake Rukwa in Tanzania; T Dodman, pers. comm.), some error can be expected

regarding the accuracy of these estimates.

Accordingly to the characteristics of their local environment, there is some

variation in the foraging habitat of pelicans. Large pelican colonies in Central

and East Africa depend on abundant fresh-water fish from large lakes along the

Rift Valley (Brown & Urban 1969), whereas fluctuating pelican populations in

West Africa feed on marine prey, targeting fish trapped in shallow channels

surrounding the sandy islands where they breed (Crivelli 1994). Also, larger

lakes and the existence of vast extensions of wetlands in central and eastern

Africa contrasts with the small wetlands of the Western Cape (Guillet & Crowe

1981) and south-eastern Europe (Crivelli et al. 1997b), where pelicans exploit

fish at fresh-water rivers and marshy areas.

��

7

In southern Africa, Great White Pelicans breed annually at three coastal

localities: Bird Rock guano platform, in Walvis Bay, Namibia; Dassen and

Vondeling Islands in the Western Cape, and St. Lucia Wetland Park in

KwaZulu-Natal, both in South Africa. Additionally, when rains have been

abundant, they breed in some inland sites: Hardap Dam, Namibia, and less

frequently in Etosha Pan, Namibia, Sua Pan and/or Lake Ngami, Botswana

(Berry et al. 1973; Berruti 1980; Crawford et al. 1981; Williams & Borello 1997;

Theron et al. 2003; Hockey et al. 2005). They have been reported breeding

occasionally at the mouth of the Zambezi River in Mozambique (R Beilfuss, pers.

comm.) and on Ihla dos Tigres off the coast of southern Angola (Dean et al. 2002).

Although not globally threatened (BirdLife International 2008), Great White

Pelicans have experienced population declines in some of areas (Crivelli et al.

1997b). In South Africa they are listed as near-threatened due to the small

number of breeding colonies (Barnes 2000). However, the Western Cape

population increased markedly over the last few decades (Crawford et al. 1995),

possibly as a consequence of the availability of agricultural offal from farms in

the Cape Town Metropolitan area (Crawford et al. 1995; de Ponte Machado &

Hofmeyr 2004).

��

The Western Cape Great White Pelican population is relatively small when

compared to other colonies elsewhere in Africa (Hockey et al. 2005; Dodman in

press). This population occupies the southernmost edge of their distributional

range and historically their foraging habitat was restricted to small wetlands

along the Atlantic coast (Guillet & Crowe 1981). The construction of farm dams

linked to agricultural expansion during the 20th century increased the biomass

of fish prey (mostly introduced Carp Cyprinus carpio) and expanded the suitable

habitat for pelicans (Crawford et al. 1995). Additionally, since the 1980s they

have increasingly foraged at refuse dumps and other artificial sites (Crawford et

��

8

al. 1995). Nonetheless, they require undisturbed sites for breeding, and they

roost next to the water in estuaries, rivers, water reservoirs, ephemeral

wetlands, offshore islands and occasionally on beaches along the coast (Guillet

1985).

Although Dassen (33°25’ S, 18°05’ E) and Vondeling (33°09’ S, 17°58’ E) Islands

are nowadays the only breeding sites for pelicans in the Western Cape since the

mid 1950s, historically they bred on other off-shore islands in the Western Cape,

including Robben Island, Dyer Island, Quoin Rock and Seal Island in False Bay,

where they were subjected to disturbance by human activities (e.g. guano

collecting, naval practices) and the encroachment of the Cape Fur Seal

population Arctocephalus pusillus. Pelicans were discouraged from breeding on

these localities during the 19th and early 20th centuries, under the claim that

they disturbed guano-producing cormorants (see detailed review in Crawford et

al. 1995). Pelicans bred for the first time on Dassen Island in 1956, and have

bred there annually since then (Crawford et al. 1995). In both 2001 and 2005 a

small satellite colony formed on Vondeling Island, 30 km north of Dassen Island

(Marine and Coastal Management, unpublished data).

Dassen Island is a flat and sandy island 10 km off-shore from Yzerfontain, and

55 km north-west of Cape Town. After Robben Island, Dassen is the second

largest of South Africa’s continental islands with an area of 222 ha. The

vegetation is shrubby, dominated by Tetragonia fruticosa and Trachyandra

divaricata. It is home to the largest population of African Penguins Spheniscus

demersus (ca. 20 000 breeding pairs in the first half of 2000s (Underhill et al.

2006), although having declined to 5 000 pairs in 2009, following alarming

downward trends across their distribution; Marine and Coastal Management,

unpublished data). Large colonies of Cape Cormorants Phalacrocorax capensis

breed on the island, fluctuating from nil to 48 000 pairs depending on the

availability of pelagic fish (Crawford 2007). Other seabird species that breed on

Dassen Island are Kelp Larus dominicanus vetula and Hartlaub’s L. Hartlaubii

Gulls, White-breasted Phalacrocorax carbo, Crowned P. coronatus and Bank P.

��

9

neglectus cormorants, Swift Terns Sterna bergii, Leach’s Storm Petrels

Oceanodroma leucorhoa, Sacred Ibises Threskiornis aethiopicus, African Black

Oystercatchers Haematopus moquini, Egyptian Geese Alopochen aegyptiacus,

four species of plovers (Charadriidae) and several land birds (Passeriformes)

(Whittington & Wolfaardt 1999). The densest population of Angulate Tortoise

Chersina angulata occurs on the island, as well as an introduced population of

European Rabbits Oryctolagus cuniculus brought by sailors in the 17th century

(Rand 1963).

The island was an important source of guano, especially during the 19th century

(Rand 1963). Most deposits were stripped to bare rock but guano continued to be

collected until 1974. Other extractive activities with an adverse effect on seabird

populations included penguin egg collection (Wolfaardt 2000). The island was

previously managed by the Guano Island Division of the Department of

Industries of the South African government. In 1987 Dassen Island was

proclaimed a Provincial Nature Reserve, and is currently managed by the

provincial conservation authority, CapeNature. Extractive activities are no

longer allowed, and public access is not permitted without authorisation to

minimise disturbance to seabirds (Wolfaardt 2000).

��

Despite being a large and charismatic species, large gaps remain in the

knowledge of the Great White Pelican in Africa. Pelican research bloomed from

the late 1960s to the early 1980s mostly in eastern and southern Africa, but

there have been few studies since then. The following summary includes the

most relevant studies for Great White Pelicans related to the topics and

questions developed in this thesis.

During the late 1970s, annual regional surveys were undertaken in the Western

Cape with the aim to estimate the size of the pelican population (Cooper 1976;

1977; 1978; 1979). Siblicide and other aspects of their breeding behaviour were

also studied (Cooper 1980b) as well as the energetic requirements of a captive-

��

10

reared juvenile (Cooper 1980a). The use of habitat by Great White Pelicans and

some aspects of their breeding ecology were thoroughly investigated in the

Western Cape (Guillet & Crowe 1981; 1983), as well as their energy

requirements (Guillet & Furness 1985). The conservation status of waterbirds

(including pelicans), as well as their breeding biology and energy consumption

were studied at Lake St Lucia, KwaZulu-Natal (Berruti 1980; 1983). Numbers of

pelicans breeding in Namibia, at the Bird Rock Platform in Walvis Bay (Berry

1975; Crawford et al. 1981; Williams in press), and at Hardap Dam (Theron et al.

2003) have also been reported. Breeding attempts by the more nomadic southern

African inland populations were recorded from Etosha Pan in 1971 (Berry et al.

1973), Lake Ngami in 1981 (Williams & Randall 1995) and at Sua Pan in 2007

(McCulloch 2008). More recently, Crawford et al. (1995) reported on the observed

increase in population size and the main conservation aspects of the Western

Cape. Updated information is available on the population trends, conservation

status and breeding history for the most south-eastern African population at

Lake St Lucia, KwaZulu-Natal (Bowker & Downs 2008a; b; c).

Relevant studies have been undertaken at other localities. Some have made

reference to population parameters or life-history traits: in West Africa and

southern Europe (Crivelli et al. 1991b; Crivelli 1994); and in East and Central

Africa (Brown & Urban 1969; Brown et al. 1982; Urban 1984; Urban & Ash

2001). The incidence of migration, routes used and energetic analyses have been

studied for the migrant European population (Izhaki 1994; Shmueli et al. 2000a;

b; c; Izhaki et al. 2002), with emphasis on conservation aspects regarding the

stop-over sites in Israel.

��

Human-induced changes may often lead to population declines. However, some

species have benefited from human activities, increasing in numbers or

expanding in range, as it has been the case with some generalist or scavenging

seabirds such as gulls Larus spp. and skuas Catharacta spp. (Harris 1970;

Furness et al. 1981; Furness & Monaghan 1987; Camphuysen 1995; Pons &

��

11

Migot 1995; Furness 2003). Great White Pelican numbers in the Western Cape

have increased in the second half of the 20th century (Crawford et al. 1995; de

Ponte Machado & Hofmeyr 2004). This increase has been driven by cessation of

disturbance to the breeding colonies, the expansion of artificial fresh water

reservoirs in agricultural areas, and most importantly, the availability of extra

food to the pelicans from chicken and pig farms (Crawford et al. 1995). Close

contact between domestic animals and wild fowl could have consequences for

public and wildlife health (Hull 1963; Botzler 1991; Wobeser 1997; Acha &

Szyfres 2001; Reed et al. 2003; Casas & Pozo 2005; Makarov et al. 2005; Acevedo-

Whitehouse 2009; United States Geological Survey 2009). Also, the increase in

numbers of subsidised species could have detrimental consequences on

threatened populations that constitute natural or alternative prey of these

subsidised organisms (Furness 1981; Bosch 1996; Martínez-Abraín et al. 2003;

Votier et al. 2004; Oro et al. 2005; Oro & Martínez-Abraín 2007). In the case of

this study, declining populations of cormorants Phalacrocorax spp., gannets

Morus capensis, and other seabirds are threatened by pelican predation, which

has become a novel avian predator in the region (du Toit et al. 2002; de Ponte

Machado 2007; Kemper et al. 2007). Pelican predation has propitiated a

conservation paradox in which a protected species, as it is the Great White

Pelican in southern Africa (Barnes 2000; du Toit et al. 2002), threatens the

recruitment and hence the conservation status of other seabird species, some of

which are listed as globally threatened following alarming signs of decline during

the 20th century and the first decade of the 21st century (Kemper et al. 2007).

A poorly explored consequence of global change is the exhibition of new or

modified behavioural patterns by adaptable species. As mentioned above, for

some opportunistic species this could consist of the exploitation of innovative

sources of food, which in turn could result in knock on effects on the ecosystem as

a whole by affecting other members of the community. Extensive research has

been carried out on these denominated ‘problem species’ and their impacts

(Furness 1981; Furness & Monaghan 1987; Regehr & Montevecchi 1997; Vidal et

al. 1998; Oro & Furness 2002). Intensified predation or competition with other

��

12

seabirds has resulted on ongoing efforts to understand, control or revert the

increases in numbers of these super-abundant species and their negative impact

on sensitive species (Bosch 1996; Vidal et al. 1998; Bosch et al. 2000; Guillemette

& Brousseau 2001; Finney et al. 2003; Neves et al. 2006; Sanz-Aguilar et al.

2009). However, these interactions are often complex and various confounding

factors influence the dynamics of both predator and prey species (Regehr &

Montevecchi 1997; Oro & Furness 2002; Votier et al. 2004; Oro & Martínez-

Abraín 2007; Oro et al. 2009).

Behavioural complexity in the light of global change is an important topic that

should be addressed in studies linking behaviour and conservation biology

(Sutherland 1998). Predicting the responses of organisms to environmental

change is crucial to maintain the processes and services of ecosystems, and

functional healthy communities (Sutherland 2006). Thus, understanding the

behaviour of problem species is necessary to make informed management

decisions, and to avoid undesirable effects as a consequence of well-intentioned

interventions to protect vulnerable species or communities.

Similarly, population genetics studies coupled with demographic and

behavioural data are critical to determine the demographic history of organisms

that are affected by these global changes (Lande 1988; Allendorf & Luikart

2007). Molecular studies can be used to predict the transference of newly

acquired skills among populations, by understanding population structure and

gene flow levels among colonies. Also, understanding the role of metapopulations

in the dynamics and conservation of organisms is critical to develop efficient

management plans (Hanski & Gilpin 1991; Stephens & Sutherland 1999).

Finally, in order to undertake effective conservation actions to restore balance in

communities and ecosystems that have been directly or indirectly altered by

human activities, there is an increasing need to transform conservation

management from an experience-based practice to an evidence-based science,

similar to the developments that constitute the base of the fields of medicine and

��

13

public health (Pullin & Knight 2001; 2003; Sutherland et al. 2004). Thus,

monitoring and evaluating the effectiveness of management interventions, as

well as an adequate dissemination of their results, are needed in order to provide

managers with a range of options and likely outcomes of the possible

interventions (Pullin et al. 2004; Pullin & Knight 2005).

��

Given the conceptual background developed above, the aims of this study are:

(1) to understand the population dynamics of Great White Pelicans in the

Western Cape, South Africa, examining the potential mechanisms driving the

increase in numbers since the 1950s; (2) to estimate the population parameters

and life-history traits necessary to build demographic models to explain the

observed trends; (3) to analyse the pattern of dispersal and/or migration of Great

White Pelicans in southern Africa; (4) to describe a novel foraging behaviour for

the species and the mechanisms of spread of this behaviour through the

population; (5) to evaluate the impact of pelican predation on other seabirds and

the energetic trade-offs of their switch in prey preferences; (6) to evaluate the

outcome of a management intervention undertaken on the islands of the West

Coast National Park in South Africa with the aim to prevent pelican predation

on selected seabird colonies; (7) to investigate the patterns of movements and

gene flow across southern African colonies, testing whether there is reproductive

isolation between pelicans from the west and east coasts of South Africa; (8) to

trace molecular evidence of a historical bottleneck followed by a population

expansion in the Western Cape population as suggested by demographic data;

(9) to confirm whether immigration from other populations (in particular

Namibia) can explain the exponential population growth experienced in the

Western Cape, or whether it can be explained by an increased reproductive

output of this population alone; (10) to identify future areas of research for the

species at a local and global level; and (11) and to propose management actions

given the population expansion of the Great White Pelican population and its

repercussions for the conservation of local seabirds due to the newly acquired

predatory behaviour in the Western Cape.

��

14

In order to achieve these aims I have combined existing historical information

with current population data using a multidisciplinary approach that

incorporates methodologies and conceptualisation from the fields of demography,

population biology, population genetics, behavioural ecology, life-history, avian

disease ecology and adaptive management (Figure 1.2).

Figure 1.2: Conceptual map of the disciplines integrated in this study.

Corresponding chapter numbers are indicated in brackets.

��

Chapter 2 examines the exponential increase of the Great White Pelican in the

Western Cape, investigating the mechanisms of such increase and determining

the importance of artificial food sources for this population. Several datasets are

analysed and combined to explain the demographic changes experienced in the

last five or six decades, as well as to evaluate the consequences of this drastic

increase in population numbers.

��

15

Chapter 3 complements the findings of Chapter 2 by examining detailed aspects

of the Great White Pelican’s life-history and other basic population parameters,

viz. timing of breeding, age at first breeding, movement patterns, causes of

mortality and most importantly survival. Survival estimates are often difficult to

obtain, as populations frequently do not validate all assumptions needed for this

type of mark-recapture analysis. Because of its small size and relative isolation,

the pelican population in the Western Cape is a good model organism for this

type of analysis, being the first time that survival estimates have been calculated

for any species of pelican using maximum likelihood and model selection

methodologies.

Chapter 4 explains the behavioural shift experienced in the foraging habits of

pelicans in the Western Cape, which increasingly rely on seabird chicks for food

during the breeding season. I describe the different foraging patterns observed

when preying on different prey species, and discuss the adaptation of cooperative

hunting techniques from water to land. I also discuss the suspected role of social

learning as a mechanism for the spread of this behaviour throughout the

population and analyse the impact of pelican predation on seabirds during the

2006/07 breeding season on four islands off the west coast of South Africa.

Chapter 5 discusses the different alternatives proposed to counteract the

negative impact of pelican predation, and subsequently I present the results of a

management intervention undertaken in the 2007/08 breeding season on two

islands of the West Coast National Park in the Western Cape, South Africa. The

actions were designed to curb the impact of pelican predation on the already

declining populations of Cape Cormorants and Cape Gannets. This intervention

is the result of the establishment of a collaboration with South African National

Parks (SANParks) and the West Coast National Park Honorary Rangers

(volunteer group).

Chapter 6 assesses gene flow among three southern African breeding

populations in using both mitochondrial and nuclear markers. I use tandem

repeats of nuclear DNA (microsatellites) to examine the population structure of

��

16

the Namibian and Western Cape populations, exploring the occurrence of

demographic bottlenecks and subsequent population expansions on both of these

populations. The findings of this chapter are used to assess the risk of exporting

predatory behaviour from the Western Cape to other areas of distribution of the

species (Chapter 4) as well as to confirm the absence of immigration into the

Western Cape in order to explain the demographic increase of this population

(Chapter 2).

Chapter 7 describes the molecular protocols used for the development and

characterisation of eight microsatellite loci for Great White Pelican, which are

used in the previous chapter (Chapter 6), as well as their potential application to

other members of the Pelecanidae. This chapter was published in the journal

Conservation Genetics. The full citation and author’s institutional addresses are:

de Ponte Machado M1,2, Feldheim KA3, Sellas AB1, Bowie RCK1 (2009)

Development and characterisation of microsatellite loci from the Great White

Pelican and widespread application to other members of the Pelecanidae.

Conservation Genetics 10, 1033–1036.

1 Department of Integrative Biology, Museum of Vertebrate Zoology, University of

California, 3101 Valley Life Science Building, Berkeley, CA 94720, USA

2 Animal Demography Unit, University of Cape Town, Rondebosch 7701, South Africa

3 The Field Museum, Pritzker Laboratory for Molecular Systematics and Evolution, 1400

S. Lake Shore Drive, Chicago, IL 60605, USA

Chapter 8 summarises the main findings included in this thesis, linking them

with broader ecological topics and highlighting the global and local importance of

this study. I enumerate what aspects of this thesis constitute a novel

contribution and present a list of future research questions relevant to the topics

and questions contained in this thesis.

Appendix A.1 contains a study on the prevalence of pathogens on the Great

White Pelican in the Western Cape. Contact between pelicans and domestic

animals could favour the transmission of diseases, with possible detrimental

consequences for the pelicans and potential economic and human health

implications. This study was primarily designed to address the risk of

��

17

transmission of pathogens between wild and domestic species, especially

considering the atypical foraging habits of pelicans in this region, eating raw

chicken offal on a pig farms in close proximity with pigs, humans and poultry.

Further investigation, especially comparisons with other pelican populations

that do not engage on artificial sources of food are required to establish a base-

line level of prevalence of pathogens and to assess the level of biosecurity risk.

This chapter was published in the Journal of Applied Animal Research. The full

citation and author’s institutional addresses are:

Assunção P4, de Ponte Machado M2, Ramírez AS4, Rosales RS4, Antunes NT4,

Poveda C4, De la Fe C4, Poveda JB4 (2007) Prevalence of pathogens in Great

White Pelicans Pelecanus onocrotalus from the Western Cape, South Africa.

Journal of Applied Animal Research 32, 29-32.

2 Animal Demography Unit, University of Cape Town, Rondebosch 7701, South Africa

4 Unidad de Epidemiología y Medicina Preventiva Facultad de Veterinaria, Universidad

de Las Palmas de Gran Canaria Trasmontaña s/n, 35416 Arucas, Spain

Appendix A.2 includes photographic evidence for some of the findings presented

on the other chapters, as well as an illustration of the settings of this study, in

order to help the reader visualise the habitat and characteristics of the species

studied.

��

Each one of the chapters of this thesis has been formatted as an independent

piece of work that will be submitted for publication, or as has already been

published (Chapter 7: de Ponte Machado et al. 2009 and Appendix A.1: Assunção

et al. 2007). As a result, each chapter contains its own introduction and reference

list, and some repetition on the background information and articles cited is

inevitable. Nevertheless, this thesis constitutes in its whole a coherent piece of

work that attempts to make a significant contribution to the understanding of

Great White Pelican population in southern Africa.

��

18

Although the participation of other people was necessary, the responsibility of

the design, data collection, analysis and writing is exclusively mine. Crucial

input from co-authors and supervisors as well as other colleagues has been

acknowledged. Some of the data (i.e. mark-recapture records and historical

counts of pelican breeding pairs and chick production on Dassen Island) were

collected as part of on-going monitoring projects by local research and

conservation agencies. For the regional censuses, ringing expeditions and

management intervention I relied on with the assistance of volunteers and other

field workers, although I was responsible for coordinating and directing such

efforts. I also was responsible for data exploration, analyses, drawing of

conclusions and writing both first drafts and final versions of the chapters. All

chapters in this thesis, except Chapter 4 (which is written in the style of a

submission to a high-impact journal) have been written to a uniform style, which

will require to be transformed into the style of the journal to which it is

submitted.

��

19

��

Acevedo-Whitehouse K (2009) The importance of disease management

programmes for wildlife conservation. Animal Conservation 12, 185–186.

Acha PN, Szyfres B (eds.) (2001) Zoonoses and communicable diseases common to

man and animals. 3rd edn. Pan American Health Organization,

Washington.

Allendorf FW, Luikart G (2007) Conservation and the genetics of populations.

Blackwell, Oxford.

American Ornithologists' Union (2007) Separate Pelecanus thagus from

P. occidentalis – SACC Proposal 271. NACC Proposals 2007-A, 31–37.

Anderson DW (1988) Dose-response relationship between human disturbance

and Brown Pelican breeding success. Wildlife Society Bulletin 16, 339–

345.

Anderson DW, Gress F, Fry DM (1996) Survival and dispersal of oiled Brown

Pelicans after rehabilitation and release. Marine Pollution Bulletin 32,

711–718.

Anderson DW, Gress F, Mais KF (1982) Brown Pelicans: influence of food supply

on reproduction. Oikos 39, 23–31.

Anderson DW, Gress F, Mais KF, Kelly PR (1980) Brown Pelicans as anchovy

stock indicators and their relationships to commercial fishing. CalCOFI

Report XXI, 54–61.

Anderson DW, Keith JO (1980) The human influence on seabird nesting success:

conservation implications. Biological Conservation 18, 65–80.

Anderson DW, Keith JO, Trapp GR, Gress F, Moreno LA (1989) Introduced small

ground predators in California Brown Pelican colonies. Colonial

Waterbirds 12, 98–103.

Anderson WB, Jurek RM, Keith JO (1977) The status of Brown Pelicans at

Anacapa Island in 1975. Californian Fish and Game 63, 4–10.

Assunção P, de Ponte Machado M, Ramírez AS, Rosales RS, Antunes NT, Poveda

C, De la Fe C, Poveda JB (2007) Prevalence of pathogens in Great White

Pelicans Pelecanus onocrotalus from the Western Cape, South Africa.

Journal of Applied Animal Research 32, 29–32.

Banks RC, Chesser RT, Cicero C, Dunn JL, Kratter AW, Lovette IJ, Rasmussen

PC, Remsen JV, Rising JD, Stotz DF, Winker K (2008) Forty-ninth

supplement to the American Ornithologist's Union check-list of North

American Birds. Auk 125, 758–768.

Barnes KN (2000) The Eskom Red Data Book of birds of South Africa, Lesotho

and Swaziland. BirdLife South Africa, Johannesburg.

Berruti A (1980) Status and review of waterbirds breeding at Lake St Lucia.

Lammergeyer 28, 1–19.

Berruti A (1983) The biomass, energy consumption and breeding of waterbirds

relative to hydrological conditions at Lake St Lucia. Ostrich 54, 65–82.

Berry HH (1975) History of the guano platform on Bird Rock, Walvis Bay, South

West Africa. Bokmakierie 27, 60–64.

Berry HH, Stark HP, van Vuuren AS (1973) White Pelicans Pelecanus

onocrotalus breeding on the Etosha Pan, South West Africa, during 1971.

Madoqua (Ser. 1) 7, 17–31.

��

20

BirdLife International (2008) Pelecanus onocrotalus. In: IUCN 2009. IUCN Red

List of Threatened Species, version 2009.1. www.iucnredlist.org.

Blus LJ (1982) Further interpretation of the relation of organochlorine residues

in Brown Pelican eggs to reproductive success. Environmental Pollution

Series A, Ecological and Biological 28, 15–33.

Boellstorff DE, Anderson DW, Ohlendorf HM, O'Neill EJ (1988) Reproductive

effects of nest-marking studies in an American White Pelican colony.

Colonial Waterbirds 11, 215–219.

Bosch M (1996) The effects of culling on attacks by Yellow-legged Gulls Larus

cachinans upon three species of herons. Colonial Waterbirds 19, 248–252.

Bosch M, Oro D, Cantos F, J, Zabala M (2000) Short-term effects of culling on the

ecology and population dynamics of the Yellow-legged Gull. Journal of

Applied Ecology 37, 369–385.

Botzler RG (1991) Epizootiology of Avian Cholera in wildfowl. Journal of Wildlife

Diseases 27, 367–395.

Bowker MB, Downs CT (2008a) Breeding incidence of the Great White Pelican

Pelecanus onocrotalus and the Pink-backed Pelican P. rufescens in south-

eastern Africa from 1933 to 2005. Ostrich 79, 23–25.

Bowker MB, Downs CT (2008b) Fluctuations in numbers of Great White Pelicans

at Lake St Lucia in response to changing water-levels. Journal of African

Ecology 46, 282–290.

Bowker MB, Downs CT (2008c) The status of habitat of Great White and Pink-

backed pelicans in north-eastern KwaZulu-Natal, South Africa: a review.

African Zoology 43, 90–98.

Brown LH, Urban EK (1969) The breeding biology of the Great White Pelican

Pelecanus onocrotalus roseus at Lake Shala, Ethiopia. Ibis 111, 199–237.

Brown LH, Urban EK, Newman K (eds.) (1982) Family Pelecanidae: Pelicans. In:

The birds of Africa, vol. I, pp. 122–129. Academic Press, London.

Burke VEM, Brown LH (1970) Observations on the breeding of the Pink-backed

Pelican. Ibis 112, 499–512.

Camphuysen CJ (1995) Herring Gull Larus argentatus and Lesser Black-backed

Gull L. fuscus feedign at fishing vessels in the breeding season:

competitive scavenging versus efficient flying. Ardea 83, 365–380.

Casas I, Pozo F (2005) SARS, Avian Influenza, and Human Metapneumovirus

Infection. Enfermedades Infecciosas y Microbiología Clínica 23, 438–448.

Catsadorakis G, Crivelli A (2001) Nesting habitat characteristics and breeding

performance of Dalmatian Pelicans in Lake Mikri Prespa, NW Greece.


Catsadorakis G, Malakou M, Crivelli AJ (1996) The effects of the 1989/1990

drought on the colonial waterbirds nesting at Lake Mikri Prespa, Greece,

with special emphasis on pelicans. Colonial Waterbirds 19, 207–218.

Collar NJ, Andrew P (1988) Birds to watch. The ICBP world checklist of

threatened birds. ICBP Technical Publication, 8. Smithsonian Institute,

Washington D.C.

Cooper J (1976) The status of White Pelicans in the southwestern Cape, first

report. Promerops 124, 3–4.

Cooper J (1977) The status of White Pelicans in the southwestern Cape, 1977

survey. Promerops 128, 1–2.

��

21


survey. Promerops 133, 3–5.


survey. Promerops 138, 6.

Cooper J (1980a) Energetic requirements for maintenance of a captive juvenile

Great White Pelican. Cormorant 8, 17–19.

Cooper J (1980b) Fatal sibling aggression in pelicans – a review. Ostrich 51, 183–

186.

Cracraft J (1985) Monophyly and phylogenetic relationships of the

Pelecaniformes: A numerical cladistic analysis. Auk 102, 834–853.

Crawford RJM (2007) Trends in numbers of three cormorants Phalacrocorax spp.

breeding in South Africa's Western Cape Province. In: Final report of the

BCLME (Benguela Current Large Marine Ecosystem) project on top

predators as biological indicators of ecosystem change in the BCLME (ed.

Kirkman SP), pp. 173–178. Avian Demography Unit, Cape Town.

Crawford RJM, Cooper J, Dyer BM (1995) Conservation of an increasing

population of Great White Pelicans Pelecanus onocrotalus in South

Africa's Western Cape. African Journal of Marine Science 15, 3342.

Crawford RJM, Cooper J, Shelton PA (1981) The breeding population of White

Pelicans Pelecanus onocrotalus at Bird Rock Platform in Walvis Bay,

1949–1978. Fisheries Bulletin of South Africa 15, 67–70.

Crivelli AJ (1994) Why do Great White Pelican chicks die suddenly on Arel

Island, Banc D'Arguin, in Mauritania? Terre et Vie 49, 321–330.

Crivelli AJ, Catsadorakis G, Hatzilacou D, Hulea D, Malakou M, Marinov M,

Mitchev T, Nazirides T, Peja N, Sarigul G, Siki M (1998a) Status and

population development of Great White Pelican Pelecanus onocrotalus and

Dalmatian Pelican P. crispus breeding in the Palearctic. In: Monitoring

and conservation of birds, mammals and sea turtles of the Mediterranean

and Black Seas. Proceedings of the 5th Medmaravis Symposium (eds.

Yésou P, Sultana J), pp. 38–45. Environment Protection Department,

Gozo, Malta.

Crivelli AJ, Catsadorakis G, Jerrentrup H, Hatzilacou D, Michev T (1991a)

Conservation and management of pelicans nesting in the Palearctic. ICBP

Technical Publication 12, 137–152.

Crivelli AJ, Catsadorakis G, Nazirides T (1997b) Pelecanus onocrotalus White

Pelican. In: Handbook of the birds of Europe, the Middle East and North

Africa. The birds of the Western Palearctic update (eds. Cramp S, Simmons

KEL), vol. 1, pp. 144–148. Oxford University Press, Oxford.

Crivelli AJ, Focardi S, Fossi C, Leonzio C, Massi A, Renzoni A (1989) Trace

elements and chlorinated hydrocarbons in eggs of Pelecanus crispus, a

world endangered bird species nesting at Lake Mikri Prespa, North-

Western Greece. Environmental Pollution 61, 235–247.

Crivelli AJ, Hatzilacou D, Catsadorakis G (1997a) The breeding biology of the

Dalmatian Pelican Pelecanus crispus. Ibis 140, 472–481.

Crivelli AJ, Hatzilacou D, Catsadorakis G (1998b) The breeding biology of the


��

22

Crivelli AJ, Jerrentrup H, Mitchev T (1988) Electric power lines: a cause of

mortality in Pelecanus crispus Bruch, a world endangered bird species, in

Porto-Lago, Greece. Colonial Waterbirds 11, 301–305.

Crivelli AJ, Leshem Y, Mitchev T, Jerrentrup H (1991b) Where do Palaearctic

Great White Pelicans Pelecanus onocrotalus presently overwinter? Terre et

Vie 46, 145–171.

Crivelli AJ, Schreiber RW (1984) Status of the Pelecanidae. Biological

Conservation 30, 147–156.

Crivelli AJ, Vizi O (1981) The Dalmatian Pelican Pelecanus crispus Bruch 1832,

a recently world-endangered bird species. Biological Conservation 20,

297–310.

de Ponte Machado M (2007) Is predation on seabirds a new foraging behaviour

for Great White Pelicans? History, foraging strategies and prey defensive

responses. In: Final report of the BCLME (Benguela Current Large Marine

Ecosystem) project on top predators as biological indicators of ecosystem

change in the BCLME (ed. Kirkman SP), pp. 131–142. Avian Demography

Unit, Cape Town.

de Ponte Machado M, Feldheim KA, Sellas AB, Bowie RCK (2009) Development

and characterization of microsatellite loci from the Great White Pelican

Pelecanus onocrotalus and widespread application to other members of the

Pelecanidae. Conservation Genetics 10, 1033–1036.

de Ponte Machado M, Hofmeyr J (2004) Great White Pelicans Pelecanus

onocrotalus: waterbirds or farm birds? Bird Numbers 13, 11–13.

Dean WRJ, Dowsett RJ, Sakko A, Simmons RE (2002) New records and

amendments to the birds of Angola. Bulletin of the British Ornithological

Club 122, 180–185.

del Hoyo J, Elliott A, Sargatal J (eds.) (1992) Family Pelecanidae (Pelicans). In:

Handbook of the birds of the world, vol. 1. Lynx Edicions, Barcelona.

Din NA, Eltringham SK (1974) Ecological separation between White and Pink-

backed Pelicans in the Ruwenzori National Park, Uganda. Ibis 116, 28–43.

Dodman T (in press) Status, estimates and trends of waterbird populations in

Africa. Wetlands International, Dakar.

Dorst J, Mougin JL (1979) Order Pelecaniformes. In: Check list of birds of the

world, vol. 1, pp. 155–193. Museum of Comparative Zoology, Cambridge.

du Toit M, Boere GC, Cooper J, de Villiers MS, Kemper J, Lenten B, Petersen

SL, Simmons RE, Underhill LG, Whittington PA, Byers OP (2003)

Conservation assessment and management plan for southern African

coastal seabirds. Avian Demography Unit & Apple Valley: Conservation

Breeding specialist group, Cape Town.

Duffy DC (1983) Environmental uncertainty and commercial fishing: effects on

Peruvian guano birds. Biological Conservation 26, 227–238.

Elliot DG (1869) A monograph of the genus Pelecanus. Proceedings Zoological

Society of London, 571–591.

Ericson PGP, Anderson CL, Britton T, Elzanowski A, Johansson US, Källersjö

M, Ohlson JI, Parsons TJ, Zuccon D, Mayr G (2006) Diversification of

Neoaves: integration of molecular sequence data and fossils. Biology

Letters 2, 543–547.

��

23

Finney S, K, Harris M, P, Keller L, F, Elston D, A, Monaghan P, Wanless S

(2003) Reducing the density of breeding gulls influences the pattern of

recruitment on immature Atlantic puffins Fratercula arctica to a breeding

colony. Journal of Applied Ecology 40, 545–552.

Furness RW (1981) The impact of predation by Great Skuas Catharacta skua on

other seabird populations at a Shetland colony. Ibis 123, 534–539.

Furness RW (2003) Impacts of fisheries on seabird communities. Scientia Marina

67, 33–45.

Furness RW, Monaghan P (1987) Seabird ecology. Blackie, Glasgow and London.

Furness RW, Monaghan P, Sheddon C (1981) Exploitation of a new food source

by the Great Skua in Shetland. Bird Study 28, 49–52.

Gee EP (1960) The breeding of the Gray or Spot-billed Pelican Pelecanus

philippensis. Journal of Bombay Natural History Society 57, 245–251.

Gress F (1995) Organochlorines, eggshell thinning and productivity relationships

in Brown Pelicans breeding in the Southern California Bight. University of

California, Davis.

Guillemette M, Brousseau P (2001) Does culling predatory gulls enhance the

productivity of breeding Common Terns? Journal of Applied

Ecology 38, 1–8.

Guillet A (1985) Biogeography and Ecology of African waterbirds. Doctoral

thesis, University of Cape Town.

Guillet A, Crowe TM (1981) Seasonal variation in group size and dispersion in a

population of Great White Pelicans. Gerfaut 71, 185–194.

Guillet A, Crowe TM (1983) Temporal variation in breeding, foraging and bird

sanctuary visitation by a Southern African population of Great White

Pelicans Pelecanus onocrotalus. Biological Conservation 26, 15–31.

Guillet A, Furness RW (1985) Energy requirements of a Great White Pelican

Pelecanus onocrotalus population and its impact on fish stocks. Journal of

Zoology 205, 573–583.

Hackett SJ, Kimball RT, Reddy S, Bowie RCK, Braun EL, Braun MJ,

Chojnowski JL, Cox WA, Han K-L, Harshman J, Hoddleston CJ, Marks

BD, Miglia KJ, Moore WS, Sheldon FH, Steadman DW, Witt CC, Yuri T

(2008) A phylogenomic study of birds reveals their evolutionary history.

Science 30, 1763–1768.

Hanski I, Gilpin ME (1991) Metapopulation dynamics – brief history and

conceptual domain. Biological Journal of the Linnean Society 42, 3–16.

Harris MP (1970) Rates and causes of increases of some British gull populations.

Bird Study 17, 325–335.

Hockey PAR, Dean WRJ, Ryan PG (2005) Roberts birds of southern Africa. 7th

edn. Trustees of the John Voelcker Bird Book Fund, Cape Town.

Hull TG (1963) Diseases transmitted from animals to man. 6th edn. C.C.

Thomas, Springfield, IL.

IUCN (2009) IUCN Red List of Threatened Species, version 2009.1.

www.iucnredlist.org.

Izhaki I (1994) Preliminary data on the importance of Israel for the conservation

of the White Pelican Pelecanus onocrotalus L. Ostrich 65, 213–217.

Izhaki I, Shmueli M, Arad Z, Steinberg Y, Crivelli A (2002) Satellite tracking of

migratory and ranging behavior of immature Great White Pelicans.

��

24


Johnsgard PA (1993) Cormorants, darters and pelicans of the world. Smithsonian

Institution Press, Washington and London.

Johnson JRF, Sloan NF (1976) The effects of human disturbance on the White

Pelican colony at Chase Lake National Wildlife Refuge, North Dakota.

Inland Bird Banding News 48, 163–170.

Keith JO (2005) An overview of the American White Pelican. Waterbirds 28, 9–

17.

Kemper J, Underhill LG, Crawford RJM, Kirkman SP (2007) Revision of the

conservation status of seabirds and seals breeding in the Benguela

Ecosystem. In: Final report of the BCLME (Benguela Current Large

Marine Ecosystem) project on top predators as biological indicators of

ecosystem change in the BCLME (ed. Kirkman SP), pp. 325–342. Avian

Demography Unit, Cape Town.

King DT, Michot TC (2002) Distribution, abundance and habitat use of American

White Pelicans in the delta region of Mississippi and along the western

Gulf of Mexico coast. Waterbirds 25, 410–416.

Knopf FL, Evans RM (2004) American White Pelican Pelecanus erythrorhynchos.

In: The birds of North America online (ed. Poole A), Cornell Lab of

Ornithology, Ithaca. http://bna.birds.cornell.edu/bna/species/057.

Lande R (1988) Genetics and demography in biological conservation. Science 241,

1455–1460.

Lanham UN (1947) Notes on the phylogeny of the Pelecaniformes. Auk 64, 65–

70.

Makarov VV, Vorob'ev AA, Bondarenko VM, Boev BV (2005) Highly pathogenic

avian influenza virus inducing influenza pneumonia in humans. Zhurnal

Mikrobiologii, Epidemiologii i Immunobiologii 3, 105–109.

Marchant S, Higgins PJ (1990) Handbook of Australian, New Zealand and

Antarctic birds. Oxford University Press, Melbourne.

Martínez-Abraín A, González-Solís J, Pedrochi V, Genovart M, Abella J, Carles,

Ruiz X, Jiménez J, Oro D (2003) Kleptoparasitism, disturbance and

predation of Yellow-legged Gulls on Audouin's Gulls in three colonies of

the western Mediterranean. Scientia Marina 67, 89–94.

McCulloch G (2008) Going the distance: the great flight of the pelican. Africa–

Birds and Birding 13, 37–41.

McNeil R, Rodriguez S, Ouellet H (1985) Bird mortality at a power transmission

line in northeastern Venezuela. Biological Conservation 31, 153–165.

Murphy EC (2005) Biology and conservation of the American White Pelican:

current status. Waterbirds 28, 107–112.

Nagulu V (1984) The feeding and breeding biology of the Spot-billed Pelican

Pelecanus phillippensis. Doctoral thesis, Osmania University.

Nagulu V, Rao JVR (1983) Survey of south Indian pelicanries. Journal of

Bombay Natural History Society 80, 141–143.

Neeklakantan KK (1949) A south Indian pelicanry. Journal of Bombay Natural

History Society 48, 656–666.

Nelson JB (2005) Pelicans, cormorants, and their relatives: the Pelecaniformes.

Oxford University Press, Oxford.

��

25

Neves VC, Panagiotakopoulos S, Furness RW (2006) A control taste aversion

experiment on predators of Roseate Tern Sterna dougallii eggs. European

Journal Wildlife Research 52, 259–264.

Ogilvie-Grant WR (1989) Steganopodes. In: Catalogue of the birds in the British

Museum, vol. 26, pp. 329–484. British Museum of Natural History,

London.

Oro D, de León A, Mínguez E, Furness RW (2005) Estimating predation on

breeding European Storm-petrels by Yellow-legged Gulls. Journal of

Zoology (London) 265, 421–429.

Oro D, Furness RW (2002) Influences of food availability and predation on

survival of Kittiwakes. Ecology 83, 2516–2528.

Oro D, Martínez-Abraín A (2007) Deconstructing myths on large gulls and their

impact on threatened sympatric seabirds. Animal Conservation 10, 117–

126.

Oro D, Pérez-Rodríguez A, Martínez-Vilalta A, Bertolero A, Vidal F, Genovart M

(2009) Interference competition in a threatened seabird community: A

paradox for a successful conservation. Biological Conservation 142, 1830–

1835.

Peja N, Sarigul G, Siki M, Crivelli A (1996) The Dalmatian Pelican, Pelecanus

crispus, nesting in Mediterranean Lagoons in Albania and Turkey.

Colonial Waterbirds 19, 184–189.

Peters JP (1931) Check-list of birds of the world. Harvard University Press,

Cambridge.

Pons J-M, Migot P (1995) Life-history strategy of the Herring Gull: changes in

survival and fecundity in a population subjected to various feeding

conditions. Journal of Animal Ecology 64, 592–599.

Pullin AS, Knight TM (2001) Effectiveness in conservation practice: pointers

from medicine and public health. Conservation Biology 15, 50–54.

Pullin AS, Knight TM (2003) Support for decision making in conservation

practice: an evidence-based approach. Journal for Nature Conservation 11,

83–90.

Pullin AS, Knight TM (2005) Assessing conservation management's evidence

base: a survey of management-plan compilers in the United Kingdom and

Australia. Conservation Biology 19, 1989–1996.

Pullin AS, Knight TM, Stone DA, Charman K (2004) Do conservation managers

use scientific evidence to support their decision-making? Biological


Pyrovetsi M (1997) Integrated management to create new breeding habitat for

Dalmatian Pelicans Pelecanus crispus in Greece. Environmental

Management 21, 657–667.

Pyrovetsi M, Papazahariadou M (1995) Mortality factors of Dalmatian Pelicans

Pelecanus crispus wintering in Macedonia, Greece. Environmental


Pyrovetsi MD, Economidis PS (1998) The diet of Dalmatian Pelicans Pelecanus

crispus breeding at Lake Mikri Prespa National Park, Greece. Israel

Journal of Zoology 44, 9–17.

��

26

Rand RW (1963) The biology of guano producing seabirds. 4. Composition of

colonies on the Cape islands. Investigational report, Department of

Commerce and Industry, Division of Fisheries of South Africa 43: 1–32.

Reed KD, Meece JK, Henkel JS, Shukla SK (2003) Birds, migration and

emerging zoonoses: West Nile Virus, Lyme Disease, Influenza A and

Enteropathogens. Clinical Medicine and Research 1, 5–12.

Regehr HM, Montevecchi WA (1997) Interactive effects of food shortage and

predation on breeding failure of Black-legged Kittiwakes: indirect effects

of fisheries activities and implications for indicator species. Marine

Ecology Progress Series 155, 249–260.

Ridgely RS, Greenfield PJ (2001) The birds of Ecuador. Cornell University Press,

Ithaca.

Sanz-Aguilar A, Martínez-Abraín A, Tavecchia G, Mínguez E, Oro D (2009)

Evidence-based culling of a facultative predator: efficacy and efficiency

components. Biological Conservation 142, 424–431.

Schreiber RW, Mock PJ (1988) Eastern Brown Pelicans: what does 60 years of

banding tell us? Journal of Field Ornithology 59, 171–182.

Shields M (2002) Brown Pelican Pelecanus occidentalis. In: The Birds of North

America online (ed. Poole A). Cornell Lab of Ornithology, Ithaca.

http://bna.birds.cornell.edu/bna/species/609.

Shmueli M, Arad Z, Izhaki I, Crivelli A (2000a) The energetic demand of the

Great White Pelican during migration in Israel and its implication to the

conflict between fish industry and pelicans. In: Proceedings of the 9th

Pan-African Ornithological Congress, Accra, Ghana, 1996 (eds. Craig

AJFK, Gordon C). Ostrich, 71 (1 & 2).

Shmueli M, Izhaki I, Arieli A, Arad Z (2000b) Energy requirements of migrating

Great White Pelicans Pelecanus onocrotalus. Ibis 142, 208–216.

Shmueli M, Izhaki I, Zinder O, Arad Z (2000c) The physiological state of captive

and migrating Great White Pelicans Pelecanus onocrotalus revealed by

their blood chemistry. Comparative Biochemistry and Physiology. Part A,

Molecular & Integrative Physiology 125, 25–32.

Sibley CG, Ahlquist J (1990) Phylogeny and classification of birds. A study in

molecular evolution. Yale University Press, New Haven.

Sibley CG, Ahlquist J, Monroe BLJ (1988) A classification of the living birds of

the world, based on DNA-DNA hybridisation studies. Auk 105, 409–423.

Simeone A, Luna-Jorquera G, Bernal M, Garthe S, Sepúlveda F, Villablanca R,

Ellenberg U, Contreras M, Muñoz J, Ponce T (2003) Breeding distribution

and abundance of seabirds on islands off north-central Chile. Revista

Chilena de Historia Natural 76, 323–333.

Stephens PA, Sutherland WJ (1999) Consequences of the Allee effect for

behaviour, ecology and conservation. Trends in Ecology & Evolution 14,

401-405.

Sutherland WJ (1998) The importance of behavioural studies in conservation

biology. Animal Behaviour 56, 801–809.

Sutherland WJ (2006) Predicting the ecological consequences of environmental

change: a review of methods. Journal of Applied Ecology 43, 599–616.

Sutherland WJ, Pullin AS, Dolman PM, Knight TM (2004) The need for

evidence-based conservation. Trends in Ecology & Evolution 19, 305–308.

��

27

Theron L, Cunningham P, Simataa E (2003) Notes on the breeding of White

Pelicans at Hardap Dam, Namibia. Lanioturdus 36, 2–7.

Underhill LG, Crawford RJM, Wolfaardt AC, Whittington PA, Dyer BM, Leshoro

TM, Ruthenberg M, Upfold L, Visage J (2006) Regionally coherent trends

in colonies of African Penguins Spheniscus demersus in the Western Cape,

South Africa, 1987–2005. African Journal of Marine Science 28, 1–8.

United States Geological Survey (2009). Society, wildlife disease and wildlife

conservation: oxymoron or evolutionary siblings? ScienceDaily. August 27,

2009, http://www.sciencedaily.com/releases/2009/08/090803210021.htm.

Urban EK (1984) Time of egg-laying and number of nesting Great White

Pelicans at Lake Shala, Ethiopia, and elsewhere in Africa. In: Proceedings

of the 5th Pan-African Ornithological Congress, Lolongwe, Malawi, 1980

(ed. Ledger L), pp. 809–823. Southern African Ornithological Society,

Johannesburg.

Urban EK, Ash JS (2001) Longevity record of a Great White Pelican, Pelecanus

onocrotalus, from Lake Shala, Ethiopia. Ostrich 72, 114–126.

van Halewyin R, Norton RL (1984) The status and conservation of seabirds in

the Caribbean. In: Status and conservation of the world's seabirds (eds.

Croxall JP, Evans PGH, Schreiber RW), ICBP Technical Publication no. 2.

Vestjens WJM (1977) Breeding behaviour and ecology of the Australian Pelican,

Pelecanus conspicillatus, in New South Wales. Australian Wildlife

Research 4, 37–58.

Vidal E, Medail F, Tatoni T (1998) Is the Yellow-legged Gull a superabundant

bird species in the Mediterranean? Impact on fauna and flora,

conservation measures and research priorities. Biodiversity and


Votier SC, Furness RW, Bearhop S, Crane JE, Caldow RWG, Catry P, Enso K,

Hamer KC, V. HA, Kalmbach E, Klomp NI, Pfeiffer S, Phillips RA, Prieto

I, Thompson DR (2004) Changes in fisheries discard rates and seabird

communities. Nature 427, 727–730.

Wetlands International (2006) Waterbird population estimates – Fourth edition.

Wetlands International, Wareningoen.

Wetmore A (1945) A review of the forms of the Brown Pelican. Auk 62, 577–586.

Whittington PA, Wolfaardt A (1999) The avifauna of Dassen Island. Bird

Numbers 8, 8–11.

Williams AJ (in press) Birds at Bird Rock Guano Platform, Walvis Bay, Namibia.

Ostrich.

Williams AJ, Borello WD (1997) White Pelican Pelecanus onocrotalus. In: The

atlas of southern African birds (eds. Harrison JA, Allan DG, Underhill LG,

Herremans M, Tree AJ, Parker V, Brown CJ), vol. 1: non-passerines, pp.

24–25. BirdLife South Africa, Johannesburg.

Williams AJ, Randall RM (1995) Pelecaniform birds in South African wetlands.

In: Wetlands of South Africa (ed. Cowan GI), pp. 147–161. Department of

Environmental Affairs and Tourism, Pretoria.

Wobeser GA (1997) Diseases of wild waterfowl. Plenum Press, New York.

Wolfaardt A (2000) Dassen Island Management Plan. Western Cape Nature

Conservation Board, Yzerfontein.

�

CHAPTER 2

��

��

��

��

29

��

��

��

��

��

Unlike many pelican populations, the Western Cape population of Great White

Pelicans Pelecanus onocrotalus increased exponentially, from 30 breeding pairs

in 1930–1950 to 659 in 2001. Since 1956, Dassen and Vondeling Islands have

been the only breeding localities for pelicans in the Western Cape. Ringing data

indicated that immigration from other colonies is negligible. The observed

population growth until the early 1980s has been explained by an increase in

breeding area on Dassen Island and the cessation of disturbance during the

breeding season. From the mid-1980s, pelicans fed mostly on surplus animal

offal, available at local farming sites. The number of pelicans in the Western

Cape increased four fold from a maximum count of 652 in 1976–1985 to 2 723 in

2004–2007. Regular counts undertaken at one pig farm from 2000 to 2009

showed that this farm became the main pelican foraging locality between April

2003 and February 2005 attracting up to 2 723 pelicans. Chick production from

1998 to 2007 was 222 ± 125 (mean ± SD) chicks/year at an average of 0.42

chicks/pair. The maximum number of breeding pairs on Dassen Island (659 in

2001) did not correspond with the largest counts at the monitored pig farm, but

other farms operating in the region also supplied food to pelicans. A decline in

the availability of food at the pig farm starting in October 2004 caused the death

of > 200 pelicans, reduced pelican numbers at this site and coincided with a 40%

decline in the number of breeding pairs in the two following seasons and a

diminished breeding success to 0.12 chicks/pair. The proportion of adults

breeding was 57% in 2004 and averaged 34% during 2004–2008. The relatively

small pelican population size before subsidisation may indicate that the biomass

��

30

and availability of fresh-water fish and amphibians in the Western Cape is

insufficient to maintain a large breeding population of pelicans. Food subsidies at

farms apparently allowed the population of pelicans to expand beyond the

natural carrying capacity. Changes in the availability of surplus food have

altered prey selection by pelicans, resulting in increased predation by pelicans on

other seabirds, a rare behaviour for the species and unique in its extent to the

Western Cape.

��

Complexity is the key feature of food webs (Polis 1994). Omnivory and trophic

promiscuity are common in nature, creating reticulated connections that affect

web dynamics (Polis & Strong 1996; Hunter 2009). Although the frequency and

ubiquity of trophic cascades is a subject for debate, it has been shown that these

can be enhanced by nutrient enrichment (Pace et al. 1999; Polis et al. 2000). One

of the mechanisms is by subsidisation of opportunistic species, resulting in

changes in the population dynamics of the subsidised species (Huxel & McCann

1998; Rose & Polis 1998; Pace et al. 1999). Prey subsidies can enable consumer

populations to achieve greater numbers than those supported by in situ

productivity. Top-down effects occur when spatially subsidised consumers affect

local resources by preying on key species and occasionally by initiating trophic

cascades (Polis & Hurd 1996; Polis et al. 1997; 2000).

Among the examples of nutrient additions are agricultural subsidies to bird

populations (Jefferies 2000). Occasionally, the capacity of these species to benefit

from surplus food results in large population increases that cause them to attain

pest status, disrupting the normal functioning of their communities and leading

to depletion of prey populations (Jefferies 2000). Among seabirds, large gulls

Larus spp., skuas Catharacta spp. and giant petrels Macronectes spp. are often

considered ‘problem’ species. Their population increases, linked to the

availability of fisheries discards or refuse tips, can affect the population

dynamics of more specialised and vulnerable species, occasionally constituting a

serious threat to their conservation (Furness 1981; Blokpoel & Spaans 1991;

��

31

Hunter & Brooke 1992; Pons & Migot 1995; Regehr & Montevecchi 1997; Vidal et

al. 1998; Weidinger 1998; Yorio et al. 1998; Bosch et al. 2000; Hernández-Matías

& Ruiz 2003; Votier et al. 2004; Oro & Martínez-Abraín 2007). In addition,

intensified predation on smaller seabird species is expected when food subsidies

are suddenly removed (Oro & Furness 2002; Votier et al. 2004; Furness 2007).

The Great White Pelican Pelecanus onocrotalus breeds in Europe and Africa (del

Hoyo et al. 1992). Although not globally threatened (IUCN 2008), Great White

Pelicans have experienced population declines in some of their areas of

distribution (Crivelli et al. 1998). In South Africa they are listed as Near-

threatened due to the small number of breeding colonies (Barnes 2000). They

breed at three coastal localities in southern Africa: Bird Rock Guano Platform in

Walvis Bay (Namibia), Dassen and Vondeling Islands in the Western Cape, and

St Lucia Wetland Park in KwaZulu-Natal (South Africa) (Hockey et al. 2005).

Additionally, when rains have been abundant, they breed at some inland sites:

Hardap Dam (Namibia), and less frequently at Etosha Pan (Namibia), and Sua

Pan and Lake Ngami (Botswana) (Berry et al. 1973; Berruti 1980; Crawford et al.

1981; Williams & Borello 1997; Theron et al. 2003).

The pelican population in the Western Cape has increased considerably. Records

of breeding pairs have been collected sporadically from 1930 (summarized by

Crawford et al. 1995). Until 1956, pelican numbers fluctuated between 20 and 30

pairs. The pelicans moved their breeding colony to Dassen Island (Figure 2.1) in

1956, reaching 174 pairs by 1978. This population expansion resulted mainly

from reduced disturbance during the breeding season, the larger breeding area

available on Dassen Island in comparison to previous breeding localities and an

increase of suitable habitat throughout the expanding agricultural areas, due to

the construction of farm reservoirs and dredging of seasonal wetlands to make

permanent waterbodies (Crawford et al. 1995). Pelicans were first detected

eating agricultural offal in the early 1990s by the discovery of chicken Gallus

gallus remains in pelican chick regurgitations (Crawford et al. 1995). From the

mid-1980s, pelicans started frequenting free-range pig farms in the agricultural

��

32

areas around Cape Town, feeding on fish and chicken offal dumped in open areas

for the pigs (Stanley and George Fisk, pers. comms). This large, predictable food

resource may explain the subsequent rapid growth of the pelican population.

Breeding numbers increased to 300 pairs by 1992, and reached 500 pairs by 1995

(Crawford et al. 1995).

Across their range of distribution, Great White Pelicans feed on fish and other

aquatic organisms (del Hoyo et al. 1992; Johnsgard 1993). Due to the physical

characteristics of the Western Cape environment, consisting of small and

sometimes ephemeral wetlands, pelicans in the Western Cape relied on a limited

amount of fresh-water fish and amphibians for food (Guillet & Crowe 1981).

Consequently, artificial food sources became of great importance as alternative

foraging sites for pelicans and food supplementation presumably increased

reproductive success and the number of breeding pairs in the region. This study

reports the population dynamics of pelicans in the Western Cape, with emphasis

on the role of food subsidisation on their population expansion. Evidence from

breeding censuses, diet samples at the breeding colony, annual chick production,

counts at the pig farm and at the main water bodies and ringing recoveries are

compared to data collected prior to food supplementation in the 1970s–80s to

identify the mechanisms underpinning the population increase. In addition, I

will discuss succinctly the response of the pelican population to the diminishing

availability of surplus food, which caused a shift in pelican diet preferences, and

introduced a conservation dilemma due to the increased predation rates on

seabird chicks on the offshore islands. This novel behaviour has the potential to

become a serious threat to seabird communities and to bring about changes in

local trophic webs.

��

33

��

��

Numbers of pelican breeding pairs on Dassen Island (33°25’ S, 18°05’ E, Figure

2.1) were counted sporadically until 1988 and then monitored annually during

the austral spring and summer between 1995 and 2007. Several counts were

made each year to sample the peak of breeding effort and to estimate chick

production. Counts of incubating and brooding birds were made using binoculars

to limit disturbance. Larger pelican chicks (28–42 days old) aggregate in crèches

until they start practising flight at around 70–75 days old (Brown et al. 1982).

Larger chicks were considered to represent one breeding pairs because as a

result of siblicide just one chick survives (Cooper 1980). Vondeling Island (33°09’

S, 17°59’ E, Figure 2.1) was also surveyed annually for breeding pelicans. In the

Western Cape, pelicans often breed in two or three overlapping ‘waves’ (Guillet &

Crowe 1983), but the maximum count of breeding pairs in a breeding season was

taken as the number of breeding pairs in the year. Annual chick production was

calculated as the sum of the numbers of fledglings produced in each sub-colony

throughout the breeding season. Historical data were extracted from the

databases of Marine and Coastal Management and CapeNature, as well as from

work published by Crawford et al. (1995).

To incorporate natural fluctuations in the number of breeding pairs as well as

smooth the effect of bias due to the different times of surveys of the colonies, a

moving percentage rate of increase (or decrease) of the breeding population over

the period 1988–2007 was obtained by fitting a weighted linear regression to the

estimates of numbers breeding for each year. For this study, the degree of

smoothing of the curve was defined as = 2.5 years (see Underhill et al. 2006 for

further details of the technique and a justification of this choice).

A Leslie matrix deterministic population model (Caswell 2001) was used to

estimate the likely maximum growth rate of the pelican population in the

Western Cape. The model parameters were an initial population size of 75

��

34

individuals in 1955, four years as age at first breeding, 0.96 annual survival rate

for adults and 0.74 for immatures (Chapter 3). Additional parameters, i.e. the

proportion of the adults breeding each year and the average number of fledglings

per pair per year, were calculated in this study.

��

��

The objective of the Coordinated Waterbird Counts (CWAC) project is to monitor

South Africa's waterbird populations and the conditions of the wetlands that are

important for waterbirds (Taylor et al. 1999). A minimum of one mid-summer

and one mid-winter census are carried out at 400 wetlands in South Africa, of

which 70 are in the Western Cape (Coordinated Waterbird Counts, Animal

Demography Unit, University of Cape Town). All sites frequented by pelicans

were included in this analysis (Figure 2.1). Additionally, pelican counts at the

Orange River estuary, Northern Cape, were included because of the large

number of pelicans occurring there (Anderson et al. 2003) and their suspected

affiliation to the Western Cape population. For the purposes of this study, CWAC

counts provide a long data-series (1992–present) to assess the relative

importance of the different wetlands and foraging sites to the pelicans in the

Western and Northern Cape.

��

Eight quarterly censuses were undertaken in October 2004, April, July and

October 2006, and January, April, July and November 2007. A combination of

designated ground counters and aerial surveys covered all localities regularly

frequented by pelicans in the Western Cape (Figure 2.1). Counts were made

shortly after sunrise (05h00–07h00 in summer and 07h00–09h00 in winter) to

avoid double counting of pelicans once they disperse from night roosts aided by

the onset of the thermals. Seven aerial surveys were undertaken to cover areas

inaccessible to ground counters and to check for accuracy and bias at sites where

both aerial and ground counts were carried out (Hone 2008). The same aircraft

��

35

(Moony M-20M ‘Bravo’), pilot and two observers were used in all the aerial

surveys. Observers sat on opposite sides of the plane, one in the back seat and

the other in the front. Flight time was ca. 4 hours, from sunrise until the

planned route had been completed. Agricultural areas were divided into sectors

and the aircraft zig-zagged between easily identifiable landmarks, while both

observers searched for pelicans on farm dams, rivers and marshlands. Once

detected, the aircraft circled above a group to allow counting or photography.

Small groups (<10) were counted, whereas digital photographs were taken of

larger groups using a NIKON D70 camera and an AF-D VR 80–400mm lens. The

positions of pelican groups were marked with a GARMIN QUEST GPS. The camera

and GPS were synchronised to assign the photographs to the exact geographical

location. Numbers of pelicans were subsequently counted from the aerial

photographs and geo-referenced. MAP SOURCE v3.06 (GARMIN Ltd) and digital

maps were used to visualize the aircraft track. The routes taken during flights

were analysed to check the degree of coverage of the region and whether any

important wetlands had been missed. In July 2006, bad weather prevented the

scheduled aerial survey. Consequently, an exhaustive two-day road survey was

conducted throughout the agricultural areas, as well as to distant sites not

covered by the regular ground counters.

Because of pelicans’ tendency of aggregating in dense groups and ground-level

obstacles to the visibility of all groups of pelicans, ground counts are generally

less accurate than aerial counts (Sutherland et al. 2004; Rodgers et al. 2005).

Therefore counts based on aerial photographs were used when available. In

aerial surveys, many factors, for example aircraft design, observer fatigue and

the behaviour of the study animal, affect the reliability with which observers

detect animals (Fleming & Tracey 2008; Southwell et al. 2008). Negative bias is

especially a problem when group size is small (less than three individuals).

Double-sampling methods reduce bias (Southwell et al. 2008). In this study, two

observers detected the target animals, combining all observations. Other factors

affecting detectability are habitat type and seasonal changes in the habitat and

behaviour of the target animals (Choquenot et al. 2008), so different degrees of

��

36

bias were expected in the pelican surveys undertaken at different times of the

year. Although pelicans did not appear to react to the flying aircraft, different

backgrounds and habitats did influence their visibility from the air.

A ranking for expected accuracy was assigned to each survey (1 to 5), taking into

account (1) the visibility and level of coverage of the region and (2) whether the

aerial and ground counts happened synchronously. All aerial and ground counts

occurred simultaneously except in October 2004, when the aerial count was done

in the afternoon of the same day due to poor weather; July 2006 (no aerial

survey); and November 2007, when the ground and aerial counts were

undertaken two weeks apart (4 and 17 November). For the November 2007

survey the results of the aerial and ground counts were combined by taking the

highest count for each site. Visibility was generally good and pelicans are easy to

detect due to their characteristic shape, white plumage, and their freshwater

habitat preference. Thick fog along the coast during the January 2007 survey

meant that only ground counts were available for Langebaan Lagoon and the

Berg River estuary, thus possibly introducing a negative bias. Also, it was easier

to locate pelicans during the dry season, when they aggregate at a small number

of sites. Counts in July 2006 (only ground count), October 2006 and July 2007

(when pelicans were highly dispersed following heavy rains) were negatively

biased.

To investigate trends between counts in the abundance of pelicans from 2004 to

2007 I used TRIM v3, a software package that estimates population trends using

a log-linear regression model, taking into account missing data (Pannekoek &

van Strien 1996). I used the software’s default parameters for the linear model

and included serial correlation corrections but did not consider overdispersion

because the data attempt to be a total census of the species in the region. The

final model examined the counts at the pig farm separately (as a covariate)

because of the marked changes in pelican numbers that occurred at this site.

��

37

��

Pelicans were counted at a pig farm between the towns of Stellenbosch and Paarl

in the Western Cape, South Africa (33°49’ S, 18°47’ E, Figure 2.1) from December

2000 to January 2009. The farm was divided into sectors and counts were

conducted by the same person (Jan Hofmeyr) for consistency. Pelican movements

during counts were taken into consideration to avoid double counts. When tight

pelican groups made counting individuals difficult, estimates were made with

the final count being the average of at least three counts. Ground counts were

done at least once or twice per month, often with a frequency of two or three

times a week, the average for each month being reported. The amount of food

accessible to the pelicans was estimated, either by observation or by discussion

with the farm managers. An average of 4 tons of offal was delivered daily from

1985 to 1990. By 2004, this quantity had increased to 20–24 tons of fresh offal,

which was brought to the farm at least five days a week. In the first half of 2005

the supply was cut in half, and from July 2005, there was an almost complete

cessation in the offal supply to the farm. This, together with the construction of

an in situ offal processing plant, meant that there was little (or no) food available

for the pelicans at this time. From the end of 2005 the supply of offal resumed,

although most of the offal was utilised on the farm and only the leftovers of the

processed food were made available to the pelicans (a maximum of

30 tons/month). From November 2006, a supply of unhatched eggs and dead

embryos replaced the chicken offal. The quantity and availability of offal during

2006–2009 fluctuated widely. From October 2008 fresh poultry offal was made

available to the birds, but the supply was discontinued from March 2009.

To evaluate the importance of agricultural offal for Great White Pelicans, diet

samples were obtained opportunistically during chick ringing trips on Dassen

Island in 2000, 2002, 2005 and 2007. These consisted of regurgitations from

large pelican chicks and were kept frozen for subsequent analysis. Food items

were identified to species level and weighed separately to the nearest 0.1 g. Diet

was expressed as percentage mass and frequency of occurrence for each prey

type.

��

38

��

A total of 728 pre-fledging chicks were captured and ringed between 2001 and

2008 on Dassen Island. An additional 404 chicks were ringed at Hardap Dam in

Namibia (24°30’ S, 17°50’ E, Figure 2.1) between May 2004 and May 2008. On

each capture attempt, a temporary enclosure was erected adjacent to the

breeding colony. Pelican chicks were herded towards the enclosure by a

coordinated team of people. Time in the enclosure was minimised. Once ringed

and measured, the pelican chicks were released near their nesting site. South

African Bird Ringing Unit (SAFRING) numbered 34 x 19 mm (internal diameter)

stainless-steel rings and a yearly-coded colour plastic ring on the opposite leg

were used from 2001 to 2003 on Dassen Island. From 2004 on Dassen Island and

2006 in Namibia, both metal and alphanumerically encoded plastic colour-rings

(Pro-Touch, Saskatoon, Canada) were fitted to enable individual identification of

pelicans in the field.

Adult pelicans were trapped using modified leg-traps (Victor # 3 Softcatch), set

on the ground near roosting or feeding sites. Modifications consisted of weaker

springs (# 1.5 coil springs) and extra padded jaws to avoid harming the birds

(King et al. 1998). Traps were camouflaged and anchored to the ground using

aircraft cable and elastic shock-cord. Immediately after the birds were caught,

they were recovered and transported to the ringing station. Both SAFRING

metal rings and engraved alphanumerical colour-rings were fitted to these birds.

Between 1970 and 2009, 1 370 pelicans were ringed in southern Africa. Besides

this study, private ringers, conservation authorities and the Southern African

Foundation for the Conservation of Coastal Birds (SANCCOB) have used

SAFRING metal rings on pelicans. The SAFRING database was queried for

records on ringed pelicans in the region. A total of 1 269 resightings and 184

recoveries of ringed pelicans was analysed for movements in and out of their

natal colonies.

��

39

��

��

Dassen and Vondeling Islands are the only breeding localities for Great White

Pelicans in the Western Cape since 1956 (Figure 2.1). Pelicans bred successfully

on Dassen Island in all years during 1995–2007, with a satellite colony on

Vondeling Island in 2001 and 2005 (nine pairs in each year). The breeding

population grew exponentially from 1955 to 2004 at a rate of 7.4 % per year on

average (Figure 2.2a; r = 0.873, p < 0.001) and then showed a marked decrease in

2005 and 2006. A weighted linear regression from 1988 to 2007 suggested that

the numbers of breeding pairs increased most rapidly between 1988 and 1990.

The growth rate slowed in the mid 1990s before again increasing from 1998 to

2000. Subsequently numbers decreased from 2002, and a marked decline

occurred from 2004, when the number of pelican breeding pairs decreased from

560 in 2004 to 323 in 2005 and 370 in 2006. They then recovered to 505 pairs in

2007 (Figure 2.2b, Appendix 2.1).

The estimated proportion of adults breeding for the period 2004–2007 was 57%

in 2004, 34% in 2005, 36% in 2006 and 49% in 2007 considering age at maturity

at four years, but 49%, 31%, 33% and 48% if pelicans breed at three years. Adult

and immature pelicans were counted separately during the census in the 1970s

(Crawford et al. 1995); from these results (number of adults in the population /

number of breeding pairs) it was estimated that 61% of the adults bred in 1978

and 53% in 1979, although the value for 1979 is probably closer to 60%, because

the reported number of breeding pairs was counted in December, two months

after the peak of breeding in all reported years (Chapter 3).

Chick production (number of chicks fledged) from Dassen Island during 1998–

2007 was 222 ± 125 [67–459] (mean ± SD [range]). Excluding 2004, breeding

success (chicks fledged per pair attempting to breed) averaged 0.45 ± 0.15 (range

0.25 in 1998 to 0.70 in 2001). In 2004 breeding success was substantially lower

(0.12).

��

40

The population growth model yielded a growth rate that was equivalent to the

observed growth (7.4% per year), when considering age at breeding as four years,

participation in breeding as 60% (see estimates for 1978–1979 and 2004,

previous to reduction in food availability), survival rates as estimated in Chapter

3 and an average reproductive success of 0.64 fledglings per pair (greater than

the mean recorded but within the range of observations). A reduction on pelicans’

age at breeding from four to three years (Chapter 3) would increase annual

growth to 7.8%. These results suggest that intrinsic population growth is

sufficient to explain the observed trend from 1955 to 2007. Furthermore,

considering that pelicans have the capacity to increase both the proportion of

breeding pairs and reproductive success, the growth rate of the pelican

population could be increased (at least theoretically) even further.

��

During the 2000s, the pig farm was the single most important site for pelicans in

the Western Cape with an aerial peak count of more than 2700 birds, equivalent

to the entire Western Cape population. Ground counts at the pig farm during

2001–2008 showed that pelican numbers there increased from April 2003 and

peaked in October 2004, averaging 742 ± 307 birds in this period (Figure 2.3).

During 2003–2008, the average number at the pig farm was greater than that at

any of the 70 CWAC wetlands in the Western and Northern Cape where

waterbirds were counted regularly (Figure 2.4). Counts at the main CWAC sites

in the period 1992–2008 showed substantial numbers of pelicans on large

waterbodies (e.g. Verlorenvlei and Orange River mouth), and noteworthy counts

were also frequent at Droëvlei farm, Berg River, De Hoop, Rietvlei, Langebaan

Lagoon and Strandfontein Sewage Works. Numbers of pelicans at CWAC sites

did not change between 1992–2000 and 2001–2008, except for counts at the Berg

River mouth, which were significantly larger in the first part of this period

(Mann-Whitney test: U = 52, T = 254, p = 0.013, n = 12, 19) (Appendix 2.2).

Census data from the seven regional surveys indicate that the total pelican

population in the Western Cape from October 2004 to July 2007 averaged

��

41

1 878 ± 337 (± SD, maximum 2363) (Table 2.1). This figure was 1 964 ± 273 (±SD)

if the ground-based count in July 2006 was excluded. The largest regional count

was recorded in October 2004 when 66% of the regional population (1550 birds)

was at the pig farm during the synoptic survey. On the afternoon of the same

day, during an aerial survey, 2 723 pelicans were counted at this site, more than

the total number of pelicans counted on any regional census. The breeding colony

on Dassen Island fluctuated between nil or small counts (in autumn, after the

end of the breeding season) and 654 in October 2004. Verlorenvlei hosted a large

number of pelicans in all regional surveys. The maximum number of pelicans at

this site (807) represented 41.7% of the total for April 2006. Simultaneous aerial

and ground counts were available for the Berg River (n = 4), the pig farm (n = 6)

and Verlorenvlei (n = 5). The correlation coefficient between the pairs of counts

at these three sites was 0.962 (n = 15, p < 0.001). The logarithmic model,

g = a 0.9315, where g was the ground count and a the aerial count, explained 85.3%

of the variance, and suggested that the ground counts were about 7% below air

counts on average.

The overall trend in pelican population size showed a moderate decline (-2.5%

per survey) from October 2004 to November 2007 (Figure 2.5). Pelican numbers

at the pig farm decreased during this period (most markedly from March 2005

when food supply fell dramatically at the pig farm) and concurrently increased at

other localities. From July 2005, the availability of offal was reduced further and

at least 200 pelicans died at the farm over a period of two weeks, apparently due

to starvation. Thereafter pelican numbers at the pig farm fluctuated between

200 and 400 birds linked to irregularities in the availability of food supply

(Figure 2.3).

No correlations were found between mean or maximum number of pelicans at

the pig farm during the breeding months of the years 2001 to 2008 and the

maximum number of breeding pairs on Dassen Island, or the annual chick

productivity for the same period. From 2001 to 2004, the number of breeding

pairs on Dassen Island reached its maximum, but the largest counts on the pig

��

42

farm were only 103 (± 127 SD; range 0–560) pelicans from December 2000 to

March 2003. Pelican numbers were not monitored on other similar artificial

feeding sites.

Chicken remains dominated food fed to pelican chicks at Dassen island from

2000 to 2008 (70% by mass). Other prey included fish remains (mostly carp

Cyprinus carpio), frogs and/or tadpoles Xenopus laevis and Kelp Gull Larus

dominicanus chicks (Table 2.2).

��

Information on pelican movements and dispersal was not previously available for

pelicans at a regional or southern African level (Crawford et al. 1995). Most

(1 404) of the resighting events and recoveries of ringed individuals occurred

within the range of the populations of origin; 0.03% of the movements reported

corresponded to dispersal outside the range of their natal colony (Figure 2.6). Of

all resightings, 87% were within the Western Cape, which is the region with

highest resighting effort. Results suggested juvenile dispersal from the colony at

Dassen Island mostly in a northward direction, with the Orange River being the

most regular northern boundary for this population. Although the Namibian

population disperses in all directions, none of the ringed pelicans reached the

Western Cape. All but one of the 36 ringed pelicans observed at the mouth of the

Orange River, on the border between South Africa and Namibia, were ringed in

the Western Cape, thus supporting the inclusion of this site in the Western Cape

regional quarterly surveys. Of these, three juvenile pelicans ringed on Dassen

Island were seen at the Orange River mouth and subsequently resighted in

wetlands within their population of origin (Western Cape) close to the time of

reaching sexual maturity (Figure 2.6). The longest distance recorded was a

pelican ringed as a chick on Dassen Island and seen at Lake Ngami, Botswana

within eight months of fledging (1 500–3 200 km depending on the route

followed; Figure 2.6). Rarely, young pelicans dispersed eastward beyond the

habitual range of the species, as exemplified by resightings of one colour-ringed

Western Cape pelican at several localities in the Eastern Cape (Figure 2.6).

��

43

��

The remarkable population expansion of the Great White Pelicans in the

Western Cape has been previously attributed to immigration from other

breeding colonies, because it was argued that the observed growth could not be

explained by intrinsic growth alone (Crawford et al. 1995). This study suggests

that a steady annual increase of the population (i.e. recruitment) is sufficient to

explain the observed population growth. But, what are the mechanisms to

explain such an increase in numbers?

Additional sources of food for pelicans (agricultural offal) apparently played an

important role in their local increase in numbers. During 2003–2004, up to two-

thirds of the pelicans of the Western Cape fed regularly at a single pig farm,

which provided surplus food (mainly chicken offal) to scavenging birds, including

pelicans. A controlled experiment to manipulate food availability at the pig farm

was not possible due to the ad hoc management style of the farm, and lack of

control over other potential alternative sites where food resources were available

to the pelicans (refuse tips, chicken farms, abattoirs, etc.). However, fluctuations

of the quantity of food at the farm created a natural experiment. When the

supply of food at the farm was interrupted in 2005, some 10% of the pelican

population died of starvation. Also, chick production fell to the lowest recorded

value (0.12 chicks/pair) and the number of breeding pairs decreased in

subsequent breeding seasons.

��

Ringing records showed no evidence of immigration into the Western Cape

population, despite the high resighting effort in this region. However, contact

between the coastal populations of Namibia and the Western Cape was observed

at the Orange River estuary (Northern Cape), and a few additional events of

long-distance dispersal were recorded. The expansion of the Western Cape

population could have produced a surplus of juveniles that dispersed into other

localities. Preliminarily, natal philopatry seems to be an important trait for the

species, as suggested by the return of three ringed pelicans to their breeding

��

44

area in the Western Cape approximately at the time they would reach sexual

maturity. However, more mark-recapture data would be necessary to test this

hypothesis.

The postulated isolation of the Western Cape population is in agreement with

the results of molecular studies. Both nuclear and mitochondrial DNA analyses

reveal lower genetic variability of this pelican population compared with other

sampled colonies, as well as no molecular evidence of immigration into this

population (Chapter 6).

��

The largest growth rate in the number of breeding pairs was recorded between

1988 and 1990, which coincides with the initial years in which surplus food, in

the form of fish and later chicken offal, was made available to pelicans. The

proportion of adults breeding was higher in the 1970s (ca. 60%) than the average

value for the period 2004–2007, suggesting that breeding effort had diminished

as the population increased. In 2004 and 2007, the proportion of adults breeding

was near 50%; whereas it was reduced to 30% in years of lower availability of

food subsides (2005 and 2006). Similarly, a variable proportion (33%–100%) of

the adult pelican population bred annually at Lake Shala, Ethiopia (Brown et al.

1982). It is unknown whether adult pelicans may refrain from breeding after

successful attempts. Non-breeding years could be explained by density-

dependent processes related to food availability (Furness & Monaghan 1987).

Adults of long-lived species should reduce their reproductive effort when food is

limited, so that they do not compromise their survival and therefore future

reproductive output (Williams 1966; Martin 1987; Stearns 1992; Erikstad et al.

1998).

Studies of experimentally supplemented populations of seabirds show that chick

production and survival increase during good food years (Kitaysky et al. 2000;

Davis et al. 2005). Abundant offal was delivered regularly to the pig farm at the

beginning of the 2004 breeding season, thus pelican numbers on this site peaked

��

45

and a large number of pelicans attempted breeding (560 pairs). When food

availability at the pig farm was reduced from October 2004, many breeding

attempts failed, and although some re-laying was observed, breeding production

in 2004 was anomalously low. Also, the number of breeding pairs in the

subsequent two breeding seasons declined by more than 40%, probably due to a

smaller proportion of adults breeding (see above). Variability in estimates of

reproductive success is common in the literature: a Great White Pelican colony in

Greece fledged 0.65 ± 0.11 chicks per nest during 1986–1996 (Crivelli et al.

1997). Two other independent studies of Great White Pelicans in Ethiopia and in

the Western Cape in 1977 estimated reproductive success at 0.8–0.9 young per

pair (Brown and Urban 1969, Guillet and Crowe 1983). Simultaneous

measurement of breeding success and environmental covariables should be

attempted to test the effect of food availability on reproduction.

Agricultural waste has played a key role in the pelican population expansion in

the Western Cape. In the late 1970s, wetlands were the main feeding site for

pelicans. At the time, Rondevlei Nature Reserve was utilised by up to 50% of the

pelican population (Guillet & Crowe 1983), whereas it attracted only 20 ± 42

pelicans during 1995–2008. Although the monitored pig farm was the most

important foraging site for pelicans between 2003 and 2008, the reported

increase in the number of breeding pairs on Dassen Island cannot be attributed

to the studied pig farm alone, but in combination with similar farms in operation

during the initial period of subsidised population growth in the 1980s. According

to CWAC data, no changes in wetland visitation by pelicans were observed since

the early 1990s. The predominance of chicken in pelican regurgitations, and

discussion with local farmers confirmed that there were at least two other farms

that provided offal to the pelicans in the mid-1980s, although no data on the

numbers of pelicans at these sites, or the quantities of offal supplied are

available.

��

Regional pelican counts from October 2004 to July 2007 showed that the pelican

��

46

population in the Western Cape consists of ca. 2500 pelicans. However, these

counts underestimated the population, as exemplified by the 2723 pelicans

counted at the pig farm in October 2004. The concentration of pelicans at the pig

farm in this survey reduced the detection bias likely to have applied during other

surveys when birds were more widely dispersed.

A comparison between the results of regional counts undertaken during 1976–

1985 (n = 10) (Guillet & Crowe 1981; Crawford et al. 1995) and this study

showed that the Western Cape population increased four fold from the maximum

count of 652 pelicans in 1978 to 2 723 in 2004. Similarly, the number of breeding

pairs was 3.2 times higher in 2004 than in 1978 (174 in 1978 and 560 in 2004).

The population censuses indicated a faster growth than that of the breeding

population, either due to changes in the proportion of breeders or to a lag in

growth of the breeding population, which would apply if the population was

growing rapidly.

A standardised survey protocol needs to be adopted to detect long-term trends in

the number of pelicans in the region. The best time of the year for surveys is in

April, when most ephemeral wetlands are dry, and fish availability increases in

the larger waterbodies (Guillet & Crowe 1981), so that pelicans concentrate in

larger groups at fewer sites. An April survey would also provide information on

juvenile dispersal. Combining aerial and ground counts is strongly

recommended, and simultaneous counts of the larger waterbodies (i.e.

Verlorenvlei, Berg River, Langebaan Lagoon and Botriviervlei) are advisable.

Although ground counts tend to underestimate pelican numbers, fog or other

adverse weather conditions could prevent effective aerial counts.

��

Great White Pelicans have a more diverse diet in the Western Cape than

elsewhere in Africa or Europe. Analyses of chick regurgitations at Hardap Dam,

Namibia, yielded 44% (by mass) Tilapia (mostly�Oreochromis mossambicus), 39%

Yellowfish Labeo spp. and the remaining 17% other fish species (n = 9, total

��

47

mass = 3 500 g; pers. obs). An adult pelican trapped at the salt pans in Walvis

Bay regurgitated over 600 g and more than 1 000 individuals of Sand Shrimp

Talaemon peringueya (Mark Boorman, unpubl. data). This is in accordance with

other data collected across Africa and Europe, where pelicans feed almost

exclusively on fish and other aquatic prey (summarized by Cramp & Simmons

1977; Brown et al. 1982; del Hoyo et al. 1992; Johnsgard 1993). In the Western

Cape, chicken offal has been a major constituent of pelican chick regurgitations

at Dassen Island since 1990 (Crawford et al. 1995). In fact, chicken intestines

may be underrepresented in the dietary analysis, because they probably are

quickly digested and are not easily identified, compared to other prey items.

Furthermore, pelican predation on other seabirds during the breeding season

was much more prevalent than the analysis of chick regurgitations indicates

(Chapter 4). However, it is unclear how many breeding pelicans are involved in

this predatory behaviour. Cannibalism was observed on at least two occasions at

the pelican colony (Chapter 3). The presence of litter in the diet has not been

reported from other localities before, and possibly results from scavenging offal

at dump sites.

Unlike the West African coastal pelican populations of Mauritania and Senegal

that exploit the shallow waters surrounding the sandy islands where they breed

(Crivelli 1994), pelicans along the west coast of South Africa seldom feed in the

sea, due to the scarcity of shallow bays and the species’ inability to dive. Most

Great White Pelican populations elsewhere in Africa forage at large inland fresh-

water bodies with a dependable and abundant supply of fish (Feely 1962; Brown

& Urban 1969; Din & Eltringham 1974; Whitfield & Blaber 1979). The biomass

and availability of fresh-water fish in the Western Cape may be insufficient to

maintain a large pelican population. Foraging range is restricted to wetlands

along the Atlantic coast and farmlands (farm dams), and dependent on a

fluctuating availability of fresh-water fish (Guillet & Crowe 1981; 1983). Thus,

the small numbers of pelicans breeding in the Western Cape in historical times

correspond to the limited carrying capacity of this Mediterranean-climate

habitat. The paucity of waterbodies and fish biomass in this region may be

��

48

comparable to the southern European populations of Great White Pelicans,

which are also small, although in their case, increasing habitat loss due to

human pressure is the most important cause of population decline (Crivelli et al.

1991; 1998).

Great White Pelican adult daily food intake has been estimated to be 800–1200 g

of fish (Brown & Urban 1969; Guillet & Furness 1985; van Zyl et al. 1995). Food

demands during the breeding period (150 days) for a pair and their chick are

about 420 kg of food (Brown & Urban 1969). Assuming that pelicans in the

Western Cape have similar food requirements to other pelican populations in

Africa, they would have consumed about 270 tonnes of fish per year in the mid

1980s (pre-pig farm). In the 2000s, after the population expansion, the 660

breeding pairs (including non-breeders and immatures) would have required

1 105 tonnes of fish. These estimates could be higher than the actual figures,

especially considering that pelicans in the Western Cape seem to be smaller in

mass than other pelicans in Africa (MdPM, unpublished data). The adjusted

values of their annual fish consumption could be 190 tonnes for the population in

the mid 1980s and 790 tonnes in the 2000s, the former being approximately the

same value as the 184 tonnes calculated by Guillet & Furness (1985) using a

detailed bioenergetics model. In any case, estimates of the amount of fish

available to pelicans in the Western Cape and any evidence of depletion of the

sources of fresh-water fish near the breeding colony require further

investigation.

��

After the pig farm had become the main foraging site for pelicans in the region,

the reduction in availability of agricultural offal had drastic consequences for

this artificially increased population, affecting reproductive performance and

adult survival. In addition, a shift in pelican foraging behaviour followed the

scarcity of food at the pig farm. The pelicans looked for other sources of food as

an alternative to chicken offal, intensifying predation on other seabirds to

unprecedented levels (de Ponte Machado 2007). This switching of prey by a

��

49

facultative scavenger has precedents in other avian top predators. In the North

Sea, Great Skuas Stercorarious skua became a potentially serious threat to

seabird communities following diminishing fisheries discard rates (Votier et al.

2004), constituting an example of the indirect effects of trophic cascades in

community dynamics (Polis et al. 2000).

Uncontrolled pelican predation on seabirds could trigger community-level

changes in local ecological communities. A well-researched case is the interaction

between Killer Whales Orcinus orca, Sea Otters Enhydra lutris and sea urchins

in the Alaskan North Pacific, which constitutes a classic example of top-down

regulation of ecological cascades (Estes et al. 1998). Reduced fish availability

forced a shift in prey preferences of Killer Whales from fish-dependent marine

mammals to Sea Otters, a keystone benthic predator. Reduced otter populations

caused an increase in sea urchin abundance and consequently affected a primary

producer: Kelp forests. However, sometimes compensatory mechanisms could

truncate a trophic cascade, preventing it from reaching the primary level

producers (Pace et al. 1999). Thus, pelican predation on seabirds could cause

declines in local seabird populations but not necessarily affect primary producers

of marine or terrestrial communities. The complexity of trophic webs, relatively

small spatial and temporal scales and confounding factors affecting the

abundance of seabirds and their natural prey (e.g. fisheries, natural fluctuation

of pelagic fish) could buffer the ecological effects of pelican predation on seabirds.

Yet, the declines experienced by some South African seabird species are

sufficiently serious to regard pelican predation as an additional threat to their

conservation.

Although the availability of offal at the pig farm has been reduced and

management interventions at some seabird colonies have been effective at

curbing pelican predation (Chapter 5), it is not certain what trajectory the

pelican population will follow. If the pelicans find another alternative food

source, the population might remain stable or even increase further, but if they

fail to find alternative food sources, the population might decrease to the level

��

50

that ‘natural’ food sources can support. It is hard to predict the effect of changes

in food supply on the population trends of opportunistic species that have the

ability to switch to alternative prey sources (Cairns 1987), and it is not within

the scope of this study to determine possible future trends in pelican population

dynamics.

��

51

��

Anderson MD, Kolberg H, Anderson PC, Dini J, Abrahams A (2003) Waterbird

populations at the Orange River mouth from 1980–2001: a re-assessment

of its Ramsar status. Ostrich 74, 159–172.







Madoqua (Ser. 1) 7, 17–31.

Blokpoel H, Spaans AL (1991) Superabundance in gulls, causes, problems and

solutions. In: Acta XX Congressus Internationalis Ornithologici, vol. 4, pp.

2359-2398. New Zealand Ornithological Congress Turst Board,

Wellington, New Zealand.

Bosch M, Oro D, Cantos F, J, Zabala M (2000) Short-term effects of culling on

the ecology and population dynamics of the Yellow-legged Gull. Journal of






Cairns DK (1987) Seabirds as indicators of marine food supplies. Biological

Oceanography 5, 261–271.

Caswell H (2001) Matrix population models. Sinauer, Sunderland.

Choquenot D, Bolton N, Woods D (2008) Evaluating helicopter-based surveys for

estimating densities of Himalayan thar. Wildlife Research 35, 358–364.

Cooper J (1980) Fatal sibling aggression in pelicans–a review. Ostrich 51, 183–

186.

Cramp S, Simmons KEL (1977) Family Pelecanidae: White Pelican. In:

Handbook of the birds of Europe, the Middle East, and North Africa: the

birds of the Western Palearctic (ed. Cramp S), vol. 1: Ostrich-Ducks, pp.

226–233. Oxford University Press, Oxford.


population of Great White Pelicans Pelecanus onocrotalus in South Africa's

Western Cape. South African Journal of Marine Science 15, 33–42.



1949–1978. Fisheries Bulletin of South Africa 15, 67–70.




Mitchev T, Nazirides T, Peja N, Sarigul G, Siki M (1998) Status and

population development of Great White Pelican Pelecanus onocrotalus and

Dalmatian Pelican P. crispus breeding in the Palearctic. In: Monitoring

and conservation of birds, mammals and sea turtles of the Mediterranean

��

52

and Black Seas. Proceedings of the 5th Medmaravis Symposium (eds.

Yésou P, Sultana J), pp. 38–45. Environment Protection Department,

Gozo, Malta.

Crivelli AJ, Catsadorakis G, Jerrentrup H, Hatzilacou D, Michev T (1991)

Conservation and management of pelicans nesting in the Palearctic. ICBP

Technical Publication 12, 137–152.

Crivelli AJ, Catsadorakis G, Nazirides T (1997) Pelecanus onocrotalus White




Davis SE, Nager RG, Furness RW (2005) Food availability affects adult survival

as well as breeding success of parasitic jaegers. Ecology 86, 1047–1056.






Unit, Cape Town.




backed Pelicans in the Ruwenzori National Park, Uganda. Ibis 116, 28–43.

Erikstad KE, Fauchald P, Tveraa T, Steen H (1998) On the cost of reproduction

in long-lived birds: the influence of environmental variability. Ecology 79,

1781–1788.

Estes JA, Tinker MT, Williams TM, Doak DF (1998) Killer Whale predation on

Sea Otters linking oceanic and nearshore ecosystems. Science 282, 473–

476.

Feely JM (1962) Observations on the breeding of the White Pelican, Pelecanus

onocrotalus, at Lake St. Lucia, Zululand, during 1957 and 1958.


Fleming PJS, Tracey JP (2008) Some human, aircraft and animal factors

affecting aerial surveys: how to enumerate animals from the air. Wildlife

Research 35, 258–267.

Furness RW (1981) The impact of predation by Great Skuas Catharacta skua on

other seabird populations at a Shetland colony. Ibis 123, 534–539.

Furness RW (2003) Impacts of fisheries on seabird communities. Scientia Marina

67, 33–45.

Furness RW (2007) Responses of seabirds to depletion of food fish stocks.

Journal of Ornithology 148, 247–252.

Furness RW, Monaghan P (1987) Seabird Ecology. Blackie, Glasgow and London.






��

53



Zoology 205, 573–583.

Hernández-Matías A, Ruiz X (2003) Predation on Common Tern eggs by the

Yellow-legged Gull at the Ebro Delta Scientia Marina 67, 95–101.



Hone J (2008) On bias, precision and accuracy. Wildlife Research 35, 253–257.

Hunter MD (2009) Trophic promiscuity, intraguild predation and the problem of

omnivores. Agricultural and Forest Entomology 11, 125–131.

Hunter S, Brooke M, De L (1992) The diet of giant petrels Macronectes spp. at

Marion Island, Southern Indian Ocean. Colonial Waterbirds 15, 56–65.

Huxel GR, McCann K (1998) Food web stability: the influence of trophic flows

across habitats. American Naturalist 152, 460–469.

IUCN (2008) 2008 IUCN Red List of Threatened Species. www.iucnredlist.org.

Jefferies RL (2000) Allochthonous inputs: integrating population changes and

food-web dynamics. Trends in Ecology & Evolution 15, 19–22.



King DT, Paulson JD, Leblang DJ, Bruce K (1998) Two capture techniques for

American White Pelicans and Great Blue Herons. Colonial Waterbirds 21,

258–260.

Kitaysky AS, Hunt GL, Flint EN, Rubega MA (2000) Resource allocation in

breeding seabirds: responses to fluctuations in their food supply. Marine


Martin CM (1987) Food as a limit on breeding birds: a life-history perspective.

Annual Review of Ecology and Systematics 18, 453–487.


survival of kittiwakes. Ecology 83, 2516–2528.


impact on threatened sympatric seabirds. Animal conservation 10, 117–

126.

Pace ML, Cole JJ, Carpenter SR, Kitchell JF (1999) Trophic cascades revealed in

diverse ecosystems. Trends in Ecology & Evolution 14, 483–488.

Pannekoek J, van Strien AJ (1996) TRIM—trends and indices for monitoring

data. Research paper no. 9634. Statistics Netherland, Voorburg,

Netherlands.

Polis GA (1994) Food webs, trophic cascades and community structure. Austral

Ecology 19, 121–136.

Polis GA, Anderson WB, Holt RD (1997) Toward an integration of landscape and

food web ecology: the dynamics of spatially subsidized food webs. Annual

Review of Ecology and Systematics 28, 289–316.

Polis GA, Hurd SD (1996) Linking marine and terrestrial food webs:

allochthonous input from the ocean supports high secondary productivity

on small islands and coastal land communities. American Naturalist 147,

396–423.

��

54

Polis GA, Sears ALW, Huxel GR, Strong DR, Maron J (2000) When is a trophic

cascade a trophic cascade? Trends in Ecology & Evolution 15, 473–475.

Polis GA, Strong DR (1996) Food web complexity and community dynamics.

American Naturalist 147, 813–846.








Rodgers, JA Jr, Kubilis, PS, Nesbit, SA (2005) Accuracy of aerial surveys of

waterbird colonies. Waterbirds, 28, 230–237.

Rose MD, Polis GA (1998) The distribution and abundance of coyotes: the effects

of allochthonous food subsidies from the sea. Ecology 79, 998–1007.

Southwell C, Paxton CGM, Borchers DL (2008) Detectability of penguins in

aerial surveys over the pack-ice off Antarctica. Wildlife Research 35, 349–

357.

Stearns SC (1992) The evolution of life histories. Oxford University Press,

Oxford.

Sutherland WJ, Newton I, Green RE (2004) Bird ecology and conservation: a

handbook of techniques. Oxford University Press, Oxford.

Taylor PB, Navarro RA, Wren-Sargent M, Harrison JA, Kieswetter SL (1999)

Total CWAC Report: Coordinated Waterbird Counts in South Africa 1992–

97. Avian Demography Unit, Cape Town, South Africa.


Pelicans Pelecanus onocrotalus at Hardap Dam, Namibia. Lanioturdus 36,

2–7.





van Zyl BJ, Hay CJ, Steyn GJ (1995) Some aspects of the reproduction biology of

Labeo capensis (Smith, 1941) (Pisces, Cyprinidae) in relation to

exploitation and extreme environmental conditions in Hardap Dam,

Namibia. South African Journal of Aquatic Science 21, 88–98.









Weidinger K (1998) Effect of predation by skuas on breeding success of the Cape

Petrel Daption capense at Nelson island, Antarctica. Polar Biology 20,

170–177.

��

55

Whitfield AK, Blaber SJM (1979) Feeding ecology of piscivorous birds at Lake St.

Lucia. Part 3: swimming birds. Ostrich 50, 10–20.





Williams GC (1966) Natural Selection, the Costs of Reproduction, and a

Refinement of Lack's Principle. The American Naturalist 100, 687–690.

Yorio P, Bertellotti M, Gandini P, Frere E (1998) Kelp gulls Larus dominicanus

breeding on the Argentine coast: population status and relationship with

coastal management and conservation. Marine Ornithology 26, 11–18.

57

Table 2.1: Number of Great White Pelicans from the quarterly surveys

undertaken in the Western Cape from October 2004 to November 2007, including

the Orange River mouth (Northern Cape). Expected accuracy for each survey is

ranked from 1 (low) to 5 (high), depending on visibility and coverage. Sp.: Spring;

Sp.+: late spring; Su.: Summer; Au.: Autumn; W.: Winter; A: aerial survey; G:

ground count. Missing data are represented by a dash. * 1 550 was the result of

the ground count in the morning of 16th October 2004. An aerial count that

afternoon yielded 2 723 pelicans.

Month of survey Oct 04 Apr

06

Jul

06

Oct

06

Jan

07

Apr

07

Jul

07

Nov

07

Nov

07

Season Sp. Au. W. Sp. Su. Au. W. Sp.+ Sp.+

Type of survey A, G A, G G A, G A, G A, G A, G G A

Expected accuracy 3 5 3 5 5 5 5 2 2

Berg River 4 48 16 101 50 153 56 2 194

Botriviervlei 0 0 12 67 61 122 222 0

Dassen Island 654 0 230 454 118 9 415 -

De Hoop vlei 0 85 45 44 39 33 155 31

Elim vlei - 9 0 6 - 116 0 0

WCNP 4 28 9 98 27 200 74 150 231

Olifants River mouth 0 2 0 0 43 79 264 0

Orange River mouth - 57 99 25 80 85 99 123

Pig farm 1 550* 706 641 659 640 355 213 220 214

Rietvlei 7 10 3 2 90 16 3 90

Rondevlei Nature

Reserve 2 9 12 21 56 74 0 12

Strandfontain

Sewage Works 62 68 74 39 74 41 16 22

Verlorenvlei 66 807 106 52 334 439 81 - 52

Zeekovlei 7 110 113 16 23 0 76 7

Farm dams 1 49 2 19 481 181 14 9 49

Other sites 6 4 0 14 28 50 28 3 8

TOTAL 2 363 1 992 1 362 1 617 2 144 1 953 1 716 669 748

58

Table 2.2: Percentage contribution by mass and frequency of occurrence of

different items in regurgitations of Great White Pelicans collected on Dassen

Island from 2000 to 2008.

2000–2001 2002–2003 2005–2006 2007–2008 Overall

mass items mass items mass items mass items

mass items

Chicken - - 69 8.8 - - 73.5 59.3 66.9 17.0

Fish 0.3 0.6 25.9 24.6 1.9 33.3 15.9 28.4 15.7 12.5

Frog 99.7 99.4 2 64.9 - - - - 1.4 66.2

Mammal

remains - - - - - - 8.4 6.2 6.9 1.6

Kelp Gull chick - - - - 96 33.3 1.5 1.2 8.0 0.6

Pelican chick - - 3.1 1.8 - - - - 0.3 0.3

Rubbish - - - - 2.2 33.3 0.7 4.9 0.7 1.6

Mass (g) 217 1 674 1 270 14 673 17 834

# items 170 57 3 81 311

# samples 1 7 1 43 52

65

Appendix 2.1: Great White pelican breeding estimates from 1955 to 2007 and

population growth rates, both crude ( = (Nt / N0) 1/t) and the modelled weighted

linear regression rates of increase or decrease (Underhill et al. 2006) at the

colony on Dassen Island in the Western Cape from 1988 to 2007. Data on

number of breeding pairs and number of chicks fledged up to 1994 was obtained

from Crawford et al. (1995). Subsequently, it was compiled from Marine and

Coastal Management and CapeNature records.

Year Breeding

pairs

Empirical

growth rate

( )

Modelled

rate of

change (%) of

breeding

population

Numbers

of chicks

fledged

1955 25 - - -

-

1978 174 1.0880 - -

-

1985 185 1.0088 - -

-

1988 260 1.1201 16.1 -

1989 265 1.0096 13.6 -

1990 407 1.2393 10.3 -

1991 434 1.0326 7.2 -

1992 306 -0.8397 5.2 -

1993 504 1.2834 3.5 -

1994 400 -0.8909 1.5 -

1995 508 1.1269 0.2 -

1996 341 -0.8193 1.3 >67

1997 429 1.1216 4.7 -

1998 419 -0.9882 8.2 106

1999 511 1.1043 9.2 265

2000 620 1.1015 6.6 366

2001 659 1.0310 1.7 459

2002 564 -0.9251 -3.7 285

2003 609 1.0391 -8 229

2004 560 -0.9589 -9.9 67

2005 332 -0.7610 -8.6 154

2006 371 1.0571 -4.2 96

2007 505 1.1667 2.2 197

�

CHAPTER 3

��

��

67

�

��

��

�

��

Great White Pelicans Pelecanus onocrotalus bred during the austral spring and

summer (August–February) at their most southerly colony in Africa, Dassen

Island in the Western Cape. The greater demand for food by the growing chicks

coincides with the shrinking of the seasonal waterbodies, providing better access

to aquatic prey for this non-diving species. Age at first breeding was confirmed at

three years. Adult survival was high; the record of longevity for the species in the

region was 39 years. Survival among first year birds was lower ( 1 = 0.74) than

for older pelicans ( 2 = 0.96). A total of 253 resightings and 164 recoveries of

ringed pelicans were recorded in southern Africa; most of the resighting effort

being in the Western Cape. Pelicans from the Western Cape were largely

philopatric; no evidence of immigration was found but some long distance

dispersal outside the range of their natal colony was observed. Of the 247

pelicans resighted in the Western Cape, 218 were seen foraging or roosting at a

farm that provided offal to scavenging birds, the same area most of the locations

received from two satellite tracked pelicans originated. Time of breeding and

demographic parameters differed between Dassen Island in the Western Cape

(during the austral summer) and Hardap Dam, Namibia. Breeding at Hardap

Dam is dependent on rainfall and during 2004–2008 happened mostly between

March and July. Mortality rates of pre-fledging chicks were much higher in

Namibia (27.5% of the 404 chicks ringed) than in the Western Cape (0.4% of the

785 chicks ringed), suggesting that different factors drive the dynamics at each

breeding locality. Both siblicide and cannibalism were observed at the colony on

Dassen Island, and are discussed in relation to the life-history of the species, and

in comparison with other pelicans of the world.

��

�

68

�

��

Understanding the dynamics of animal populations requires estimates of

population parameters (Lebreton et al. 1992). In addition, a detailed knowledge

of the life-history traits of organisms enables the formulation and testing of

interesting hypotheses in evolutionary biology, as well as facilitating the design

of population biology studies and the collection of relevant ecological data

(Stearns 1992; Ricklefs 2000). Pelicans Pelecanus, like many other seabirds

sensu lato, display life-history traits that are consistent with long-lived species:

small clutches, deferred maturity and a low adult mortality rate (Furness &

Monaghan 1987; Hamer et al. 2002). All pelicans are colonial breeders that

forage mostly on fish and are dependent on aquatic habitats (Johnsgard 1993).

Often, food supply is located at a distance from the breeding colony and because

food availability is unpredictable, chick growth is relatively slow. These factors

result in low reproductive output, which combined with high life expectancy of

adults, defines them as K-selected organisms (Furness & Monaghan 1987;

Stearns 1992).

Great White Pelicans Pelecanus onocrotalus are one of the largest and heaviest

flying birds. They breed in sub-Saharian Africa and parts of Eurasia (del Hoyo et

al. 1992; Johnsgard 1993). In temperate regions they breed annually, forming

large dense colonies on the ground (Urban 1984). They lay 1–3 eggs, usually two,

which are incubated by both parents for 29–38 days (Brown & Urban 1969;

Cooper 1980). Chicks are nidiculous, hatching naked and helpless. They are

brooded continuously for at least three weeks (Brown & Urban 1969; del Hoyo et

al. 1992). Siblicide is common and by day 60 after hatching all nests contain only

one chick (Brown et al. 1982). After 20–25 days their legs are strong enough to

hold them upright and they start aggregating in crèches when both adults are

absent from the colony. They fledge and disperse away from the breeding colony

at 70–75 days old (Brown et al. 1982). Despite this protracted chick growth

period, development and growth of pelican chicks is faster than other species in

the Pelecaniformes (Nelson 2005). The breeding season is long, often involving

two or more overlapping waves of breeding, in which some relaying is probable

��

�

69

�

(Berruti 1980; Guillet & Crowe 1983; Crivelli 1994). Chicks thus leave the colony

in different waves during the breeding phase (Berruti 1980). Age at maturity is

not known but thought to breed for the first time at four years old (Johnsgard

1993). Survival is high for adults, which can live up to 23 years (Cooper 1980;

Urban & Ash 2001) or perhaps 28 years (Tree 1997).

The Western Cape Great White Pelican population is one of three regular

breeding localities for the species in southern Africa. Regionally, the population

has been estimated at 6 000 breeding pairs (Williams & Borello 1997; du Toit et

al. 2002; Dodman in press; Hockey et al. 2005). Breeding censuses and repeated

counts of pelicans in the Western Cape show that this most southern pelican

population has experienced an exponential increase in numbers since the 1950s

(Crawford et al. 1995; Chapter 2). Surplus food in the form of agricultural

subsidies probably accounts for this growth in numbers (Chapter 2). The

breeding history at Hardap Dam, Namibia, is quite recent. The dam was built in

1962 and pelicans were first reported breeding there in 1968 (Urban 1984).

Breeding at this locality is sporadic, conditional on abundant rainfall, but

occurred annually during 2004–2008 following good rains. Numbers at Hardap

Dam have increased in the last few decades (Theron et al. 2003). The average

breeding population has been estimated at 250–300 pairs in the 2000s

(H Kolberg, pers. comm.); although up to 500 chicks were counted in May 2006

(pers. obs.).

Despite being large and charismatic birds Great White Pelicans have not been

studied extensively in southern Africa. Detailed studies have been undertaken at

some African colonies (Brown & Urban 1969; Berry et al. 1973; Berry 1975;

Urban 1984; Guillet 1985; Crivelli 1994; Crawford et al. 1995; Bowker & Downs

2008; see Chapter 1 for a more exhaustive review), but many basic population

parameters have not been estimated for the species. This study provides a better

understanding of life-history traits of Great White Pelicans in the Western Cape,

and provides an update of some basic reproductive, temporal and spatial

characteristics for the species in the region.

��

�

70

�

Reproductive parameters (e.g. breeding effort, chick production, proportion of

adults breeding, etc.) are of great importance to calculate demographic patterns

in populations (see Chapter 2). However, for long-lived bird species adult

survival is often more important than breeding success or other reproductive

parameters to explain long-term population growth (Stearns 1992; Caswell

2001). Capture-recapture data have been used to assess population status and

trends in many vertebrate populations (Lebreton et al. 1992). Here they provide

detailed information from individual histories that allow the estimation of

survival rates for different age classes. Ringing data also provide understanding

of the use of habitat, dispersal movements and degree of interconnectedness

between pelican breeding colonies in the region. Finally, behavioural

observations at the breeding colony contribute to further our understanding of

the natural history of the species, and could contribute to shape future research.

��

��

Timing of breeding on Dassen Island (33°25’ S, 18°05’ E) was monitored each

austral spring and summer between 1995 and 2007. Data were extracted from

the databases of Marine and Coastal Management and CapeNature. Several

counts of the numbers breeding were made each year to estimate peak breeding

effort and chick production. The location of each breeding colony on the island

was mapped and changes between sub-colonies during the breeding season

recorded. In the Western Cape, pelicans often breed in two or three overlapping

‘waves’ (Guillet & Crowe 1983). For this study, the maximum count of breeding

pairs in a breeding season was taken as the number of breeding pairs in the

year. Annual chick production was calculated as the sum of the numbers of

fledglings produced in each sub-colony throughout the breeding season. The

beginning and end of the breeding season was not recorded accurately in all

years. No observations were made in 1997, except for a count of the number of

breeding pairs in the month of October.

��

�

71

�

Behavioural observations of the breeding colony on Dassen Island were made

from a distance of c. 25 m using 10�42 binoculars and a 30–60� telescope in

September and October 2004. The breeding colony was approached slowly to

minimize disturbance. Great White Pelicans are sensitive to disturbance while

nesting; sometimes abandoning breeding attempts (Berry et al. 1973; Berruti

1980; Bowker & Downs 2008). Also, Kelp Gulls Larus dominicanus prey on eggs

and unattended chicks (Hockey et al. 2005). No significant disturbance events

were recorded as a consequence of these observations.

��

Between 2001 and 2008, 791 pelicans were ringed in the Western Cape, using

elliptical 34 � 19 mm (internal diameter) numbered stainless-steel rings from the

South African Bird Ringing Unit (SAFRING) and yearly combinations of colour

rings. Of these, 727 were pre-fledging chicks ringed on Dassen Island (33°25’ S,

18°05’ E) and 57 were fully grown birds captured at the main foraging locality in

the Western Cape, a pig farm between the towns of Stellenbosch and Paarl in the

Western Cape, South Africa (33°49’ S, 18°47’ E). Seven additional pelicans were

ringed opportunistically across their distribution range. Fully grown pelicans

were trapped using modified leg-traps, set on the ground in commonly used

pathways near feeding areas. Modifications consisted of weaker springs (#1.5 coil

springs) and extra padded jaws to avoid harming the birds (King et al. 1998).

Traps were camouflaged and anchored to the ground using aircraft cable and

elastic shock-cord. Immediately after the birds were caught, they were ringed

and measured, and later released.

From 2001 to 2003, a combination of a metal ring with the SAFRING code and a

colour plastic ring identifying the year (cohort-specific) were used. Since 2004,

the plain colour ring was substituted with an alphanumerically engraved plastic

colour-ring (Pro-Touch, Saskatoon, Canada) to allow individual identification of

ringed pelicans (Table 3.1).

��

�

72

�

At total of 416 pelicans was ringed in Namibia from 2001. Of these, 404 were

chicks from the colony at Hardap Dam (24° 30’ S, 17° 50’ E) ringed between May

2004 and May 2008; four were chicks from Bird Rock Guano Platform in Walvis

Bay and eight additional fully grown birds ringed in the coastal stretch between

Walvis Bay and Swakopmund. Engraved colour rings were fitted to 98 of the

Hardap Dam chicks and five adults or immatures from the coastal population

(Table 3.1),

Resightings of ringed pelicans were recorded at foraging and roosting sites and a

database with life resightings and dead recoveries was compiled from the onset

of the colour ringing programme in 2001 until June 2009. Because rings of

different colour were used for each annual cohort of pelican chicks on Dassen

Island from 2001 to 2008, the age of ringed pelicans could be identified

immediately in the field. However, the sighting of a ringed bird in the proximity

of the breeding colony did not necessarily mean that the individual in question

was breeding, because it could be an immature or non-breeder loafing in the

periphery of the colony. Other more obvious signs of breeding such as full

breeding plumage or breeding behaviour (including nest building, copulation,

incubation or brooding) were required.

��

Survival of Great White Pelicans was calculated using combined capture-mark-

resighting and a dead-recovery models (Burnham 1993). These methods account

for the fact that not all surviving individuals were seen each year, and not all

individuals were found if they died (Lebreton et al. 1992). The models consist of

four sets of parameters: survival ( ); resighting rate (the probability of

encountering an individual given that it is alive and in the population, P);

recovery rate (the probability of a dead bird being found and reported, r); and

fidelity rate (the probability of staying in the study area, F). All model

components can independently depend on covariates, such as age, time or ring

type. I ran all models using the maximum likelihood techniques implemented in

the programme MARK 5.1 (White & Burnham 1999).

��

�

73

�

Although ringing effort was not exclusive to the Western Cape colony, the

survival analysis was confined to the Western Cape population because most

ringing and resighting effort has been carried out in this region. This population

shows a high level of philopatry, with little dispersal out of the region (Chapter 2

and 6). All individuals ringed at the Western Cape colony since 2001 were

included in the study, totalling 791 pelicans, 811 resighting events and 47

recoveries.

The most general model (Model 7, Table 3.2) allowed for differences in survival

among three age classes: Age 1 (0–12 months post-fledging), Age 2 (13–24

months) and Age 3 (older than 24 months). For recovery and fidelity rates, I

distinguished between two age classes, age 1 vs. older. This model further

allowed resighting probabilities to differ between the ring types because we

expected the engraved plastic rings to be much easier to be read in the field than

the metal rings (metal rings were read occasionally using a telescope). Finally,

all model components except fidelity were fully time dependent. I used this

model as a starting point and subsequently simplified model components

following Lebreton et al. (1992). I based model selection on the sample-size

adjusted Akaike’s Information Criterion (AIC) (Burnham & Anderson 2002). The

lowest AIC value indicates the most parsimonious model, that is the model that

fits the data best and at the same time requires fewest parameters to explain the

structure of the data. A larger number of parameters would improve the fit of the

model but also decrease the precision of parameter estimates. The set of

candidate models analysed here was compiled by considering the known life-

history of pelicans in the region, and only models that would make sense

biologically were considered (Lebreton et al. 1992; Anderson & Burnham 1999).

The approach used here also assumes that individuals in the same class age, or

yearly cohort are similar in their survival, resighting, recovery, and fidelity

rates, and that the resighting occasions are short compared to the survival

intervals (Burnham 1993). Violation of these assumptions leads to

��

�

74

�

overdispersion. I tested goodness-of-fit for the most general model, and estimated

the level of overdispersion in the data using the median- approach in program

MARK.

��

On 1 July 2004 two PTT-100 satellite transmitters (Microwave Telemetry, Inc.)

were deployed on two pelicans at the main foraging locality in the Western Cape,

the pig farm. Pelicans were trapped using the modified leg traps described above.

A battery-powered device (30 g; 63 � 18 � 17 mm) was deployed on one bird and a

solar-powered one (95 g; 94 � 33 � 30 mm) on the other. Both were <1.5% of the

mass of the birds, less than the 3% maximum recommended for flying birds

(Phillips et al. 2003). Both devices were fitted with a 150 mm antenna, and

attached with a Velcro extension to the bird’s back using epoxy resin ‘Araldite’.

The birds were kept at the ringing station until the epoxy had dried completely

and they were released at the site of capture. The transmitters were set to

transmit a signal every 5 min, for 8 hrs on and 8 hrs off. The position of the

pelicans was obtained from ARGOS at weekly intervals until they stopped

transmitting, which was estimated at 6–8 weeks after deployment for the

battery-powered device and for as long as the device remained on the pelican

back for the solar-powered device.

ARGOS platforms are located by using the Doppler Effect, by computing the shift

on the transmitter signals while a satellite passes over the transmitting platform

(ARGOS 2008). This system also provides information on the accuracy of each

position estimate, depending on the number and configuration of signals received

on each satellite pass for each location event. Classes 0, 1, 2, and 3 indicate that

the location was obtained with four messages or more. The accuracy of each class

ranges from a radius of 0–250 m for class 3, 250–500 m for class 2, 500–1 500 m

for class 1 to > 1 500 m (and no upper limit) estimated error for class 0 locations.

Accuracy cannot be estimated for classes A and B, although this does not mean

that the position is erroneous, or the location inaccurate. I checked all locations

in class 0 for inconsistencies in accuracy using speed between consecutive

��

�

75

�

locations, and improbable movements during the night. Only positions in the

classes 0–3 were included in the analyses.

��

��

On Dassen Island, the breeding season typically starts in July, with pelicans

initiating display and nest building during this month in at least seven of the 12

years documented (Figure 3.1). In three of the 12 years pelicans were already on

nests at the beginning of July. The end of the breeding cycle was more

inconsistent, varying from February (in 1996, 2003 and 2006) to April (1998,

2001 and 2007). Pelicans normally were absent from the island for 3–4 months

each year (from March/April to June/July). The peak of breeding (measured as

the maximum single count of breeding pairs) was October in eight years and

September in four (Figure 3.1). The duration of the breeding season was 7–8

months in the Western Cape, although the interval between egg laying and

fledging was between 100 days and four months for each wave of breeding.

On Dassen Island, the main breeding colony was located at Boom Point, the

north-western tip of the island, which was also the first site where pelicans

gathered and started breeding each year. In eight of the 13 years of the study,

subsequent waves of breeding pelicans broke off in smaller colonies distributed

across the island as the breeding season advanced (Figure 3.2). The average

number of sub-colonies was 6.9, ranging from one (in 1996) to 13 (when the

population peaked in 2001).

Birds bred synchronously within each sub-colony, although the larger groups

generally were occupied during the whole of the breeding season by overlapping

waves of breeders. In years when breeding effort was large, they tended to break

off in smaller sub-colonies. The total number of sub-colonies on the island

correlated positively with chick production (r2 = 0.85, p = 0.004, n = 10) and with

the number of breeding pairs (r2 = 0.23, p = 0.03, n = 20).

��

�

76

�

��

On 21 September 2004, a three year-old pelican was observed incubating a nest

with two eggs (Photo 10 in Appendix A.2). Two other three-year-old pelican were

observed on Dassen Island in September 2008, one was incubating a nest

(contents not visible) and the other near the colony together with its potential

mate, both in full breeding condition (L Upfold, pers. comm.) (Photo 11 in

Appendix A.2). Overall, a small proportion of breeding pelicans displayed rings

on the breeding colony on Dassen Island. Despite 34 additional resightings of

ringed pelicans on Dassen Island (ranging from two to seven years-old), there

were no other observations of unambiguous breeding. Three more three-year-old

pelicans were seen at foraging sites in full breeding plumage, including the pre-

egg laying knob, crest and bright colours around the eye and on the chest

(J Hofmeyr, pers. comm.).

��

The oldest ringed Great White Pelican recorded in the region was 36 years and

five months old (ringed as a chick in Walvis Bay in December 1972, and re-

trapped in May 2009, also in Walvis Bay). Another pelican ringed during the

same season (1972) in Walvis Bay was resighted in the same locality in

November 2000, at age 28. A solitary un-marked pelican was reported regularly

at the Keiskamma River Estuary, Eastern Cape, since 1969, being resighted at

least annually until January 2009, thus being at least 39 years-old (Tree 1997;

AJ Tree, pers. comm.).

��

After correcting for overdispersion ( = 1.12, SE = 0.02), the model that best

fitted the pelican mark-recapture dataset (Table 3.2) assumed two survival

classes: pelicans in their first year, Age 1, and pelicans that were two year-old

and older, Age 2. It also assumed fixed resighting (P) rates for non-engraved

rings and a year-effect for engraved ones. Recovery (r) and fidelity (F) rates were

allowed to differ for pelicans in their first, second and third age classes. The

QAICc for the next best model (Model 2) differed by 2.7, and had approximately

four times less support than the best model (ratio of QAICc weights).

��

�

77

�

During 2002–2008, survival rates averaged 0.74 ± 0.17 (SE) for first year

pelicans (Figure 3.3). Survival estimates for the second age class were not

estimable for the first year of the study (2002) because there were no ringed

individuals in that age-class at the time. The most reliable survival estimates for

pelicans in the second age class were those calculated for the period 2006–2008

(0.96 ± 0.04 (SE)), because (1) the unusually high mortality experienced by

pelicans in 2005, caused by anthropomorphic changes in the availability of food

subsidies (Chapter 2), (2) the extremely low resighting effort for metal rings

(resighting rate P = 0.04 ± 0.03) and non-engraved colour rings (P = 0.02 ± 0.01),

and (3) individually-identifiable ringed pelicans joined the second age-class only

from 2006. I do not report survival and resighting probabilities for 2009 because

these two parameters were confounded in the fully time-dependent model that

was best supported by the data (Lebreton et al. 1992).

The model estimated different fidelity rates between one and two year-old

pelicans (F = 0.77 ± 0.09 (SE)) and older than two years-old (F = 0.89 ± 0.10

(SE)), which possibly corresponded to the observed pattern of young birds

dispersing further from the ringing site (Chapter 2).

� ��

Ringing effort in southern Africa was concentrated on chicks at Dassen Island in

the Western Cape (South Africa), the Bird Island Guano Platform in Walvis Bay

(Namibia) and Hardap Dam near Mariental (Namibia) (Table 3.1). Once-off

ringing events were undertaken in Lake Ngami (Botswana) and Etosha Pan

(Namibia) in 1971. Resightings and recoveries of ringed birds were concentrated

in four main centres, which corresponded mostly with the ringing localities.

These centres were (1) the south-western coast of the Western Cape including

inland agricultural areas, from the Olifants River estuary on the West Coast and

along the south coast to De Hoop Vlei; (2) the inland population in Namibia

around Hardap Dam, Mariental, (3) the coastal population of Namibia centered

on Walvis Bay; and (4) the Orange River estuary (Figure 3.4). Additionally,

although pelicans are rare in the Eastern Cape, repeated resightings of a ringed

��

�

78

�

individual in the estuaries and coastal habitats near Algoa Bay provided a

relatively high density of resightings at this locality. Intermediate sites showed

sporadic number of resightings and/or recoveries of pelicans (Figure 3.4, grey

circles).

A total of 253 resightings and 164 recoveries of ringed pelicans were recorded

from the onset of the colour ringing programme in 2001 to June 2009 (Table 3.3).

Of the overall total of resightings/recoveries in southern Africa, 37.5%

corresponded to pelicans ringed in the Western Cape and 30.4% in Namibia. The

number of resighting events in the Western Cape was 200-fold larger than in

Namibia, although chick mortality at the Hardap Dam breeding colony (ca. 30–

40% of the chicks ringed in 2004, 2005 and 2006 died before fledging) was 10-fold

larger than mortality at Dassen Island (Table 3.3). A large part of the resighting

effort in the Western Cape was undertaken at one foraging site, the pig farm,

where 218 ringed pelicans were resighted in 981 resighting events since 2002.

Although the causes of mortality of recovered pelicans were not ascertained in all

cases, electrocution or collision with power lines was the reason for at least eight

of the 52 recoveries of ringed pelicans reported in the Western Cape.

Additionally, a shortage in food availability at the pig farm was responsible for

the death of more than 200 pelicans in July 2005, of which 21 had rings.

Observations in the weeks before this catastrophic mortality event indicated that

they were desperately searching for food. Fresh carcasses were 25% under the

average weight for the species for both males and females, and the results of two

post-mortem examinations revealed that the stomachs of both birds were filled

with soil, possibly swallowed while feeding on the water run-off from the farm

processing plant (pers. obs.). Four pre-fledging chicks were found dead near the

breeding colony on Dassen Island and four more shortly after fledging on the

coast of the mainland closest to Dassen Island. Five ringed fledglings were taken

to the Southern African Foundation for the Conservation of Coastal Birds

(SANCCOB) and were successfully rehabilitated.

��

�

79

�

The battery-powered PTT device was deployed on a juvenile male pelican (ringed

as H01215), and the solar-powered one was deployed on an adult female pelican

(H01083). Both pelicans remained within 40 km of the deployment site, in the

agricultural areas of the south-western Cape, for the time they were tracked

(Figure 3.5). The adult female pelican exhibited incipient breeding plumage;

however during the three months that the signal was received, and besides

having been captured at the beginning of the breeding season, it did not visit the

breeding colony on Dassen Island, indicating that either it was a late breeder in

2004 or it did not breed at all.

The latest position received from the battery-powered device was on 11 August

2004, 42 days after deployment. The solar-powered device was last positioned on

7 September 2004, 69 days after deployment. The last two readings received

from the battery-powered satellite transmitter were from the same position, and

within eight days of each other, indicating either that the pelican had died or the

device fell off the bird. The battery was expected to last for six to eight weeks, so

battery depletion cannot be discarded as an explanation for subsequent cessation

of readings received. The pelican which had the solar-powered transmitter was

seen again in 2005, alive but with no device attached, so presumably it fell from

the bird some time after the last position was received.

��

During observations at the colony in September 2004, sibling aggression was

observed between chicks that were pink and naked, and therefore under three

days of age (Brown & Urban 1969). Siblicide was observed at three nests located

at the periphery of the colony while chicks were being brooded by the adult. Both

chicks were of similar sizes, although in each case the chick that repeatedly

attacked the sibling was slightly larger and a darker pink colour. The adult

remained indifferent to the behaviour of its offspring, which continued until the

wounded chick was dead or extremely weak. The interaction among the chicks

consisted of repeated pecking and holding on the neck and head of the younger

chick. This activity continued for more than 1 hr at a time. In two cases, when

��

�

80

�

the chick was seemingly dead, the adult swallowed it. In one case, after 5 min

the adult regurgitated the chick and tried repeatedly to feed it to its sibling,

unsuccessfully during the observation period (Photo 12 in Appendix A.2).

I also observed cannibalism of pelican chicks on two separate occasions. In both

incidents, a brooding adult pelican grabbed a black downy chick (estimated age,

14–28 days (Brown & Urban 1969) from a nearby nest and swallowed it. No

apparent provocation was observed, and it was not possible to identify whether

the chick belonged to an unattended nest or whether it was guarded by its

parents. Nonetheless, no evident reaction was observed from any other pelican

(chick or adult) in the surrounding colony.

Kelp Gulls and occasionally Sacred Ibis Threskiornis aethiopicus preyed on eggs

and small pelican chicks at the Dassen Island colony (Photo 15 in Appendix A.2).

Opportunistic predation by gulls was observed in the absence of obvious

disturbance, often during nest relief or while the incubating pelican stood on the

nest to stretch or change position.

��

As it is common of birds in temperate climates (Lack 1954), pelicans on Dassen

Island synchronised the most energetically-demanding phase of the pelican

breeding cycle with the end of the Western Cape’s rainy season. Although the

breeding season lasted between 7–8 months in the 11 years studied, incubation

and chick-rearing took nearly four months, and only one successful breeding

attempt was possible per pair. During the summer-drought, water levels drop in

wetlands and dams so that fish from deeper waters become available to foraging

pelicans (Guillet & Crowe 1981; 1983). The Western Cape population requires

between 800 and 1 100 tons of fish per year (Chapter 2). Foraging intensity of

adult pelicans increases during November–December (Guillet & Crowe 1983),

when pelican chicks need to acquire a large biomass in a relatively short period

of time. By the end of summer many ephemeral wetlands dry out and fish stocks

��

�

81

�

in the remaining waterbodies are depleted (Guillet & Crowe 1983), suggesting

that food supply could be limited towards the end of the reproductive period

(Furness & Birkhead 1984).

In contrast, pelicans at Hardap Dam, Namibia, bred mainly in the first half of

the year, following abundant rains in this late summer rainfall region. However,

Urban (1984) reported that pelicans breed during June–September at this

locality, as confirmed also by the presence of chicks in August 2008 (D. Heinrich,

in litt.). During drier years, the islands where pelicans breed at Hardap Dam are

joined to the mainland and allow access to land predators, which deters pelicans

from attempting to breed (del Hoyo et al. 1992). In most tropical breeding

localities across Africa, Great White Pelicans usually breed during the dry

season (e.g. Lake Shala, Ethiopia; St Lucia, South Africa), linked to greater fish

abundance; whereas, at localities in which the inaccessibility of the breeding site

by land predators is dependent on heavy rainfall they breed predominantly

during the wet season (e.g. Lake Rukwa, Tanzania) (Brown & Urban 1969). In

Banc d’Arguin, Mauritania, Great White Pelicans breed from the beginning of

July to the end of February, coinciding with the arrival of warm water and

related shallow-water marine fishes from the Gulf of Guinea (Crivelli 1994). In

tropical localities pelicans often breed more than once a year, as has been

observed at Hardap Dam (Theron et al. 2003), or even continuously in a

protracted breeding season, e.g. St Lucia (Berruti 1980) and Lake Shala,

Ethiopia (Brown & Urban 1969).

Great White Pelicans are a gregarious species and many aspects of their life

cycle including reproduction, foraging, roosting and flying require social

interaction (Brown et al. 1982). The onset of breeding is well synchronised,

possibly requiring some amount of social stimuli (Brown & Urban 1969; Guillet

& Crowe 1983). During the breeding cycle it was common to observe two or more

‘waves’ of breeding activity. In the Western Cape, these waves have been

described as discrete; their onset separated by about two months (Guillet &

Crowe 1983), and they could correspond to replacement clutches as well as to

��

�

82

�

new individuals that acquired breeding condition later in the breeding season.

Although most pelican chicks could be assigned to a discrete breeding ‘wave’,

sometimes chicks of intermediate sizes were found, especially when multiple

sub-colonies formed on the island. As the colony grew, more waves were likely to

occur and breeding attempts became less discrete. This behaviour of forming

smaller sub-colonies seems to be common among Great White Pelicans (Brown &

Urban 1969; Crivelli 1994), as well as for other large pelicans like the American

White Pelican P. erythrorhynchos (Knopf 1979) and Dalmatian Pelican P. crispus

(Crivelli et al. 1998b). Among both American and Great White Pelicans new

nests were appended to already existing groups of incubating or brooding birds

(Knopf 1979; Crivelli 1994, this study). Due to space constraints, Australian

Pelicans P. conspicillatus occupy nests immediately after they are vacated due to

failure or once the chick is no longer brooded (Vestjens 1977). Synchronisation of

laying dates is common across the genus, and could provide an adaptive

advantage by preventing territorial harassment of earlier chicks by incubating

adults, although no evidence of a decline in chick production was found in non-

synchronised groups of American White Pelicans (Knopf 1979). Alternative

explanations could be minimising predation risk of chicks and eggs, and

preventing disturbance caused by older chicks moving through the colony.

Great White Pelicans exhibit obligate siblicide. Brood reduction has been

reported on Dassen Island in the Western Cape (Cooper 1980) and in the

Rhukwa Valley, Tanzania (Vesey-Fitzgerald 1957), but as also noted by Cooper

(1980), it has been overlooked in some studies (Brown & Urban 1969; Cramp &

Simmons 1977; Crivelli 1994). It is a common trait of the genus (Table 3.4) and

other Pelecaniformes (Drummond 1987). Brood reduction occurs generally

through ‘sibling-murder’ in the Pink-backed Pelican P. rufescens (Burke & Brown

1970; Din & Eltringham 1974); by differential starving in the Brown Pelican P.

occidentalis (Schreiber 1976); and by starvation or exposure due to sibling

aggression in the American White Pelican (Knopf 1980; Cash & Evans 1986).

Because siblicide occurs independently of food availability, obligate siblicide has

often been explained by the ‘insurance-egg’ hypothesis (Dorward 1962), which

��

�

83

�

states that parents produce more eggs that the number of offspring they can

successfully raise as insurance against early loss of one of the eggs or chicks. The

insurance hypothesis has been supported experimentally for American White

Pelicans (Cash & Evans 1986; Evans 1996), where hatching asynchrony (0–4

days) determines dominance. The older chick has a greater chance of survival,

although occasionally this species does successfully raise two chicks (Knopf 1980;

Evans 1996). The incidents of siblicide reported here happened immediately

after hatching. Brood reduction sometimes occur weeks after hatching (nests

with two siblings between 20–30 days of age were observed), providing some

insurance against chick losses. Hatching asynchrony in the Great White Pelican

is 24–36hrs (Brown & Urban 1969), providing limited competitive advantage to

the older chick. However, no information exists in this case on the differential

egg investment by females (Slagsvold et al. 1984; Williams 1994). A larger

sample size and more extensive research are required to explain this behaviour

from an adaptive perspective.

Cannibalistic behaviour had not been previously reported for Great White

Pelicans. Observations at breeding colonies have been sporadic, yet it seems this

is a rare phenomenon. Although speculative, the causes could be linked to

changes in foraging preferences of pelicans, which have become habitual seabird

predators in the Western Cape (Chapter 4 and 5). Frequently, cannibalism has

been explained as a density-dependence regulator of population size (Polis 1981);

however whether this is the case for pelicans remains unanswered. Cannibalistic

behaviour is not uncommon among wild and captive animals (Polis 1981;

Cloutier et al. 2002) and an isolated instance of an Australian Pelican fledgling

eating pelican nestlings as well as chicks of Australian White Ibis Threskiornis

molucca has been reported (Smith & Munro 2008).

Pre-fledging mortality during the late stage of the breeding period on Dassen

Island was much lower than at Hardap Dam. Besides the three ringed chicks

found dead on Dassen Island since 2001, four additional fledglings were

recovered dead on the mainland beaches adjacent to the breeding colony of

��

�

84

�

Dassen Island. Also, five chicks (two at the time of ringing and three found in

weak condition soon thereafter) were taken to the Southern African Foundation

for the Conservation of Coastal Birds (SANCCOB) to be rehabilitated. In the

absence of human intervention, at least 12 pelicans would have been included in

the figures of pre-fledging mortality for the Western Cape. Still, the contrast

with the levels of mortality at Hardap Dam is evident. Possible causes of

recurring high mortality rates at Hardap Dam could be starvation,

abandonment, heat waves or disease. Fish are not in short supply (van Zyl et al.

1995; CJ Hay, pers. comm.), thus starvation is unlikely, but other hypotheses

require further investigation. Other Great White Pelican colonies experience

massive mortality of chicks, e.g. in the Banc d’Arguin, Mauritania, seasonal

upwelling at the end of the breeding season causes a reduction in the availability

of the shallow-water fish species that are the main prey of pelicans at this

locality, and hundreds of chicks die when adults move away from the breeding

colony (Crivelli 1994).

Age of maturity is thought to be between three and four years-old for Great

White Pelicans (Brown et al. 1982; del Hoyo et al. 1992; Johnsgard 1993), similar

to other pelican species (Table 3.4). The observation of three three-year-old

pelicans breeding on Dassen Island constitutes the first unequivocal

confirmation of breeding at this age for Great White Pelicans. Empirical

relationships among life-table variables of birds calculated by Ricklefs (2000)

reveal that age of maturity in pelicans is earlier than the expected value given

their long life span and low adult mortality, although they fit the trend of

shorter-deference of first breeding by inshore-feeding seabirds (Nelson 2005).

Limited observations of ringed pelicans were made at the breeding colonies;

Dassen Island is one of the few colonies in southern Africa which is readily

accessible, thus research effort should be emphasised if conducted with caution

to avoid disturbance. Because of the increasing proportion of colour-coded pelican

cohorts and of individually-ringed pelicans, observations of the breeding colonies

could yield further information on proportions of cohorts breeding at different

ages, frequency of breeding, incidence and success of re-laying attempts,

��

�

85

�

hatching and fledging success, parental investment and others. Investigating

this link between breeding biology, life history and environmental interactions is

crucial to understand population dynamics for the species.

As expected, pre-reproductive annual survival was proportional to adult

survival, roughly fitting the empirical relationships described by Ricklefs (2000).

However, the empirical relationships between fecundity and mortality or

survival (Ricklefs 1977; 2000) were not in agreement with the data presented

here, indicating that this population was not stationary but was increasing

without apparent density-dependence mechanisms restricting fecundity or pre-

reproductive mortality. Although changes in the availability of food for long-lived

species are expected to have a stronger effect on reproductive parameters than

on survival (Stearns 1992), the sudden reduction in food subsidies negatively

affected survival in 2005, mostly at pre-reproductive age (two years-old). The

limited number of cohorts ringed did not enable me to perform a more detailed

analysis of the age classes involved.

Survival for pelicans in the first year of life may have been underestimated

during 2002–2004, due to the small sample size of individually-identifiable

recapture records; the estimates being calculated almost exclusively from

recoveries of dead birds. On the other hand, both survival estimates were high in

comparison to those obtained for other members of the genus (Table 3.4).

However the methods used to calculate survival for other pelican species did not

take into account recapture probability, resulting in possible underestimates of

survival rates (Strait & Sloan 1974; Ryder 1981; Schreiber & Mock 1988). For

species with such a large distributional range it may be costly to establish

successful mark-recapture programmes, yet the possibilities of modelling

accurate survival estimates and testing biological hypotheses would not be

possible with other methods (Lebreton et al. 1992).

The longevity record of 37–40 years greatly extended the previous record of 23

years and 5 months from Lake Shala, Ethiopia, placing it within the same range

��

�

86

�

(30–40 years) of other pelicans (Table 3.4), and just below that of other long-lived

seabird species like frigatebirds and albatrosses (Clapp & Hackman 1969; Hagen

1982; Douglas & Fernández 1997; Cooper et al. 2003; Kimkiewicz 2008).

The main centres of distribution of pelicans identified in Figure 3.4 corresponded

to the sites in which ringed pelicans have been seen or recovered. Therefore, this

is not a true measure of the pelican distributional range, because it includes

confounding factors like resighting effort and the potential differential spatial

distribution of ringed individuals. Variation in human population density,

concentration of birders near the urban centres, remoteness of the sites

frequented by pelicans and different resighting effort affected the probability of

reporting ringed pelicans. Also, because the colour-ringing program started in

2001, resightings could be biased towards the distribution of younger pelicans.

Despite these considerations, the figure compares quite consistently with the

known distribution of the species; for example with the results of the Southern

African Bird Atlas Project (Figure 6.1; Williams & Borello 1997).

Due to lower resighting effort in Namibia I could not verify whether pelicans

from the Western Cape population dispersed north into Namibia. However, the

more intensive resighting effort in the Western Cape suggested that there was

no immigration into this population, as corroborated by molecular data (Chapter

6). A question that remains unanswered is why pelicans (with the exception of a

limited number of anecdotal resightings of solitary pelicans) do not visit the

seemingly suitable waterbodies along the southern coast of South Africa,

creating a disjunct distribution between De Hoop Nature Reserve in the Western

Cape and Durban in KwaZulu-Natal. With the exception of the islands in Algoa

Bay in the Eastern Cape, appropriate breeding sites (free from land predators

and human disturbance) are not available on this stretch of coast. However,

some coastal estuaries and wetlands in the area sustain fish populations that

may support the communal-foraging habits of the species (S Low, pers. comm.).

Fidelity to the area of breeding was lower for younger birds than ringed adults.

Although more data are needed, young birds seem to disperse further than

��

�

87

�

adults and return to their natal area close to their reproductive age, similar to

findings for American White Pelicans (Ryder 1981).

A larger resighting effort in Namibia is needed in order to estimate survival for

this population, and compare it with the estimates for the subsidised population

in the Western Cape. A continued ringing programme should be maintained for

both colonies.

Great White Pelicans are mostly vulnerable to predation while breeding. When

breeding on mainland sites they are vulnerable to terrestrial predators, such as

mongooses, caracals, jackals, leopards and hyenas. On the West Coast of

southern Africa (i.e. Western Cape, Namibia and Angola), pelicans rely on

offshore islands to breed due to the lack of large waterbodies with inaccessible

islands to breed. Here, their most prominent predators are avian species, viz.

Kelp Gulls and Sacred Ibises, which prey on eggs and small chicks, especially,

but not only, in the event of disturbance to the breeding colony. At other

breeding localities across Africa, pelicans are susceptible of predation by a larger

array of species. Mammalian predators are common on the inland colonies, and

have been reported as the cause of breeding failure at Etosha Pan (Berry et al.

1973). Egyptian Vultures Neophron percnopterus and storks Leptoptilos spp.

have been reported also to prey on eggs and small chicks (Brown & Urban 1969;

Berry et al. 1973).

The demography of long-lived species is sensitive to changes in adult mortality

(Stearns 1992). The main threats affecting adult pelican survival are collision

with power lines, industrial or agricultural pollution, urban run-off, reduction of

fish populations and destruction of habitat (Crivelli & Schreiber 1984; Crivelli et

al. 1998a; Shields 2002; Knopf & Evans 2004; IUCN 2009). Because their

increase in numbers (Chapter 2) and especially due to their recent behavioural

shift into an avian predator of seabird chicks (Chapter 4), the conservation of

Great White Pelicans in the Western Cape currently does not constitute a reason

for concern. On the contrary, pelican predation has created a conservation

��

�

88

�

paradox, in which a protected species threatens the persistence of other

threatened species (Chapter 4 and 5). Nevertheless, I conclude by summarising

key factors in the conservation this charismatic apex predator species in

southern Africa: (1) the protection of breeding sites, free of land predators and

human disturbance; (2) reduction of factors (listed above) that negatively

influence adult survival; (3) and most importantly, the conservation of wetlands

used for foraging and roosting. These measures would be beneficial also for a

large number of other species, thus emphasising the flagship species role of

pelicans for conservation.

��

�

89

�

��

Anderson CK, Burnham KP (1999) Understanding information criteria for

selection among capture-recapture or ring recovery models. Bird Study

46, S14–S21.

ARGOS (2008) User's Manual. CLS/Service Argos, Toulóuse, France.


Lammergeyer 28, 1-19.





Madoqua (Ser. 1) 7, 17–31.

Bowker MB, Downs CT (2008) Breeding incidence of the Great White Pelican

Pelecanus onocrotalus and the Pink-backed Pelican P. rufescens in south-

eastern Africa from 1933 to 2005. Ostrich 79, 23–25.





Burke VEM, Brown LH (1970) Observations on the breeding of the Pink-backed

Pelican. Ibis 112, 499–512.

Burnham KP (1993) A theory for combined analysis of ring recovery and

recapture data. In: Marked individuals in the study of bird populations

(eds. Lebreton J-D, North PM), pp. 199–213. Birkhäuser Verlag, Basel,

Switzerland.

Burnham KP, Anderson DR (2002) Model selection and multimodel inference: a

practical information-theoretic approach. 2nd edn. Springer, New York.

Cash KJ, Evans RM (1986) Brood reduction in the American White Pelican

Pelecanus erythrorhynchos. Behavioral Ecology and Sociobiology 18, 413–

318.

Caswell H (2001) Matrix population models. Sinauer, Sunderland.

Clapp RB, Hackman CD (1969) Longevity record of a breeding Great Frigatebird.

Bird Banding 40, 47.

Cloutier S, Newberry RC, Honda K, Alldredge JR (2002) Cannibalistic behaviour

spread by social learning. Animal Behaviour 63, 1153–1162.

Cooper J (1980) Fatal sibling aggression in pelicans– a review. Ostrich 51, 183–

186.

Cooper J, Battam H, Loves C, Milburn PJ, Smith LE (2003) The oldest known

banded Wandering Albatross Diomedea exulans at the Prince Edward

Islands. African Journal of Marine Science 25, 525–527.

Cramp S, Simmons KEL (1977) Family Pelecanidae: White Pelican. In:

Handbook of the birds of Europe, the Middle East, and North Africa: the

birds of the Western Palearctic (ed. Cramp S), vol. 1: Ostrich–Ducks, pp.

226–233. Oxford University Press, Oxford.

��

�

90

�



Africa's Western Cape. African Journal of Marine Science 15, 33–42.




Mitchev T, Nazirides T, Peja N, Sarigul G, Siki M (1998a) Status and

population development of Great White Pelican Pelecanus onocrotalus

and Dalmatian Pelican P. crispus breeding in the Palearctic. In:

Monitoring and conservation of birds, mammals and sea turtles of the

Mediterranean and Black Seas. Proceedings of the 5th Medmaravis

Symposium (eds. Yésou P, Sultana J), pp. 38–45. Environment Protection

Department, Gozo, Malta.

Crivelli AJ, Hatzilacou D, Catsadorakis G (1998b) The breeding biology of the


Crivelli AJ, Schreiber RW (1984) Status of the Pelecanidae. Biological





backed Pelicans in the Ruwenzori National Park, Uganda. Ibis 116, 28–

43.

Dodman T (in press) Status, estimates and trends of waterbird populations in

Africa. Wetlands International, Dakar.

Dorward DF (1962) Comparative biology of the White Booby and the Brown

Booby Sula spp. at Ascension. Ibis 103, 174–200.

Douglas HD, Fernández P (1997) A longevity record for the Waved Albatross.

Journal of Field Ornithology 68, 224–227.

Drummond H (1987) A review of parent-offspring conflict and brood reduction in

the Pelecaniformes. Colonial Waterbirds 10, 1–15.





Breeding specialist group, Cape Town.

Evans RM (1996) Hatching asynchrony and survival of insurance offspring in an

obligate brood reducing species, the American White Pelican. Behavioral

Ecology and Sociobiology 39, 203–209.

Furness RW, Birkhead TR (1984) Seabird colony distributions suggest

competition for food supplies during the breeding season. Nature 311,

655–656.

Furness RW, Monaghan P (1987) Seabird Ecology. Blackie, Glasgow and London.

Guillet A (1985) Biogeography and Ecology of African Waterbirds. Doctoral

thesis, University of Cape Town.



��

�

91

�




Hagen Y (1982) Migration and longevity of Yellow-nosed Albatrosses Diomedea

chlororhynchos banded on Tristan da Cunha in 1938. Ornis

Scandinavica, 247–248.

Hamer KC, Schreiber EA, Burger J (2002) History-environment interactions in

seabirds. In: Biology of marine birds (eds. Schreiber EA, Burger J), CRC

Press, Boca Raton, Florida.



IUCN (2009) IUCN Red List of Threatened Species, version 2009.1.

www.iucnredlist.org.



Kimkiewicz MK (2008) Longevity records of North American Birds, version

2008.1. Patuxent Wildlife Reserach Center, Bird Banding Laboratory,

Laurel.

King DT, Paulson JD, Leblanc DJ, Bruce K (1998) Two capture techniques for

American White Pelicans and Great Blue Herons. Colonial Waterbirds

21, 258–260.

Knopf FL (1979) Spatial and temporal aspects of colonial nesting of White

Pelicans. Condor 81, 353–363.

Knopf FL (1980) On the hatching interval of White Pelican eggs. Proceedings

Oklahoma Academy of Science 60, 26–28.

Knopf FL, Evans RM (2004) American White Pelican Pelecanus erythrorhynchos.

In: The birds of North America Online (ed. Poole A), Cornell Lab of

Ornithology, Ithaca. http://bna.birds.cornell.edu/bna/species/057.

Lack D (1954) The natural regulation of animal numbers. Oxford University

Press, London.

Lebreton J-D, Burnham KP, Clobert J, Anderson DR (1992) Modeling survival

and testing biological hypotheses using marked animals: a unified

approach with case studies. Ecological Monographs 62, 67–118.



Phillips RA, Xavier JC, Croxall JP, Burger AE (2003) Effects of satellite

transmitters on albatrosses and petrels. Auk 120, 1082–1090.

Polis GA (1981) The evolution and dynamics of intraspecific predation. Annual

Review of Ecology and Systematics 12, 225-251.

Ricklefs RE (1977) On the evolution of reproductive strategies in birds:

reproductive effort. American Naturalist 111, 453–478.

Ricklefs RE (2000) Density dependence, evolutionary optimization, and the

diversification of avian life histories. Condor 102, 9–22.

Ryder RA (1981) Movements and mortality of White Pelicans fledged in

Colorado. Colonial Waterbirds 4, 72–76.

Schreiber RW (1976) Growth and development of nestling Brown Pelicans. Bird

Banding 47, 19–39.

��

�

92

�

Schreiber RW, Mock PJ (1988) Eastern Brown Pelicans: what does 60 years of

banding tell us? Journal of Field Ornithology 59, 171–182.

Shields M (2002) Brown Pelican Pelecanus occidentalis. In: The Birds of North

America Online (ed. Poole A). Cornell Lab of Ornithology, Ithaca.


Slagsvold TJ, Sandvik J, Rofstad g, Lorentsen Ö, Husby M (1984) On the

adaptive value of intraclutch egg-size variation in birds. Auk 101, 685–

697.

Smith ACM, Munro U (2008) Cannibalism in the Australian Pelican Pelecanus

conspicillatus and Australian White Ibis Threskiornis molucca.


Stearns SC (1992) The evolution of life histories. Oxford University Press,

Oxford.

Strait LE, Sloan NF (1974) Life table analysis for the White Pelican. Inland Bird

Banding News 46, 20–28.


Pelicans Pelecanus onocrotalus at Hardap Dam, Namibia. Lanioturdus

36, 2–7.

Tree AJ (1997) The Keishama White Pelican. Bee-Eater 48, 45–46.

Urban EK (1984) Time of egg-laying and number of nesting Great White

Pelicans at Lake Shala, Ethiopia, and elsewhere in Africa. In:

Proceedings of the Fifth Pan-African Ornithological Congress, Lolongwe,

Malawi, 1980 (ed. Ledger L), pp. 809–823. Southern African

Ornithological Society, Johannesburg.

Urban EK, Ash JS (2001) Longevity record of a Great White Pelican, Pelecanus

onocrotalus, from Lake Shala, Ethiopia. Ostrich 72, 114–126.

van Zyl BJ, Hay CJ, Steyn GJ (1995) Some aspects of the reproduction biology of

Labeo capensis (Smith, 1941) (Pisces, Cyprinidae) in relation to

exploitation and extreme environmental conditions in Hardap Dam,

Namibia. South African Journal of Aquatic Science 21, 88–98.

Vesey-Fitzgerald D (1957) The breeding of the White Pelican Pelecanus

onocrotalus in the Rukwa Valley, Tanganyika. Bulletin of the British

Ornithological Club 77, 127–129.

Vestjens WJM (1977) Breeding behaviour and ecology of the Australian Pelican,

Pelecanus conspicillatus, in New South Wales. Australian Wildlife

Research 4, 37–58.





Williams TD (1994) Intraspecific variation in egg size and egg composition in

birds: effects on offspring fitness. Biological Review 68, 35–59.

White GC, Burnham KP (1999) Program MARK: survival estimation from

populations of marked animals. Bird Study 46, S120-S139.

93

Table 3.1: Great White Pelicans ringed in southern Africa from 1970 to 2008.

Cohort-based colour rings were fitted from 2001 on Dassen Island in the Western

Cape. Engraved plastic rings were used from 2004 on Dassen Island and from

2006 at Hardap Dam in order to allow individual identification of pelicans.

Ringing locality Season Ring colour Leg

(colour)

Numbe

r

ringed

Age

(number)

Metal rings:

Southern Africa 1970–2000 only metal n/a 161 a C (155), A (6)

South Africa 2001–2008 only metal n/a 44 C (37), A (7)

Namibia 2001–2008 only metal n/a 313 b C (310), A (3)

Colour rings:

Dassen Is., South Africa 2001 Blue - 100 C

Dassen Is., South Africa 2002 Red - 155 C

Dassen Is., South Africa 2003 Yellow - 53 C

Engraved rings:

Dassen Is., South Africa 2004 Green Left 39 C

Pig farm, South Africa * 2004 White Left 44 A

Dassen Is., South Africa 2005 Black Left 75 C

Dassen Is., South Africa 2005 White Right 36 C

Dassen Is., South Africa 2006 Pink Left 65 C

Dassen Is., South Africa * 2007 Blue/yellow - 100 C

Dassen Is., South Africa 2008 Green + White Right 54 c C

Dassen Is., South Africa 2008 Green + Pink Right 27 c C

Hardap dam, Namibia * 2006, 2008 Red/green - 98 C

Coastal, Namibia 2006, 2008 Red/green - 5 A

Chatanoga, Botswana 2007 White + Pink - 1 C

TOTAL 1370

C: chicks; A: adult / immature

(*) additional individuals were fitted metal rings only; these are included in the metal ring totals.

a 34 at Lake Ngami, Botswana; 13 at Etosha Pan, Namibia; 50 at Bird Rock, Walvis Bay,

Namibia.

b 306 were ringed at Hardap Dam as pre-fledgling chicks.

c small chicks were only fitted one colour (engraved) ring.

98

Figure 3.2: Dassen Island, showing the annual location of breeding colonies of

Great White Pelicans from 1995 to 2007 and a summary for all years combined.

The maximum nest count each year is indicated in parentheses.

�

�

�� !��"��#�#��$%��%�$�� %&��$$�$�� %��%��$��%��

��$%��$��%��'��()��*��$%�'��$��#�#��!��( � ��

��$"��$��'��!��* ��$"��#��$��#�#��%��

%�� $%�'��$$��$�+�,-�!�$��"�� %��$$��$�+�./��

*++/0*++1�!��%%��"��!��%%��"��%��%�� $�� $�

!��$��%��$"��$��'��%�!��$2��$"��

99

Figure 3.4: Density of resighting events (n = 1453) and recoveries (n = 184) of

ringed Great White Pelicans in southern Africa from 1971 to 2009. Darker

shaded areas indicate higher density of resightings/recoveries and grey circles

represent individual resighting or recovery events. Numbers are referred to in

the text.

Km 1

3

2

4

��3��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

� !�"##$%�&�'��(��

��

��4

5��

�&)'��)��(��

�� $"

��

��*��+� ��

��

�� )��

�#��

� 3##

��&,-�./� �� !�"!��'!��

��

��

��

�� &��6

� 3##

��%�,-�./� ��#'�0�� +��1��

� �� %��

��

��

��

1��

�

CHAPTER 4

��

��

��

103

��

��

��

Behavioural flexibility typical of opportunistic species allows animals to

adjust their foraging strategies to adapt to fluctuating environments

(Timmermans et al. 2000; Lefebvre et al. 2004). Innovation rate in the wild

has been suggested as a good proxy for behavioural plasticity (Wyles et al.

1983; Lefebvre et al. 1997; 2001; Timmermans et al. 2000). Once an

innovative feeding strategy has been acquired, social learning constitutes

the optimal mechanism for spreading the new behaviour because it

enables rapid responses to environmental changes (Giraldeau et al. 1994;

Lefebvre 1995b; Galef & Giraldeau 2001; Duckworth 2009). On the other

hand, human-induced changes, including food subsidisation to

opportunistic species, may affect the functioning of ecosystems and could

lead to changes in food web dynamics (Pace et al. 1999; Jefferies 2000).

Agricultural subsidies can enable consumer populations to achieve greater

numbers than those supported by in situ productivity, which in turn may

increase predation pressure on prey species, occasionally leading to

depletion of prey populations (Polis et al. 1997; Jefferies 2000; Votier et al.

2004). Pelicans typically feed on fish and other aquatic organisms,

occasionally scavenging on carcasses or fish offal (del Hoyo et al. 1992).

However, an increasing proportion of Great White Pelicans Pelecanus

onocrotalus in South Africa’s Western Cape exploit novel sources of food:

agricultural offal (Crawford et al. 1995; Chapter 2), and as shown here,

seabird chicks. Locally, pelicans increasingly rely on avian prey during the

breeding season, switching from a mostly piscivorous menu to a

��

104

terrestrial-based diet. Predation on seabird chicks has escalated in extent,

geographical coverage and number of prey-species targeted. An artificially

subsidised pelican population, limited abundance of fish prey, and

abundance of ground-nesting seabirds likely explain this behaviour, which

in turn threatens local populations of seabirds. This study summarises the

different hunting methods utilised by pelicans to target different species of

seabirds, adapting cooperative techniques used to herd fish to target

seabird chicks on land, and suggests that cultural transmission could play

a key role in the spread of this behaviour.

Great White Pelicans are specialist feeders, eating fish and aquatic organisms

throughout their distributional range (del Hoyo et al. 1992) in fresh-water lakes,

estuaries and shallow marine bays (Brown & Urban 1969; Crivelli 1994; Crivelli

et al. 1997). South Africa’s Western Cape pelican population occupies the

southern limit of their range and has been historically dependent on fish from

estuaries and small, sometimes ephemeral, waterbodies (Guillet & Crowe 1983;

Guillet & Furness 1985).

The Western Cape population grew exponentially from 25 pairs in the 1950s to

more than 600 in the 2000s (Crawford et al. 1995; Chapter 2). Availability of fish

near the breeding colony probably limited population size and the four-fold

increase in numbers between 1980 and 2005 has been attributed to

anthropogenic food subsidies (Box 1). In this period, up to 65% of the pelican

population regularly fed on chicken Gallus gallus offal, until a halt in

availability in 2005 forced pelicans to find alternative food resources (Chapter 2).

Pelican predation on seabirds had been observed occasionally in south-western

Africa from the 1800s. Incidents were isolated, involved solitary pelicans, and

were limited to a few localities adjacent to the pelican’s breeding colony in

western South Africa (Crawford et al. 1995; de Ponte Machado 2007) and at

Walvis Bay, Namibia (Berry 1976). Anecdotal reports of avian predation by

solitary pelicans also exist for Australian P. conspicillatus (Marchant & Higgins

��

105

1990; Smith & Munro 2008) and Brown Pelicans P. occidentalis (Mora 1989;

Shields 2002; Knechtel 2003).

Cooperative hunting of seabird chicks by pelicans was first observed on Dassen

Island in the 1990s, targeting Cape Cormorants Phalacrocorax capensis and

Kelp Gulls Larus dominicanus (Crawford et al. 1995). The number of pelicans

involved in these predation sprees was small and their impact limited. Although

the trigger for this behavioural shift is unknown, food shortage near the breeding

colony due to the population’s rapid expansion is plausible explanation.

Following the cessation of chicken subsidies in 2005, the consumption of seabird

chicks by coordinated groups of pelicans escalated in the Western Cape,

becoming a frequent behaviour during the breeding season. Pelicans targeted

novel sites in search of seabird chicks, and ate additional species not previously

consumed (de Ponte Machado 2007).

The mechanism of spread of this innovative feeding behaviour appears to be

cultural transmission, either by contagion (instinctive behaviour that is

triggered by observing others performing it) or through local or social

enhancement (Giraldeau et al. 1994; Galef & Giraldeau 2001; Cloutier et al.

2002). The transmission of cannibalistic or predatory behaviour through social

learning has been demonstrated experimentally in captive chickens by a

combination of stimulus enhancement and observational conditioning (Cloutier

et al. 2002). Social transmission could explain the spread of cannibalistic

behaviour in wild populations, given adequate environmental conditions and

sufficient cognitive aptitudes (Cloutier et al. 2002). It has been used to explain

the spread of behaviours like the opening of milk bottles by Blue Tits Parus

caeruleus (Fisher & Hinde 1949; Lefebvre 1995b), lobtail feeding in Humpback

Whales Megaptera novaeangliae (Weinrich et al. 1992), beach hunting by Killer

Whales Orcina orca (Guinet & Bouvier 1995) and a diversity of foraging

behaviours in primates (Kawai 1965; Lefebvre 1995a). Pelicans are unique in

that the learned behaviour is itself a cooperative behaviour.

��

106

Cooperative hunting is commonly used by Great White Pelicans to capture fish

(del Hoyo et al. 1992; Saino et al. 1994). When preying on colonies of seabirds,

pelicans exhibit a similar degree of coordination, suggesting that communal

hunting techniques have been adapted from water to the land. They adopt

different foraging strategies depending on the location and density of prey

species and the defensive responses of adults and chicks (Figure 4.1; Table 4.1).

Targeted seabirds are generally ground-nesting, colonial breeders. Although

more than 10 species were eaten, Cape Cormorants Phalacrocorax capensis, Kelp

Gulls Larus dominicanus, Cape Gannets Morus capensis and Swift Terns Sterna

bergii were their main prey. Up to 13 pelicans (mean 4.7 ± 2.9 (SD), n= 94

observations) surrounded clusters of Cape Cormorant nests (Figure 4.1a),

capturing chicks from under adults or chasing larger mobile chicks on the

ground, whereas up to 77 individuals (14.0 ± 11.7, n = 88 observations) hunted

for Kelp Gull chicks. Tight groups of pelicans walked among gull nests (Figure

4.1b), searching for chicks in crevices and hiding places. Frequently, they aligned

in an advancing line or in V-formation (Figure 4.1c), similar to the configuration

used to herd fish in shallow water. Adult Kelp Gulls displayed the most

aggressive responses towards pelicans, occasionally inflicting injuries or

successfully chasing them away. Due to high breeding density of Cape Gannet

colonies, pelicans targeted chicks on the periphery of the colony (Figure 4.1d;

4.1e). Pelican group size averaged 7.9 ± 1.2 (max = 20, n = 211 observations) and

pelicans were cautious of adult gannets, even though they did not show defensive

responses towards pelicans. A different technique was observed when preying on

Swift Terns chicks, which aggregate in crèches five days after hatching (del Hoyo

et al. 1992). Pelicans herded and encircled groups of chicks (pelican group size ca.

15–20), synchronously catching and swallowing them (Figure 4.1f) while adult

terns flew above. Occasionally, pelicans were observed chasing and capturing

free-swimming tern chicks in the water near the natal colony. Foraging

approaches largely differed when targeting mobile prey (gulls and terns) and

more sedentary chicks (cormorants and gannets); the former benefited from

group hunting and the latter do not require herding but may require a few

pelicans to displace the parents.

��

107

Pelicans’ time of breeding is consistent with Lack’s prediction (1954) that states

that birds in temperate climates breed at times when food availability is largest.

Pelican reproduction in the Western Cape is synchronised with the period of

maximum fish availability during the aestival drought (Guillet & Crowe 1983;

Chapter 3). This also coincides with the peak breeding period for most species of

seabirds in the southern Benguela region (Figure 4.2). The number of pelicans

involved in predation on three islands in Saldanha Bay fluctuated from October

to January (mean = 89, range 13–232; Figure 4.3) moving between islands

according to the availability of seabird chicks of the appropriate age classes.

Pelicans caused complete breeding failure of Cape Cormorants and Kelp Gulls on

Jutten and Schaapen Islands during the 2006/07 breeding season and ate up to

8 400 Cape Gannet chicks (Table 4.2). Pelicans disturbed Bank Cormorants

Phalacrocorax neglectus nesting on Dassen Island and probably ate chicks on

Jutten Island. Predation on seabirds occurred predominantly when pelicans were

breeding, possibly linked to their larger energetic demands and/or depletion of

other prey resources near the breeding colony (Furness & Birkhead 1984; Birt et

al. 1987; Lewis et al. 2001). Cape Cormorants undertook a second breeding

attempt on Jutten Island in March–April 2007 (Figure 4.2); a maximum of 22

pelicans visited the colony and had little impact on breeding success (15 015

chicks fledged from 8 356 nests = 1.8 fleglings per nest).

It is not known what proportion of Western Cape pelicans were involved in

seabird predation. About 15% of pelicans were individually ringed in 2006, but

insufficient observations of marked pelicans feeding on seabird chicks were made

to draw conclusions. Following the cessation of food subsidies to pelicans in 2005,

predatory behaviour increased faster than would be expected from genetic

transmission or individually acquired learning (Lefebvre 1995a; b). The

cooperative hunting methods of pelicans imply cognitive capacities and learning

abilities that favour the spread of novel foraging behaviour (Sol et al. 2005;

Marino et al. 2007; Lefebvre & Sol 2008), yet this is the first time that communal

feeding has been observed in a terrestrial environment.

��

108

In conclusion, pelicans in the Western Cape have developed a unique foraging

behaviour, both in their choice of prey and in their displayed cooperative

strategies. Dietary subsidization in the form of agricultural offal facilitated

population growth and led to a prey switch to chicks of ground-nesting colonial

seabirds, targeted mostly during pelican’s breeding period. Behavioural

plasticity and social learning may have triggered the spread of innovative

foraging behaviour in the population. Pelican predation could have detrimental

mid-term effects for local seabird populations.

��

��

Pelican predation was monitored on four islands off the west coast of South

Africa: Dassen (33°25’S, 18°05’E), Malgas (33°03’S, 17°55‘E), Jutten (33°05’S,

17°57‘E) and Schaapen (33°05’S, 18°01’ E) between 2004 and 2007. Intensive

observations (221 hrs) were carried out over 32 days during the austral summer

of 2006/07. Observations were made using 10�42 binoculars and a 30–60�

telescope and documented by means of still photography and video footage. Focal

group methodologies (Altmann 1974) were used to quantify number, size and

species of chicks eaten, pelican group size, predation rate, and agonistic and

defensive behaviour by the pelicans. Scans of the total number of pelicans on the

island were undertaken at 60 min intervals. Seabird breeding success was

monitored by regular counts of nests, as well as brooding chicks in selected

colonies and fledglings over the whole island.

� ��

Pelican numbers on Malgas, Jutten and Schaapen Islands were monitored daily

from 1 October 2006 to 31 January 2007, and frequently until May 2008.

Predatory pelicans involved in predation events on their breeding island

��

109

(Dassen) or other sites were not included in these estimates. On Malgas and

Jutten Islands, they were counted from advantage points on top of one of the

buildings or from the highest hill, respectively. Schaapen Island was monitored

from the stretch of coast close to the island (400–800 m); only the eastern side

was visible. The maximum simultaneous count for each day on each island was

considered for the estimates. To account for missing counts, I used the software

package TRIM with the default options (Pannekoek & van Strien 1996) to

impute missing data values and I plotted the estimated daily numbers of

pelicans on each island, and the estimated total counts.

��

In order to calculate the impact of pelican predation on seabirds (see Table 4.2), I

used estimates of breeding success for each species as summarised by Hockey et

al. (2005). Cape Cormorant breeding success fluctuates among years depending

on environmental factors, mostly on pelagic fish availability and distribution. In

years of good food availability, 0.87 of the eggs hatch and 0.91 of the chicks

fledge. During 2006/07, the section of the Cape Cormorant colony nesting on the

eastern side of Malgas Island was protected from pelican predation due to the

presence of researchers on the island which deterred pelicans from approaching

these nests. Observed chick production at this Cape Cormorant colony was 1.8

fledglings per nest (L Pichegru, pers. comm.). This was the estimated used for

the calculations on Table 4.2. The maximum expected chick production for Cape

Cormorants [number of nests � 2.82 (modal clutch size) � 0.87 (hatching success)

� 0.91 (fledgling success)] would be 1.2% higher than the value calculated from

the protected colony. Crowned Cormorant expected chick production is 1.3

fledglings per nest. Kelp Gull productivity varies from 0.16–1.33 fledglings per

nest (here I used the average value, 0.7). Cape Gannet breeding success is 0.42

fledglings per nest (RJM Crawford, unpublished data) and Swift Tern breeding

productivity is 0.48 fledglings per nest. Other major causes of mortality were

excluded by inspecting breeding colonies for signs of disease or abandonment

��

110

(cadavers of adults and/or chicks). Direct observations of pelican predation and

proportion of Cape Cormorants fledged on Malgas Island (see above) were also

used to confirm whether pelican was predation the cause of breeding failure on

the monitored islands during 2006/07.

��

A total of 28 partially digested stomach contents were collected opportunistically

from a pelican roost on Jutten Island in November 2006, as well as 20 pellets

from Jutten, Malgas and Schaapen Islands. The number of prey items (by

species), size and weight to the nearest 0.1 g were quantified.

��

A total of 728 chicks and 57 adult pelicans were ringed between 2001 and 2006

in the Western Cape, South Africa. The percentage of pelicans individually

ringed that were present in the population during the studied period in 2006 was

calculated considering survival rates of 0.74 for the first year of age and 0.96 for

the second year and older (Chapter 3), and a total population in 2006 of 2 500

birds (Chapter 2).

��

�

111

��

Altmann J (1974) Observational study of behavior: sampling methods. Behaviour

49, 227–267.

Berry HH (1976) Physiological and behavioural ecology of the Cape Cormorant

Phalacrocorax capensis. Madoqua (Ser. 4) 9, 5–55.

Birt VL, Birt TP, Goulet D, Cairns DK, Montevecchi WA (1987) Ashmole's halo:

direct evidence for prey depletion by a seabird. Marine Ecology Progress

Series 40, 205–208.


Pelecanus onocrotalus roseus at Lake Shala, Ethiopia. Ibis 111, 199-237.

Cloutier S, Newberry RC, Honda K, Alldredge JR (2002) Cannibalistic behaviour

spread by social learning. Animal Behaviour 63, 1153–1162.



Western Cape. African Journal of Marine Science 15, 33–42.



Crivelli AJ, Catsadorakis G, Nazirides T (1997) Pelecanus onocrotalus White









Unit, Cape Town.



Duckworth R (2009) The role of behavior in evolution: a search for mechanism.

Evolutionary Ecology 23, 513–531.

Furness RW, Birkhead TR (1984) Seabird colony distributions suggest

competition for food supplies during the breeding season. Nature 311,

655–656.

Galef BG, Giraldeau L-A (2001) Social influences on foraging in vertebrates:

causal mechanisms and adaptive functions. Animal Behaviour 61, 3–15.

Giraldeau L-A, Caraco T, Valone TJ (1994) Social foraging: individual learning

and cultural transmission of innovations. Behavioral Ecology 5, 35–43.






Zoology 205, 573–583.



��

112



Knechtel HA (2003) Brown Pelican disturbances at two central Californian

Common Murre colonies. Pacific seabirds 30, 42.

Fisher J, Hinde RA (1949) The opening of milk bottles by birds. British Birds, 42:

347–357.

Guinet C, Bouvier J (1995) Development of intentional stranding hunting

techniques in Killer Whale Orcina orca calves at Crozet Archipelago.

Canadian Journal of Zoology 73, 27–33.

Kawai M (1965) Newly-acquired pre-cultural behavior of the natural troop of

Japanese monkeys on Koshima islet. Primates 6, 1–30.

Lack D (1954) The natural regulation of animal numbers. Oxford University

Press, London.

Lefebvre L (1995a) Culturally-transmitted feeding behaviour in primates:

Evidence for accelerating learning rates. Primates 36, 227–239.

Lefebvre L (1995b) The opening of milk bottles by birds: Evidence for

accelerating learning rates, but against the wave-of-advance model of

cultural transmission. Behavioural Processes 34, 43–54.

Lefebvre L, Juretic N, Nicolakakis N, Timmermans S (2001) Is the link between

forebrain size and feeding innovations caused by confounding variables? A

study of Australian and North American birds. Animal Cognition 4, 91–97.

Lefebvre L, Reader SM, Sol D (2004) Brains, innovations and evolution in birds

and primates. Brain, Behavior and Evolution 63, 233–246.

Lefebvre L, Sol D (2008) Brains, lifestyles and cognition: are there general

trends? Brain, Behavior and Evolution 72, 135-144.

Lefebvre L, Whittle P, Lascaris E, Finkelstein A (1997) Feeding innovations and

forebrain size in birds. Animal Behaviour 53, 549–560.

Lewis S, Sherratt T, N, Hamer K, C, Wanless S (2001) Evidence of intra-specific

competition for food in a pelagic seabird. Nature 412, 816–819.

Marchant S, Higgins PJ (1990) Handbook of Australian, New Zealand and

Antarctic birds. Oxford University Press, Melbourne.

Mora MA (1989) Predation by a Brown Pelican at a mixed-species heronry.

Condor 91, 742–743.

Musangu, M (2007) The effect of Great White Pelican predation on breeding

seabirds at Dassen Island, South Africa. MSc thesis. University of Cape

Town.

Marino L, Connor RC, Fordyce RE, Herman LM, Hof PR, Lefebvre L, Lusseau D,

McCowan B, Nimchinsky EA, Pack AA, Rendell L, Reidenberg JS, Reiss

D, Uhen MD, Van der Gucht E, Whitehead H (2007) Cetaceans have

complex brains for complex cognition. PLoS Biology 5, e139.





Netherlands.

Polis GA, Anderson WB, Holt RD (1997) Toward an integration of landscape and

��

113

food web ecology: the dynamics of spatially subsidized food webs. Annual

Review of Ecology and Systematics 28, 289–316.

Saino N, Fasola M, Waiyaki E (1994) Do White Pelicans Pelecanus onocrotalus

benefit from foraging in flocks using synchronous feeding? Ibis 137, 227–

230.

Shields M (2002) Brown Pelican Pelecanus onocrotalus. In: The Birds of North

America Online (ed. Poole A). Cornell Lab of Ornithology, Ithaca.


Smith ACM, Munro U (2008) Cannibalism in the Australian Pelican Pelecanus

conspicillatus and Australian White Ibis Threskiornis molucca. Waterbirds

31, 632–635.

Sol D, Duncan RP, Blackburn TM, Cassey P, Lefebvre L (2005) Big brains,

enhanced cognition, and response of birds to novel environments.

Proceedings National Academy of Sciences 102, 5460–5465.

Souci SW, Fachmann W, Kraut H (2000) Food composition and nutrition tables.

Medpharm Scientific Publishers, Stuttgart.

Timmermans S, Lefebvre L, Boire D, Basu P (2000) Relative size of the

hyperstriatum ventrale is the best predictor of feeding innovation rate in

birds. Brain, Behavior and Evolution 56, 196–203.





Weinrich MT, Schilling MR, Belt CR (1992) Evidence for acquisition of a novel

feeding behaviour: lobtail feeding in Humpback Whales, Megaptera

novaeangliae. Animal Behavior 44, 1059–1072.

Wyles JS, Kinkel JG, Wilson AC (1983) Birds, behavior and anatomical

evolution. Proceedings National Academy of Sciences 80, 4349–4397.

��

115

Box 1: Discontinuity of offal increases pelican predation on seabirds

The Western Cape Great White Pelican population was estimated at 2 700 individuals

in 2004. Pelican daily food requirements are 0.8–1.2 kg, about 10% of adult mass (Brown

& Urban 1969). The breeding period consists of 150 days on average, during which time

a pair and their chick would require 420 kg of fish (Brown & Urban 1969). A breeding

population of 660 pairs and their chicks would consume 214–277 tons of fish. Food

consumption for breeders and non-breeders is 790–1 105 tons of fish per year. Assuming

a purely piscivorous diet and a calorific value of fish of 4 kJ g-1 (Guillet & Furness 1985),

the current population requires 3.16 x 109–4.42 x 109 kJ per year. Fish biomass

available to pelicans in the Western Cape may be insufficient to sustain the current

pelican population, but chicken offal has provided an excess of food to scavenging

pelicans since the mid 1980s. Up to 18 tons of offal were available daily at a single site

from 2000, which amounts to 1.58 x 108 kJ (times 365 per year) for calorific values of 8.8

kJ for chicken (Souci et al. 2000). Given a field metabolic rate per pelican per day of

3 300 kJ (Guillet & Furness 1985), offal provided enough sustenance for the 1 200

pelicans that visited the site daily. Since 2005, offal was reduced to a maximum of 200

kg per week (Chapter 2) and pelicans increasingly relied on seabirds during the

breeding season. During the austral summer of 2006/07, an average of 89 pelicans

[range 13–232] were observed feeding exclusively on seabirds, giving an energetic

requirement of 3.5 x 106 kJ (that is, 3 300 kJ x 89 pelicans x 120 days) during the four

months pelicans relied on seabirds for maintenance and activity. Considering 10.9 kJ g-1

as average caloric content of seabirds (Votier et al. 2004), 89 pelicans require 3 230 kg of

seabirds from October to January. Pelican regurgitations showed that a pelican could

eat up to nine Cape Cormorant chicks, or 23 Swift Terns (Table 4.3). Bird prey sizes

fluctuated between 50 g (small cormorant chicks) to 2 kg (large gannet chicks) and the

number of chicks preyed ond varied in function of the prey species and the stage of

breeding (Box 1 Table 1).

�

116

Box 1 Table 1: Estimated number of chicks eaten by Great White Pelicans

according to their energetic requirements on the islands of Saldanha Bay during

the breeding season 2006/07.

Prey species

Estimated

average prey

weight (g)

Observed

pelican

numbers

Required

prey mass /

day (g)

Number

chicks

eaten/day

Predation

(days)

Estimated

number of

chicks

eaten

Cape Cormorant * 90 150 45 412.8 505 30 15 138

Kelp Gull * 200 150 45 412.8 227 30 6 812

Cape Gannet † 250 56 16 954.1 68 60 4 069

Swift Tern § 100 22 6 660.6 67 15 999

* Cape Cormorant and Kelp Gull figures correspond to the island of Jutten during October and

November 2006, respectively. † Cape Gannet figures correspond to chicks eaten during November

and December 2006 on Malgas Island. § Swift Tern figures correspond to the chicks eaten during

March–April 2007 on Schaapen Island.

117

Table 4.1: Great White Pelican main prey species showing their nesting

characteristics and defensive responses to intruders (based on Hockey et al.

(2005) and personal observations).

Prey species

(conservation

category and

trends) †

Nest location Nest density Adult defensive

response

Chick defensive

response

Cape

Cormorant

(NT; declining)

Mostly at

ground level, on

rocks and

artificial

structures

Dense colonies,

often in large

clusters

(3.1–5.5 / m2)

Shaking head, back

and forth movement

of neck, pointing to

intruder with bill.

Groaning noises

None when small.

Large chicks run

away en masse

from intruders

Kelp Gull

(LC; stable)

On the ground,

flat areas and

hills

Sparsely

distributed

(1–4 / m2)

Dive-bombing

intruders, attacking

with legs and beak.

Loud alarm call

Hiding among

bushes and rock

crevices

Cape Gannet

(VU, declining)

On flat

sandy/muddy

ground

Dense colonies

(2.75–6.25 / m2)

Aggressive to

conspecifics,

including pointing

with beak, thrusting

neck and pecking

neighbours in close

proximity

None

Swift Tern

(LC, stable)

On flat open

ground

Tightly packed

colonies

(6.1–8.2 / m2)

Loud alarm calls,

flying in circles over

colony. Diving at

intruders not well

developed

None when small.

Larger chicks

move rapidly away

from intruders,

often in tight

groups

Bank

Cormorant

(EN, declining)

On top of big

boulders close to

the shore

Sparse to

moderate

densities.

Extending neck and

bill forward towards

intruder

Not observed

† LC: least concern; NT: near threatened; VU: vulnerable; EN: endangered (IUCN categories)

Ta

ble

4.2

: Im

pa

ct o

f p

eli

ca

n p

red

atio

n o

n s

ea

bir

ds’ b

reed

ing p

rod

uctio

n o

n t

he m

on

itored

isla

nd

s o

f th

e w

est c

oa

st o

f S

ou

th

Afr

ica

in

th

e s

ea

son

20

06

/07

.

§ M

arin

e a

nd

Coa

sta

l M

an

agem

en

t,

un

pu

bli

sh

ed

da

ta

; * 1

80

fle

dged

Ca

pe C

orm

ora

nts c

orresp

on

ded

to a

con

trol

colo

ny o

f 1

00

nests w

here p

eli

ca

ns w

ere e

xclu

ded

on

Ma

lga

s I

sla

nd

(1.8

fle

dgli

ngs/n

est,

L P

ich

agru

pers.

com

m.)

; M

usa

ngu

(2

00

7);

† E

stim

ated

nu

mb

er o

f ga

nn

et c

hic

ks e

aten

by p

eli

ca

ns o

n a

breed

ing s

ea

son

: ~

2.5

ch

ick

s/p

eli

ca

n/d

ay �

30

–5

6 p

eli

ca

ns

� 6

0 d

ays =

4 5

00

–8

40

0 c

hic

ks;

Secon

d b

reed

ing a

ttem

pt n

ot i

nclu

ded

B

re

ed

ing

pa

irs

Ex

pe

cte

d

ch

ick

pro

du

cti

on

Ch

ick

s

fle

dg

ed

% e

xp

ecte

d

ch

ick

pro

du

cti

on

Bre

ed

ing

su

cce

ss

(# f

led

ge

d

/ne

st)

Gre

at

Wh

ite

Pe

lica

n p

re

da

tio

n r

ate

s.

Resu

lts i

n i

ta

lics r

efe

r t

o f

oca

l grou

p o

bserva

tio

ns.

Oth

erw

ise r

esu

lts r

efe

r t

o

ma

xim

um

cou

nts o

f p

eli

ca

ns o

n e

ach

isla

nd

or t

o i

ncid

en

ta

l ob

serva

tio

ns o

f

pred

atio

n.

CA

PE

CO

RM

OR

AN

TS

Ju

tten

Is.

7 4

00

1

3 3

20

4

2

0.3

0

.01

7

.3 p

red

ati

on

s/h

r (1

87

even

ts;

25

.5 h

rs;

ob

serv

ed m

ean

# p

elic

an

=

17

; ra

nge

3–5

7);

Mea

n n

um

ber o

f p

eli

ca

ns o

n i

sla

nd

= 8

5 (

ra

nge 4

3–

17

6;

n=

14

da

ys)

Ma

lga

s I

s.

2 2

06

§

4 6

08

3

41

8

.6

0.1

5 *

O

bserved

, n

ot q

ua

ntif

ied

Sch

aa

pen

Is.

2 5

60

§

3 9

70

1

5

0.3

0

.01

2

3 p

red

atio

ns o

bserved

op

portu

nis

tic

all

y

To

tal

Sa

lda

nh

a i

sla

nd

s

12

16

6

21

90

0

39

8

1.8

0

.03

Da

ssen

Is.

17

3

31

1

0

0

0

Peli

ca

n p

red

atio

n s

usp

ected

CR

OW

NE

D C

OR

MO

RA

NT

S

Ju

tten

Is.

72

9

4

97

1

03

.6

1.3

0

No p

red

atio

ns o

bserved

(n

estin

g o

n t

rees)

Ma

lga

s I

s.

88

§

11

4

- -

- N

o p

red

atio

ns o

bserved

(n

estin

g o

n t

rees a

nd

hig

h s

tru

ctu

res)

Sch

aa

pen

Is.

14

0

18

2

30

1

6.5

0

.20

7

pred

atio

ns o

bserved

op

portu

nis

tic

all

y;

>5

0 s

usp

ected

To

tal

Sa

lda

nh

a i

sla

nd

s

24

7

32

1

- -

-

Da

ssen

Is.

24

0

31

2

18

5

.8

0.0

8

7 p

red

atio

ns o

bserved

; 1

34

nests d

estroyed

by p

eli

ca

ns

KE

LP

GU

LL

S

Ju

tten

Is.

2 5

28

1

88

3

60

3

.2

0.0

2

8.9

pre

da

tion

s/h

r (2

28

even

ts;

25

.5 h

rs;

ob

serv

ed m

ean

# p

elic

an

=

17

, ra

nge

3– 5

7)

Mea

n n

um

ber o

f p

eli

ca

ns o

n i

sla

nd

= 8

5 (

ra

nge 4

3–

17

6;

n=

14

da

ys)

Ma

lga

s I

s.

15

6

11

6

19

1

6.3

0

.12

7

pre

da

tion

even

ts /

84

hrs

; o

bse

rved

# p

elic

an

/ h

r =

16

.7;

ran

ge1

2–2

5

Mea

n n

um

ber o

f p

eli

ca

ns o

n i

sla

nd

= 3

8 (

ra

nge 3

0–

56

; n

= 1

0 d

ays)

Sch

aa

pen

Is.

5 2

01

3

87

5

9

0.2

0

.00

N

ot q

ua

ntif

ied

, h

un

dred

s o

bserved

op

portu

nis

tic

all

y

To

tal

Sa

lda

nh

a i

sla

nd

s

7 8

85

5

87

4

88

1

.5

0.0

1

Da

ssen

Is.

5 1

15

3

81

1

18

0 §

4

.7

0.0

4

15

9 p

red

atio

ns o

bserved

op

portu

nis

tic

all

y

CA

PE

GA

NN

ET

S

Ma

lga

s I

s.

29

25

2 §

1

2 2

85

§

- -

- 6

.4 p

red

ati

on

s/h

r (5

37

even

ts;

84

hrs

; ob

serv

ed m

ean

# p

elic

an

= 1

6.7

; ra

nge1

2–2

5)

†

Mea

n n

um

ber o

f p

eli

ca

ns o

n i

sla

nd

= 3

8 (

ra

nge 3

0–

56

; n

= 1

0 d

ays)

SW

IFT

TE

RN

S

Sch

aa

pen

Is.

2 9

90

1 4

35

1

04

0

72

.5

0.3

5

>5

4 p

red

atio

ns o

bserved

op

portu

nis

tic

all

y;

22

peli

ca

ns (

ma

xim

um

ob

served

);

Peli

ca

ns w

ere c

ha

sed

aw

ay f

rom

colo

ny w

hen

possib

le

120

Figure 4.1: Great White Pelican hunting methods for the different seabird prey

species: small group surrounds a cluster of Cape Cormorant nests (a); walking

through a Kelp Gull colony in a disorderly group (b), and in V-formation (c);

patrolling the edge of a Cape Gannet colony (d); stalking a suitable Cape Gannet

chick (e); and a tight circle of pelicans surrounding a group of nidifugous Swift

Tern chicks (f).

�

��

��

��

��

Fig

ure 4

.3: N

um

ber o

f p

eli

can

s i

nvolv

ed

in

pred

atory a

ctiv

itie

s o

n t

hree i

sla

nd

s o

f th

e W

est C

oast N

atio

nal

Park

du

rin

g t

he

su

mm

er o

f 2006/0

7. H

istogram

rep

resen

t d

ail

y c

ou

nts o

n a

ll t

hree i

sla

nd

s a

nd

lin

es s

how

cou

nts f

or e

ach

isla

nd

,

incorp

oratin

g m

od

ell

ed

estim

ates f

or t

he d

ays c

ou

nts w

ere n

ot u

nd

ertak

en

(P

an

nek

oek

& v

an

Strie

n 1

996).

�

CHAPTER 5

��

��

��

��

�

123

� ��

��

��

��

��

Because the scale of Great White Pelican predation on seabird chicks threatened to

have a non-sustainable impact on breeding seabirds on the islands off the west coast

of South Africa, a management intervention was implemented in 2007/08 on two

islands of the West Coast National Park. The objective was to reduce pelican

predation on selected seabird breeding colonies. The whole of Malgas Island and the

southern half of Jutten Island were designated as pelican-exclusion (or managed)

areas, while the other half of Jutten Island was used as a control (non-managed

area). The intervention consisted of chasing pelicans from the managed areas by

walking towards them. Monitoring of seabird colonies and direct observations of

predation showed that pelican predation caused near-zero recruitment in the non-

managed area, in which pelicans were allowed to forage, whereas breeding success

of Cape Cormorants was 0.71 fledglings/nest on the managed area of Jutten Island.

Pelican numbers on Malgas Island were lower than during the same period the year

before in the absence of management intervention. No pelicans visited Malgas

Island from mid December, despite the continued availability of gannet chicks of

appropriate sizes. This intervention demonstrated that active chasing of pelicans is

an effective measure to prevent pelican predation and to enable natural levels of

recruitment of ground-breeding seabirds on the islands off the coast of South Africa.

However, this measure may not be easily implemented at all localities and testing of

alternative management measures may be necessary. Also, pelican predation is not

the only factor affecting the population dynamics of seabird species and a holistic

approach to ecosystem management is required for the persistence of declining

populations of seabirds in the Benguela ecosystem.

��

�

124

��

All seabirds are dependent on land to breed. Their choice of breeding sites is in

part determined by the location of food sources, but also by the availability of

adequate breeding grounds, which provide protection for the progeny and the

adults during a time they are land-bound and thus most vulnerable to predation

(Furness & Monaghan 1987). Seabirds choice of breeding sites (cliffs, islands,

rockeries, etc.) are not readily accessible to most ground predators. Scavenging

and carnivorous avians such as large gulls Larus spp., skuas (family

Sterorariidae) and Giant Petrels Macronectes spp are also natural predators of

seabirds. In some cases, these species have become superabundant as a

consequence of human-induced changes in communities and ecosystems, for

example by benefiting from fisheries discards, surplus food at rubbish tips and/or

agricultural offal (Furness et al. 1981; Boutin 1990; Blokpoel & Spaans 1991;

Garrott et al. 1993; Goodrich & Buskirk 1995; Pons & Migot 1995; Vidal et al.

1998; Tasker et al. 2000; Oro & Furness 2002; Oro et al. 2005; Duhem et al.

2008). Furthermore, subsequent shortages of food discards can cause increased

predation on seabirds by opportunistic predators (Regehr & Montevecchi 1997;

Votier et al. 2004). Management of these populations has been often deemed

necessary to control their population expansion and to curb the impact on

recruitment and/or survival of vulnerable prey species (Furness & Monaghan

1987; Garrott et al. 1993; Goodrich & Buskirk 1995; Cote & Sutherland 1997).

Great White Pelicans Pelecanus onocrotalus have become a novel predator of

seabirds on the islands off the west coast of the Western Cape, South Africa

(Chapter 4). The eastern-boundary Benguela Current Upwelling Ecosystem

(15°–17°S) is characterised by nutrient-rich waters that sustain diverse and rich

trophic webs. Up to 12 species of seabirds breed in the Benguela ecosystem,

feeding on its abundant marine biomass (Berruti 1989). Pelican predation was

first observed in large scale near their breeding colony on Dassen Island in the

1990s (Crawford et al. 1995; de Ponte Machado 2007). It intensified from 2005,

subsequent to the reduction of surplus food available to pelicans in agricultural

areas (Chapters 2 and 4). An artificially-expanded pelican population and the

��

125

opportunity presented by large colonies of ground-nesting seabirds facilitated the

spread and intensification of this novel behaviour (Chapter 4). On the other

hand, a reduction on the availability of pelagic fish to breeding seabirds is the

main cause of their current decline in numbers (Crawford 2007a; Crawford et al.

2007d; 2008a; b; Okes et al. 2009) and this relationship has been particularly

manifest for the three most numerous species: Cape Gannet Morus capensis,

Cape Cormorant Phalacrocorax capensis and African Penguin Spheniscus

demersus. Markedly since 2005, pelican predation has become an additional

factor threatening the recruitment and affecting population dynamics of seabird

species in the region; mostly affecting Cape Cormorants, Cape Gannets, Kelp

Gulls L. dominicanus, Crowned Cormorants P. coronatus and Swift Terns Sterna

bergii, although pelicans have been also recorded eating Haurtlaub’s Gulls L.

haurtlabii, African Penguins, and eating or disturbing several shore- and water-

birds (Crawford et al. 1995; 1997; Wolfaardt 2000; de Ponte Machado 2007;

Chapter 4). Paradoxically, pelicans are classified as Near-Threatened in South

Africa (Barnes 2000; du Toit et al. 2003) due to the small number of breeding

sites (limited to two: one in the Western Cape and one in St Lucia, KwaZulu-

Natal). This has created a conservation dilemma, in which pelicans, a protected

species, threaten the conservation of other threatened species, some of which

have shown alarming signs of decline during the 20th century and the first

decade of the 21st century (du Toit et al. 2003; Kemper et al. 2007; Table 5.1).

Recent developments in the field of conservation management have emphasised

the need to move from an experience-based practice to a more rigorous evidence-

based conservation, in which effective management decisions are based on

proven scientific experiments or on systematic reviews of evidence (Pullin &

Knight 2003; Pullin et al. 2004; Sutherland et al. 2004). In addition, actions

taken should be adequately monitored, the effectiveness of the measures

evaluated and their results appropriately disseminated (Pullin & Knight 2001).

Removal of predators to prevent predation on vulnerable bird populations has

been a widely used management strategy (Garrott et al. 1993; Smith & Carlile

1993; Goodrich & Buskirk 1995; Bosch 1996; Cote & Sutherland 1997; Bosch et

��

�

126

al. 2000; Guillemette & Brousseau 2001; Finney et al. 2003; Crawford et al.

2007c; Oro & Martínez-Abraín 2007; Makhado et al. 2009; Sanz-Aguilar et al.

2009), often considered as effective, especially when aimed at removing non-

native predators (Cote & Sutherland 1997). However, an increasing number of

studies show that indiscriminate culling of avian predators (e.g. gulls) may not

have the desirable long-term effects on prey species’ populations (Goodrich &

Buskirk 1995; Cote & Sutherland 1997; Oro & Martínez-Abraín 2007). Although

pelicans are a native species to the Western Cape, the expression of mass scale

predatory behaviour is a recent acquisition (Chapter 4) and as such is exerting

novel pressure on seabird populations. The striking scale and extent of pelican

predation (Table 4.2) (de Ponte Machado 2007; Musangu et al. unpublished

manuscript) motivated the need for immediate management intervention. Here I

discuss the potential management measures considered by conservation

authorities and seabird scientists, and show the results of an experimental

management action implemented on two islands of Saldanha Bay during the

breeding season 2007/08.

��

��

The islands of Malgas (33� 03’S, 17� 55‘E) and Jutten (33° 05’S, 17° 57‘E) are

located on the north and south sides, respectively, of the entrance to Saldanha

Bay, on the west coast of South Africa. Schaapen Island (33�05’S, 18�01’ E) is

situated at the northern end of the coastal embayment known as Langebaan

Lagoon (Whitfield 2005) that extends south from Saldanha Bay (Figure 5.1).

They are managed by South African National Parks (SANParks) as part of the

West Coast National Park and have been proclaimed as Marine Protected Areas,

MPAs (South African Government Gazette, 19 December 2000). All three islands

are also catalogued as Important Bird Areas (Barnes et al. 2001) and belong to

the Langebaan Lagoon Ramsar Site (Cowan 1995), because of the importance of

the seabirds breeding on the islands.

��

127

Malgas Island is a small rocky island of 19 ha. The interior of the island is flat

and sandy. During the breeding season (from July to February of most years) it

is occupied by dense colonies of Cape Gannets (area 1.02 ha in 1990 to 1.99 ha in

1994; 19 000–56 000 pairs; Crawford et al. 2007a). Cape Cormorants breed

opportunistically on boulders along the periphery of the island and among semi-

abandoned buildings on the eastern side. Crowned Cormorants breed on rocky

outcrops, enlarged by the accumulation of nest material over the years, in the

midst of the gannet colony, as well as on the roof of buildings, the bushes

surrounding these, and on an old abandoned jetty. Other seabird species such as

Bank Cormorants P. neglectus and African Penguins breed near the shore on top

of large boulders and among rocks, respectively (Crawford 2007b; Crawford et al.

2007d). Kelp Gulls breed sparsely along the periphery of the gannet colony and

form denser colonies in the west and south-west parts of the island (Crawford

2007c). The island is home to a breeding population of African Black

Oystercatchers Haematopus moquinii and occasionally to ground-breeding

colonies of Swift Terns (Hockey 1997; Hockey et al. 2005; Crawford 2009).

Jutten Island lies 4.3 km south of Malgas Island, is double the size (44 ha) and

comprises more diverse habitat types. Two hills (60 m above sea level at the

highest point) are situated in the centre of the island above the flat plains and

fields of rocky boulders. Old stone walls demarcate the island into sections and a

group of old buildings are clustered in the eastern side of the island. The number

of Cape Cormorant breeding pairs on Jutten Island fluctuated between 200 and

29 800 pairs in 1978–2006 (Crawford et al. 2007b), depending on the abundance

and distribution of pelagic fish near the colony. They occupy rocky outcrops in

the flat plains, the slope of the hills, artificial walls and the boulder fields along

the shore. Crowned Cormorants breed on trees among the buildings, having

increased from ca. 30 pairs in 1990–1994 to ca. 80 in 2003–2006 (Crawford

2007b). Kelp Gulls are spread throughout the island, in parts forming dense

colonies (in the period 1978–2006 they increased from 700 pairs to a maximum of

2 600; Crawford et al. 2007c; Marine and Coastal Management, unpublished

data). Bank Cormorants breed on top of smooth boulders near the shore on the

��

�

128

eastern side of the island, their population fluctuating between 30 and 60 pairs

from 1978 to 2006 (Crawford 2007b). Other species recorded breeding on Jutten

Island are African Black Oystercatchers, African Penguins, Caspian Terns

Sterna caspia, Swift Terns, Hartlaub’s Gulls, Sacred Ibis Threskiornis

aethiopicus and Leach’s Storm Petrels Oceanodroma leucorhoa. (Hockey 1997;

Hockey et al. 2005; Underhill et al. 2002; Crawford et al. 2007d; Crawford 2009;

Marine and Coastal Management, unpublished data).

Schaapen Island (40 ha) is located between the Postberg Peninsula and the town

of Langebaan, at the northern end of the Langebaan Lagoon. It is a flat island

scattered with rocky outcrops. It has a sandy beach on its northern perimeter

and low cliffs in the northern and western shores. No buildings, walls or other

artificial structures exist on the island. During the winter and spring, the island

is covered by annual plant species that provide some protection to ground-

breeding seabirds. The number of Cape Cormorant breeding pairs fluctuated

from a few hundred to more than 2 000 from 1978–2006 (Crawford et al. 2007b).

A few dead trees provide an elevated platform for Crowned Cormorant nests,

although they also breed on colonies on the ground and cliffs or rocky outcrops in

the periphery of the island, reaching a maximum of 260 breeding pairs on the

island (Crawford 2007b). The number of Kelp Gulls nesting on the island has

increased from ca. 2 000 in the 1970s to ca. 6 000 in the 2000s (Crawford et al.

2007c), the highest breeding density in South Africa for this species. Other

species breeding here are African Black Oystercatchers, Swift Terns, Haurtlaub’s

Gulls and Sacred Ibises (Hockey 1997; Hockey et al. 2005; Crawford 2009;

Marine and Coastal Management, unpublished data).

��

To counteract the detrimental effect of pelican predation on recruitment of

seabirds, triggered or exacerbated by the anthropogenic-related increase in

numbers of the pelican population, a management intervention was

recommended to curb the impact of pelican predation. I compiled a series of

potential management options (Figure 5.2) including: (a) removal of predatory

��

129

pelicans by culling; (b) preventing pelicans from breeding; (c) active chasing of

pelicans in order to prevent predation; (d) conditioned taste aversion for pelicans;

(e) building artificial structures and enclosures to protect prey species; and

(f) reduction of artificial sources of food (Appendix 5.1). After discussions with

management authorities and local seabird scientists, the decision was taken to

test the effectiveness of dissuading pelicans from eating chicks at seabird

colonies by chasing them before implementing other type of actions, especially

those involving lethal control of pelicans.

The intervention was first implemented during the prey species’ breeding season

from October 2007 to January 2008. The activities involved active chasing of

pelicans from Malgas Island and from the southern half of Jutten Island. Old

stone walls and natural marks were used to draw the boundary between the

pelican-exclusion area (or managed area; southern half) and the control area

(non-managed area; northern half; Figure 5.3). During the second week of

November the intensity of pelican chasing was increased on Jutten Island to

cover the whole island.

��

Training of park rangers and volunteers (mostly formal members of the

Honorary Rangers association) consisted on practical demonstrations on:

• how to approach seabird colonies in order to chase pelicans without

causing disturbance to breeding seabirds;

• monitoring the number of pelicans present on the islands;

• scanning pelicans for the presence of rings, in an attempt to identify

predatory pelicans;

• monitoring breeding success of selected study colonies.

I visited each island at the beginning and the end of the intervention and at least

weekly during this period to assess the success of the intervention and to

monitor both the consistency of the intervention activities and the recruitment of

breeding seabirds.

��

�

130

��

The numbers of pelicans on Malgas, Schaapen and Jutten Islands were

monitored from the beginning of October 2006 to the end of January 2007. On

Malgas and Jutten Island, pelicans were counted hourly from 5h00 to 19h00.

Schaapen Island was monitored from the stretch of coast closest to the island

(400–800 m) with a frequency of at least three times a day (at 7h00, 9h00 and

15h00). However, only the eastern and central parts of the island were visible

from there. The maximum simultaneous count for each day on each island was

considered for the estimates. To account for missing counts, I used the software

package TRIM (Pannekoek & van Strien 1996), which uses a log-linear

regression model to estimate missing values. I selected the software’s default

parameters for the linear model and included corrections for overdispersion.

Modelled data values were used for plotting the numbers of pelicans on each

island when no counts were available.

During the scans to count pelicans on the non-managed area of Jutten Island,

the number of predation events (defined as pelicans actively capturing or

swallowing chicks) were counted. No attempt was made to identify the species

eaten or to estimate predation rates.

In order to identify whether a specialised group of pelicans was responsible for

the predatory behaviour observed on the islands, pelicans were continuously

scanned for rings or distinguishable individual features during all years (2006–

2008) in which observations were made on Jutten, Malgas and Schaapen Islands.

��

All colonies of Crowned and Bank Cormorants and the study colonies of Cape

Cormorants on Jutten and Malgas Islands were monitored weekly. The colonies

were mapped and the numbers of nests and chicks counted. Whenever detected,

the incidence of potential causes of mortality: disease, disturbance or predation

was recorded. All but one of the Crowned Cormorant breeding sites and all Bank

Cormorant colonies were included in the pelican-exclusion area.

��

131

Cape Cormorant study colonies with similar densities and nesting types were

selected in both designated areas of Jutten Island: 464 and 577 nests on the

managed and non-managed areas, respectively (Figure 5.3). Both areas are

affected by the same environmental conditions (food availability, climatic events,

etc.). From an age of ca. five weeks, Cape Cormorant chicks are left alone on the

nest and/or form crèches (Hockey et al. 2005). For monitoring purposes, large

chicks that had moved away from the nests were counted separately. In order to

include them into the total number of active nests I divided the number of chicks

in crèches by 2.45, obtained by multiplying the mean clutch size for the species

(2.82; Berry 1976), times the mean hatching success (0.87; Hockey et al. 2005).

The resulting figure was rounded to the closest integer and added to the number

of active nests. To estimate the maximum expected chick production I used

average estimates of clutch size (2.81 eggs/nest), hatching success (0.87) and

fledgling success (0.91) for years of good food availability (Hockey et al. 2005).

Kelp Gull predation on conspecifics discouraged the demarcation of Kelp Gull

nests to monitor breeding success, however an estimate of number of breeding

pairs at the beginning of the reproductive period and counts of fledglings on both

areas of the island were done at the end of the study period. To estimate the

impact of pelican predation on Kelp Gull’s breeding success I used the mean

value of breeding productivity (0.75), which fluctuates between 0.16 and 1.33

fledglings per nest (Hockey et al. 2005). In most cases, the large differences in

estimates of breeding success may be attributed to the high incidence of

intraspecific predation on eggs and chicks. Crawford et al. (2007c) calculated

breeding success for Kelp Gulls as 1.63 chicks per pair, according to mean clutch

size and chick survival values. Because of the increased human activity on the

island due the management intervention, it is justifiable that we used a smaller

value than the maximum estimate for expected breeding success.

��

�

132

��

��

A maximum of 188 pelicans was involved in predatory activities on Malgas,

Jutten and Schaapen Islands during the intervention period in 2007/08 (Figure

5.4). The increased intensity of chasing activities on Jutten Island during the

second week of November reduced the number of pelicans on the island to less

than 30. This period coincided with a reduced availability of chicks on the island

because almost all Cape Cormorant nests on the non-managed area had been

destroyed by pelican predation and Kelp Gull chicks were not available as they

were still guarded by adults at the nest. During the last week of November

pelican numbers on Jutten Island increased again to more than 100 (maximum

124) due to the increased mobility of Kelp Gull chicks, which were not brooded

continuously and wandered in close proximity of their nest.

The maximum number of pelicans on Malgas Island was 54 on 29 November

2007 (Figure 5.4). No pelicans visited Malgas Island from mid December, despite

the continued availability of gannet chicks of appropriate sizes until the third

week of January. The mean number of pelicans on the three islands of Saldanha

Bay was 65 (range 4–219) for the study period.

A total of 29 ringed pelicans was observed on Malgas and Jutten Island during

the intervention period, and 43 since 2006 (Table 5.2). Six of the engraved rings

were read, which allowed identifying individual pelicans. However, it was not

possible to confirm whether a specialized group of pelicans was responsible for

the predation incidents observed at the West Coast National Park.

� ��

Direct observation of predation events indicated that Cape Cormorant’s breeding

failure on the non-managed area of Jutten Island was caused by pelican

predation on chicks. More than 900 predation events were counted

��

133

opportunistically on Jutten Island during the hourly counts of pelican numbers,

although this figure comprised both Cape Cormorants and Kelp Gulls. A total of

62 chicks of the 5 019 Cape Cormorant nests located in the northern half of

Jutten Island (non-managed area) survived pelican predation. The expected

chick production was 11 166 (2.81 (eggs/nest) � 0.87 (hatching success) � 0.91

(fledgling success). The number of Cape Cormorant chicks that survived

predation in the non-managed area was 0.01% of the expected chick production.

Chicks consumed ranged from small naked chicks brooded by adults (1–2 days

old, estimated 25–30 g) to chicks of nearly adult size aggregated in crèches (>

800 g). From the 30 October to the 6 November, 20–28 days after the start of the

management intervention, the number of active nests remaining in the

monitored colonies of the non-managed area was reduced from 135 to eight by

pelican predation (Figure 5.5). Chicks were then about two-weeks-old and still

brooded by the adults. Pelicans were observed systematically eating chicks at

these colonies until no chicks were left in the area and all adults had abandoned

the nesting attempt. On 11 November 2008 the total chick count for the Cape

Cormorant study colonies in the non-managed area was zero.

The Cape Cormorant study colonies in the managed area were occupied by

nesting cormorants until the end of the study on 14 December 2008, nine weeks

after the start of the management intervention (Figure 5.5). Because some chicks

had already fledged by mid December, 332 large fledglings counted on 8

December 2008 were considered as the overall chick production for the study

colonies, yielding a breeding success of 0.72 fledglings per nest. No pelicans were

observed landing near, or eating chicks in the selected study colonies, which

were situated at the core of the managed area. Importantly, no disturbance-

related predation to the Cape Cormorant study colonies was registered due to

the monitoring or management activities.

Other causes of mortality or breeding failure (abandonment, disease or human

disturbance) were excluded because no noteworthy mortality of cormorants

��

�

134

(adults or chicks) was observed in any colony on the island, and human presence

in the non-managed area was reduced to 10 visits during the study period (in

which no disturbance or disturbance-related predation was observed on the

cormorant colonies).

Pelicans were dissuaded from approaching Bank and Crowned Cormorant

nesting areas, predation by pelicans was not detected, and recruitment was

within the average values for each species on both islands. Pelicans were

observed eating ground-nesting Cape Cormorant chicks in close proximity to the

tree-nesting Crowned Cormorant colony in the non-managed area of Jutten

Island. However, no predation attempts to this Crowned Cormorant colony were

observed, and 17 chicks fledged from 12 nests (1.4 fledglings/nest). In three

occasions pelicans were seen in the proximity of Bank Cormorant nests on Jutten

Island and were carefully chased away.

A total of ca. 1 200 Kelp Gull active nests were counted on Jutten Island at the

beginning of the study period in October 2007 (Marine and Coastal Management,

unpublished data). Of these, approximately 500 were located in the managed

area and 700 in the non-managed northern half of the island. At the end of the

monitored period a total of 35 Kelp Gull fledglings were counted in the non-

managed area (7% of the expected chick production for that half of the island),

whereas 292 fledglings were counted on the managed area (83% of the expected

chick production).

Pelicans repetitively landed and ate chicks of Kelp Gull and Cape Cormorants

situated in the periphery of the managed area. These areas were not rapidly

accessible in order to chase pelicans and some seabird colonies close to the shore

were not visible from the observation points. On the third week after the start of

the intervention, pelicans were reported foraging on seabirds in the managed

areas during the night (already preying on chicks at dawn and after sunset) and

also accessing colonies within the managed area by swimming onto the island

from the sea.

��

135

��

Chasing pelicans from selected seabird colonies successfully reduced the impact

of predation on recruitment at Jutten and Malgas Islands. Approximately the

same numbers of pelicans (ca. 220–230) were involved in predatory activities

during the reported intervention in 2007/08 and the previous season 2006/07,

when no co-ordinate management intervention was in place (Chapter 4).

However, the maximum number of pelicans visiting Malgas Island was reduced

from 110 during 2006/07 (Chapter 4) to 54 during 2007/08. Chasing pelicans on

Jutten Island indirectly reduced pelican numbers on Malgas Island, because they

used the former as a night roost and most pelicans visiting Malgas Island flew in

from Jutten Island (pers. obs.). In addition, because pelicans were chased from

Malgas Island soon after they were detected, often immediately after landing on

the island, the impact of pelican predation on gannet recruitment during 2007/08

was minimal.

The design of the intervention on Jutten Island demonstrated that, if left

unmanaged, pelicans can reduce the recruitment of Cape Cormorants and Kelp

Gulls to values close to zero. Chasing pelicans from selected colonies resulted in

unequivocally larger recruitment rates for Cape Cormorants, Kelp Gulls and

expectedly Cape Gannets. However, this type of intervention is labour-intensive

and requires continuous monitoring of the number of pelicans on all

colonies/islands at which the protection of seabird colonies would be desirable.

This may not be possible on localities of difficult access, or that lack basic

infrastructure for long-term human occupation, as is the case with Schaapen,

Meeuw and Vondeling Islands (Figure 5.1). Unchecked human activities during

monitoring or chasing activities could cause severe disturbance and nest losses

due to abandonment and predation of nest contents by opportunistic Kelp Gulls

and Sacred Ibis. Extreme caution should be taken at all times to avoid

disturbance to breeding colonies.

��

�

136

The spatial coverage of this intervention was restricted to two islands, which

represented a variable proportion of seabird species affected by pelican

predation. Malgas Island held ca. 27% of the world population of Cape Gannets

in the period 1978–2006 (Crawford et al. 2007a). Pelicans first targeted gannets

in 2005, and only at this locality (de Ponte Machado 2007). The management

intervention implemented in 2007/08 was thus successful to reduce its impact on

the global gannet population. However, other opportunistic predators (Kelp Gulls

and Cape Fur Seals Arctocephalus pusillus) also have a major impact on gannet

recruitment by predating on eggs and/or chicks (David et al. 2003; Makhado et

al. 2006; Crawford et al. 2007a).

Jutten and Malgas Islands held ca. 8% of the Cape Cormorant’s global

population during 1988/89–2006/07, whereas all islands in which pelican

predation on Cape Cormorants occurred (Dassen, Vondeling, Malgas, Jutten,

Schaapen and Meeuw) accounted for ca. 19% of the global population, and ca.

48% of the South African population (Crawford et al. 2007b). Consequently, this

intervention alone may be insufficient to protect Cape Cormorants from pelican

predation. Large unpredictability from year to year in the location of their main

colonies may make difficult to plan intervention measures that involve deploying

people on all affected islands. On the other hand, the easterly displacement of

the population of Cape Cormorants following a shift in the distribution of their

pelagic fish prey (van der Lingen et al. 2005; Crawford et al. 2008a) has made

Dyer Island the main breeding locality for the species, holding 70–75% of the

Western Cape population during 2004–2006 (Crawford et al. 2007b). Pelicans are

not common visitors to Dyer Island and predation has not been reported from

this locality.

Malgas and Jutten Islands held ca. 28% of the Western Cape’s Crowned

Cormorant population, and ca. 8% of their global population (RJM Crawford,

unpublished data). They were not targeted by pelican predation on these

localities due to their habit of breeding on high structures and/or in trees, but

they are particularly vulnerable to pelican predation when nesting on the

��

137

ground. Localities at which Crowned Cormorants nested on the ground (Dassen,

Caspian, Schaapen and Meeuw Islands; B Dyer, pers. comm.) accounted for ca.

50% of the Western Cape and ca. 16% of the global breeding population,

respectively (RJM Crawford, unpublished data). The lack of infrastructure on

Meeuw and Caspian Islands hinders the application of active chasing as a

management strategy on these islands. Although Schaapen Island is readily

accessible from the town of Langebaan, its lack of infrastructure and the large

density of Kelp Gulls during the breeding season persuade against chasing

pelicans on this locality. Walking through the island would produce high levels of

disturbance and increased opportunistic predation by Kelp Gulls on the

cormorant colonies. Dassen Island holds ca. 22% of the Western Cape population

and the majority of Crowned Cormorant nests are highly susceptible to pelican

predation (Musangu et al. unpublished manuscript; J Visagie, pers. comm.);

hence management actions to prevent pelican predation are necessary and active

chasing of pelicans on this island is recommended as a plausible method to

prevent predation by pelicans.

Kelp Gull populations have increased in the Western Cape from 6 500 pairs in

the late 1970s to more than 16 000 in the late 1990s, and decreased to 13 000 in

the mid 2000s (Crawford et al. 2007c). The increase has been driven mostly by

the increased availability of food in the form of fisheries discards and

agricultural and urban waste. Pelican predation on Kelp Gulls on Dassen Island,

first detected in the 1990s, reduced breeding productivity to 0.3 chicks/nest, and

an estimated 80–85% of the chicks that hatched eaten by pelicans (Crawford et

al. 2007c). On the monitored islands of the West Coast National Park, pelican

predation reduced recruitment to 0.12 chicks per pair on Malgas Island, 0.02 on

Jutten island and zero on Schaapen Island in 2006/07 (Chapter 4). On the other

hand, because of the predation impact that the increasing gull population exerts

on other seabirds, management measures have been suggested: (1) to reduce

sources of food available to gulls and (2) to apply some form of population control

of adults or chicks (Whittington et al. 2006; Altwegg et al. 2007; Crawford et al.

2007c). Pelican predation on gulls has been suggested as a ‘natural’ method to

��

�

138

control gull populations by limiting recruitment (Crawford et al. 2007c); however,

they breed in close proximity to cormorants and gannets and may be difficult to

selectively chase pelicans from selected colonies and still allow them to eat gull

chicks.

Although predation on Bank Cormorants was not observed during this study,

pelicans were reported disturbing nesting individuals. Predation on this species

was suspected on Jutten Island during the previous season (2006/07; Chapter 4)

and during 2004/05 (pers. obs.). Due to the delicate conservation status of this

species, large emphasis should be made to prevent predation on all colonies

targeted by pelicans, which at present are Dassen, Jutten, Malgas, Vondeling,

and potentially, Marcus Islands.

Chasing pelicans has proven also an effective measure to prevent predation on

Swift Tern and Hartlaub’s Gulls on Schaapen Island (pers. obs. and P Nel,

unpublished data). Smaller numbers of pelicans (ca. 20) visited the area during

the breeding period of this species (in both 2007 and 2008 it occurred from

February to May; pers. obs.). Kelp Gulls are not breeding during these months

(Hockey et al. 2005) and hence gull predation due to human presence was not of

concern.

Other alternatives to prevent pelican predation on vulnerable species may be

used in combination with active chasing. These may include the provision of

nesting exclosures (Isaksson et al. 2007) or refuge supplements for chicks (Prieto

et al. 2003). However, research is needed to design and test adequate structures

for local ground-breeding species. In 2006/07, a fenced perimeter was erected to

measure Kelp Gull breeding success on Dassen Island, which was significantly

higher inside the enclosure (0.28 chicks/pair of a total of 25 nests) than outside

(zero chicks of 49 nests; Musangu et al. unpublished manuscript). Other non-

lethal methods to reduce predation by birds, such as conditioned taste aversion,

have been used successfully for ravens and starlings (Nicolaus et al. 1983;

Dimmick & Nicolaus 1990; Avery et al. 1995; Cowan et al. 2000; Neves et al.

��

139

2006). This approach could be tested for deterring pelicans and/or gulls from

preying on gannet eggs and chicks, although research must be applied to test

what substances to use, the response of the prey species and the time retention

of the conditioned behaviour for pelicans.

Improved waste and offal management is required to prevent opportunistic

species from increasing their numbers and acquiring pest status (Smith &

Carlile 1993; Oro & Martínez-Abraín 2007). Pons & Migot (1995) measured body

condition, fecundity and adult survival rates in a population of Herring Gulls

Larus argentatus after the closure of a refuse tip that provided abundant food for

the species. They found a sharp decrease in fecundity while adult survival rates

remained unchanged, which suggests that this measure will not reduce

predation pressure immediately. On the contrary, in the short-term, the ability

of some opportunistic species to switch prey may increase predation on nesting

seabirds, as it has been demonstrated for Great Skuas Stercorarius skua (Votier

et al. 2004), large gulls (Regehr & Montevecchi 1997) and Great White Pelicans

(de Ponte Machado 2007; Chapter 4). Therefore, although decreasing availability

of food in landfills and fisheries discards is a worthwhile recommendation, it

should not be implemented in isolation but combined with other measures

(Smith & Carlile 1993).

Culling pelicans to prevent predation is not recommended due to the availability

of other measures and the legal, ethical and conservation-related implications of

culling a charismatic and protected species (Goodrich & Buskirk 1995). Although

removing avian predators from an area has been proven as successful to increase

reproductive success of threatened seabirds in the short-term (Guillemette &

Brousseau 2001; Sanz-Aguilar et al. 2009), long-term evaluations of the effects of

these measures are contradictory. In some cases, the problem is exported to

neighbouring sites (Vidal et al. 1998; Bosch et al. 2000), implementing agencies

have incurred excessive costs in terms of labour and material resources (Bosch et

al. 2000), or the intervention has failed to reduce the populations of the

opportunistic predators (Oro & Martínez-Abraín 2007) or to increase the

��

�

140

breeding populations of vulnerable species object of conservation management

measures (Cote & Sutherland 1997). Attempts at identifying ‘rogue’ pelicans

during this study were not successful due to the small number of marked

pelicans involved in predation events.

Further research is needed to estimate the additive impact of predation by Kelp

Gulls (Altwegg et al. 2007; Crawford et al. 2007c), Great White Pelicans (Chapter

4; this study) and Cape Fur Seals (Marks et al. 1997; David et al. 2003; Makhado

et al. 2006) on the population dynamics of seabird prey species. A combination of

methods is needed to reduce the impact of predation on seabirds to prevent

habituation, and these must be implemented for a number of years to be effective

(Smith & Carlile 1993; Goodrich & Buskirk 1995; Morrison & Allcorn 2006;

Soldatini et al. 2008).

Finally, the reduced availability of pelagic fish is one the most serious threats to

the persistence of seabird populations in the Benguela Upwelling System

(Underhill et al. 2006; Crawford et al. 2007d; Mullers et al. 2009; Okes et al.

2009; Pichegru et al. in press). Therefore, the protection and restoration of their

habitat is most important conservation priority (Goodrich & Buskirk 1995; Cote

& Sutherland 1997). A holistic approach to ecosystem management is required in

order to provide adequate food supplies to seabirds in the Benguela Upwelling

System (Game et al. 2009; Pichegru et al. in press). Because predation could be

an additional factor contributing to the observed declines, the coordinated action

of the relevant conservation agencies, policy makers and governmental

departments is essential.

��

141

��

Altwegg R, Crawford RJM, Underhill LG, Martin AP, Whittington PA (2007)

Geographic variation in reproduction and survival of Kelp Gulls Larus

dominicanus vetula in southern Africa. Journal of Avian Biology 38, 580–

586.

Avery ML, Pavelka MA, Bergman DL, Decker DG, Knittle CE, Linz GM (1995)

Aversive conditioning to reduce raven predation on California Least Tern

eggs. Colonial Waterbirds 18, 131–138.



Barnes KN, Johnson DJ, Anderson MD, Taylor PB (2001) Important Bird Areas

in Africa and associated islands: priority sites for conservation. BirdLife

Conservation Series No.11 (eds. Fishpool LDC, Evans MI), pp. 793–876.

BirdLife International, Cambridge.

Berruti A (1989) The Benguela ecosystem – Part VI Seabirds. Oceanography and

Marine Biology Annual Review 27, 273–335.



Blokpoel H, Spaans AL (1991) Superabundance in gulls, causes, problems and

solutions. In: Acta XX Congressus Internationalis Ornithologici, vol. 4, pp.

2359–2398. New Zealand Ornithological Congress Trust Board,

Wellington.

Bosch M (1996) The effects of culling on attacks by Yellow-legged Gulls Larus

cachinans upon three species of herons. Colonial Waterbirds 19, 248–252.

Bosch M, Oro D, Cantos F, J, Zabala M (2000) Short-term effects of culling on the

ecology and population dynamics of the Yellow-legged Gull. Journal of


Boutin S (1990) Food supplementation experiments with terrestrial vertebrates:

patterns, problems, and the future. Canadian Journal of Zoology 68, 203–

220.

Cote IM, Sutherland WJ (1997) The effectiveness of removing predators to

protect bird populations. Conservation Biology 11, 395–405.

Cowan DP, Reynolds JC, Gill EL (2000) Manipulation predatory behaviour

through conditioned taste aversion: can it help endangered species? In:

Behaviour and conservation (eds. Gosling LM, Sutherland WJ, Avery ML),

pp. 281–299. Cambridge University, Cambridge.

Cowan GI (1995) Wetlands of South Africa. Department of Environmental

Affairs, Pretoria.

Crawford RJM (2007a) Food, fishing and seabirds in the Benguela upwelling

system. In: Final Report of the BCLME (Benguela Current Large Marine

Ecosystem) Project on Top Predators as Biological Indicators of Ecosystem

Change in the BCLME (ed. Kirkman SP), pp. 229–234. Avian Demography

Unit, Cape Town.

Crawford RJM (2007b) Trends in numbers of three cormorants Phalacrocorax

spp. breeding in South Africa's Western Cape Province. In: Final report of

the BCLME (Benguela Current Large Marine Ecosystem) project on top

��

�

142

predators as biological indicators of ecosystem change in the BCLME (ed.

Kirkman SP), pp. 173–178. Avian Demography Unit, Cape Town.

Crawford RJM (2009) A recent increase of Swift Terns Thalasseus bergii off

South Africa: the possible influence of an altered abundance and

distribution of prey. Progress in Oceanography. 10.1016/

jpocean.2009.07.21



Western Cape. African Journal of Marine Science 15, 33–42.

Crawford RJM, Dundee BL, Dyer BM, Klages NTW, Meyër MA, Upfold L

(2007a) Trends in numbers of Cape Gannets Morus capensis 1956/57–

2005/06, with a consideration of the influence of food and other factors.

ICES Journal of Marine Science 64, 169–177.

Crawford RJM, Dyer BM, Kemper J, Simmons RE, Upfold L (2007b) Trends in

numbers of Cape Cormorants Phalacrocorax capensis over a 50-year

period, 1956/57 to 2006/07. Emu 107, 253–261.

Crawford RJM, Nel DN, Williams AJ, Scott A (1997) Seasonal patterns of

abundance of Kelp Gulls Larus dominicanus at breeding and non-breeding

localities in southern Africa. Ostrich 68, 37–41.

Crawford RJM, Sabarros PS, Fairweather TP, Underhill LG, Wolfaardt A

(2008a) Implications for seabirds off South Africa of a long-term change in

the distribution of sardine. African Journal of Marine Science 30, 177–

184.

Crawford RJM, Tree AJ, Whittington PA, Visagie J, Upfold L, Roxburg KJ,

Martin AP, Dyer BM (2008b) Recent distributional changes of seabirds in

South Africa: is climate having an impact? African Journal of Marine

Science 30, 189–193.

Crawford RJM, Underhill LG, Altwegg R, Dyer BM, Upfold L (2007c) The

influence of culling, predation and food on Kelp Gulls Larus dominicanus

off western South Africa In: Final report of the BCLME (Benguela Current

Large Marine Ecosystem) project on top predators as biological indicators

of ecosystem change in the BCLME (ed. Kirkman SP), pp. 181–187. Avian


Crawford RJM, Underhill LG, Upfold L, Dyer BM (2007d) An altered carrying

capacity of the Benguela upwelling ecosystem for African Penguins

Spheniscus demersus. ICES Journal Marine Science 64, 570–576.

David JHM, Cury P, Crawford RJM, Randall RM, Underhill LG, Meÿer MA

(2003) Assessing conservation priorities in the Benguela ecosystem, South

Africa: analysing predation by seals on threatened seabirds. Biological






change in the BCLME (ed. Kirkman SP), pp. 131-142. Avian Demography

Unit, Cape Town.

��

143

Dimmick CR, Nicolaus LK (1990) Efficiency of conditioned aversion in reducing

depredation by crows. Journal of Applied Ecology 27, 200–209.





Breeding Specialist Group, Cape Town.

Duhem C, Roche P, Vidal E, Tatoni T (2008) Effects of anthropogenic food

resources on Yellow-legged Gull colony size on Mediterranean islands.

Population Ecology 50, 91–100.

Finney S, K, Harris M, P, Keller L, F, Elston D, A, Monaghan P, Wanless S

(2003) Reducing the density of breeding gulls influences the pattern of

recruitment on immature Atlantic Puffins Fratercula arctica to a breeding

colony. Journal of Applied Ecology 40, 545–552.

Furness RW, Monaghan P (1987) Seabird ecology. Blackie, Glasgow and London.

Furness RW, Monaghan P, Sheddon C (1981) Exploitation of a new food source

by the Great Skua in Shetland. Bird Study 28, 49–52.

Game ET, Grantham HS, Hobday AJ, Pressey RL, Lombard AT, Beckley LE,

Gjerde K, Bustamante R, Possingham HP, Richardson AJ (2009) Pelagic

protected areas: the missing dimension in ocean conservation. Trends in

Ecology & Evolution 24, 360–369.

Garrott RA, White PJ, White CAV (1993) Overabundance: an issue for

conservation biologists? Conservation Biology 7, 946–949.

Goodrich JM, Buskirk SW (1995) Control of abundant native vertebrates for

conservation of endangered species. Conservation Biology 9, 1357–1364.

Guillemette M, Brousseau P (2001) Does culling predatory gulls enhance the

productivity of breeding Common Terns? Journal of Applied Ecology 38,

1–8.

Hockey PAR (1997) African Black Oystercatcher: between the tides. Africa–

Birds and Birding 2 (5), 28–34.



Isaksson D, Wallander J, Larsson M (2007) Managing predation on ground-

nesting birds: the effectiveness of nest exclosures. Biological Conservation

136, 136–142.

Kemper J, Underhill LG, Crawford RJM, Kirkman SP (2007) Revision of the






Makhado AB, Crawford RJM, Underhill LG (2006) Impact of predation by Cape

Fur Seals Arctocephalus pusillus pusillus on Cape Gannets Morus

capensis at Malgas Island, Western Cape, South Africa. African Journal of

Marine Science 28, 681–687.

��

�

144

Makhado AB, Meÿer MA, Crawford RJM, Underhill LG, Wilke C (2009) The

efficacy of culling seals seen preying on seabirds as a means of reducing

seabird mortality. African Journal of Ecology 47, 335–340.

Marks MA, Brooke RK, Gildenhuys AM (1997) Cape Fur Seal Arctocephalus

pusillus predation on Cape Cormorants Phalacrocorax capensis and other

birds at Dyer Island, South Africa. Marine Ornithology 25, 9–12.

Morrison P, Allcorn RI (2006) The effectiveness of different methods to deter

large gulls Larus spp. from competing with nesting terns Sterna spp. on

Coquet Island RSPB reserve, Northumberland, England. Conservation

Evidence 3, 84–87.

Mullers RHE, Navarro RA, Crawford RJM, Underhill LG (2009) The importance

of lipid-rich fish prey for Cape Gannet chick growth: are fishery discards

and alternative? ICES Journal of Marine Science 66. doi: 10.1093/

icesjms/fsp210.

Musangu M, de Ponte Machado M, Ryan PG (unpublished manuscript) Breeding

seabirds at Dassen Island, South Africa: can they survive Great White

Pelican predation?

Neves VC, Panagiotakopoulos S, Furness RW (2006) A control taste aversion

experiment on predators of Roseate Tern Sterna dougallii eggs. European

Journal Wildlife Research 52, 259–264.

Nicolaus LK, Cassel JF, Carlson RB, Gustavson CR (1983) Taste-aversion

conditioning of crows to control predation on eggs. Science 220, 212–214.

Okes NC, Hockey PAR, Pichegru L, van der Lingen CD, Crawford RJM,

Grémillet D (2009) Competition for shifting resources in the southern

Benguela upwelling: seabirds versus purse-seine fisheries. Biological


Oro D, de León A, Mínguez E, Furness RW (2005) Estimating predation on

breeding European Storm-petrels by Yellow-legged Gulls. Journal of

Zoology (London) 265, 421–429.


survival of Kittiwakes. Ecology 83, 2516–2528.


impact on threatened sympatric seabirds. Animal Conservation 10, 117–

126.



Netherlands.

Pichegru L, Ryan PG, Le Bohec C, van der Lingen CD, Navarro RA, Petersen SL,

Lewis S, van der Westhuizen J, Grémillet D (in press) Spatial and

temporal overlap between vulnerable top predators feeding hotspots and

fisheries in the Benguela upwelling system: implications for the design of

marine protected areas. Marine Ecology Progress Series. doi:

10.3354/meps08283.




��

145

Prieto J, Jover L, Ruiz X (2003) Effect of refuge supplement on Audouin's Gull

chick survival. Scientia Marina 67, 103–108.

Pullin AS, Knight TM (2001) Effectiveness in conservation practice: pointers

from medicine and public health. Conservation Biology 15, 50–54.

Pullin AS, Knight TM (2003) Support for decision making in conservation

practice: an evidence-based approach. Journal for Nature Conservation 11,

83–90.

Pullin AS, Knight TM, Stone DA, Charman K (2004) Do conservation managers

use scientific evidence to support their decision-making? Biological






Sanz-Aguilar A, Martínez-Abraín A, Tavecchia G, Mínguez E, Oro D (2009)

Evidence-based culling of a facultative predator: efficacy and efficiency

components. Biological Conservation 142, 424–431.

Smith GC, Carlile N (1993) Methods for population control within a Silver Gull

colony. Wildlife Research 20, 219–226.

Soldatini C, Albores-Barajas YV, Torricelli P, Mainardi D (2008) Testing the

efficacy of deterring systems in two gull species. Applied Animal

Behaviour Science 110, 330–340.

Sutherland WJ, Pullin AS, Dolman PM, Knight TM (2004) The need for

evidence-based conservation. Trends in Ecology & Evolution 19, 305–308.

Tasker ML, Camphuysen CJ, Cooper J, Garthe S, Montevecchi WA, Blaber SJM

(2000) The impacts of fishing on marine birds. ICES Journal of Marine

Science 57, 531–547.

Underhill LG, Crawford RJM, Camphuysen CJ (2002) Leach's storm petrels

Oceanodroma leucorhoa off southern Africa: breeding and migratory

status, and measurements and mass of the breeding population.

Transactions of the Royal Society of South Africa 57, 43–46.





van der Lingen CD, Coetzee JC, Demarcq H, Drapeau L, Fairweather TP,

Hutchings L (2005) An eastward shift in the distribution of southern

Benguela sardine. Global Ocean Ecosystem Dynamics 11, 17–22.









��

�

146

Whitfield, A.K. 2005. Langebaan: a new type of estuary? African Journal of

Aquatic Science 30 (2), 207–209.

Whittington PA, Martin AP, Klages NTW (2006) Status, distribution and

conservation implications of the Kelp Gull Larus dominicanus vetula

within the Eastern Cape region of South Africa. Emu 106, 127–139.

Wolfaardt A (2000) Dassen Island Management Plan. Western Cape Nature

Conservation Board, Yzerfontein.

147

Table 5.1: Distribution and conservation status of Great White Pelicans and the

avian prey species targeted by pelicans on the monitored islands according to

IUCN red data list categories (EN: endangered; VU: vulnerable; NT: near-

threatened; LC: least concern) (Barnes 2000; du Toit et al. 2002).

Species Distribution Conservation status

Bank Cormorant

Phalacrocorax neglectus Endemic southern Africa EN (global)

Cape Cormorant

P. capensis Endemic to Benguela NT (global)

Cape Gannet

Morus capensis Endemic southern Africa VU (global)

Crowned Cormorant

P. coronatus

Endemic west coast southern

Africa NT (global)

Great White Pelican

Pelecanus onocrotalus

Widespread Africa, Europe and

Asia NT (southern Africa)

Hartlaub’s Gull

Larus hartlaubii Widespread LC (global)

Kelp Gull

Larus dominicanus Widespread southern oceans LC (global)

Sacred Ibis

Threskiornis aethiopica Widespread LC (global)

Swift Tern

Sterna bergii Widespread LC (southern Africa)

148

Table 5.2: Observations of ringed Great White Pelicans involved in predatory

activities on the monitored islands in the Western Cape during 2006/07 and

2007/08, and pelican age at the time of the sighting. * Ring alphanumeric codes

in parenthesis.

Ring colour * Date Island Age when seen

(years old)

Blue on left, no metal 13 Oct 2006 Schaapen 4.7

Green on left, metal 08 Nov 2006 Jutten 1.8

Blue on left 10 Nov 2006 Jutten 4.8

Red on left, no metal 10 Nov 2006 Jutten 3.8


2 x red on left, no metal 8–12 Jan 2007 Malgas 4.0

Hole in pouch 08 Jan 2007 Malgas -

Black on left (CS) 09 Jan 2007 Malgas 1.0

Black on left (other) 09 Jan 2007 Malgas 1.0

Blue on left, no metal 14 Mar 2007 Schaapen 5.1

Hole in pouch, no metal 14 Mar 2007 Schaapen -

Blue on left, no metal 15 Mar 2007 Schaapen 5.1

Blue on left, no metal 04 May 2007 Schaapen 5.3

Blue on left, no metal 16 May 2007 Jutten 5.3

Blue on left 17 May 2007 Schaapen 5.3

Red on left, no metal 25 Sep 2007 Jutten 4.7

2 x blue on left, no metal 25 Sep 2007 Jutten 5.7

Black on left 25 Sep 2007 Jutten 1.8

2 x blue on left, no metal 25 Sep 2007 Malgas 5.7

2 x blue on left, no metal 10 Oct 2007 Jutten 5.7

2 x red on left, no metal 10 Oct 2007 Jutten 4.8

Black on left (HK) 10 Oct 2007 Jutten 1.8

Blue on left, no metal 18 Oct 2007 Jutten 5.7

Yellow on right + metal 18 Oct 2007 Jutten 3.8

Metal only 24 Oct 2007 Jutten -

Black on left (AP) 26 Oct 2007 Jutten 1.8

Black (HK) 27 Oct 2007 Jutten 1.8

Black (HP) 27 Oct 2007 Jutten 1.8

Red on left, no metal 27 Oct 2007 Jutten 4.8



White on left (BS) 28 Oct 2007 Jutten -

Black on left 28 Oct 2007 Jutten 1.8

Green on left 30 Oct 2007 Jutten 2.8


Blue on left, no metal 03 Nov 2007 Jutten 5.8



Blue on left, no metal 06 Nov 2007 Jutten 5.8


Green on right 13 Nov 2007 Jutten 2.8

Metal on left 28–30 Nov 2007 Malgas -

149

Figure 5.1: Saldanha Bay islands off the west coast of South Africa.

Monitoring seabird colonies was implemented on Malgas, Jutten, Schaapen

and Dassen Islands. The management intervention was undertaken on the

islands of Malgas and Jutten at the West Coast National Park.

150

Figure 5.2: Potential management options to reduce Great White Pelican

predation on seabird chicks on the islands of the Western Cape, ordered from

higher level of intervention to the least aggressive of the possibilities. The black

arrow indicates the approach implemented on Jutten and Malgas Islands in

2007/2008. See Appendix 5.1 for more detail on these management interventions.

151

Figure 5.3: Aerial view of Jutten Island showing the division into two

management zones: pelican-exclusion or managed area in the southern half and

a control area, non-managed, in the north. The locations of Cape Cormorant

study colonies for each treatment are circled.

Fig

ure 5

.4: N

um

ber o

f G

reat W

hit

e P

eli

can

s i

nvolv

ed

in

pred

atory a

ctiv

itie

s o

n t

hree i

sla

nd

s o

f th

e W

est C

oast N

atio

nal

Park

du

rin

g t

he s

um

mer o

f 2007/0

8. H

istogram

rep

resen

t d

ail

y c

ou

nts o

n a

ll t

hree i

sla

nd

s a

nd

lin

es s

how

cou

nts f

or e

ach

isla

nd

,

incorp

oratin

g m

od

ell

ed

estim

ates f

or t

he d

ays c

ou

nts w

ere n

ot u

nd

ertak

en

(P

an

nek

oek

& v

an

Strie

n 1

996).

153

Figure 5.5: Numbers of Cape Cormorant active nests and chicks (in italics in

figure) at Jutten Island study colonies during the months of October–December

2007. Colonies were located in: (a) pelican-exclusion or managed area, where

pelicans where chased to protect seabird colonies; (b) non-managed or control

area in which pelican predation happened at natural levels without human

interference.

a)

b)

�

CHAPTER 6

��

��

��

��

155

��

��

��

��

Since 1990, predation on seabird chicks by cooperative-hunting Great White

Pelicans Pelecanus onocrotalus increased to the extent that recruitment of endemic

seabird species decreased drastically, generating conservation concerns. Until 2009,

this foraging behaviour was confined to the Western Cape population, but there is a

risk it could spread to other populations as a consequence of pelican movements and

dispersal. In order to establish the degree of gene flow among southern African

breeding populations, I used eight microsatellite loci and one mitochondrial DNA

locus. Significant genetic structure was found among breeding populations in

Namibia, the Western Cape and KwaZulu-Natal. Differentiation was consistent for

both mtDNA and microsatellites ( ST = 0.109, p < 0.005; RST = 0.055, p < 0.0001).

Gene flow coefficients were low, indicating limited gene flow among populations,

possibly nil from Namibia towards the Western Cape population (m = 0.01), and low

to moderate in the northward direction (m = 0.16). The southern African Great

White Pelican populations exhibited a typical metapopulation structure. Both

Namibian and the Western Cape populations were small and signals of a

demographic bottleneck or founder effect were encountered in both populations. The

population expansion observed in the Western Cape from 1956 to 2004 appeared to

mask the effects of the bottleneck, and possibly triggered changes in foraging habits

(i.e. increased predation on seabirds) and outgoing dispersal rates. Molecular data

also confirmed the potential for exporting this learned behaviour via emigration into

the Namibian population, which constitutes an important fraction of the breeding

range of regionally-endemic seabird species. Pelican predation on seabirds could

exert top-down regulation by affecting the recruitment of top-predator seabirds,

which could in turn trigger trophic cascades through the highly-productive yet

currently threatened Benguela Upwelling System.

��

156

��

The human-induced population expansion of Great White Pelicans Pelecanus

onocrotalus in South Africa’s Western Cape Province has caused unpredictable

ecological consequences (Chapters 2 and 5). Of particular concern is the

development of newly acquired behaviour that has resulted in unprecedented

large-scale predation by pelicans on chicks of endangered or vulnerable colonial

seabirds (de Ponte Machado 2007; Chapter 4). In view of current predation rates

that could have a detrimental effect on the persistence of breeding colonies of

endemic seabirds, it has become an urgent priority to develop a management

plan for the southern African population of Great White Pelicans; a species

which in turn is classified as near-threatened in the region (Barnes 2000).

Great White Pelicans breed annually in three distinct areas in southern Africa:

one on the continent’s east coast at Lake St Lucia, and two on the west coast,

with a breeding colony on Dassen Island in the Western Cape and two breeding

sites in Namibia, at Bird Rock Guano Platform near Walvis Bay and on an island

in Hardap Dam near Mariental (Williams & Randall 1995, Figure 6.1).

Additional breeding sites are used occasionally by nomadic populations when

conditions are favourable at Lake Ngami and Sua Pan in Botswana, and Lake

Oponono and Etosha in Namibia (Crawford et al. 1994; Hockey et al. 2005).

Great White Pelicans are capable of flying long distances to feed their chicks and

are believed to undertake seasonal migration or dispersal (del Hoyo et al. 1992).

Re-sightings and recoveries of banded individuals suggest a low degree of

dispersal among the three discrete breeding colonies, with no immigration

detected into the Western Cape population (Chapters 2 and 3). The Southern

African Bird Atlas Project (Williams & Borello 1997, Figure 6.1) documented few

distribution records of pelicans between the Western Cape and Lake St Lucia,

suggesting that there is limited dispersal between the western and eastern

coastal populations.

��

157

Demographic data on the Western Cape pelican population shows an increase in

pelican numbers from the mid-1950s to the 2000s (Figure 2.2). Prior to the

1950s, the number of breeding pairs was presumably reduced by human

persecution, as exemplified by the recorded maximum estimate of 30 pairs from

1930 to 1950 (Crawford et al. 1995). Since 1956, pelicans were afforded better

protection at their breeding colony on Dassen Island and have benefited from

improved availability of food in agricultural water reservoirs. Furthermore, since

the 1980s, their food sources have been further supplemented in the form of

agricultural discards on pig farms (Crawford et al. 1995; Chapter 2).

Consequently, the local pelican population has continued to expand, reaching a

maximum of 659 breeding pairs in 2001 (Figure 2.2), and a census population

size of at least 2 700 individuals in 2004 (Table 2.1).

Despite seemingly low dispersal rates, there is concern that pelicans from the

Western Cape colony on Dassen Island could disperse to another of the discrete

breeding colonies in southern Africa and spread a novel cooperative foraging

behaviour, consistent on communally hunting colonial seabird chicks (Chapter

4). The spread of this novel foraging behaviour, which may occur by social

learning, may lead to high mortality of colonial seabird nestlings through much

of their breeding distribution, and possibly trigger ecosystem-wide changes,

intensifying the declining trend of targeted seabird species.

The primary objective of this study was to measure the extent of gene flow

among the breeding colonies of southern African pelicans, particularly with

respect to the Western Cape population acting as a potential source population of

seabird-eating pelicans. To complement the results of mark-recapture data from

southern African pelican populations, a molecular approach to measure gene

flow is particularly appropriate if a management plan is to be developed in a

short-time frame. In addition, I test for the presence of the putative historical

bottleneck followed by population expansion in the Western Cape population

suggested by demographic data (Crawford et al. 1995), as well as the hypothesis

��

158

that immigration from other populations (in particular Namibia) could explain

the exponential population growth experienced in the Western Cape population.

To address these questions I used eight microsatellite loci and one mitochondrial

DNA locus.

��

��

Blood samples were collected from 100 pelicans across three breeding sites in

southern Africa: 63 from Dassen Island in the Western Cape (South Africa) in

three consecutive breeding seasons (20 in 2005; 23 in 2006 and 20 in 2007) and

35 from Hardap Dam, Namibia, in 2006 (Figure 6.1). The prolonged drought

experienced at the most eastern population (Lake St Lucia) prevented the

collection of fresh samples from this population. Two blood samples were

obtained from captive pelicans of confirmed origin kept in the National Zoo in

Johannesburg, South Africa. Additionally, 22 museum skins and egg membranes

from St Lucia pelicans were sampled from the Durban Natural Science Museum

collection, as well as one skin preserved at St Lucia Wetland Park, KZN Wildlife.

DNA was extracted from blood samples using a PureGene DNA purification kit

(Gentra Systems). DNA concentration was measured and standardised to

approximately 10ng/�l before genotyping. For the museum specimens I used both

a DNeasy tissue Extraction Kit (Qiagen Inc.) and the standard Phenol-

Chloroform extraction method. DNA was re-amplified for some low yielding

samples using the GenomiPhi TM DNA amplification kit (Amersham Biosciences)

to increase the amount of genomic DNA available.

All individuals sampled from the Namibian and Western Cape populations were

pre-fledging chicks, captured at their breeding colonies. Great White Pelicans

produce a maximum of one offspring per breeding attempt, due to obligate

siblicide (Cooper 1980). This sampling scheme avoided the potential sampling of

siblings or of a chick and its parents.

��

159

��

Primers L5204 (5’- GCTAACAAAGCTATCGGGCCCAT) and H6312 (5’- CTTATT

TAAGGCTTTGAAGGCC) (Cicero & Johnson 2001) were used to amplify

1 041 bp of the NADH subunit 2 gene (ND2) for 85 individuals from the three

demographic populations (35 from Namibia, 40 from the Western Cape, and 10

from Lake St Lucia, KwaZulu-Natal).

The concentration of DNA from the 23 museum specimen samples was adequate

(above 200ng/�l for some samples). Three internal primer pairs were designed for

fragments of 300-400 bp of ND2, L5204/H2 (5’-CCTGAAGACCCTTAGTCTTCAC-

3’), L1 (5’-CTGTCTCCTACTAACATCAGCAAT)/H1B (5’-CTCGGAACCAACACA

ATT TC-3’) and L2B (5’-CTACTACTGTATTCCTCGCC)/H6312.

PCR amplification was performed using standard conditions (40 cycles:

denaturation at 94°C, annealing at 53°C, and extension at 72°C). PCR products

were directly sequenced in both directions with ABI Big Dye Terminator

chemistry and run on an AB 3730 DNA Analyser. All sequences were aligned

using SEQUENCHER 3.1 (Gene Codes Corporation, Inc.), and verified for accuracy

and the lack of any indels or stop codons.

Eight microsatellite loci described for this species (PEL86, PEL149, PEL175,

PEL188, PEL190, PEL207, PEL265 and PEL304) were PCR-amplified using

published primers and protocols (de Ponte Machado et al. 2009; Chapter 7). The

dataset included 27 pre-fledgling pelican chicks from Hardap Dam (Namibia), 63

from Dassen Island (Western Cape), collected over three consecutive breeding

seasons: 2005 (n = 20), 2006 (n = 23) and 2007 (n = 20) and two from the captive

adult pelicans from KwaZulu-Natal. Allele sizes of fluorescently-labelled

fragments were determined using the size standard GSLIZ-500 on an AB 3730

DNA Analyser, followed by fragment analysis and genotyping using

GENEMAPPER 4.0 (Applied Biosystems).

��

160

��

Mitochondrial DNA variation such as haplotype diversity (h), nucleotide

diversity (�), FST and ST were estimated using ARLEQUIN 3.1 (Excoffier et al.

2005). Relationships between haplotypes were reconstructed in TCS 1.21

(Clement et al. 2000). Neutrality indices (Tajima’s D and Fu’s FS) and mismatch

distributions were also calculated using ARLEQUIN 3.1 to assess population size

trends. To discern whether the neutrality tests were an indication of population

expansion or of selection I used the McDonald-Kreitman (1991) test as

implemented in DNASP 4.50 (Rozas et al. 2003), using the Pink-backed Pelican

Pelecanus rufescens as an outgroup.

The eight microsatellite loci were tested individually for Hardy–Weinberg

equilibrium (HWE) and linkage disequilibrium using GENEPOP 3.3 (Raymond &

Rousset 1995) and ARLEQUIN 3.1 (Excoffier et al. 2005), respectively. Detection of

genotyping errors (null alleles and allelic dropout) were estimated using MICRO-

CHECKER 2.2.3 (van Oosterhout et al. 2004). Allelic frequencies, allelic diversity

as well as observed and expected heterozygosity values were calculated using

GENALEX 6.1 (Peakall & Smouse 2006). A measure of population differentiation

(FST) was estimated using ARLEQUIN 3.1 across all populations, assuming allele

frequencies under the infinite-alleles model. Slatkin’s estimate RST was

calculated using the program RST CALC (Goodman 1997), which incorporates the

stepwise-mutation model and uses the distance method developed by Slatkin

(1995). GENALEX 6.1 was also used to test for matching genotypes and the level

of relatedness among individuals.

A Bayesian clustering method, implemented in STRUCTURE 2.2 (Pritchard et al.

2000; Falush et al. 2003) was used to investigate the presence of current

population structure, assigning individuals into K-populations and identifying

recent migrants and admixed individuals. I used the admixture model, which

assumes that each individual draws some fraction of its genome from each of the

K-populations, without previous information on population identity, and both

��

161

independent and correlated allele frequency models. I ran STRUCTURE for all the

data, including individuals from Lake St Lucia, and also for different subsets of

the sample. The algorithm was run 105 times after a burn-in period of 104 for

each K-value (K = 1 to 4), for five independent iterations, making sure that the

Markov chains converged before commencing sampling from the posterior

distribution.

The programme BAYESASS 1.3 (Wilson & Rannala 2003) was used to estimate

recent migration rates, m (± SD [95% CI]) in the last several generations

between populations and to assign individuals of unknown population affinity to

potential source populations. This model was also used to estimate an inbreeding

coefficient (FIS) using the ‘Identity by Descent’ (IBD) paradigm. The Markov

Chain Monte Carlo (MCMC) chains were run for 30 � 106 iterations with a

sampling frequency of 2 000 and a burn-in of 106 iterations. The prior values for

allele frequency, migration rate and inbreeding were set at 0.20, 0.15 and 0.24,

respectively, to ensure optimal mixing of the Markov chains and maximum

values for the log-likelihood after testing different model parameters. Due to the

small sample size (n = 2) of the KwaZulu-Natal population I excluded it from this

analysis.

Populations that have experienced a recent reduction of their effective

population size show a reduction in the number of alleles and a ‘heterozygosity

excess’ (He) at polymorphic loci. This is because such bottlenecks cause rare

alleles to become less abundant than alleles with intermediate frequencies, but

because rare alleles contribute relatively little to expected heterozygosity, there

will be an excess of observed heterozygosity when compared to a population at

equilibrium with an equivalent number of alleles (Maruyama & Fuerst 1985;

Luikart & Cornuet 1998; Allendorf & Luikart 2007). In order to determine if the

reported bottleneck experienced by the Western Cape pelican population

between the 1930 and 1950s (Crawford et al. 1995) left a footprint in the genetic

variation of the Great White Pelican, I used three statistical tests implemented

��

162

in the software BOTTLENECK 1.2.02 (Cornuet & Luikart 1996; Piry et al. 1999).

The deviation between the expected and observed He was tested using a

nonparametric Wilcoxon test, following recommendations by Piry et al. (1999), as

well as a ‘mode-shift’ test and a ‘sign test’. Most microsatellites appear to evolve

under a mutation model with a large incidence of single-step mutations and a

small percentage of multi-step changes (Luikart & Cornuet 1998). Hence, I used

the two-phase model (TPM; Di Rienzo et al. 1994) with three different settings:

70%, 90% and 95% of alleles following a single-mutation model (SMM), allowing

the remaining percentage to behave under an infinite-alleles model (IAM) or K-

alleles model (KAM).

The heterozygosity excess is maintained for only a few generations, in particular

if the bottleneck was not recent, severe enough, or it did not last long in time

(Luikart & Cornuet 1998). In these cases, other methods than implemented in

BOTTLENECK may be more accurate. Garza & Williamson (2001) developed a

method designed to detect a reduction in population size based on the

distribution of allele frequencies; an approach which the authors claim is able to

detect a reduction in population size even after a complete post-bottleneck

demographic recovery. They developed an algorithm, implemented in the

software M-RATIO that uses the total number of alleles in a population (k) and

the overall range in allele size (r) to calculate the mean ratio M = k / r (Garza &

Williamson 2001). When a population is reduced in size, it will lose rare alleles

more rapidly than reducing the overall allele size range; hence, the ratio M will

be smaller in recently reduced populations than for populations in equilibrium.

The program CRITICAL_M (Garza & Williamson 2001) calculates a critical value

(MC) below which a population is assumed to have experienced a demographic

bottleneck. This value is dependent on the mutational model chosen; I used a

two-phased mutation model with the proportion of single-step mutations set at

ps = 90% and 70% for successive runs. The effective population size at

equilibrium before the bottleneck was considered as Ne = 100, 200, 500 and 1 000

for the Western Cape and Ne = 400 and 1 500 for Namibia. Other model

��

163

parameters were as recommended by the authors: 10 000 replicates; average size

of non one-step mutations, �g = 3.5; and mutation rate, � = 5 � 10-4.

��

��

I recovered 10 mtDNA haplotypes from the 85 pelicans sequenced for ND2 (Table

6.1, Figure 6.2). Due to the highly fragmented nature of the DNA from the 23

museum skins and egg membranes I obtained six whole copies of ND2, and two

samples with two fragments each (861 bp for sample MPM324, and 746 bp for

sample MPM322) from the Lake St Lucia, KwaZulu-Natal, population. The

Western Cape population was represented by a single haplotype (PEL01NAM,

n = 40), which was also the most frequent haplotype for both the Namibian (28 of

the 35 samples) and KwaZulu-Natal (7 of the 10 samples) populations. Six other

haplotypes were present in Namibia (five restricted to single individuals and one

haplotype shared by two pelicans), and three haplotypes were exclusive to the

KwaZulu-Natal population. Although the Namibian population had the largest

number of polymorphic sites (six), the haplotype diversity estimate (H) was

larger for the KwaZulu-Natal population (0.533 ± 0.180) than for the Namibian

population (0.363 ± 0.104) given the smaller sample size (Table 6.2).

Overall FST and ST values for mtDNA were estimated at 0.096 (p < 0.005) and

0.109 (p < 0.005), respectively, which is suggestive of population differentiation,

with a relatively low rate of maternal gene flow among the three southern

African pelican populations. Pairwise population ST values indicated that the

differences are primarily between the Namibian and Western Cape populations

( ST = 0.043; p < 0.0001) and between the Western Cape and KwaZulu-Natal

populations ( ST = 0.325; p < 0.0001), although a significant difference was still

recovered between the Namibian and KwaZulu-Natal populations ( ST = 0.121; p

= 0.04; Table 6.3a).

��

164

Mismatch profiles that adjust to a one-tailed Poisson distribution for the

Namibian and KwaZulu-Natal populations indicated a recent population

expansion. The values of goodness-of-fit statistic Hrag were 0.21 (ns) for Namibia

and 0.40 (ns) for KwaZulu-Natal, meaning that neither population was static,

and suggesting the possibility of recent population expansion. Negative and

significant values of Tajima’s D and Fu’s Fs indicated non-equilibrium, either

due to a population expansion or to selection, which was particularly pronounced

for the Namibian population (DTaj = -1.63, p < 0.05; Fu’s F = -4.45, p < 0.001)

(Table 6.2). Neither test could be implemented for the Western Cape population

because it was represented by a single haplotype. The differences between and

within species in the ratio of synonymous and non-synonymous substitutions

obtained with the McDonald-Kreitman test for selection were not significant

(G = 0.155; p = 0.69), suggesting that the significant results of Tajima’s D and

Fu’s F tests were indicative of a recent population expansion rather than

selection on the mtDNA locus.

��

Attempts to amplify the museum samples for the microsatellite loci failed

because the genotypes obtained were not consistent across runs and hence were

excluded from the analyses. All other microsatellites amplified consistently,

producing unambiguous peaks (Table 6.4; Appendix 6.1). All loci were

polymorphic; the number of alleles ranged from two to 20. A total of 66 alleles

were observed among all geographic locations across the eight loci. HO and HE

ranged from 0.250–0.963 (Appendix 2 and 3) and 0.219–0.929, respectively.

PEL149 was the most diverse locus across all populations, driving most of the

signal found in this analysis. The Namibian population was the most

polymorphic, displaying 12 unique alleles among four of the eight loci (Table 6.4).

No linkage disequilibrium was detected for the eight loci after applying the

Bonferroni correction for multiple comparisons (Rice 1989), and all loci were in

agreement with HWE expectations. No evidence of null alleles or other

genotyping errors were detected.

��

165

No matching multilocus genotypes were found using Match analysis in GenAlex

6.1, indicating that all samples could be individually identified using these eight

loci. A combination of three loci was enough to distinguish all individuals, and

the probability of two individuals sharing the same genotype from a random

sample using the eight loci is 10-7. At least 58 individuals shared at least one

allele at all loci with another individual, although most of the matches occurred

within the Western Cape population (n = 50 matches). This seems to indicate

that there is some degree of relatedness. The sampling protocols exclude the

possibility of sampling parents and offspring, or siblings from the same pair and

same year, but not siblings from different years. Alternatively, there could be

high levels of extra-pair paternity, or the population exhibited a high level of

shared alleles due to a reduced population size in the past as predicted from

historical demographic records (Crawford et al. 1995). The probability of identity

of siblings (PIsibs) is 10-3, which indicates that additional markers would be

needed in order to do any forensic or relatedness studies.

��

Overall genetic differentiation for the microsatellite loci was estimated to be

FST = 0.044 (p < 0.0001) when all the populations were included, and FST = 0.041

(p < 0.0001) when the KwaZulu-Natal population was excluded on the bases of

small sample size (n = 2; museum samples and egg membranes did not amplify).

Pairwise population FST values were significant between Namibia and the

Western Cape (FST = 0.044; p < 0.0001; Table 6.3a). No differentiation was found

among sampling years (2005, 2006, 2007) for the Western Cape population, but

significant differences were found between each year and the Namibian

population (Table 6.3b). RST values between Namibian and Western Cape

populations were also significant (RST = 0.055, p < 0.0001) and estimates of gene

flow were moderate to low, inferring the number of migrants per generation (Nm)

as 4.32.

��

166

��

One panmictic population was identified when using the independent allele

frequency model with the program STRUCTURE. This model assumes that

populations diverged under a scenario of little to null gene flow and therefore

predicts very different allelic frequencies among populations (Falush et al. 2003).

However, when assuming the correlated allele frequency model, which allows

some degree of gene flow among populations, and is concordant with the mark-

recapture data available (Chapter 3), the highest posterior probability of the

number of populations (k), was identified as k = 2 (Figure 6.3). The two detected

populations largely corresponded to the expected distributional pattern and were

analogous with the sampling localities of the Western Cape and Namibia. These

results remained consistent whether the KwaZulu-Natal individuals were

included or not, although the sample size (n = 2) was not sufficient to consider

the KwaZulu-Natal population as a distinct demographic unit.

Mean immigration rates obtained with BAYESASS, although low, showed a

marked asymmetry in the direction of dispersal, being almost exclusively from

the Western Cape to Namibia (m = 0.16 ± 0.07 [0.03–0.32]), whereas dispersal in

the opposite direction was negligible (m = 0.01 ± 0.008 [0.0002-0.03]). A large

proportion of individuals derived from their own population (84% ± 7% for

Namibia and 99% ± 0.8% for the Western Cape) suggesting isolation between

both breeding populations, at least with respect to first- and second-generation

migrants.

BAYESASS individual assignments were compared to the genotype sorting

proportions produced with STRUCTURE. Four individuals from Namibia expressed

two distinct sources of alleles according to STRUCTURE: individuals H1999,

H1994, H1995 and H00719 had a 78.7%, 32.7%, 63.9% and 41.4% probability of

assignment to their sampling population, respectively. These same individuals

were identified by BAYESASS as individuals of mixed origin with assignments of

12.4%, 5.0%, 11.3% and 3.1%, respectively. Mostly, they were identified as

��

167

second-generation immigrants (offspring of a Namibian pelican and a migrant

from the Western Cape), except for pelican H1994, which was assigned as a first-

generation immigrant with a 65% probability (Figure 6.4a). This was in

concordance with STRUCTURE that suggested that 32.7% of H1994’s admixed

genotype corresponded to allelic diversity in its source population. Individual

H1539 was assigned less than 30% of the times to its source population

(Namibia) by BAYESASS (Figure 6.4a), but showed a likelihood of 86.5% of

belonging to the Namibian population in STRUCTURE. An additional five

individuals sampled in Namibia showed less than 50% assignment to the

population from which they were sampled according to BAYESASS (data not

shown), although according to STRUCTURE all five individuals belonged to the

sampled population (> 90%).

All Western Cape pelicans were unequivocally assigned to their own source

population according to BAYESASS. However, STRUCTURE suggested that at least

four admixed individuals (H1383, H1389, H1421, H00893) had low probability of

belonging to their sampled population (18.1%, 15.1%, 15.8% and 11.9%,

respectively). BAYESASS recognised nearly 15% of the genome to be of mixed

origin in individual H1421 but the other three individuals could be assigned with

a confidence greater than 95% to their sampled population (Figure 6.4b). In

addition, five other individuals sampled in the Western Cape showed some

genetic mixing according to STRUCTURE (Figure 6.3), but > 92% of assignment to

the sampled population according to BAYESASS.

��

The Wilcoxon sign rank test implemented in the program BOTTLENECK was

significant for the Western Cape population (p = 0.02) when using a two-phased

model of mutation (TPM) with 70% of mutational dynamics being accounted for

by single-step mutations (SSM). However, the allele frequency distribution ‘mode

shift’ graphical detection method did not indicate a recent bottleneck in the

Western Cape population irrespective of the SSM contribution (90% or 95%).

��

168

Unexpectedly, I detected a stronger signal of a bottleneck for the Namibian

population: significant with both a 70% of SSM contribution (p = 0.004) and with

a 95% of SSM (p = 0.04), even though the two other tests implemented in the

program (sign rank test and the ‘mode-shift’ graphical method) did not show

evidence of a bottleneck for the Namibian population.

Similarly, the M-ratio test did not provide evidence for a population bottleneck in

either the Western Cape or Namibian pelican populations, as the observed M-

ratios (0.855 ± 0.031 and 0.909 ± 0.012, respectively) were well-above the

expected ratio at equilibrium, and the critical value Mc was smaller than the

observed values of M for all parameter scenarios tested and under different

mutational models. Garza & Williamson (2001) reported that values of M higher

than 0.82 are characteristic of stable populations that have not suffered a known

reduction in size, according to the results of simulations and of datasets of

natural populations analysed.

No evidence of inbreeding was found for either population (Namibia FIS = 0.055 ±

0.097; Western Cape FIS = 0.036 ± 0.021) with BAYESASS (Wilson & Rannala

2003). Collectively these results suggested that the observed demographic

bottleneck in the Western Cape population was of relatively short duration and

hence this population was able to retain nuclear diversity relative to mtDNA due

to the differences in effective population size between the markers.

��

��

Significant population structure was found between the three pelican

populations in southern Africa. Differentiation was consistent for both mtDNA

and microsatellites ( ST = 0.109, p < 0.005; RST = 0.055, p < 0.0001). Gene flow

coefficients were low, indicating limited gene flow among populations. The

pattern of pairwise FST comparisons indicated that structure was found between

��

169

the Western Cape and Namibian populations (FST = 0.043, p < 0.0001), and

between the Western Cape and KwaZulu-Natal (FST = 0.325, p < 0.0001), at least

with regards to mtDNA. Comparisons between Namibia and KwaZulu-Natal

suggested that there was a larger degree of maternal gene flow between these

populations, as exemplified by the lower significance levels observed (p = 0.04).

Microsatellite data indicated a similar result, denoting population structure and

negligible gene flow between the Western Cape and Namibia. The small sample

size of the KwaZulu-Natal population (n = 2) was insufficient to extract any

conclusive result, other than that these individuals seemed to be more closely

assigned to the Namibian population, suggesting potential contact of these two

populations through northern routes. However, the existence of distinct mtDNA

haplotypes in both the Namibian and KwaZulu-Natal populations (Figure 6.2)

also suggested that these two populations do not belong to a panmictic

population. A larger sampling effort is needed from KwaZulu-Natal to be able to

infer any conclusive results from the nuclear markers for this population.

Although we attempted to sample this population several times, the protracted

drought in Lake St Lucia caused pelicans to vacate the Lake since 2005 and

prevented them from breeding in 2006–2008 (Bowker & Downs 2008; R Taylor,

pers. comm.). Sampling this population should be considered a future priority.

Although the distribution of pelicans along the west coast of southern Africa

clustered around two clearly distinct breeding areas, the observation of pelicans

at intermediate sites (i.e. Orange River mouth) was frequent (Hockey et al.

2005), which suggests some degree of movement among or to and from the

breeding colonies. Temporally-spaced re-sightings of the same individuals

confirmed that some young birds dispersed northwards from the breeding colony

in the Western Cape and returned to their breeding area after one or two years

(Chapter 2). Most pelican ringing effort took place in the Western Cape (728

ringed chicks since 2001), and four cohorts of pelican chicks (n = 404) were

ringed in Namibia in 2004–2008. Additionally, 66 and 14 adults were ringed in

the Western Cape and Namibia, respectively. Despite this, not one of the ringed

��

170

pelicans from the Western Cape has been recovered or re-sighted among the

coastal Namibian population, or vice versa, yet two pelicans from Hardap Dam

(Namibian inland colony) were recovered further north from their colony, one in

the Eastern Caprivi Strip and another near Maun in Botswana (Figure 6.1). In

addition, one young bird travelled from the ringing site in the Western Cape to

northern Botswana (>3 200 km; Figure 6.1), seen there eight months after

fledgling. It presumably followed the coastal route up to the Orange River and

then north through Namibia (Chapter 2), and confirmed at least occasional

contact between coastal pelicans and other more northern inland populations.

Genetic results confirmed the observed movement patterns detected by mark-

recapture studies. For instance, the existence of a single mtDNA haplotype in

the Western Cape population and microsatellite-based assignment tests both

suggested that there has been no recent dispersal from Namibia towards the

Western Cape (m = 0.01), but small to moderate gene flow from the Western

Cape to Namibia (m = 0.16). This pattern of asymmetrical gene flow could result

from the population expansion experienced by the Western Cape population,

which would function as a source population exporting recruits to Namibia.

However, due to the design of sampling protocols, all individuals with admixed

genotypes must be either second-generation migrants (offspring of a migrant and

a local) or perhaps the offspring of two migrants, but not the actual migrants,

because sampling included only pre-fledgling chicks from each of the discrete

breeding colonies. Despite sporadic dispersal, the coastal populations of the

Western Cape and Namibia are mostly philopatric (also suggested by Crawford

et al. 1994; Nelson 2006), whereas the inland populations are nomadic, breeding

infrequently at different localities according to favourable environmental

conditions (Crawford et al. 1994; Williams & Randall 1995).

In summary, the significant structure found in the molecular data suggests that

gene flow among southern African pelican populations is restricted. Further, it

appears that some of the observed northward movement along Africa’s west

��

171

coast from the Western Cape towards Namibia corresponds to juvenile dispersal,

and apparently, pelicans return to their natal colony in the Western Cape to

breed (Figure 2.6, Chapter 2). My results suggest that the southern African

Great White Pelican populations exhibits a typical metapopulation structure, in

which local-discrete populations are connected by infrequent dispersing

individuals and experience long-term patterns of local extinction and

recolonisation. Similar metapopulation dynamics have been described for

populations of American White P. erythrorhynchos (Anderson & King 2005) and

Brown Pelicans P. occidentalis (Anderson et al. 2007).

��

The recovery of a single haplotype of mitochondrial DNA from the Western Cape

population, coupled with a reduced number of alleles, lower observed

heterozygosity and lower number of private alleles in the microsatellite data

compared to the Namibian sample indicated reduced genetic diversity in the

Western Cape population. This could be explained by either a recent bottleneck

or a founder event. The former explanation is preferred because it is concordant

with the known demographic history of the population. Pelicans inhabited the

region prior to the arrival of European settlers in the early 1600s (Crawford et al.

1995). From the beginning of the 20th century, pelicans were persecuted and

chased off the islands, mostly by collectors of guano, but also because of the

expansion of the Cape Fur Seal Arctocephalus pusillus population and by naval

activities that utilised the islands as military targets (Crawford et al. 1995).

Historical data suggested that between 1930 and 1950, a maximum of 20–30

pairs of pelicans bred in the Western Cape (Crawford et al. 1995), and the species

was considered rare in the region. Surprisingly, molecular data did not provide

conclusive evidence of a recent bottleneck in the Western Cape population; only

the Wilcoxon test indicated a bottleneck signal with a 70% SSM model. Only

bottlenecks that occurred 0.25–2.5 * 2Ne generations ago are readily detected by

BOTTLENECK, because changes in allelic frequencies are transient (Cornuet &

Luikart 1996). The time elapsed since the bottleneck occurred (ca. 3 generations;

��

172

where generation time, T, is calculated as T = + [s / (1 – s)]; being , age of

first breeding and s, adult survival; Saether et al. 2005) is too short to be take

this as an explanation for not finding molecular evidence of bottleneck. Likely,

the effects of the extensive post-bottleneck demographic expansion, as detected

by mismatch distributions, have masked the excess of heterozygosity that is

indicative of this type of reduction in population size.

Pelicans breed at two localities in Namibia. The samples used in this study were

collected at Hardap Dam, an artificial water body built in 1962. Pelicans breed

on this site intermittently: only when rain has been abundant and an island is

formed in the middle of the dam (Williams & Randall 1995; Theron et al. 2003).

During such favourable times, colonies of 250–300 breeding pairs (Holger

Kolberg, in litt.) formed but no long-term demographic data are available for this

site (Theron et al. 2003; Simmons & Brown 2009). From 1949, Great White

Pelicans were recorded breeding at Bird Rock platform, Walvis Bay, built in the

1930s to provide breeding area for the commercial exploitation of guano (Berry

1975). Pelican numbers have remained low, fluctuating between a maximum of

400 pairs in the early 1970s and less than 100 subsequently (Crawford et al.

1981; Simmons & Brown 2009; Williams, in press). Also, pelicans used to breed

further south on an island in Sandwich Harbour, until it was linked to the

mainland by a sandbank in 1947; the arrival of terrestrial predators marking the

end of the breeding colony on this island (Crawford et al. 1981; Simmons &

Brown 2009). Given these fluctuations and documented local population

extinction, it is not entirely surprising to find some evidence of heterozygote

excess in the Namibian population, possibly due to founder effects following local

extinction or severe reductions in size.

M-RATIO did not suggested a reduction in population size for either the Western

Cape or Namibian population, despite the claim that this method is able to

detect bottlenecks that occurred up to 125 generations post-bottleneck (Garza &

Williamson 2001). The failure of M-RATIO to detect any bottlenecks could be

��

173

influenced by the small number of polymorphic microsatellite loci. Three of the

loci used in this study had four alleles and one locus only two alleles, yielding a

M-ratio of K / r = 1 for four of the loci and the resulting loss of statistical power.

Garza & Williamson (2001) also report that the method does not perform well

when � is small in the starting population (transition from a small population, in

which drift is already important, to an even smaller population), which could be

the case for pelicans in the Western Cape, and perhaps in Namibia. A third

plausible explanation, that new alleles were introduced into the Western Cape

population through immigration, was discarded due to evidence from mark-

recapture studies (Chapter 2 and 3) and the small migration rates obtained from

these genetic analyses. Similarly, Whitehouse & Harley (2001) analysed nine

polymorphic microsatellite loci of a southern African Elephant Loxodonta

africana population that had experienced a confirmed reduction in population

size (down to 11 animals in 1931), and failed to detect the expected bottleneck

effects with the same methods implemented here (He excess, shift distribution

and M-ratio). Collectively, the results from Whitehouse & Harley (2001) and this

study warn about the application of these methods for populations with unknown

demographic history, especially in the case of threatened populations.

On the other hand, comparative measures of allelic diversity (between the

suspected bottlenecked population and another population thought to be

demographically stable) have proven to be one of the most powerful methods for

detecting population bottlenecks (Garza & Williamson 2001). In this study, the

bottleneck suffered by the Western Cape population was apparent from the

reduced number of alleles relative to the Namibian population, and the existence

of a single haplotype in the Western Cape for the mitochondrial gene ND2. Also,

the pattern of low or no variability in mtDNA and the greater levels of variability

in nuclear DNA observed in this study are shared with several populations of

vertebrates in southern Africa (Grant & Leslie 1993). This is likely a

consequence of metapopulation dynamics, in which frequent local extinctions and

recolonisations due to the paleo-climatic history of the region have caused a

��

174

greater reduction in variability in mtDNA due to its smaller effective population

size than in nuclear DNA.

Mismatch distributions indicated a recent population expansion in both the

Western Cape and Namibia. Census data estimates for 2004–2007 were ca. 2 500

pelicans in the Western Cape (Chapter 2) and between 1 500 and 3 000 pelicans

in Nambia (Namibian Ministry of Environment and Tourism, unpublished data;

MdPM, unpublished data). The Western Cape population increased roughly four-

fold from 600 pelicans in 1976–1985 (Crawford et al. 1995; Chapter 2). Because

immigration into the Western Cape population did not play a role in explaining

population growth, it is implied that this population expansion must be a result

of increased productivity and/or reduced mortality. Changes in local conditions

(better protection and larger breeding area, and mostly an increased availability

of food) explain the increase in pelican numbers in the Western Cape

(Chapter 2). Furthermore, the population expansion could be responsible for the

change in foraging habits (increased predation; Chapter 4), and likely to have

triggered the dispersal of pelicans to other areas further north.

The limited sample size from KwaZulu-Natal prevents us from inferring

demographic trends for this population based on genetic markers. This

population experiences severe episodes of drought, markedly since 2005 (R.

Taylor, KZN Wildlife, in litt.) causing drastic reductions in the area occupied by

Lake St Lucia and diminished water levels. Pelican numbers in the region have

experienced large fluctuations because of changes in the availability of fish. In

September 2005 the number of pelicans feeding in the lake reached a maximum

of 6 000 birds. Soon afterwards, their numbers dropped to a few tens, as the lake

dried out almost completely (Bowker & Downs 2008). Subsequently, large flocks

of pelicans appeared in the Okavango region in northern Botswana, following

seasonal floods (P. Hanckock, in litt.). Thus, it seems reasonable to assume that

the pelicans that left Lake St Lucia after depleting the fish stocks were the same

that appeared in Botswana. The inland population of pelicans in southern Africa

��

175

seems to be highly dispersive (Crawford et al. 1994; Williams & Randall 1995),

presumably responding to extreme fluctuations in their environment (Berry et al.

1973; Berruti 1983; Williams & Randall 1995; Bowker & Downs 2008). Even

though the small wetlands characteristic of the Western Cape are less productive

than the larger inland lakes available to the central and eastern African pelican

populations, more stable winter rainfall conditions and lower incidence of multi-

annual cycles of drought and floods have selected for a relatively small and more

sedentary pelican population in the Western Cape. Sampling of nomadic pelican

flocks could confirm the hypothesis of nomadism among northern and eastern

pelican populations.

��

Pelican predation of seabird chicks in the Western Cape constitutes a novel

threat to endangered and vulnerable species, many of which are endemic to the

Benguela coastal region (Chapter 4 and 5). Pelicans are aquatic feeders with a

limited range of foraging behaviour and food types (mostly fish, frogs and some

crustaceans). Predation on other seabirds represents an adaptation of foraging

techniques from water to land (Chapter 4). Novel foraging techniques can be

culturally transmitted among groups of animals (Giraldeau et al. 1994; Galef &

Giraldeau 2001). Pelican predation on seabirds was reported anecdotally from

Namibia (Berry 1976) but it has never reached the scale observed in the Western

Cape. This study confirms that pelicans from the Western Cape occasionally

disperse northwards, thus allowing for the possibility of exporting this

cooperative predatory behaviour to northern localities (e.g. Namibia), and

therefore increasing the level of threat to regionally endemic seabird species,

many of which are included in IUCN threat categories (Kemper et al. 2007).

Even at a population level, changes in top-down regulation of ecosystems have

been reported to trigger trophic cascades (Polis 1994; Estes et al. 1998; Pace et al.

1999; Polis et al. 2000). The juxtaposition of two habitats differing in productivity

as it is the case of the highly productive Benguela Upwelling Ecosystem and the

��

176

nutrient-poor terrestrial southern African west coast favours the transfer of

energy and biomass from the sea to the land (Polis & Hurd 1996), mostly

through shore drift and seabird colonies (Polis & Hurd 1995; Anderson & Polis

1999). Seabird populations in the Benguela have decreased as a result of

multiple factors (e.g. direct exploitation, pollution, competition with fisheries,

habitat modifications and possibly climate change) (du Toit et al. 2002; Crawford

2007; Crawford et al. 2007). Pelican predation places additional pressure on

these populations, being especially severe on offshore islands where terrestrial

predators have not previously occurred. Managers of both the Western Cape and

now also the Namibian (results of this study) seabird populations should be

concerned that pelican predation could affect recruitment of colonial seabird

species and thus trigger changes in the dynamics of local marine and terrestrial

ecosystems.

��

177

��

Allendorf FW, Luikart G (2007) Conservation and the genetics of populations.

Blackwell, Oxford.

Anderson DW, Henny CJ, Godinez-Reyes C, Gress F, Palacios EL, Santos del

Prado K, Bredy J (2007) Size of the California Brown Pelican

metapopulation during a non-El Niño year. U.S. Geological Survey,

Reston, Virginia.

Anderson DW, King DT (2005) Introduction: biology and conservation of the

American White Pelican. Waterbirds 28, 1–8.

Anderson WB, Polis GA (1999) Nutrient fluxes from water to land: seabirds

affect plant nutrient status on Gulf of California islands. Oecologia 118,

324–332.



Berruti A (1983) The biomass, energy consumption and breeding of waterbirds

relative to hydrological conditions at Lake St Lucia. Ostrich 54, 65–82.







Madoqua (Ser. 1) 7, 17–31.

Bowker MB, Downs CT (2008) Fluctuations in numbers of Great White Pelicans

at Lake St Lucia in response to changing water-levels. Journal of African

Ecology 46, 282–290.

Cicero C, Johnson NK (2001) Higher-level phylogeny of New World vireos (Aves:

Vireonidae) based on sequences of multiple mitochondrial DNA genes.

Molecular Phylogenetics and Evolution 20, 27-40.

Clement M, Posada D, Crandall KA (2000) TCS: a computer program to estimate

gene genealogies. Molecular Ecology 9, 1657–1659.

Cooper J (1980) Fatal sibling aggression in pelicans – a review. Ostrich 51, 183–

186.

Cornuet JM, Luikart G (1996) Description and power analysis of two tests for

detecting recent population bottlenecks from allele frequency data.

Genetics 144, 2001–2014.

Crawford RJM (2007) Food, fishing and seabirds in the Benguela upwelling

system. In: Final report of the BCLME (Benguela Current Large Marine



Unit, Cape Town.



Africa's Western Cape. South African Journal of Marine Science 15, 33–

42.

��

178



1949/78. Fisheries Bulletin of South Africa 15, 67–70.

Crawford RJM, Fairweather T, Underhill LG, Wolfaardt A (2007) Implications

for seabirds of an unfavourable, long-term change in the distribution of

prey: a South African experience. In: Final report of the BCLME (Benguela

Current Large Marine Ecosystem) project on top predators as biological

indicators of ecosystem change in the BCLME (ed. Kirkman SP), pp. 243–

249. Avian Demography Unit, Cape Town.

Crawford RJM, Dyer BM, Brooke RK (1994) Breeding nomadism in Southern

African seabirds– constraints, causes and conservation. Ostrich 65, 231–

246.






Unit, Cape Town.

de Ponte Machado M, Feldheim KA, Sellas AB, Bowie RCK (2009) Development

and characterization of microsatellite loci from the Great White Pelican

Pelecanus onocrotalus and widespread application to other members of the

Pelecanidae. Conservation Genetics 10, 1033–1036.



Di Rienzo A, Peterson AC, Garza JC, Valdes AM, Slatkin M, Freimer NB (1994)

Mutational processes of simple-sequence repeat loci in human populations.

Proceedings National Academy of Sciences 91, 3166–3170.

du Toit M, Boere GC, Cooper J, de Villiers MS, Kemper J, Lenten B, Petersen S,

Simmons RE, Underhill LG, Whittington PA, Byers OP (2003)


coastal seabirds. In: Workshop report, Cape Town, South Africa.

Estes JA, Tinker MT, Williams TM, Doak DF (1998) Killer Whale predation on

Sea Otters linking oceanic and nearshore ecosystems. Science 282, 473–

476.

Excoffier L, Laval G, Schneider S (2005) ARLEQUIN v 3.0: an integrated software

package for population genetics data analysis. Evolutionary

Bioinformatics Online 1, 47–50.

Falush D, Stephens M, Pritchard JK (2003) Inference of population structure

using multilocus genotype data: linked loci and correlated allele

frequencies. Genetics 164, 1567–1587.

Galef BG, Giraldeau L-A (2001) Social influences on foraging in vertebrates:

causal mechanisms and adaptive functions. Animal Behaviour 61, 3–15.

Garza JC, Williamson EG (2001) Detection of reduction in population size using

data from microsatellite loci. Molecular Ecology 10, 305–318.

Giraldeau L-A, Caraco T, Valone TJ (1994) Social foraging: individual learning

and cultural transmission of innovations. Behavioral Ecology 5, 35–43.

��

179

Goodman SJ (1997) RST CALC: a collection of computer programs for calculating

estimates of genetic differentiation from microsatellite data and

determining their significance. Molecular Ecology 6, 881–885.

Grant WS, Leslie RW (1993) Effect of metapopulation structure on nuclear and

organellar DNA variability in semi-arid environments of southern Africa.

South African Journal of Marine Science 89, 287–293.



Kemper J, Underhill L, G, Crawford R, J, M, Kirkman SP (2007) Revision of the






Luikart G, Cornuet JM (1998) Empirical evaluation of a test for identifying

recently bottlenecked populations from allele frequency data. Conservation

Biology 12, 228–237.

Maruyama T, Fuerst PA (1985) Population bottlenecks and non-equilibrium

models in population genetics. II. Number of alleles in a small population

that was formed by a recent bottleneck. Genetics 111, 675–689.

McDonald JH, Kreitman M (1991) Adaptive protein evolution at the Adh locus in

Drosophila. Nature 351, 652–654.





Peakall R, Smouse PE (2006) GENEALEX 6: Genetic analysis in Excel. Population

genetic software for teaching and research. Molecular Ecology Notes 6,

288–295.

Piry S, Luikart G, Cornuet JM (1999) BOTTLENECK: A computer program for

detecting recent reduction in the effective population size using allele

frequency data. Journal of Heredity 90, 502–503.

Polis GA (1994) Food webs, trophic cascades and community structure. Austral

Ecology 19, 121–136.

Polis GA, Hurd SD (1995) Extraordinarily high spider densities on islands: flow

of energy from the marine to terrestrial food webs and the absence of

predation. Proceedings National Academy of Sciences 92, 4382–4386.

Polis GA, Hurd SD (1996) Linking marine and terrestrial food webs:

allochthonous input from the ocean supports high secondary productivity

on small islands and coastal land communities. American Naturalist 147,

396–423.

Polis GA, Sears ALW, Huxel GR, Strong DR, Maron J (2000) When is a trophic

cascade a trophic cascade? Trends in Ecology & Evolution 15, 473–475.

Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure

using mulitlocus genotype data. Genetics 155, 945–959.

��

180

Raymond ML, Rousset F (1995) An exact test for population differentiation.

Evolution 49, 1280–1283.

Rice WR (1989) Analyzing tables of statistical tests. Evolution 43, 223–225.

Rozas J, Sánchez-DelBarrio JC, Messenguer X, Rozas R (2003) DNASP, DNA

polymorphism analyses by the coalescent and other methods.

Bioinformatics 15, 174–175.

Sæther B-E, Lande R, Engen S, Weimerskirch H, Lillegard M, Altwegg R, Becker

PH, Bregnballe T, Brommer JE, McCleery RH, Merila J, Nyholm E,

Rendell W, Robertson RR, Tryjanowski P, Visser ME (2005) Generation

time and temporal scaling of bird population dynamics. Nature 436, 99–

102.

Simmons RE, Brown CJ (2009) Birds to watch in Namibia: red, rare and endemic

species. National Biodiversity Programme and Namibia Nature

Foundation, Windhoek, Namibia.

Slatkin M (1995) A measure of population subdivision based on microsatellite

allele frequencies. Genetics 139, 457–462.


Pelicans Pelecanus onocrotalus at Hardap Dam, Namibia. Lanioturdus 36,

2–7.

van Oosterhout CV, Hutchinson WF, Wills DPM, Shipley P (2004)

MICROCHECKER: Software for identifying and correcting genotyping errors

in microsatellite data. Molecular Ecology 4, 535–538.

Whitehouse AM, Harley EH (2001) Post-bottleneck genetic diversity of elephant

populations in South Africa, revealed using microsatellite analysis.

Molecular Ecology 10, 2139–2149.

Williams AJ (in press) Birds at Bird Rock Guano Platform, Walvis Bay, Namibia.

Ostrich.



Herremans M, Tree AJ, Parker V, Brown CJ), vol. 1: Non-passerines, pp.


Williams AJ, Randall RM (1995) Pelecaniform birds in South African wetlands.

In: Wetlands of South Africa (ed. Cowan GI), pp. 147–161. Department of

Environmental Affairs and Tourism, Pretoria.

Wilson GA, Rannala B (2003) Bayesian inference of recent migration rates using

multilocus genotypes. Genetics 163, 1177–1191.

�

181

�

Table 6.1: Great White Pelican mitochondrial NADH2 (1 041 bp) haplotypes with

reference to the most common haplotype PEL01, present in all three populations

(Western Cape, WC; Namibia, Nam; KwaZulu-Natal, KZN). Vertical numbers above

columns indicate the position in the base pair sequence at which the variable site was

detected. Haplotypes PEL11 and PEL12 ,recovered from samples from historical

museum specimens, were not completely resolved.

126 157 436 483 504 555 858 876 891 915 1026

PEL01 T A T A A T G A T

PEL02 (Nam) . . . .

A

.

G

. G . A . .

PEL03 (Nam) . . . . . . G A . C

PEL04 (Nam) . . . . . . . . . . C

PEL05 (Nam) . . . . . . . . . G .

PEL06 (Nam) . G . . . . . . . . .

PEL07 (Nam) . . . G . . . . . . .

PEL09 (KZN) . . . . G . . C . . .

PEL11 (KZN) C . ? ? ? ? . C . . .

PEL12 (KZN) . . C . . A ? ? ? ? ?

�

182

�

Table 6.2: Variation in mtDNA sequences and measures of selective neutrality from the

three sampling areas: (n) sample size, (H) haplotype diversity, (�) nucleotide diversity,

(DTaj,) Tajima’s D and (Fu’s Fs ) Fu’s statistics. The Western Cape Great White Pelican

population is represented by one haplotype, hence the zero values for H, � and the

inability to perform the neutrality tests.

Table 6.3: (a) Pairwise FST values for microsatellite markers (above the diagonal) and

ST values (Tamura and Nei method) for mtDNA (below) for the three sampling areas;

(b) FST comparisons among sampling years (2005, 2006, 2007) for the Western Cape

(WC) Great White Pelican population. P-values significantly different from zero are

indicated by asterisks (*p < 0.05; **p < 0.0001; after Bonferroni corrections for multiple

tests). Due to the small sample size (n =2) for the KwaZulu-Natal (KZN) population it is

difficult to interpret the results from this population with any confidence.

Localities n Polymorphic

sites

No.

haplotypes

(private)

H

(sd)

� * 10-3

(sd)

DTaj

(p)

Fu’s Fs

(p)

Namibia 35 6 7 (6) 0.363

(0.104)

0.584

(0.528)

-1.61

(0.03)

-4.45

(<0.001)

Western Cape 40 0 1 (0) 0 0 N/A N/A

KwaZulu-Natal 10 5 4 (3) 0.533

(0.180)

0.598

(0.909)

-1.11

(0.19)

-4.59

(<0.001)

(a) FST (msats)/ ST (mtDNA) (b) FST (msats)- WC yearly

Namibia WC–combined KZN WC 2005 WC 2006 WC 2007

Namibia - 0.044** 0

0.044** 0.039** 0.042**

WC: all years 0.043** - 0.034 - - -

KZN 0.121* 0.325** - 0.062 0.021 0.034

WC 2005 - 0 0

WC 2006 - 0.004

WC 2007 -

� � Ta

ble

6.4

: S

um

ma

ry s

ta

tis

tic

s o

f m

icrosa

tell

ite l

oci

typ

ed

for 9

2 i

nd

ivid

ua

ls f

rom

th

ree p

op

ula

tio

ns o

f G

rea

t W

hit

e P

eli

ca

ns:

Na

mib

ia

(n =

27

), t

hree c

on

secu

tiv

e b

reed

ing s

ea

son

s f

or t

he W

estern

Ca

pe p

op

ula

tio

n (

n =

63

) a

nd

Kw

aZ

ulu

-Na

ta

l (n

= 2

). N

A,

nu

mb

er o

f a

llele

s;

b,

nu

mb

er o

f p

riv

ate a

llele

s;

Ho a

nd

He

rep

resen

t t

he o

bserved

an

d e

xp

ected

heterozygosit

y,

resp

ectiv

ely

. *M

ore d

eta

ils a

bou

t t

he

mic

rosa

tell

ites (

rep

ea

t m

otif

, P

CR

con

dit

ion

s a

nd

prim

ers)

ca

n b

e f

ou

nd

in

de P

on

te M

ach

ad

o e

t a

l. (

20

09

).

N

am

ibia

(n =

27

)

We

ste

rn

Ca

pe

20

05

(n =

20

)

We

ste

rn

Ca

pe

20

06

(n =

23

)

We

ste

rn

Ca

pe

20

07

(n =

20

)

We

ste

rn

Ca

pe

co

mb

ine

d

(n =

63

)

Kw

aZ

ulu

-Na

tal

(n =

2)

Lo

cu

s *

N

A

(b)

HO

H

E

N

A

(b)

HO

H

E

N

A

(b)

HO

H

E

N

A

(b)

HO

H

E

N

A

(b)

HO

H

E

NA

(b)

HO

H

E

PE

L0

86

1

1 (

2)

0.9

63

0

.87

3

7

0

.82

6

0.8

18

9

0.8

00

0

.83

6

7

0

.73

7

0.7

77

9

0.7

90

0

.82

8

3

0.5

00

0

.62

5

PE

L1

49

2

0 (

7)

0.9

26

0

.92

9

1

0

0.8

70

0

.84

1

1

2

0.8

00

0

.76

9

1

0

0.7

00

0

.70

8

1

2 (

1)

0.7

94

0

.78

9

4

1.0

00

0

.75

0

PE

L1

75

6

0

.80

8

0.7

43

6

0.8

26

0

.77

9

7

0

.78

9

0.7

74

7

0.8

00

0

.77

1

7

(1

) 0

.80

6

0.7

82

2

(1

) 1

.00

0

0.5

00

PE

L1

88

4

0

.77

8

0.7

06

4

0.7

39

0

.63

6

4

0

.55

0

0.6

59

4

0.8

00

0

.64

4

4

0

.69

8

0.6

50

2

0

.50

0

0.3

75

PE

L1

90

4

0

.70

4

0.6

87

3

0.4

35

0

.58

5

4

0

.33

3

0.3

56

4

0.7

00

0

.56

9

4

0

.49

2

0.5

30

3

1

.00

0

0.6

25

PE

L2

07

2

0

.25

9

0.2

78

2

0.2

61

0

.26

1

2

0

.22

7

0.1

45

2

0.2

50

0

.21

9

2

0

.22

6

0.2

00

1

0

0

PE

L2

65

5

(1

) 0

.63

0

0.6

68

4

0.4

78

0

.47

8

4

0

.68

3

0.6

43

4

0.5

50

0

.60

1

4

0

.54

8

0.6

59

4

1

.00

0

0.7

50

PE

L3

04

1

1 (

2)

0.7

78

0

.82

0

6

0

.73

9

0.7

39

7

0.7

71

0

.72

0

7

0

.65

0

0.7

90

8

0.7

26

0

.77

5

3

0.5

00

0

.62

5

Figure 6.1: Map of southern Africa showing reporting rates for Great White Pelicans

and the location of their most frequent breeding sites (1: Dassen Island, Western Cape;

2: Bird Rock guano platform, Walvis Bay; 3: Hardap Dam, Mariental; 4: St Lucia

Wetland, KwaZulu-Natal). The symbols represent other localities mentioned in the text

(open square: Lake Oponono, Namibia; open diamond: Etosha, Namibia; open star: Lake

Ngami, Botswana; open triangle: Maun, Botswana; filled star: Caprivi Strip, Namibia;

open circle: Sua Pan, Botswana; open trapezoid: Orange River mouth, Namibia and

South Africa). Map reproduced with permission of the editors of the atlas of southern

African birds (Williams and Borello 1997).

�

��

Figure 6.2: Statistical parsimony network derived for the ND2 gene for three

populations of Great White Pelican. WC: Western Cape; Nam: Namibia; KZN: KwaZulu-

Natal. Haplotype names (PEL01 to PEL12) correspond to the ones presented in Table 1.

Connections between haplotypes represent one mutational step and small white circles

symbolise intermediate haplotypes not sampled in the population. The topology of

haplotypes 11 and 12 (individuals sampled from historical museum specimens) is not

completely resolved (multiple connections) as the sequences are incomplete (746 bp and

861 bp fragments respectively).

�

��

�

Figure 6.3: Population structure of the Great White Pelican as estimated with

STRUCTURE 2.2 for K=2 (value with greatest mean likelihood and posterior probability).

Each individual is represented by a vertical bar, which is partitioned into K-colours

corresponding to the individual’s membership. Individuals are ordered according to

sampling locality (1, Namibia; 2, Western Cape; 3, KwaZulu-Natal). Colours correspond

to one of two ancestral populations (k = 2). Average likelihood values for k=1 to 4 were

-2 248.5, -2 212.8, -2 461.9 and -2 520.6, respectively.

��

(a)

(b)

Figure 6.4: BAYESASS posterior probability distribution of non-immigrant, first- and

second-generation immigrant ancestry of (a) five Great White Pelicans from the

Namibian population, and (b) four individuals from the Western Cape population.

Ap

pen

dix

6.1

: G

reat W

hit

e P

eli

can

raw

mic

rosatell

ite s

cores f

or a

ll s

am

ple

s o

f th

e N

am

ibia

n (

Nam

), W

estern

Cap

e (

WC

) an

d K

waZ

ulu

-

Natal

(KZ

N)

pop

ula

tio

ns.

Sa

mp

le

Po

pn

. Y

ea

r

PE

L8

6

PE

L1

49

P

EL

17

5

PE

L1

88

P

EL

19

0

PE

L2

07

P

EL

26

5

PE

L3

04

MP

M0

47

W

C

20

05

1

82

1

86

3

25

3

25

2

54

2

66

2

27

2

31

2

51

2

51

1

17

1

17

MP

M0

61

W

C

20

05

1

55

1

86

3

25

3

25

2

54

2

64

2

27

2

31

2

51

2

51

2

16

2

19

1

14

1

23

2

16

2

16

MP

M0

80

W

C

20

05

1

55

1

63

3

25

3

25

2

27

2

31

2

49

2

51

2

19

2

19

1

14

1

20

2

10

2

16

MP

M0

85

W

C

20

05

1

67

1

67

3

46

3

62

2

54

2

66

2

27

2

31

2

51

2

51

2

19

2

19

1

17

1

17

2

16

2

16

MP

M1

63

W

C

20

05

1

55

1

82

3

29

3

58

2

54

2

54

2

29

2

31

2

49

2

49

2

19

2

19

1

20

1

23

2

13

2

13

MP

M1

64

W

C

20

05

1

58

1

90

3

25

3

46

2

54

2

66

2

31

2

31

2

51

2

51

2

16

2

19

1

17

1

20

2

13

2

16

MP

M1

67

W

C

20

05

1

67

1

90

3

25

3

38

2

54

2

54

2

29

2

31

2

16

2

19

1

17

1

17

1

95

2

10

MP

M1

69

W

C

20

05

1

58

1

86

3

25

3

38

2

54

2

66

2

31

2

31

2

19

2

19

1

14

1

17

2

13

2

16

MP

M1

82

W

C

20

05

1

90

1

90

3

46

3

85

2

52

2

60

2

27

2

29

2

51

2

51

2

19

2

19

1

20

1

23

2

13

2

16

MP

M1

83

W

C

20

05

1

58

1

90

3

46

3

50

2

54

2

56

2

27

2

31

2

51

2

51

2

19

2

19

1

17

1

17

2

10

2

13

MP

M1

84

W

C

20

05

1

55

1

63

3

25

3

38

2

54

2

64

2

27

2

27

2

49

2

51

2

19

2

19

1

17

1

17

2

10

2

16

MP

M1

85

W

C

20

05

1

67

1

86

3

38

3

46

2

60

2

66

2

25

2

27

2

51

2

51

2

19

2

19

1

17

1

23

2

16

2

31

MP

M1

86

W

C

20

05

1

58

1

86

3

25

3

38

2

52

2

54

2

31

2

31

2

51

2

51

2

19

2

19

1

14

1

17

1

95

2

13

MP

M1

87

W

C

20

05

1

55

1

55

3

25

3

50

2

54

2

64

2

27

2

27

2

51

2

51

2

19

2

19

1

20

1

23

2

10

2

16

MP

M1

89

W

C

20

05

1

67

1

67

3

25

3

85

2

60

2

64

2

29

2

29

2

49

2

51

2

19

2

19

1

17

1

23

2

13

2

13

MP

M1

91

W

C

20

05

1

67

1

86

3

13

3

50

2

60

2

60

2

27

2

31

2

51

2

51

2

19

2

19

2

10

2

16

MP

M1

93

W

C

20

05

1

58

1

90

3

22

3

25

2

52

2

66

2

27

2

27

2

51

2

51

2

19

2

19

1

20

1

23

2

10

2

13

MP

M1

94

W

C

20

05

1

67

1

90

3

50

3

58

2

64

2

64

2

27

2

27

2

51

2

55

2

19

2

19

1

17

1

17

2

16

2

19

MP

M1

95

W

C

20

05

1

82

1

90

3

25

3

25

2

52

2

54

2

29

2

31

2

49

2

51

2

19

2

19

1

17

1

17

2

10

2

16

MP

M1

97

W

C

20

05

1

55

1

58

3

25

3

29

2

56

2

66

2

29

2

29

2

49

2

51

2

19

2

19

1

17

1

23

2

10

2

13

Ap

pen

dix

6.1

: (c

on

tin

ued

)

Sa

mp

le

Po

pn

. Y

ea

r

PE

L8

6

PE

L1

49

P

EL

17

5

PE

L1

88

P

EL

19

0

PE

L2

07

P

EL

26

5

PE

L3

04

MP

M2

31

W

C

20

06

1

58

1

58

3

38

3

46

2

54

2

56

2

29

2

31

2

49

2

51

2

19

2

19

1

14

1

17

1

98

2

16

MP

M2

35

W

C

20

06

1

63

1

90

3

25

3

58

2

52

2

54

2

27

2

27

2

51

2

51

2

19

2

19

1

20

1

23

1

95

2

13

MP

M2

36

W

C

20

06

1

67

1

90

3

13

3

62

2

52

2

54

2

31

2

31

2

49

2

51

2

19

2

19

1

17

1

17

2

10

2

16

MP

M2

37

W

C

20

06

1

58

1

90

3

25

3

25

2

54

2

54

2

27

2

31

2

51

2

51

2

16

2

19

1

20

1

20

2

10

2

13

MP

M2

42

W

C

20

06

1

63

1

97

3

38

3

58

2

56

2

64

2

29

2

31

2

51

2

51

2

16

2

19

1

23

1

23

1

95

2

16

MP

M2

43

W

C

20

06

1

67

1

90

3

22

3

50

2

60

2

62

2

27

2

31

2

53

2

55

2

19

2

19

1

20

1

20

2

13

2

16

MP

M2

44

W

C

20

06

1

55

1

90

3

25

3

46

2

54

2

62

2

27

2

31

2

51

2

51

2

19

2

19

1

14

1

17

2

10

2

13

MP

M2

48

W

C

20

06

1

90

1

90

3

17

3

22

2

54

2

64

2

27

2

29

2

51

2

53

2

16

2

19

1

17

1

20

2

10

2

19

MP

M2

49

W

C

20

06

1

67

1

90

3

25

3

25

2

56

2

60

2

27

2

31

2

51

2

51

2

19

2

19

1

17

1

17

2

10

2

16

MP

M2

50

W

C

20

06

1

55

1

58

3

25

3

29

2

52

2

52

2

27

2

31

2

49

2

51

2

19

2

19

1

20

1

20

2

16

2

16

MP

M2

52

W

C

20

06

1

86

1

90

3

29

3

85

2

54

2

54

2

27

2

31

2

51

2

51

2

19

2

19

1

14

1

17

1

95

2

13

MP

M2

54

W

C

20

06

1

67

1

67

3

50

3

89

2

60

2

64

2

27

2

31

2

53

2

53

2

16

2

19

1

17

1

20

2

16

2

19

MP

M2

55

W

C

20

06

1

58

1

67

3

50

3

50

2

54

2

60

2

29

2

29

2

49

2

53

2

19

2

19

1

17

1

17

2

13

2

13

MP

M2

57

W

C

20

06

1

67

1

90

3

25

3

46

2

54

2

60

2

27

2

31

2

49

2

51

2

16

2

19

1

20

1

20

2

10

2

13

MP

M2

58

W

C

20

06

1

55

1

90

3

25

3

29

2

64

2

66

2

27

2

31

2

53

2

53

2

19

2

19

1

17

1

17

1

98

2

19

MP

M2

59

W

C

20

06

1

58

1

86

3

25

3

46

2

52

2

54

2

25

2

31

2

49

2

53

2

16

2

19

1

20

1

23

2

13

2

13

MP

M2

60

W

C

20

06

1

90

1

94

3

29

3

38

2

52

2

52

2

25

2

27

2

49

2

49

2

19

2

19

1

17

1

23

2

13

2

13

MP

M2

61

W

C

20

06

1

67

1

86

3

25

3

50

2

54

2

60

2

31

2

31

2

51

2

53

2

19

2

19

1

17

1

20

2

10

2

13

MP

M2

62

W

C

20

06

1

63

1

67

3

29

3

46

2

54

2

64

2

31

2

31

2

51

2

51

2

19

2

19

1

17

1

17

2

19

2

19

MP

M2

64

W

C

20

06

1

86

1

90

3

25

3

85

2

54

2

64

2

27

2

27

2

51

2

51

2

19

2

19

1

17

1

23

1

95

2

16

MP

M2

66

W

C

20

06

1

86

1

86

3

25

3

46

2

54

2

64

2

27

2

31

2

51

2

51

2

19

2

19

1

23

1

23

2

13

2

13

MP

M2

67

W

C

20

06

1

82

1

86

3

38

3

46

2

54

2

64

2

29

2

31

2

51

2

55

2

19

2

19

1

20

1

20

2

13

2

31

MP

M2

68

W

C

20

06

1

82

1

90

3

25

3

38

2

60

2

66

2

27

2

31

2

51

2

51

2

19

2

19

1

20

1

23

2

13

2

16

Ap

pen

dix

6.1

: (c

on

tin

ued

)

Sa

mp

le

Po

pn

. Y

ea

r

PE

L8

6

PE

L1

49

P

EL

17

5

PE

L1

88

P

EL

19

0

PE

L2

07

P

EL

26

5

PE

L3

04

MP

M3

57

W

C

20

07

3

25

3

25

2

54

2

60

2

31

2

31

2

51

2

51

2

19

2

19

1

17

1

17

2

13

2

16

MP

M3

58

W

C

20

07

1

63

1

82

3

13

3

25

2

62

2

64

2

27

2

31

2

51

2

55

2

19

2

19

1

17

1

20

2

10

2

13

MP

M3

59

W

C

20

07

1

55

1

90

3

25

3

25

2

54

2

66

2

27

2

31

2

49

2

51

2

16

2

19

1

17

1

17

2

16

2

31

MP

M3

60

W

C

20

07

1

55

1

86

3

50

3

58

2

54

2

54

2

27

2

31

2

49

2

51

2

16

2

19

1

17

1

23

2

13

2

16

MP

M3

61

W

C

20

07

1

90

1

90

3

17

3

50

2

52

2

64

2

27

2

29

2

49

2

51

2

19

2

19

1

17

1

20

2

10

2

10

MP

M3

63

W

C

20

07

1

55

1

63

3

38

3

58

2

54

2

64

2

27

2

29

2

49

2

51

2

16

2

19

1

17

1

17

2

19

2

19

MP

M3

64

W

C

20

07

1

55

1

67

3

25

3

62

2

54

2

60

2

27

2

31

2

49

2

49

2

19

2

19

1

14

1

20

2

16

2

19

MP

M3

65

W

C

20

07

1

63

1

90

3

46

3

50

2

64

2

66

2

25

2

31

2

51

2

51

2

19

2

19

1

17

1

20

1

95

2

10

MP

M3

66

W

C

20

07

1

90

1

90

3

25

3

29

2

52

2

64

2

29

2

29

2

49

2

51

2

19

2

19

1

17

1

20

2

13

2

13

MP

M3

67

W

C

20

07

1

55

1

86

3

25

3

25

2

54

2

54

2

27

2

31

2

51

2

51

2

16

2

19

1

17

1

17

1

95

2

10

MP

M3

68

W

C

20

07

1

90

1

90

3

25

3

38

2

64

2

66

2

27

2

31

2

51

2

51

2

19

2

19

1

14

1

20

2

10

2

10

MP

M3

69

W

C

20

07

1

82

1

90

3

25

3

58

2

54

2

62

2

31

2

31

2

51

2

53

2

19

2

19

1

17

1

23

2

16

2

31

MP

M3

70

W

C

20

07

1

63

1

90

3

25

3

38

2

54

2

54

2

27

2

29

2

49

2

51

2

19

2

19

1

17

1

17

1

95

2

16

MP

M3

71

W

C

20

07

1

67

1

86

3

25

3

58

2

54

2

66

2

27

2

31

2

51

2

53

2

19

2

19

1

17

1

17

2

13

2

13

MP

M3

72

W

C

20

07

1

67

1

90

3

25

3

25

2

54

2

56

2

29

2

31

2

49

2

51

2

19

2

19

1

17

1

17

2

10

2

13

MP

M3

73

W

C

20

07

1

58

1

90

3

17

3

25

2

64

2

66

2

29

2

31

2

49

2

53

2

19

2

19

1

14

1

17

2

16

2

16

MP

M3

75

W

C

20

07

1

55

1

55

3

25

3

25

2

60

2

64

2

31

2

31

2

49

2

51

2

19

2

19

1

20

1

20

1

95

2

10

MP

M3

76

W

C

20

07

1

63

1

63

3

38

3

58

2

54

2

54

2

29

2

31

2

51

2

53

2

16

2

19

1

17

1

23

2

10

2

16

MP

M3

77

W

C

20

07

1

58

1

90

3

22

3

38

2

60

2

64

2

27

2

31

2

49

2

51

2

19

2

19

1

17

1

23

2

16

2

16

MP

M3

86

W

C

20

07

1

55

1

90

3

25

3

25

2

52

2

62

2

27

2

31

2

51

2

51

2

19

2

19

1

23

1

23

2

07

2

10

Ap

pen

dix

6.1

: (c

on

tin

ued

)

Sa

mp

le

Po

pn

. Y

ea

r

PE

L8

6

PE

L1

49

P

EL

17

5

PE

L1

88

P

EL

19

0

PE

L2

07

P

EL

26

5

PE

L3

04

MP

M1

22

N

am

2

00

6

15

8

16

3

36

6

38

9

25

2

25

4

22

5

22

9

25

1

25

3

21

9

21

9

11

4

11

7

21

0

22

2

MP

M1

23

N

am

2

00

6

15

8

19

4

37

4

37

7

25

2

25

4

22

5

23

1

25

1

25

5

21

9

21

9

12

3

12

3

20

7

21

0

MP

M1

24

N

am

2

00

6

14

9

15

8

33

8

35

4

25

4

25

4

22

7

22

9

24

9

25

1

21

6

21

9

11

0

11

4

21

3

21

3

MP

M1

25

N

am

2

00

6

15

8

19

7

32

2

37

4

25

4

26

4

22

7

23

1

25

1

25

3

21

6

21

9

11

7

12

3

21

3

22

8

MP

M1

26

N

am

2

00

6

18

2

18

6

37

4

37

7

25

2

26

6

22

9

23

1

25

1

25

1

21

9

21

9

11

4

11

7

19

8

21

3

MP

M1

27

N

am

2

00

6

16

3

19

0

32

2

34

6

25

2

26

4

22

9

22

9

25

1

25

1

21

6

21

9

11

7

11

7

21

0

21

9

MP

M1

28

N

am

2

00

6

15

5

19

0

35

0

36

6

26

4

26

6

22

7

22

9

24

9

25

3

21

9

21

9

11

7

11

7

19

5

21

0

MP

M1

29

N

am

2

00

6

14

9

15

8

36

2

36

2

25

4

26

6

22

7

22

9

25

1

25

3

21

9

21

9

11

7

11

7

21

0

21

3

MP

M1

31

N

am

2

00

6

15

8

16

3

35

8

36

2

25

2

26

4

22

7

22

9

25

3

25

5

21

9

21

9

11

7

12

3

21

0

21

3

MP

M1

32

N

am

2

00

6

15

5

18

6

32

2

36

2

25

2

26

6

22

9

22

9

24

9

25

1

21

6

21

9

11

4

11

7

19

5

21

0

MP

M1

33

N

am

2

00

6

16

7

19

4

36

2

36

6

26

4

26

4

22

7

22

7

24

9

25

3

21

9

21

9

11

7

11

7

19

5

21

9

MP

M1

34

N

am

2

00

6

15

8

19

7

32

2

38

5

25

2

25

4

22

9

23

1

25

3

25

3

21

9

21

9

11

4

11

4

21

0

21

6

MP

M1

35

N

am

2

00

6

15

5

16

3

37

7

38

5

25

2

26

2

23

1

23

1

25

1

25

3

21

9

21

9

11

7

12

3

21

0

21

0

MP

M1

36

N

am

2

00

6

18

2

19

0

33

4

38

1

25

2

26

4

22

7

23

1

25

1

25

3

21

6

21

9

12

3

12

3

21

9

21

9

MP

M1

98

N

am

2

00

6

16

3

19

4

32

5

34

2

25

4

26

6

22

9

22

9

25

1

25

1

21

9

21

9

11

4

11

7

22

8

23

1

MP

M2

19

N

am

2

00

6

16

3

16

7

32

2

35

8

26

0

26

6

22

7

22

9

25

1

25

1

21

9

21

9

11

4

12

3

19

8

21

3

MP

M2

20

N

am

2

00

6

16

3

16

3

35

0

36

2

25

2

25

4

22

5

22

7

24

9

25

1

21

9

21

9

11

4

11

7

19

5

20

1

MP

M2

21

N

am

2

00

6

15

5

19

0

33

8

35

0

25

2

26

4

22

9

23

1

25

1

25

1

21

9

21

9

11

7

11

7

21

0

21

3

MP

M2

22

N

am

2

00

6

15

8

16

3

32

9

36

9

25

2

26

6

22

9

23

1

25

1

25

3

21

9

21

9

11

7

12

3

21

0

23

1

MP

M2

23

N

am

2

00

6

15

8

16

7

32

5

33

8

25

4

25

4

22

7

23

1

24

9

25

3

21

6

21

9

11

4

12

3

19

8

21

3

MP

M2

24

N

am

2

00

6

16

3

19

0

38

1

38

1

25

4

25

4

22

7

23

1

24

9

25

3

21

9

21

9

11

7

12

0

19

5

19

5

MP

M2

25

N

am

2

00

6

18

2

20

2

34

6

40

0

25

4

26

6

22

9

23

1

24

9

25

1

21

9

21

9

11

4

11

4

19

8

21

0

MP

M2

26

N

am

2

00

6

16

7

19

4

32

2

33

8

25

2

26

6

22

7

22

9

24

9

24

9

21

9

21

9

11

7

12

3

21

0

21

0

MP

M3

78

N

am

2

00

8

15

5

15

8

31

7

36

6

22

5

22

9

24

9

25

3

21

6

21

9

11

4

11

7

20

7

21

6

MP

M3

80

N

am

2

00

8

16

3

19

0

34

6

35

0

25

2

25

4

22

7

23

1

24

9

25

1

21

9

21

9

12

0

12

3

20

7

21

3

MP

M3

81

N

am

2

00

8

16

7

19

7

34

6

38

1

25

2

25

4

22

9

22

9

24

9

24

9

21

6

21

6

11

7

11

7

21

0

21

9

MP

M3

87

N

am

2

00

8

15

5

19

4

32

2

38

9

25

4

25

4

22

5

23

1

25

3

25

5

21

9

21

9

11

7

12

3

21

0

21

0

MP

M3

82

K

ZN

16

3

16

3

34

2

36

2

25

2

25

8

22

9

22

9

24

9

25

1

21

9

21

9

12

0

12

3

21

0

22

8

MP

M3

83

K

ZN

15

8

19

0

32

9

35

0

22

9

23

1

25

1

25

3

21

9

21

9

11

4

11

7

19

5

19

5

193

�

Appendix 6.2: Great White Pelican microsatellite allele frequencies for each

locus and sampling location. (WC: Western Cape).

Locus Allele/n Namibia WC

2005

WC

2006

WC

2007

WC

combined

KwaZulu-

Natal

PEL86 N 27 20 23 19 62 2

149 0.037 0.000 0.000 0.000 0.000 0.000

155 0.111 0.175 0.065 0.211 0.145 0.000

158 0.185 0.150 0.130 0.053 0.113 0.250

163 0.204 0.050 0.065 0.158 0.089 0.500

167 0.093 0.200 0.196 0.079 0.161 0.000

182 0.056 0.075 0.043 0.053 0.056 0.000

186 0.037 0.150 0.152 0.079 0.129 0.000

190 0.111 0.200 0.304 0.368 0.290 0.250

194 0.093 0.000 0.022 0.000 0.008 0.000

197 0.056 0.000 0.022 0.000 0.008 0.000

202 0.019 0.000 0.000 0.000 0.000 0.000

PEL149 N 27 20 23 20 63 2

313 0.000 0.025 0.022 0.025 0.024 0.000

317 0.019 0.000 0.022 0.050 0.024 0.000

322 0.130 0.025 0.043 0.025 0.032 0.000

325 0.037 0.425 0.304 0.500 0.405 0.000

329 0.019 0.050 0.109 0.025 0.063 0.250

334 0.019 0.000 0.000 0.000 0.000 0.000

338 0.074 0.125 0.109 0.125 0.119 0.000

342 0.019 0.000 0.000 0.000 0.000 0.250

346 0.074 0.125 0.152 0.025 0.103 0.000

350 0.074 0.100 0.109 0.075 0.095 0.250

354 0.019 0.000 0.000 0.000 0.000 0.000

358 0.037 0.050 0.043 0.125 0.071 0.000

362 0.111 0.025 0.022 0.025 0.024 0.250

366 0.074 0.000 0.000 0.000 0.000 0.000

369 0.019 0.000 0.000 0.000 0.000 0.000

374 0.056 0.000 0.000 0.000 0.000 0.000

377 0.056 0.000 0.000 0.000 0.000 0.000

381 0.074 0.000 0.000 0.000 0.000 0.000

385 0.037 0.050 0.043 0.000 0.032 0.000

389 0.037 0.000 0.022 0.000 0.008 0.000

400 0.019 0.000 0.000 0.000 0.000 0.000

PEL175 N 26 19 23 20 62 1

252 0.288 0.105 0.152 0.075 0.113 0.500

254 0.346 0.368 0.370 0.375 0.371 0.000

256 0.000 0.053 0.065 0.025 0.048 0.000

258 0.000 0.000 0.000 0.000 0.000 0.500

260 0.019 0.132 0.152 0.100 0.129 0.000

262 0.019 0.000 0.043 0.075 0.040 0.000

264 0.154 0.158 0.174 0.225 0.185 0.000

266 0.173 0.184 0.043 0.125 0.113 0.000

194

�

Appendix 6.2 (continued)

Locus Allele/n Namibia WC

2005

WC

2006

WC

2007

WC

combined

KwaZulu-

Natal

PEL188 N 27 20 23 20 63 2

225 0.093 0.025 0.043 0.025 0.032 0.000

227 0.259 0.400 0.370 0.300 0.357 0.000

229 0.389 0.200 0.130 0.200 0.175 0.750

231 0.259 0.375 0.457 0.475 0.437 0.250

PEL190 N 27 18 23 20 61 2

249 0.259 0.194 0.174 0.300 0.221 0.250

251 0.407 0.778 0.587 0.575 0.639 0.500

253 0.278 0.000 0.196 0.100 0.107 0.250

255 0.056 0.028 0.043 0.025 0.033 0.000

PEL207 N 27 19 23 20 62 2

216 0.167 0.079 0.130 0.125 0.113 0.000

219 0.833 0.921 0.870 0.875 0.887 1.000

PEL265 N 27 19 23 20 62 2

110 0.019 0.000 0.000 0.000 0.000 0.000

114 0.241 0.105 0.065 0.075 0.081 0.250

117 0.463 0.526 0.391 0.575 0.492 0.250

120 0.037 0.158 0.348 0.200 0.242 0.250

123 0.241 0.211 0.196 0.150 0.185 0.250

PEL304 N 27 19 23 20 62 2

195 0.111 0.053 0.087 0.100 0.081 0.500

198 0.074 0.000 0.043 0.000 0.016 0.000

201 0.019 0.000 0.000 0.000 0.000 0.000

207 0.056 0.000 0.000 0.025 0.008 0.000

210 0.333 0.237 0.152 0.275 0.218 0.250

213 0.185 0.289 0.370 0.200 0.290 0.000

216 0.037 0.368 0.217 0.275 0.282 0.000

219 0.093 0.026 0.109 0.075 0.073 0.000

222 0.019 0.000 0.000 0.000 0.000 0.000

228 0.037 0.000 0.000 0.000 0.000 0.250

231 0.037 0.026 0.022 0.050 0.032 0.000

�

CHAPTER 7

��

��

��

��

��

197

��

��

��

��

��

We describe the isolation of 10 microsatellite loci from the Great White Pelican

Pelecanus onocrotalus using an enrichment protocol. All loci were variable with

the number of alleles ranging from 2 to 19, and observed heterozygosity ranging

from 0.261–0.913. Two loci were out of Hardy-Weinberg equilibrium, although in

each case this was restricted to one of the two populations screened. None of the

loci were linked. Eight to ten loci PCR-amplified in three related species: Brown

P. occidentalis, American White P. erythrorhynchos and Pink-backed P. rufescens

pelicans. These loci should prove to be widely applicable to studies of

phylogeography and demography in pelicans globally.

��

Great White Pelicans Pelecanus onocrotalus are a widely distributed species in

Africa, capable of long distance dispersal, but there are few studies describing

movements of individuals among populations. In southern Africa they breed at

three sites and it has been suggested that there is no intermixing between the

western and eastern coastal colonies. Here, we develop microsatellite markers to

determine the degree of gene flow among colonies, as well as investigate the

extent of a reported bottleneck in western South Africa during the 1950s (20–30

pairs; Crawford et al. 1995).

��

198

��

Microsatellites were developed using an enrichment protocol (Glenn & Schable

2005). Genomic DNA (gDNA) from one individual was digested with RsaI and

XmnI, and SuperSNX24 linkers were ligated onto the ends of gDNA fragments.

Biotinylated probes [(AAG)8, (AAT)12, (ACT)12, (AAC)6, (ATC)8, (AAGT)8,

(AAAT)8, (AAAG)8, (ACAT)8, (AGAT)8, (AACT)8] were hybridised to gDNA.

Streptavidin-coated magnetic beads (Invitrogen) were added, and this mixture

was washed twice with 2�SSC, 0.1% SDS and four times with 2�SSC, 0.1% SDS

at 55°C. Between washes, a magnetic particle-collecting unit was used to capture

the magnetic beads, which are bound to the biotin-gDNA complex. Enriched

fragments were removed from the biotinylated probe by denaturing at 95°C and

precipitated with 95% ethanol and 3M sodium acetate. To increase the amount of

enriched fragments, a ‘recovery’ PCR was performed in a 25 μl reaction

containing 1�PCR buffer (10 mM Tris–HCl, 50 mM KCl, pH 8.3), 1.5 mM MgCl2,

0.16 mM of each dNTP, 10XBSA, 0.52 μM of the SuperSNX24 forward primer,

1 U Taq DNA polymerase, and approximately 25 ng of the enriched gDNA.

Thermal cycling was as follows: 95°C for 2min followed by 25 cycles of 95°C for

20s, 60°C for 20s, and 72°C for 90s, and a final elongation step of 72ºC for 30min.

PCR fragments were cloned using the TOPO-TA Cloning® kit following the

manufacturer’s protocol (Invitrogen). Bacterial colonies containing a vector with

gDNA were used as a template for subsequent PCR in a 25 μl reaction containing

1XPCR buffer, 1.5 mM MgCl2, 0.12 mM of each dNTP, 10XBSA, 0.25 μM of the

M13 primers, and 1 U Taq DNA polymerase. Thermal cycling was as follows:

95ºC for 7min followed by 35 cycles of 95°C for 20s, 50°C for 20s, and 72°C for

90s. DNA sequencing of cleaned PCR products was performed using the

BigDye™ Terminator v3.1 Kit (Applied Biosystems), and sequences were run on

an ABI3730 DNA Analyser. Primers flanking core microsatellite repeats were

developed using PRIMER 3.

��

199

��

Ten loci amplified consistently, producing unambiguous peaks (Table 7.1). PCR

reactions (10 μl) consisted of 1�PCR Buffer (10 mM Tris–HCl, 1.5 mM MgCl2,

50 mM KCl, pH 8.3), 0.6 μM each primer, 200 μM each dNTP, 0.6 U Taq and

approximately 10 ng/μl genomic DNA. Thermal cycling was as follows: 94°C for

2 min; 30 cycles at 94°C for 15s, Ta°C for 15s (Table 7.1) and 72°C for 15s. PCR

products were mixed with the GSLIZ500 size standard and formamide and run

on an ABI3730 DNA Analyser. Fragment analysis and genotyping were carried

out using the programme GENEMAPPER 4.0 (Applied Biosystems).

A total of 46 Great White Pelicans were screened: 23 from Namibia and 23 from

the South African population. All loci were polymorphic with the number of

alleles ranging from two to 19 (Table 7.1). HO and HE ranged from 0.261–0.913

and 0.232–0.947, respectively. No linkage disequilibrium was detected for the 10

loci after applying Bonferroni correction for multiple comparisons (Rice 1989).

Loci were tested for Hardy-Weinberg equilibrium (HWE) and linkage

disequilibrium using GENEPOP 3.3 (Raymond & Rousset 1995). Locus PEL185

deviated significantly (P < 0.01) from HWE in the South African population and

locus PEL221 deviated significantly (P < 0.05) from HWE in the Namibia

population. MICRO-CHECKER 2.2.3 (van Oosterhaut et al. 2004) suggested that

PEL185 and PEL190 showed significant homozygote excess in the South African

population, indicating the potential for null alleles at these loci. Neither of the

loci that deviated from HWE showed evidence of sex-linkage, as confirmed using

genetic sexing methods (Fridolfsson & Ellegren 1999). Cross-species

amplification was successful for three other pelican species, showing diverse

heterozygosity levels and unique alleles (Table 7.2). As such, we expect that

some of these loci will have broad utility for phylogeographic and demographic

studies in all eight of the world’s pelican species.

��

200

��



Africa's Western Cape. South African Journal Marine Science 15, 33–42.

Fridolfsson AK, Ellegren H (1999) A simple and universal method for molecular

sexing of non-ratite birds. Journal Avian Biology 30, 116–121.

Ginot F, Bordelais I, Nguyen S, Gyapay G (1996) Correction of some genotyping

errors in automated fluorescent microsatellite analysis by enzymatic

removal of one base overhangs. Nucleic Acids Research 24, 540–541.

Glenn TC, Schable NA (2005) Isolating microsatellite DNA loci. Methods in

Enzymology 395, 202–222.

Marshall TC, Slate J, Kruuk LEB, Pemberton JM (1998) Statistical confidence

for likelihood-based paternity inference in natural populations. Molecular

Ecology 7, 639–655.

van Oosterhout CV, Hutchinson WF, Wills DPM, Shipley P (2004) MICRO-

CHECKER: software for identifying and correcting genotyping errors in

microsatellite data. Molecular Ecology 4, 535–538.

Raymond ML, Rousset F (1995) An exact test for population differentiation.

Evolution 49, 1280–1283.

Rice WR (1989) Analyzing tables of statistical tests. Evolution 43, 223–225.

�

CHAPTER 8

��

��

�

�

203

��

��

Prior to this thesis, there were many aspects of the ecology and life-history of

Great White Pelican Pelecanus onocrotalus that had not been studied at all

throughout their distribution. The paucity of studies in southern Africa, and

particularly in the Western Cape, called for a detailed ecological study of the

species, especially due to the drastic changes in the habitat and also in the

population dynamics of pelicans in the region. Although the many unknowns and

the lack of the necessary baseline data to formulate and test new hypothesis

constituted a limitation, I took it as an opportunity to describe and investigate

important aspects of pelican’s ecology, demographics, life-history, population

structure, behaviour and conservation.

One of my main interests is the effect of human activities on natural systems,

and most importantly the chains of effects that often occur as a consequence of

these activities, which can offer a unique opportunity to understand ecological

interactions. Despite the advances, the complexity of these interactions and the

innumerable simultaneous levels in which these occur still present many

challenges. An example of the impact of human activities is illustrated in

Chapter 2 by the role played by agricultural subsidies in the expansion of the

Great White Pelican population in the Western Cape (Figure 8.1), and the

consequences that these agricultural practices have brought about for the

pelicans themselves and also for coexisting seabird species, eaten by an

increasing pelican population.

Other consequences of the food subsidisation to pelicans (only lightly explored in

this study in Appendix A.1) are the potential risk of transmission of diseases and

pathogens from pelicans (or other scavenging species, such as gulls, ibises,

ravens, etc.) to domestic animals, and vice versa. Birds are one of the

��

204

predominant hosts of emerging zoonoses (Reed et al. 2003). There are numerous

examples of confirmed cross-species transmission of pathogens (e.g. H5N1

strains of Influenza A virus, Brucellosis, SARS, and HIV-AIDS); some of these

diseases posing a significant threat to both human and animal health, and can

produce detrimental socio-economic and ecological consequences (e.g. Friend

2006; Kuiken et al. 2006; Melville & Shortridge 2006; Rocke 2006a; b; Samuel

2006; Shortridge & Melville 2006). In this case, pelicans feeding on raw chicken

offal at pig farms and rubbish tips had the potential for transmission of disease,

but this did not materialise as an actual threat during this study.

Figure 8.1: Simplified trophic web showing the marine and terrestrial origin of

food sources for Great White Pelicans in the Western Cape, and emphasising the

reinforcement effect of agricultural subsidies into the fresh-water habitat, and

specifically to the pelicans through the consumption of chicken offal, as well as

the subtraction of the intermediate predator level from the marine environment

by the pelagic fishery.

��

205

Other human activities, such as fishing of pelagic fish, severely affect the

abundance of predators that depend on these fish species (Figure 8.1). A growing

number of studies have illustrated the changes in the trophic webs caused by the

collapse of fisheries due to over-exploitation (Pauly et al. 1998; Steneck 1998;

Cury et al. 2000; Tasker et al. 2000; Jackson et al. 2001; Heithaus et al. 2008). In

south-western Africa, the Benguela Upwelling System sustains rich trophic webs

and abundant biomass of seabirds that feed on pelagic fish. Anthropogenic

changes during the 20th century have altered the structure and composition of

these food webs and it has caused the decline of top predator populations

(Crawford et al. 1992; 2007b; d; Crawford 2007; Grémillet et al. 2008; Mullers et

al. 2009; Okes et al. 2009; Pichegru et al. in press). In addition to fisheries-

induced mortality and competition for food, predation by other organisms that

benefit from food subsidisation like Kelp Gulls Larus dominicanus (Crawford et

al. 2007c; d), and also seabird predation by an increasing populations of Cape

Fur Seals Arctocephalus pusillus (Marks et al. 1997; David et al. 2003; Makhado

et al. 2006) contribute to the ongoing decline in seabird numbers.

Although pelican’s shift in foraging strategies to overcome the shortage in the

availability of agricultural offal was expected (de Ponte Machado & Hofmeyr

2004), the scale of their impact on local populations of seabirds was

unanticipated. Pelican predation on seabirds has become an additional factor

contributing to the declining trends of seabird populations (Chapter 4), although

the question remains of whether pelican predation adds up to these declines, or

whether the lack of food and mortality caused by fishing operations has reduced

the carrying capacity to such extent that pelican predation on chicks does not

cause further decline on seabird populations. In any case, urgent actions are

required to assure that there is a provision for the food required by seabirds and

marine mammals in the Benguela ecosystem through the establishment of

escapement quotas of fish for seabirds or the implementation of closure zones or

offshore Marine Protected Areas (Game et al. 2009; Halpern et al. 2009; Pichegru

et al. in press).

The spread of predatory behaviour by pelicans to other localities where both

��

206

Great White Pelicans and colonial seabirds occur would be a reason for concern.

The findings of the genetic study (Chapter 6) presented in this thesis, in

agreement with data from the colour ringing program (Chapter 3), showed that

pelicans disperse north and enter into contact with the Namibian population.

The spread of this behaviour by cultural transmission could occur quickly if a

certain number of pelicans that have learned the behaviour emigrates or enter

into contact with pelicans of the coastal population of Namibia. Also, there is a

possibility that the expanding population of Great White Pelicans would extend

its range towards the eastern coastal region of South Africa. Historically, the

eastern limit of distribution of the Western Cape pelicans has been at De Hoop

Nature Reserve in the Agulhas Plains, with the exception of one or two solitary

individuals recorded in the proximity of Algoa Bay in the Eastern Cape. Because

of the changes in pelagic fish abundance and distribution (van der Lingen et al.

2005), most species of seabirds feeding on these resources (viz. Cape Gannet

Morus capensis, African Penguin Spheniscus demersus and Cape Cormorants

Phalacrocorax capensis) have shifted east, following the shifting centre of gravity

of the distribution of pelagic fish species (Crawford et al. 2007a; 2008a). Thus,

the eastern breeding sites are becoming increasingly important in breeding

numbers and contribution to the overall populations of the mentioned species,

whereas the colonies further north and west (Namibia and South African west

coast) have experienced steady declines (Crawford et al. 2008b).

Consequently, if pelicans were increasingly to move east, and bring their

seabird-predator behaviour with them, they would increase the level of threat to

the targeted seabird prey species. Although pelican numbers are small in the

south and east coasts, they could increase if new sources of food were to become

available. Pelicans were observed scouting the dense Cape Cormorant colonies

on Dyer Island in 2005 (LJ Waller, pers. comm.), one of the sites where pelicans

bred in the first half of the 20th century (Crawford et al. 1995). This island is

nowadays the stronghold of Cape Cormorants as well as an important breeding

site for African Penguins (many of which are surface nesters on Dyer Island). All

efforts should be made to discourage pelicans from landing on Dyer Island and

��

207

the nearby Quoin Rock. On the other hand, besides these localities there are no

other adequate breeding sites for pelicans (free of disturbance and land

predators) on this stretch of coast, so it will be unlikely that large number of

pelicans move to the area. Along the south and east coast of South Africa, the

next potential breeding site for pelicans would be Bird Island in Algoa Bay. One

single pelican from the Western Cape (ringed with a red plastic ring) has been

observed on numerous occasions between 2004 and 2007 on this island in the

middle of the gannet colony. Although it has been recorded fishing in estuaries in

the adjacent mainland coast, on once occasion it was observed preying on a

gannet chick (L Pichegru, pers. comm.).

The results of the management intervention on the islands of Saldanha Bay in

the West Coast National Park (Chapter 5) showed that chasing pelicans from

selected seabird colonies is an effective measure to prevent pelican predation and

increase recruitment. However, this is an effort-intensive intervention. Other

favoured options, which could be used in combination with active chasing, would

be to provide nesting structures that provide protection from pelicans to

cormorants and other ground nesting species, although research is needed to

design and test these. Also, the spatial coverage of this management

intervention was restricted to two islands, and this represents a small proportion

of the breeding colonies of seabirds that are affected by pelican predation.

��

The novel contributions of this study to the knowledge and understanding of the

pelican population in the Western Cape, and in some cases of southern Africa,

are listed below:

o This study includes the first survival analysis using mark-recapture data

for any species of pelican, revealing remarkable low mortality rates for

both adults and juveniles (survival rates are 0.74 for first year of age and

0.96 subsequently). Resighting rates have been larger than 75% for some

cohorts, allowing for accurate estimation of survival rates.

o It contributes with a detailed analysis of breeding effort and breeding

��

208

productivity, as well as the compilation of life-history parameters (e.g. age

at first breeding and longevity), which are essential for the construction of

population models of the studied population.

o It reports on the importance of dietary subsidisation for the pelican

population in the Western Cape. Although the increase in numbers by

opportunistic species is not new itself (see the prolific literature on gulls,

skuas and others cited in Chapters 2 and 5) it is a novel occurrence for any

pelican species, and even for any member of the Pelecaniformes, which are

generally fish-eaters.

o The development and optimisation of microsatellite primers is a first for

pelicans. Besides allowing the development of questions of gene flow and

intraspecific variability, it is a first step to allow fine-scale studies of

relatedness, paternity, forensics, and others. These microsatellite primers

could be used in similar studies on other members of the genus Pelecanus.

o This is the first detailed population genetics study for any seabird (or even

large terrestrial bird species) in the region. It confirmed that immigration

has not played a role in the increase in numbers of the Western Cape

pelican population and limited gene flow occurs among southern African

populations, in particular into the Western Cape. However, some

exchange may be possible among northern and eastern populations. No

sign of inbreeding and/or molecular bottlenecks were found on the

Namibian or Western Cape populations, despite reported small population

sizes during the 20th century. Ten new haplotypes for the mitochondrial

gene NADH2 were isolated and will be submitted to GeneBank at the time

of publishing these results.

o A total of 1 200 pelicans have been ringed since 2001 in southern Africa.

In 2009, about 12% of the population of the Western Cape was

individually identifiable with engraved colour-rings, discounting the birds

��

209

that died (according to estimates of survival calculated in this study).

Ringing has provided information on long-distance dispersal of pelicans

for the fist time in southern Africa. Some highlighted examples are the

dispersal from the Western Cape, South Africa to Lake Ngami, Botswana;

from the Western Cape to Algoa Bay, Eastern Cape; from Hardap Dam in

Namibia, to Maun, Botswana and from Hardap Dam to the Eastern

Caprivi Strip, northern Namibia.

o This study reports on a novel foraging behaviour of Great White Pelicans,

adapting cooperative techniques, commonly used by the species while

capturing prey in lakes and estuaries, from the water to the land. I

propose that social learning could be the mechanism of transmission of the

new behaviour in the population and report on the impact of predation on

the prey species, some of which are threatened endemic species and whose

populations are experiencing alarming declines.

o Finally, this study documents a management intervention implemented

on two islands of the West Coast National Park in the Western Cape,

South Africa, aimed to prevent pelican predation at selected seabird

colonies. It proved to be a successful strategy to minimise the impact on

seabird populations. It also provided an example of a successful

collaborative work between scientists and managers, bridging the gap

between the disciplines of ecology and conservation management and

contributing to the emerging field of conservation evidence.

��

210

��

As it is often the case, finding answers to research questions triggers the

formulation of new hypothesis and further research. Here I make some

suggestions of potential future lines of investigation:

o Analysis of Great White Pelican diet using stable isotopes: estimating the

proportion of the different food items in the diet of pelicans, assessing the

importance of food supplementation and of predation of seabirds for the

Western Cape population, and comparing to other southern African

populations. Also, if the signature of the individual food items is

sufficiently diverse, sampling pelican chicks would contribute to answer

whether breeding pelicans depend on seabirds as food during the

reproductive period.

o Communication systems in pelicans and foraging ecology: how pelicans

use visual and/or environmental cues to locate food resources in a patchy

environment, as well as the importance of colonies and aggregations of

individuals.

o Expansion of the survival analysis reported in this study, including

covariates such as sex and body condition of chicks at the moment of

ringing.

o Modelling of pelican’s prey species population dynamics (i.e. seabirds, in

particular Cape Gannets) to include the effects of predation by

opportunistic predators, Kelp Gulls, Great White Pelicans and Cape Fur

Seals.

o Microbiological studies: comparing the pathogenic load of pelicans feeding

on offal and other agricultural waste (mainly at the Western Cape

population) with natural feeding populations (e.g. Namibian population).

��

211

o To continue the colour ringing programme initiated in 2001 to obtain

ongoing information on dispersal and migration. I strongly recommend

using engraved plastic rings. About 19% of the Western Cape Great White

Pelican population is ringed (taking into account all individuals ringed

since 2001 and adjusting for survival rates); however, most pelicans with

rings are relatively young, considering they may reach up to 40 years of

age. Pelicans are large and visible and the ringing programme has proven

to be highly efficient. Larger resighting effort at breeding colonies (i.e.

Dassen Island) and during predation at seabird colonies would be crucial

to answer questions like the percentage of pelicans breeding at different

ages, incidence and frequency of breeding of individual pelicans,

proportion of the breeding pelican population involved in predation

activities and whether there is any specialisation regarding the different

food sources (fish, offal and seabirds) in the pelican population. The nature

of the materials used for rings and their permanence on the wild also need

to be examined to correct for ring losses in the long-term.

o To continue monitoring breeding effort and chick production at the colony

on Dassen Island. Counts of number of breeding pairs and chicks should

be done with a frequency of at least once a month, and the approximate

age of the chicks should be annotated in order to identify the overall chick

production from the different breeding waves.

o To calculate the interrelations between food availability and reproduction

and other life-history parameters. To investigate whether there are any

density-dependence mechanisms driving the population dynamics of Great

White Pelicans in the Western Cape.

o On going projects such as CWAC (Coordinated Waterbird Counts) are

helpful to assess use of habitat of the Great White Pelican population by

estimating numbers foraging regularly on the monitored sites. However,

an estimate of the total population size in the Western Cape can only be

��

212

achieved by a coordinated census covering the whole area of distribution.

This is achievable for the species by simultaneous aerial and ground

counts. It is best done in the month of April and it is recommended to

count adults and juveniles separately.

o To expand the population genetics study by increasing the number of

sampling sites in southern Africa, in particular the eastern South African

population at St Lucia Lake and the inland nomadic populations

(Botswana and northern Namibia). Sampling in West Africa, East and

Central Africa and southern Europe/Asia would be needed to better

understand the population structure and broader geographical movements

among populations.

o Remote tracking with satellite or GPS devices to collect information on the

movements of pelicans at both local and regional levels. Locally, this data

could be used to establish use of habitat and requirements related to

parental investment and other reproductive parameters (nest attendance,

distance to food sources, etc.). Regionally it would contribute to our

understanding of large scale use of habitat, dispersal of young, migration,

foraging ecology and to establish important sites for the conservation of

waterbirds.

o Establishing the impact of linkages between anthropogenic, fresh-water

and marine ecosystems, in particular regarding the amount of

allochthonous materials derived from agriculture, that are moving into

natural habitats and their contribution in changing the structure of the

ecological communities and trophic dynamics (Jefferies 2000). An example

that illustrates the interaction between marine and terrestrial systems is

provided by opportunistic species (in this case pelicans and gulls) serving

as a link between human waste (agricultural offal, rubbish tips and

fisheries discards) and the population dynamics of oceanic seabirds.

��

213

o To further investigate the importance of pelicans as apex predators and

the potential for trophic cascades as a consequence of the multi-factor

declines of top predators (i.e. seabirds) in the Benguela system.

o To monitor population changes of other opportunistic species, such Kelp

Gulls. Their increase in numbers has been linked to their capacity to feed

on surplus food from human activities and they could acquire the status of

pest species if their predatory pressure on seabirds is maintained

(Williams & Parsons 2004; Whittington et al. 2006; Altwegg et al. 2007). A

coordinated response towards managing the ecosystem and its different

trophic levels is needed, yet challenging. Ongoing research on the

combined impact of human activities and population dynamics of natural

populations is required.

Finally, I would like to emphasise the importance of continued studies and long-

term monitoring efforts. This thesis has established a baseline of information

against which changes can be measured. The population dynamics of the Great

White Pelican in the Western Cape are genuinely dynamic! There is a rare

opportunity here for ongoing studies that will enable us to understand broader

ecological processes and be able to respond in an informed manner to future

changes in the ecosystems.

��

214

��

Altwegg R, Crawford RJM, Underhill LG, Martin AP, Whittington PA (2007)

Geographic variation in reproduction and survival of Kelp Gulls Larus

dominicanus vetula in southern Africa. Journal of Avian Biology 38, 580–

586.

Crawford RJM (2007) Food, fishing and seabirds in the Benguela upwelling

system. In: Final report of the BCLME (Benguela Current Large Marine



Unit, Cape Town.



Africa's Western Cape. African Journal of Marine Science 15, 33–42.

Crawford RJM, Dundee BL, Dyer BM, Klages NTW, Meyër MA, Upfold L

(2007b) Trends in numbers of Cape Gannets Morus capensis 1956/1957–

2005/2006, with a consideration of the influence of food and other factors.

ICES Journal of Marine Science 64, 169–177.

Crawford RJM, Fairweather T, Underhill L, G, Wolfaardt A (2007a) Implications

for seabirds of an unfavourable, long-term change in the distribution of

prey: a South African experience. In: Final report of the BCLME (Benguela

Current Large Marine Ecosystem) project on top predators as biological

indicators of ecosystem change in the BCLME (ed. Kirkman SP), pp. 243–

249. Avian Demography Unit, Cape Town.

Crawford RJM, Sabarros PS, Fairweather TP, Underhill LG, Wolfaardt A

(2008a) Implications for seabirds off South Africa of a long-term change in

the distribution of sardine. African Journal of Marine Science 30, 177–

184.

Crawford RJM, Tree AJ, Whittington PA, Vissage J, Upfold L, Roxburg KJ,

Martin AP, Dyer BM (2008b) Recent distributional changes of seabirds in

South Africa: is climate having an impact? African Journal of Marine

Science 30, 189–193.

Crawford RJM, Underhill LG, Altwegg R, Dyer BM, Upfold L (2007c) The

influence of culling, predation and food on Kelp Gulls Larus dominicanus

off western South Africa In: Final report of the BCLME (Benguela Current

Large Marine Ecosystem) project on top predators as biological indicators

of ecosystem change in the BCLME (ed. Kirkman SP), pp. 181–187. Avian


Crawford RJM, Underhill LG, Raubenheimer CM, Dyer BM, Martin J (1992) Top

predators in the Benguela ecosystem: implications of their trophic

position. South African Journal of Marine Science 12, 675–687.

Crawford RJM, Underhill LG, Upfold L, Dyer BM (2007d) An altered carrying

capacity of the Benguela upwelling ecosystem for African Penguins

Spheniscus demersus. ICES Journal Marine Science 64, 570–576.

Cury P, Bakun A, Crawford RJM, Jarre A, Quinones RA, Shannon LJ, Verheye

HM (2000) Small pelagics in upwelling systems: patterns of interaction

and structural changes in ‘wasp-waist’ ecosystems. ICES Journal Marine

Science 57, 603–618.

��

215

David JHM, Cury P, Crawford RJM, Randall RM, Underhill LG, Meÿer MA

(2003) Assessing conservation priorities in the Benguela ecosystem, South

Africa: analysing predation by seals on threatened seabirds. Biological


de Ponte Machado M, Hofmeyr J (2004) Great White Pelicans Pelecanus

onocrotalus: waterbirds or farm birds? Bird Numbers 13, 11–13.

Friend M (2006) Evolving changes in diseases of waterbirds. In: Waterbirds

around the world (eds. Boere GC, Galbraith CA, Stroud DA), pp. 412-416.

The Stationery Office, Edinburgh.

Game ET, Grantham HS, Hobday AJ, Pressey RL, Lombard AT, Beckley LE,

Gjerde K, Bustamante R, Possingham HP, Richardson AJ (2009) Pelagic

protected areas: the missing dimension in ocean conservation. Trends in

Ecology & Evolution 24, 360–369.

Grémillet D, Pichegru L, Kuntz G, Woakes AG, Wilkinson S, Crawford RJM,

Ryan PG (2008) A junk-food hypothesis for gannets feeding on fishery

waste. Proceedings of the Royal Society B 275, 1149–1156.

Halpern BS, Kappel CV, Selkoe KA, Micheli F, Ebert CM, Kontgis C, Crain CM,

Martone RG, Shearer C, Teck SJ (2009) Mapping cumulative human

impacts to California Current marine ecosystems. Conservation Letters 2,

138–148.

Heithaus MR, Frid A, Wirsing AJ, Worm B (2008) Predicting ecological

consequences of marine top predator declines. Trends in Ecology &

Evolution 23, 202–210.

Jackson JBC, Kirby MX, Berger WH, Bjorndal KA, Botsford LW, Bourque BJ,

Bradbury RH, Cooke R, Erlandson J, Estes JA, Hughes TP, Kidwell S,

Lange CB, Lenihan HS, Pandolfi JM, Peterson CH, Steneck RS, Tegner

MJ, Warner RR (2001) Historical overfishing and the recent collapse of

coastal ecosystems. Science 293, 629–637.



Kuiken T, Fouchier RAM, Rimmelzwaan GF, Osterhaus ADME (2006) Emerging

viral diseases in waterbirs. In: Waterbirds around the world (eds. Boere

GC, Galbraith CA, Stroud DA), pp. 418-421. The Stationery Office,

Edinburgh.

Makhado AB, Crawford RJM, Underhill LG (2006) Impact of predation by Cape

Fur Seals Arctocephalus pusillus pusillus on Cape Gannets Morus

capensis at Malgas island, Western Cape, South Africa. African Journal of

Marine Science 28, 681–687.

Marks MA, Brooke RK, Gildenhuys AM (1997) Cape Fur Seal Arctocephalus

pusillus predation on Cape Cormorants Phalacrocorax capensis and other

birds at Dyer Island, South Africa. Marine Ornithology 25, 9–12.

Melville DS, Shortridge KF (2006) Migratory waterbirds and Avian Influenza in

the East-Asian-Australasian flyway with particular reference to the 2003-

2004 H5N1 outbreak. In: Waterbirds around the world (eds. Boere GC,

Galbraith CA, Stroud DA), pp. 432-438. The Stationery Office, Edinburgh.

��

216

Mullers RHE, Navarro RA, Crawford RJM, Underhill LG (2009) The importance

of lipid-rich fish prey for Cape Gannet chick growth: are fishery discards

and alternative? ICES Journal of Marine Science 66. doi:10.1093/

icesjms/fsp210.

Okes NC, Hockey PAR, Pichegru L, van der Lingen CD, Crawford RJM,

Grémillet D (2009) Competition for shifting resources in the southern

Benguela upwelling: seabirds versus purse-seine fisheries. Biological


Pauly D, Christensen V, Dalsgaard J, Froese R, Torres F, Jr. (1998) Fishing

down marine food webs. Science 279, 860–863.

Pichegru L, Ryan PG, Le Bohec C, van der Lingen CD, Navarro RA, Petersen SL,

Lewis S, van der Westhuizen J, Grémillet D (in press) Spatial and

temporal overlap between vulnerable top predators feeding hotspots and

fisheries in the Benguela upwelling system: implications for the design of

marine protected areas. Marine Ecology Progress Series. doi:

10.3354/meps08283.




Rocke TE (2006a) Disease emergence and impacts in migratory waterbirds.

Workshop introduction. In: Waterbirds around the world (eds. Boere GC,

Galbraith CA, Stroud DA), pp. 410-411. The Stationery Office, Edinburgh.

Rocke TE (2006b) The global importance of Avian Botulism. In: Waterbirds

around the world (eds. Boere GC, Galbraith CA, Stroud DA), pp. 422-426.

The Stationery Office, Edinburgh.

Samuel MD (2006) Snow Goose Chen caerulescens overabundance increase Avian

Cholera mortality in other species. In: Waterbirds around the world (eds.

Boere GC, Galbraith CA, Stroud DA), pp. 443-445. The Stationery Office,

Edinburgh.

Shortridge KF, Melville DS (2006) Domestic poultry and migratory birds in the

interspecies transmission of Avian Influenza viruses: a view from Hong

Kong. In: Waterbirds around the world (eds. Boere GC, Galbraith CA,

Stroud DA), pp. 427-431. The Stationery Office, Edinburgh.

Steneck RS (1998) Human influences on coastal ecosystems: does overfishing

create trophic cascades? Trends in Ecology & Evolution 13, 429–430.

Tasker ML, Camphuysen CJ, Cooper J, Garthe S, Montevecchi WA, Blaber SJM

(2000) The impacts of fishing on marine birds. ICES Journal of Marine

Science 57, 531–547.

van der Lingen CD, Coetzee JC, Demarcq H, Drapeau L, Fairweather TP,

Hutchings L (2005) An eastward shift in the distribution of southern

Benguela sardine. Global Ocean Ecosystem Dynamics 11, 17–22.

Whittington PA, Martin AP, Klages NTW (2006) Status, distribution and

conservation implications of the Kelp Gull Larus dominicanus vetula

within the Eastern Cape region of South Africa. Emu 106, 127–139.

Williams AJ, Parsons N (2004) Cholera catastrophes: are Kelp Gull culprits?

Bird Numbers 13, 8–10.

�

APPENDIX A.1

��

��

��

��

217

��

��

��

��

Great White Pelicans Pelecanus onocrotalus breed in Africa, Europe and Asia

from Greece to Vietnam. In Africa, the Western Cape is probably the only place

in the world where pelican numbers have shown a sustained increase over the

past few decades. Nothing is known regarding the prevalence of pathogens

present in these populations. Therefore, 50 Great White Pelicans from Western

Cape were tested for the presence of various bacteria and viruses by polymerase

chain reaction (PCR). It was observed that 49 pelicans were positive for

Mycoplasma spp., 22 for Salmonella spp. and three for NDV, making these a

potential risk to domesticated avian species, as well as human beings. This must

be considered for management decisions regarding artificial food sources that

bring pelicans in contact with domestic animals.

��

Great White Pelicans Pelecanus onocrotalus breed in Africa, Europe and Asia

from Greece to Vietman. The world’s population is believed to be about 85 000

pairs, of which about 80% are in Africa (del Hoyo et al. 1992). Although the

population of Great White pelican is on decline worldwide, their numbers have

increased substantially in the Western Cape in the last decades. Traditionally,

they feed at freshwater wetlands and at estuaries (Crawford et al. 1995). The

construction of a plethora of farm dams and the practice of stocking them with

freshwater Carp Cyprinus carpio has increased food availability for the pelicans.

Also, in recent years, they have started feeding on offal at pig and chicken farms

in the Greater Cape Town area. This behaviour places them at risk to mass

poisoning incidents. On the other hand, wild birds are a serious threat to public

��

218

health because they carry emerging zoonotic pathogens, either as a reservoir

host or by dispersing infected arthropod vectors (Reed et al. 2003).

In order to better understand and conserve the population of Great White

Pelican in the Western Cape and for ecological reasons, it is essential to study,

not only the impact that dumped offal is having on breeding productivity, but

also the prevalence of pathogens present in these animals.

� ��

In the present study we investigated the prevalence of Mycoplasma spp.,

Salmonella spp., Chlamydia psittacci, Avian Infectious Laryngotracheitis Virus

(ILTV, Herpesvirus), Newcastle Disease Virus (NDV, Paramixovirus type I),

Infectious Bronchitis Virus (IBV, Coronavirus), and Avian Influenza Virus (AIV,

Influenza, A type) by polymerase chain reaction (PCR) in a community of Great

White Pelicans in the Western Cape, South Africa, capturing 50 birds using

modified leg traps (King et al. 1998) on a pig farm.

A total of 50 tracheal and 50 cloacal samples were taken from apparently healthy

pelicans using sterile swabs. The samples were preserved in a cell growth

medium RPMI-1640 (Roswell Park Memorial Institute, Sigma, St. Louis, MO,

USA) with 10% horse serum.

DNA extractions were performed following the protocol developed by Tola et al.

(1997), RNA extractions were performed with Tripure isolation reagent (Roche,

IN, USA) and the RNA was transcribed to cDNA with the iScript cDNA

synthesis kit (Biorad, CA, USA) according to the instructions of the

manufacturers. A 5 μl of DNA sample was amplified in a 25 μl reaction mixture

containing 10 pM of each primer and 12.5 μl of iQ SYBR Green supermix

(Biorad, CA, USA). The primers and conditions for the PCR were followed

according to the protocols described previously (Table A.1). PCR was performed

in a MyiQ Single-Color Real-Time detection system (Biorad, CA, USA).

��

219

��

The PCR results showed that 49 (98%) pelicans were positive to Mycoplasma

spp., 22 (44%) to Salmonella spp. and 3 (6%) to NDV. All birds were negative for

Chlamydia psittacci, ILTV, IBV and AIV.

Regarding avian mycoplasmas, only four species (M. gallisepticum, M. synoviae,

M. meleagridis and M. iowae) are known to be pathogenic (Bradbury 2005).

Although we have not determined yet which Mycoplasma species were present in

the studied population, it is known that the adaptation of, for example, M.

gallisepticum to a free flying avian host presents potential problems for the

control of mycoplasmosis in commercial poultry (Luttrell et al. 2001).

NDV, a pathogen of social and economical importance, is known to be

disseminated throughout the world by migratory and wild birds (Soares et al.

2005). There is also a general concern that NDV wild strains, circulating among

wild birds, can become highly pathogenic in poultry (Shengqing et al. 2002). The

main threat comes from Psittacine species and racing pigeons (Cross 1991), but

in some regions, other wild birds may represent an important role in spreading

the disease. Pfitzer et al. (2000) reported the presence of NDV antibodies in wild

aquatic birds in an intensive ostrich farming area in South Africa and it was

suggested that these birds could represent an important reservoir of NDV in the

area.

In this context, the high percentage of Salmonella and Mycoplasma spp. and the

presence of NDV shown by this Great White Pelican population indicated that

they were an important reservoir for these organisms.

Further sampling will reveal whether there are differences in the prevalence of

pathogens between the three main southern African populations, as could be

expected due to the different foraging behaviours observed. A comparison

between the natural- and offal-feeding pelicans will help assess the risk of

��

220

transmission of diseases from domestic animals to scavenging pelicans in the

Western Cape.

It is concluded that though some pelican populations are on decline worldwide,

the high density of Great White Pelicans foraging at farms in the Western Cape

places them at risk, as well as to domesticated avians and human beings, for

spread of pathogens. This must be considered for management decisions

regarding artificial food sources that bring pelicans in contact with domestic

animals.

��

221

�

��

Bradbury JM (2005) Gordon Memorial Lecture. Poultry mycoplasmas:

sophisticated pathogens in simple guise. British Poultry Science 46, 125–

136.



Africa’s Western Cape. South African Journal Marine Science 15, 33–42.

Cross GM (1991) Newcastle disease. The Veterinary Clinics of North America.

Small Animal Practice. 21, 1231–1239.


Handbook of the birds of the world, vol. 1. Lynx Editions, Barcelona.

Hewinson RG, Griffiths PC, Bevan BJ, Kirwan SE, Field ME, Woodward MJ,

Dawson M (1997) Detection of Chlamydia psittaci DNA in avian clinical

samples by polymerase chain reaction. Veterinarian Microbiology 54, 155–

166.

King DT, Paulson JD, Leblang DJ, Bruce K (1998) Two capture techniques for

American White Pelicans and Great Blue Herons. Colonial Waterbirds, 21,

258–260.

Liu HJ, Lee LH, Shih WL, Lin MY, Liao MH (2003) Detection of Infectious

Bronchitis virus by multiplex polymerase chain reaction and sequence

analysis. Journal of Virological Methods 109, 31–37.

Luttrell MP, Stallknecht DE, Kleven SH, Kavanaugh DM, Corn JL, Fischer JR

(2001) Mycoplasma gallisepticum in House Finches Carpodacus mexicanus

and other wild birds associated with poultry production facilities. Avian

Diseases 45, 321– 329.

Malorny B, Hoorfar J, Bunge C, Helmuth R (2003) Multicenter validation of the

analytical accuracy of Salmonella PCR: towards an international

standard. Applied Environmental Microbiology 69, 290–296.

Pang Y, Wang H, Girshick T, Xie Z, Khan MI (2002) Development and

application of a multiplex polymerase chain reaction for avian respiratory

agents. Avian Diseases 46, 691–699.

Pfitzer S, Verwoerd DJ, Gerdes GH, Labuschagne AE, Erasmus A, Manvell RJ,

Grund C (2000) Newcastle disease and Avian Influenza A virus in wild

waterfowl in South Africa. Avian Diseases 44, 655– 660.




Scholz E, Porter RE, Guo P (1994) Differential diagnosis of Infectious

Laryngotracheitis from other avian respiratory diseases by a simplified

PCR procedure. Journal of Virological Methods 50, 313–321.

Shengqing Y, Kishida N, Ito H, Kida H, Otsuki K, Kawaoka Y, Ito T (2002)

Generation of velogenic Newcastle disease viruses from a non-pathogenic

waterfowl isolate by passaging in chickens. Virology 301, 206–211.

��

222

Soares PB, Demetrio C, Sanfilippo L, Kawanoto AH, Brentano L, Durigon EL

(2005) Standardization of a duplex RT-PCR for the detection of Influenza

A and Newcastle disease viruses in migratory birds. Journal of Virological

Methods 123, 125–130.

Spackman E, Senne DA, Bulaga LL, Myers TJ, Perdue ML, Garber LP, Lohman

K, Daum LT, Suarez DL (2003) Development of real-time RT-PCR for the

detection of Avian Influenza virus. Avian Diseases 47, 1079–1082.

Spergser J, Aurich C, Aurich JE, Rosengarten R (2002) High prevalence of

mycoplasmas in the genital tract of asymptomatic stallions in Austria.

Veterinary Microbiology 87, 119–129.

Tola S, Angioi A, Rocchigiani AM, Idini G, Manunta D, Galleri G, Leori G (1997)

Detection of Mycoplasma agalactiae in sheep milk samples by polymerase

chain reaction. Veterinary Microbiology 54, 17– 22.

Table

A.1

: P

rim

er s

equ

en

ces a

nd

PC

R c

on

dit

ion

s f

or t

he s

creen

ing o

f G

reat W

hit

e P

eli

can

trach

eal

an

d c

loacal

sam

ple

s.

* A

n i

nit

ial

cycle

of

95

°C

for 2

min

wa

s p

erfo

rm

ed

for a

ll P

CR

.

PC

R c

on

dit

ion

s (

40

cy

cle

s)*

P

ath

og

en

P

rim

ers

Ta

rg

et

Re

fere

nce

s

Den

atu

ra

tio

n

An

nea

lin

g

Exten

sio

n

Myco

pla

sm

a s

pp

. 5

’-C

CA

GA

T T

CC

TA

C G

GG

CA

-3’

16

S r

ibosom

e

Sp

ergser

95

°C

6

4°C

7

2°C

5

’-T

GC

GA

G C

AT

AC

T A

CT

CA

G G

C-3

’ su

bu

nit

et

al.

(2

00

2)

for 4

0 s

ec

for 4

0 s

ec

for 1

min

Sa

lm

on

ella

sp

p.

5’-G

CA

AG

A C

AC

TC

C T

CA

AA

G C

C-3

’ In

vA

gen

e

Ma

lorn

y

95

°C

6

0°C

7

2°C

5

’-C

CT

TC

C C

AC

AT

A G

TG

CC

A T

C-3

’

et a

l. (

20

03

) fo

r 4

0 s

ec

for 4

0 s

ec

for 4

0 s

ec

Ch

la

myd

ia

5

’-G

TG

AA

A T

TA

TC

G C

CA

CG

T T

CG

GG

C A

A-3

’ O

mp

A g

en

e

Hew

inson

9

5°C

6

2°C

7

2°C

psitta

cci

5’-T

CA

TC

G C

AC

CG

T C

AA

AG

G A

AC

C-3

’

et a

l. (

19

97

) fo

r 4

0 s

ec

for 4

0 s

ec

for 4

0 s

ec

ILT

V

5’-C

GT

GG

C T

TC

AC

C A

GC

AA

-3’

Th

ym

idin

e

Sch

olz

9

5°C

5

7°C

7

2°C

5

’-C

GA

GT

A A

GT

AA

T A

GG

CT

-3’

kin

ase g

en

e

et a

l. (

19

94

) fo

r 4

0 s

ec

for 4

0 s

ec

for 1

min

AIV

(A

Ty

pe

) 5

’-A

GA

TG

A G

TC

TT

C T

AA

CC

G A

GG

TC

G-3

’ M

atrix

gen

e

Sp

ack

ma

n

95

°C

60

°C

7

2°C

5

’-T

GC

AA

A A

AC

AT

C T

TC

AA

G T

CT

CT

G-3

’

et a

l. (

20

03

) fo

r 3

0 s

ec

for 3

0 s

ec

for 4

0 s

ec

ND

V

5’- G

GA

GG

A T

GT

TG

G C

AG

CA

T T

-3’

S1

gly

cop

rotein

P

an

g

95

°C

6

0°C

7

2°C

5

’- G

TC

AA

C A

TA

TA

C A

CC

TC

A T

C-3

’ gen

e

et a

l. (

20

02

) fo

r 4

0 s

ec

for 4

0 s

ec

for 4

0 s

ec

IBV

5

’-T

GG

TT

G G

CA

TT

T A

CA

CG

GG

G-3

’ F

0 f

usio

n p

rotein

L

iu e

t a

l.

95

°C

6

0°C

7

2°C

5

’-C

AA

TG

G G

TA

AC

A A

AC

AC

-3’

cod

ing s

eq

uen

ce

(20

03

) fo

r 4

0 s

ec

for 4

0 s

ec

for 4

0 s

ec

�

APPENDIX A.2

��

225

Photos 1–4: (Top left) Great White Pelican colony at Dassen Island, South Africa,

showing two breeding groups or subcolonies. (Top right) Breeding colony in Boom

Point, Dassen Island, during the 2004 breeding season. The wreck of the

Namaqua is on the background. (Bottom left): General view of the breeding

colony of Great White Pelicans on an island at Hardap Dam, Namibia in May

2006. (Bottom right): Bird Island Guano Platform, Namibia, built in the 1930s

order to increase the breeding area of guano-producing species (i.e. cormorants).

Pelicans use the platform as a breeding site since 1949.

226

Photo 5–9: (Top left) Pelicans at a free range pig farm in 1990 (photo by Jan Hofmeyr);

(top right and middle) Pelicans waiting for food and eating chicken guts at the

monitored pig farm during the abundance offal in 2004. (Bottom) Monitored pig farm

during period of scarcity of food in July 2005. Starved pelicans tolerated people at close

proximity in their attempts to feed and often presented injuries and burns. About 10% of

the total population died of starvation during this time.

227

Photo 10: Three-

year-old Great

White Pelican

incubating two eggs

at the breeding

colony on Dassen

Island in October

2004.

Photo 11: Three-year-

old male (left) and

female of unknown

age (right) on Dassen

Island in September

2008. Notice the black

engraved ring on the

left leg of the male

pelican (photograph by

Leshia Upfold).

Photo 12: Sibbling

aggression between

two chicks (1–3 days-

old) of Great White

Pelican on Dassen

Island in October

2004.

228

Photo 13–14: (Left) Ringed Great White Pelican adult at the Orange River

estuary in January 2008 (photograph by Mark Anderson). (Right): Ringed black-

downy chick on Dassen Island (photograph by Madga Remisiewicz).

Photo 15 Kelp Gull

transporting a pelican egg

pilfered from the breeding

colony on Dassen Island

(photograph by Tom

Peschak).

Photo 16–17: (Left) A pelican on the hand... supervisors Les Underhill and Peter

Ryan during a ringing expedition to Dassen Island in January 2007.(Right):

West Coast National Park rangers at our return from one of the many trips to

Malgas and Jutten Islands in Saldanha Bay.

229

Photo 18–20: Great White Pelican predation on seabird chickss on Jutten Island,

West Coast National Park, during 2006 and 2007. (Top left) Group foraging on

the plains of Jutten Island, targeting Kelp Gulls Larus dominicanus chicks.

(Top right) Defending from a Kelp Gull attack while preying on Cape

Cormorants Phalacrocorax capensis. (Bottom) Eating Cape Cormorant fledglings.

230

Photo 21–22: Great White Pelican predation on Cape Gannets Morus capensis on

Malgas Island, West Coast National Park, in 2007. (Top) A group of pelicans

fighting for a downy chick. (Bottom): Agonistic interaction between two pelicans

preying on Cape Gannets.

Never the words of Nixon Lanier Merritt (1879–1972) in this

well-known limerick The Pelican were truer than on the islands

off the west coast of South Africa. They kept returning to my

mind as I observed pelicans eating seabird chicks:

A wonderful bird is the pelican

His bill will hold more than his belican

He can take in his beak

Enough food for a week

I’m damned if I know how the helican

Documents

POPULATION DYNAMICS OFPopulation Ecology, Life-History, Population Genetics, Behavioural Ecology, Avian Disease Ecology and Conservation. More specifically, it employed techniques