Comparative Ecology of the Noisy Philemon corniculatus (Latham 1790) and the carunculata (Shaw 1790) in Central Eastern

by

Anthony S. J. Saunders

A thesis presented to the University of Western Sydney as partial fulfilment of the requirements for the degree of Doctor of Philosophy

December, 2004

„A.S.J. Saunders

i ACKNOWLEDGEMENTS

I would like to thank the many people and agencies who provided assistance and support during this project. I would especially like to thank Shelley Burgin who supervised this work. Her patience and encouragement is greatly appreciated. I would also like to thank Hugh Ford and Stephen Ambrose for helpful advice during the early stages of planning this project. Richard Turner provided assistance with plant identification on Kings Tableland. Warren Sweeney provided companionship during some of the field work and has shown an active interest in the project since its inception. John Best (CSIRO Mathematical and Information Sciences, Macquarie University), Per Brockhoff (Royal Veterinary and Agricultural University, Copenhagen) and Hugh Jones provided very helpful advice and assistance with statistical analysis of the data. The following people provided valuable constructive criticism as referees for manuscripts arising out of this project, Hugh Ford, Harry Recher, A. Nicholls, Don Franklin as well as other anonymous referees, as did the following journal editors, Ian Rowley and Annette Cam. The New South Wales National Parks and Wildlife Service (Licence No. B897) and State Forests of New South Wales State Forests (Permit No’s. 04558, 04729, 04787, 04900) provided permission to carryout field work in their respective reserves. Phil Deamer and Bill Nethery of State Forests, and Dave Crust, Amanda Johnson and Sherrie-lee Evans of NPWS provided access assistance and advice for their local reserves. The Brothers of St Marys Towers, Douglas Park also kindly provided access to their property. The Australian Museum, through Walter Boles and Max Moulds, provided access to museum specimens of cicadas and . I would also like to thank my mother, Della, and my wife, Noelene, for their patience and support, especially through the final stages of the thesis.

ii Format of Thesis

Many of the chapters of this thesis are presented here as discrete studies as several chapters have been published as papers in refereed journals. For this reason, each data chapter contains a discrete introduction, methods, results and discussion. The general introduction (Chapter 1) puts the study into perspective and general discussion (Chapter 8) provides a synthesis of the research and its implications. The methods section in each chapter describes methods specific to the data collection and analysis for that chapter and not methods already described in chapter two, which describes the location of study sites, transects and plots, and the period over which the data were collected. Chapters which have been published are listed below with their corresponding published paper.

Chapter 4. The importance of eucalypt nectar in the diets of large honeyeaters. was published as: Saunders, A.S.J., Burgin, S. & Jones, H. 2003 The importance of eucalypt nectar in the diet of large honeyeaters. Corella 27: 1-12

Chapter 6. Selective foliage foraging by Red Wattlebirds A. carunculata and Noisy P. corniculatus was published as: Saunders, A.S.J. & Burgin, S. 2001 Selective foliage foraging by Red Wattlebirds, Anthochaera carunculata, and Noisy Friarbirds, Philemon corniculatus. Emu 101: 163-166

Chapter 7. Gape width and prey selectivity in the P. corniculatus and the Red Wattlebird A. carunculata was published as:

iii Saunders, A.S.J, Ambrose, S.J. & Burgin, S. 1995 Gape width and prey selectivity in the Noisy Friarbird Philemon corniculatus and the Red Wattlebird Anthochaera carunculata. Emu 95: 297-300

The following paper, based on data collected prior to the commencement of this study, was used as the basis for site selection and general study design for this thesis:

Saunders, A.S.J. 1993 Seasonal variation in the distribution of the Noisy Friarbird Philemon corniculatus and the Red Wattlebird Anthochaera carunculata in eastern New South Wales. Australian Watcher 15: 49-59.

Reprints of all these papers are provided in the appendices.

iv

TABLE OF CONTENTS

Acknowledgements ...... ii Format of Thesis ...... iii Statement of Authenticity ...... v List of Tables ...... x List of Figures ...... xii List of Appendices ...... xiv Abstract ...... xvii 1. General Introduction 1.1 Food resource patterns: their role in distribution and foraging behaviour ...... 1 1.2 Food resource patterns: effects on niche overlap and competition ...... 6 1.3 The general ecology of honeyeaters ...... 9 1.4 Some comparisons with other nectarivores ...... 11 1.5 The basis for selection of species for this study ...... 13 1.6 The general ecology of the Noisy Friarbird and the Red Wattlebird ...... 14 1.7 The questions for this study ...... 18 2. Location, topography and climate of study sites ...... 20 3. Floristics and structure in forest sites along an east-west transect in central eastern New South Wales 3.1 Methods ...... 26 3.3 Results 3.3.1 Forest Structure ...... 28 3.3.2 Floristic Associations ...... 33 4. The importance of eucalypt nectar in the diets of large honeyeaters

vi 4.1 Introduction ...... 40 4.2 Methods 4.2.1 Study sites ...... 43 4.2.2 Tree counts and species composition ...... 44 4.2.3 Bird Counts ...... 45 4.2.4 Analysis of bird count data ...... 46 4.2.5 Analysis of time budget data ...... 47 4.3 Results 4.3.1 Flowering trees ...... 48 4.3.2 counts at the regional scale ...... 48 4.3.3 Honeyeater counts at the local scale ...... 55 4.3.4 Correlations between bird behaviour and densities of flowering trees ...... 55 4.4 Discussion 4.4.1 Correlations between flowering trees and honeyeaters ...... 55 4.4.2 Aggression ...... 61 5. Comparative behaviour of two large honeyeaters: a landscape perspective 5.1 Introduction ...... 64 5.2 Methods 5.2.1 Data collection ...... 66 5.2.2 Statistical analysis ...... 67 5.2.3 Niche breadth and overlap ...... 69 5.3 Results 5.3.1 General patterns in foraging substrate selection ..... 69 5.3.2 Spatial and temporal variation in substrate selection ...... 74 5.3.3 Niche overlap and aggression ...... 79

vii 5.4 Discussion 5.4.1 Foraging patterns and relationships with morphological differences ...... 85 5.4.2 Spatial and temporal variation in foraging patterns 86 5.4.3 Niche variation and relationship with aggression .... 87 5.4.4 Landscape perspectives of foraging studies ...... 88 6. Selective foliage foraging by Red Wattlebirds A. carunculata and Noisy Friarbirds P. corniculatus 6.1 Introduction ...... 91 6.2 Methods ...... 93 6.3 Results ...... 94 6.4 Discussion ...... 98 7. Gape width and prey selectivity in the Noisy Friarbird P. corniculatus and the Red Wattlebird A. carunculata 7.1 Introduction ...... 100 7.2 Methods ...... 101 7.3 Results ...... 102 7.4 Discussion ...... 104 8. Discussion 8.1 The nature of food resources selected by Noisy Friarbirds and Red Wattlebirds in the landscape and their effects on distribution and foraging behaviour ...... 106 8.2 Niche overlap and competition between Noisy Friarbirds and Red Wattlebirds ...... 114 8.3 The ecological role of Noisy Friarbirds and Red Wattlebirds in the landscape ...... 115 8.4 The relevance of scale to studies of nomadic generalists in unpredictable environments ...... 118 8.5 Movement of in response to the distribution of food

viii resources ...... 126 9. References ...... 134 10. Appendices ...... 171

ix LIST OF TABLES

1.1 Summary of food resource patterns and corresponding foraging behaviours ...... 3 2.1 Location, altitude, climate and transects at study sites ...... 22 3.1 Forest canopy structure and floristic associations of transects at each site ...... 29 3.2 Nested ANOVA of tree heights for plots nested within transects, nested within sites ...... 32 4.1 Total counts of P. corniculatus and A. carunculata by site and season ...... 52 4.2 Analysis of deviance summary for counts of P. corniculatus and A. carunculata for all sites, times and seasons ...... 54 4.3 Analysis of deviance summary for alternative models for honeyeater counts at Goobang National Park in spring 1994 .. 57 4.4 Spearman Rank correlations between honeyeaters, flowering trees and behaviours ...... 58 5.1 Categories used for time-budgeting behaviours ...... 68 5.2 Number of bouts for each honeyeater in each height class ...... 72 5.3 Number of bouts within each substrate selected by each honeyeater for sites pooled where both species present ...... 72 5.4 Number of bouts for foraging substrates selected in each height class for each honeyeater ...... 72 5.5 Summary of analysis of deviance for hierachical loglinear models for species, substrate and height interactions ...... 73 5.6 Number of foraging bouts and non-foraging bouts for each honeyeater ...... 76 5.7 Number of bouts at each substrate selected for sites where both honeyeaters were found ...... 76

x 5.8 Summary of analysis of deviance for hierachical loglinear models for species, substrate and site interactions where both honeyeaters were present ...... 77 5.9 Number of bouts observed for each honeyeater on different substrates by year and site ...... 78 5.10 Number of bouts observed for each honeyeater on different substrates for each year at Goobang National Park ...... 79 5.11 Niche breadth and niche overlap based on foraging heights and substrates selected ...... 80 5.12 Analysis of variance for spatial and temporal variation in niche overlap ...... 81

xi LIST OF FIGURES

2.1 Location map and cross-section of sites, topographic regions and range of A. carunculata and P. corniculatus ...... 21 2.2 Mean precipitation, minimum and maximum temperatures for study sites ...... 23 3.1 Correlations between altitude and tree density across sites ...... 30 3.2 Mean tree heights ± S.E. for each plot ...... 31 3.3 Percent tree composition for tree groups at each site ...... 34 3.4 Percent tree composition of eucalypt subgenera at each site ..... 34 3.5 Correspondence analysis between tree species’ densities and transects ...... 38 3.6 Number of tree species at plots and transects for each site ...... 39 4.1 Density of trees species at each site ...... 49 4.2 Seasonal densities of flowering trees, A. carunculata and P. corniculatus at sites from 1992 to 1993 ...... 50 4.3 Seasonal densities of flowering trees,A. carunculata and P. corniculatus at Goobang National Park from 1993 to 1996 ... 51 4.4 Poisson regression models for counts of A. carunculata and P. corniculatus against flowering trees at Goobang National Park in spring 1994 ...... 56 4.5 Correlations between honeyeaters, flowering trees and behaviours ...... 59 5.1 Percent foraging time at various substrates by A. carunculata and P. corniculatus (data pooled) ...... 71 5.2 Activity time budgets for A. carunculata and P. corniculatus (data pooled) ...... 75 5.3 Spatial and temporal variation in niche overlap ...... 82 5.4 Direction of aggression in honeyeaters ...... 83

xii 5.5 Correlation between niche overlap and interspecific aggression ... 84 6.1 Expected and observed use of foliage-gleaned trees by A. carunculata ...... 96 6.2 Expected and observed use of foliage-gleaned trees by P. corniculatus ...... 97 7.1 Gape width of A. carunculata and P. corniculatus, and prey widths of P. moerens and P. plaga ...... 103

xiii LIST OF APPENDICES

10.1 Percent foraging time at various substrates by A. carunculata at Royal National Park ...... 172 10.2 Percent foraging time at various substrates by A. carunculata and P. corniculatus at Clandulla State Forest in 1992 ...... 173 10.3 Percent foraging time at various substrates by A. carunculata and P. corniculatus at Clandulla State Forest in 1993 ...... 174 10.4 Percent foraging time at various substrates by A. carunculata and P. corniculatus at Munghorn Gap Nature Reserve in 1992 175 10.5 Percent foraging time at various substrates by A. carunculata and P. corniculatus at Munghorn Gap Nature Reserve in 1993 176 10.6 Percent foraging time at various substrates by A. carunculata and P. corniculatus at Goobang National Park in 1993 ...... 177 10.7 Percent foraging time at various substrates by A. carunculata and P. corniculatus at Clandulla State Forest in 1994 ...... 178 10.8 Percent foraging time at various substrates by A. carunculata and P. corniculatus at Clandulla State Forest in 1995 ...... 179 10.9 Percent foraging time at various substrates by A. carunculata and P. corniculatus at Back Yamma State Forest in 1992 ...... 180 10.10 Percent foraging time at various substrates by A. carunculata and P. corniculatus at Back Yamma State Forest in 1993 ..... 181 10.11 Percent time at various activities for A. carunculata at Royal National Park ...... 182 10.12 Percent time at various activities for A. carunculata and P. corniculatus at Clandulla State Forest ...... 183 10.13 Percent time at various activities for A. carunculata and P. corniculatus at Goobang National Park in 1993 ...... 184 10.14 Percent time at various activities for A. carunculata and

xiv P. corniculatus at Goobang National Park in 1994 and 1995 . 185 10.15 Percent time at various activities for A. carunculata and P. corniculatus at Back Yamma State Forest ...... 186 10.16 Percent time in aggression between A. carunculata , P. corniculatus and other honeyeaters at Royal National Park ...... 187 10.17 Percent time in aggression between A. carunculata , P. corniculatus and other honeyeaters at Clandulla State Forest ...... 188 10.18 Percent time in aggression between A. carunculata , P. corniculatus and other honeyeaters at Munghorn Gap Nature Reserve ...... 189 10.19 Percent time in aggression between A. carunculata , P. corniculatus and other honeyeaters at Goobang National Park ...... 190 10.20 Percent time in aggression between A. carunculata , P. corniculatus and other honeyeaters at Back Yamma State Forest ...... 191

Saunders, A.S.J., Burgin, S. & Jones, H. 2003 The importance of eucalypt nectar in the diet of large honeyeaters Corella 27: 1-12 ...... 192

Saunders, A.S.J. & Burgin, S. 2001 Selective foliage foraging by Red Wattlebirds, Anthochaera carunculata, and Noisy Friarbirds, Philemon corniculatus Emu 101: 163-166 ...... 204

Saunders, A.S.J, Ambrose, S.J. & Burgin, S. 1995 Gape width and prey selectivity in the Noisy Friarbird Philemon corniculatus and the Red

xv Wattlebird Anthochaera carunculata. Emu 95: 297-300 ...... 208

Saunders, A.S.J. 1993 Seasonal variation in the distribution of the Noisy Friarbird Philemon corniculatus and the Red Wattlebird Anthochaera carunculata in eastern New South Wales. Australian Bird Watcher 15: 49-59...... 212

xvi ABSTRACT Densities and behaviour of Noisy Friarbirds Philemon corniculatus and Red Wattlebirds Anthochaera carunculata were measured during 1992 and 1993 at six sites along an east-west transect through central New South Wales from the Central Coast through to the Central Western Slopes, and at Goobang National Park from 1993 to 1996.

Both P. corniculatus and A. carunculata were found to be mostly canopy foragers with occasional forays into the shrub layer when food resources became available there. Flower-probing and foliage gleaning comprised the bulk of foraging behaviour. Most of the food resources used by these honeyeaters were seasonally unreliable and unpredictable, so that they needed to shift between foods and track them over hundreds of kilometres. A strong relationship was found between the densities of honeyeaters and the density of flowering trees at both regional and local scales. Sites on the western slopes were more important for nectar during winter and spring, while sites on the tablelands provided a greater diversity of foods over all seasons. When foraging at foliage, both honeyeaters were found to preferentially select . Time-budgeting revealed that both species spent little time in aggression and that most of this was intraspecific. Niche overlap was found to be positively correlated with interspecific aggression between them and suggested that some degree of competition exists. Because concentrations of shared food resources appeared to be short- lived and because these two honeyeaters switch between unreliable resources frequently, it is unlikely that competition shapes their foraging niche. It is more likely to be shaped by the temporal and spatial variability of their food resources.

Foraging substrate selectivity was found to vary with species, site, year and

xvii season, and food resources were found to vary considerably and often unpredictably at regional and local scales in this study. The distribution of these two honeyeaters over space and time was highly variable and their nomadic movements are likely to be linked to the variability of their food resources. It was found that large spatio-temporal scales of the order of hundreds of kilometres and several years were required to interpret the behaviour and distribution of large nomadic honeyeaters in forests, and it is likely that such scales will apply to other nomadic generalists.

Compared with P. corniculatus, A. carunculata was found to spend more time foraging and a larger portion of their foraging time, gleaning foliage for alternative carbohydrates. This is probably why A. carunculata was able to remain on tableland sites during winter, whereas P. corniculatus left and most likely sought rich nectar sites on the western slopes. Although both honeyeaters tracked nectar sources, P. corniculatus appeared to require richer resources as it was often absent at sites when lower nectar supplies were being exploited by A. carunculata. These differences can be accounted for by comparing the bills of these two honeyeaters. The bill of P. corniculatus is significantly larger both in length and gape width. The smaller bill size of A. carunculata may have allowed it to glean foliage more efficiently and thus enabled it to persist on tableland sites during the winter when P. corniculatus was forced to move-on. However, the larger bill size in P. corniculatus enabled them to ingest large cicadas that were too large for A. carunculata. The larger body mass and smaller bill size of A. carunculata may explain why this honeyeater spent more time foraging than P. corniculatus.

These two honeyeaters are very mobile and appear to cope within the fragmented landscape. Their ecological role as plant pollinators, seed dispersers and population limiters may have increased proportionally

xviii due to habitat fragmentation compared with less mobile species with similar ecological roles. Hence they may also have become more important in maintaining habitat patch quality.

xix 1. GENERAL INTRODUCTION

1.1 Food resource patterns: their role in animal distribution and foraging behaviour. The role food resources play in the distribution of is complex. It depends on the interaction between food availability at many scales and a forager’s mobility, foraging strategies and flexibility in response to environmental unpredictability (Ford 1989, Wiens 1989, Krebs 1994, Begon et al. 1996). Food sources may be constant, seasonal or unreliable, as well as limiting or non-limiting for foragers. In response to food resource patterns, foragers may be specialists or generalists, and may be resident, migratory or nomadic (Keast 1968, Ford 1989, Recher 1990, Krebs 1994, Mac Nally, R.C. & McGoldrick, J.M. 1997). The relationship between a food source and a forager will depend on the particular combination of these variables.

A forager’s perception of food availability will depend on constancy or regularity of supply, distribution within space, rate of supply compared with rate of consumption, and detectability. A food source would be described as constant when its availability fluctuates very little over time (eg. ants, Swart et al. 1999), although very few foods are likely to follow this pattern. Many food sources are seasonal (eg. fruit, Forde 1986), having a regular seasonal occurrence during a particular time of the year. Other food sources are unreliable or irregular. They may be unpredictable from year to year (eg. nectar, Mac Nally & McGoldrick 1997) or have cycles that include a variable number of years between years of relatively abundant supply (eg. flowering, Law et al. 2000). Even when such food sources have some seasonality, the relative abundance from year to year may be highly variable (eg. flowering, Law et al. 2000). A food resource may also be limiting i.e. limit the size of a population of foragers (Meagher 1991) because the resource is readily depleted (eg. as food for lizards, Roughgarden

1 1986), or non-limiting because the resource self-replenishes as it is consumed (eg. nectar, Carpenter 1978) or because rate of production exceeds rate of consumption (eg. nectar, Carpenter 1987). Food resources may also vary in distribution from localised and concentrated to evenly spread throughout a habitat, patch or region. Some foods may also vary in their detectability. Foods that are more difficult to detect (eg. worms, Goss-Custard 1977) may require considerably search time and effort compared with readily detectable foods. The combination of food resource characteristics will affect choice by foragers.

Foragers may also differ in their response to food resource characteristics. Mobility of foragers, their flexibility in response to food resource patterns and morphological adaptations to different foraging strategies will also affect food choice (Paton & Collins 1989). Foragers may be resident i.e. present all year round, generally with established feeding and/or breeding territories and with offspring dispersing from these territories. Some foragers are migratory and have regular seasonal movements between latitudes, altitudes or between regions (eg. waders, van de Kam et al. 2004), while others are nomadic, having irregular movements in response to erratic climatic conditions or unpredictable food resources (eg. snails, Bennett & Kitchens 2000, nectar, Mac Nally 2000). Coincidentally, foragers may be specialists, using a narrow range of resources (Recher 1990) where feeding is restricted to one or a few types of food sources (eg. ants, Swart et al. 1990, cockles, Norris & Johnstone 1998), or they may be generalists, where they have broad diets or are able to switch between various food sources (eg. nectar, fruit & insects, Brown & Hopkins 1996).

Food resource patterns and corresponding foraging behaviour patterns are summarised in Table 1.1. Food supply rates and food source reliability are matched to foraging behaviours which are more likely to provide superior

2 Table 1.1 Summary of food resource patterns and corresponding foraging behaviours.*

Food Food Correspondin Explanation Examples in the Supply reliability g Foraging Literature Rate Behaviour Limiting Constant Resident Foragers switch between Wolf 1970 generalist food sources as they McFarland & Ford 1991 become locally depleted or Jackson 1998 or move away until their Hester et al. 1999 particular food source is nomadic replenished. specialist

Seasonal Migratory Foragers will move into Ward 1965 specialist area when preferred food Hejl & Verner 1990 sources are in supply or Miles 1990 or will switch between Recher et al. 1991 different food sources as Cale 1994 resident they become available. Kelly & Wood 1996 generalist Barreto & Herrera 1998 Unreliable Nomadic Foragers will switch Collins 1985 generalist between food sources as McFarland & Ford 1991 they become available Mac Nally 2000 and/or move to areas when alternative food sources are available.

Non- Constant Resident Foragers are able to Swart et al. 1999 limiting specialist specialise when food Norris & Johnstone sources are non-limiting 1998 and constant in availability as there is no need to leave the area.

Seasonal Migratory Foragers will move into van de Kam 2004 specialist area when preferred food sources are in supply.

Unreliable Nomadic When food sources are Collins 1985 specialist unpredictable but plentiful Brown & Hopkins 1996 for short periods foragers Mac Nally 2000 or will switch between food Atkinson et al. 2002 sources as they become Bennetts & Kitchens nomadic available or move to areas 2000 generalist where their preferred food or Paton 2004 alternative food sources are available.

*For definitions refer to text in section 1.1 Pp 1-2.

adaptation to that particular food resource pattern. These foraging

3 behaviours are not necessarily discrete and they may be extremes along a continuum. In addition, foragers may be able to switch between several foraging strategies as resource characteristics change within an area over time (Cale 1994, Barreto & Herrera 1998) or between different areas (Mac Nally 2000). When food resources are limiting, foragers will either need to move or switch between different food sources (Brown & Hopkins 1996, Mac Nally 2000). Even when food sources are non-limiting, foragers need to move when their food sources are seasonal or unreliable (Mac Nally & McGoldrick 1997). Hence the ability to move between patches and/or the ability to switch between food sources is a prerequisite for many foragers (eg. Collins 1985, van de Kam et al. 2004), except for specialists utilising non- limiting food sources. Factors other than food sources, such as breeding requirements and predator avoidance (Alcock 1989, Ford 1989) are likely to limit resident specialists utilising non-limiting food sources.

Food resource availability in the landscape will be affected by patch quality, patch size and patch connectivity in a fragmented landscape (Forman & Gordon 1986, Krebs 1994). The influence of these patch variables on animals will depend on their mobility and foraging behaviours. Some animals require large areas of habitat (eg. deer, Hester et al. 1999, owls, Fleay 1968) while others are able to utilise small patches within the landscape (eg. hummingbirds, Wolf & Hainsworth 1971). Some animals are able to move between distant habitat patches (eg. honeyeaters, Mac Nally 2000) while others require continuous habitat (eg. some birds, Enoksson et al. 1995, birds in Table 3 p131, Ford & Barrett 1995). In other words, the requirements from a patch will differ for residents and migrants or nomads. Although the same variables may be important, their relative importance are likely to differ. For resident species a patch must provide all its needs, while for migrants or nomads a patch must be able to provide sufficient resources to enable part of their life histories to be completed or to enable them to gain

4 sufficient energy to be able to move between patches. Thus for migrants and nomads distance between patches and inter-patch habitat characteristics also become important.

Generalists are less likely to be affected by habitat quality within a patch than specialist foragers, at least above some threshold value for resource density, as they can consume a mixture of their preferred foods that are available in sufficient density and switch between them as each becomes exhausted (Cale 1994, Barreto & Herrera 1998). Nomadic foragers will be able to move between a series of patches if each patch has sufficient food resources to allow efficient foraging and if patches are sufficiently close so that not all of the energy obtained from the last patch is depleted before reaching the next patch (Mennechez et al. 2003). If either of these conditions is not met then even nomadic generalist foragers may not be able to persist in the landscape. Highly mobile foragers that can move long distances between suitable patches over various unsuitable habitat are more likely to persist in a fragmented landscape. However, if patches become too small or too widely dispersed then such species may perish.

The scales at which food resources and the animals that depend upon them vary will determine the appropriate scale of a study of the relationship between a forager and its food sources (Wiens et al. 1986, Mac Nally 1996, Malizia 2001, Morris & Wooller 2001, Kie et al. 2002). For resident species it may be sufficient to examine the relationship between food resource fluctuations and density and behaviour of foragers at a local site (eg. hummingbirds, Wolf 1970). For those utilising constantly available or regular seasonal food sources short time periods may also be appropriate (eg. Swart 1990, Sun & Moermond 1997, Jackson 1998). However, for foragers that are nomadic and which depend on unreliable food sources, relationships need to be studied at much larger regional scales and over long

5 time scales (eg. Mac Nally & McGoldrick 1997). A small scale study may provide particular insight into the behaviour of a species at one site, but relationships at this site may not necessarily be those that apply at other sites within the distribution range of that species (Wiens 1989). The distribution of migratory or nomadic foragers and their response to food resources is likely to vary over their range. Hence landscape-scale studies, rather than local- area studies, that employ large temporal and spatial scales are required to provide insights into the relationships that exist between highly mobile foragers and resources that vary over the foragers range. Also, when large scale studies are required, experimental manipulation of resources at a patch scale may be irrelevant and are difficult at the landscape scale (Haynes & Cronin 2004).

The Australian landscape is characterised by an unpredictable climate where food resources are largely unpredictable through space and over time (Mac Nally 2000). Being a nomadic generalist in such an environment may provide advantages over other foraging strategies. They may be more able to persist within the landscape under severe conditions than resident specialists and even migratory species because they can switch between different food sources. Nomadic foragers may be able to build up knowledge of resource rich areas throughout their range or may wander randomly, taking advantage of local rich food sources as they become available (Bennetts & Kitchens 2000). Nomadic generalists may also combine both strategies in dealing with unpredictable food resources in the landscape.

1.2 Food resource patterns: effects on niche overlap and competition. In the long-held general principle, known as ‘The Competitive Exclusion Principle’, two species may coexist in a stable environment only if they are ecologically dissimilar. If two ecologically identical species were to coexist for a length of time their niches would either diverge or they would move

6 away from each other to avoid competition (Krebs 1972). Competition is indicated when one species suffers a reduction in population or perishes in the presence of another species (Begon et al. 1996). Niche differentiation is the mechanism by which communities are structured so that coexistence is possible and would evolve to reduce competition where two or more species share limiting food resources (Cody 1974). Niche overlap has often been used as a measure of competition between species (eg. MacArthur & Levins 1967, Cody 1974, Leviten 1978, Pacala & Roughgarden 1982, James & Poulin 1998, McKnight & Hepp 1998). It has been found to vary with resource availability; with high overlap at abundant non-limiting food resources and a shift to low overlap as resources become scarce (Schoener 1986). Niche expansion, when potential competitors are absent, or niche differences between areas of allopatry and sympatry of two suspected competitors, have also been used to indicate interspecific competition (Cody 1974). Similar species can have narrower niche breadths with less niche overlap in their sympatric range compared with areas where the species occur separately (Cody 1974). Evidence for competition can be detected when there is a niche shift or change in the populations of suspected competitors during field experiments where resources or competitors are manipulated (eg. Carpenter 1979, Creese & Underwood 1982, Pacala & Roughgarden 1985, Abramsky et al. 1986, Aho et al. 1997).

However, several reviews have criticised the ability of niche measures to quantify interspecific competition (MacArthur 1972, Dayton 1973, Connell 1983, Schoener 1983, Underwood 1986). Niche shifts in the presence/absence of a competitor can only be used as evidence if this is the only important difference between sites or time intervals being compared (Underwood 1986, Wiens 1989) and are not always consistent with theoretical expectations (eg. Nour et al. 1997, Hester et al. 1999). It is necessary to know what factors limit their distribution in order for this to

7 provide clear evidence of competition between two species (Underwood 1986). Other factors, such as food choice, breeding requirements and predator avoidance are also likely to influence species distribution and may also be important in determining coexistence (Alcock 1989, Ford 1989). High foraging niche overlap can occur within communities where species differentiate along other niche dimensions (Cody 1974). It has also been argued that potential competitors may never reach densities in natural communities which induce interspecific competition (Krebs 1994, Begon et al. 1996). Intraspecific competition may also be stronger than interspecific competition and thus limit population density of competitors, allowing species to coexist (Underwood 1986). Field experiments that manipulate populations of suspected competitors may also interfere with natural interactions (Underwood 1986). Niche differentiation could also result from evolution of different life histories or adaptations to different food types and hence not result from competition (Alcock 1989). Heterogeneous and unpredictable environments may never reach equilibrium and the balance may constantly shift. Under such conditions, competitive exclusion or niche shift may never occur and coexistence within communities may be fostered (Krebs 1994, Begon et al. 1996).

If foraging species share non-limiting resources then they have no reason to competitively exclude one another or cause niche differentiation (Johnson 1966, Underwood 1986). In addition, many food resources can be short- lived, seasonally unreliable and unpredictable over time and space. Species foraging together at short-lived, abundant and often unpredictable food resources will not be constantly interacting and there would not be sufficient time for competitive exclusion or niche differentiation to take place (Wiens 1989, Mac Nally 2000). Niche expansion or compression in relation to variable resource density will only apply to resident species and not to migratory or nomadic species (Feinsinger & Swarm 1982).

8 Residents utilising limiting food resources and migrants foraging at reliable food resources at one time and place on a regular basis are more likely to exhibit niche differentiation between them than nomadic generalists which may turn up anywhere for relatively short periods to share locally abundant but short-lived, unpredictable food resources. Hence, it is more likely that variability in food resources, rather than competition, shapes the niche and foraging behaviours of nomadic generalists.

Nectarivores are particularly suited to the study of competition between species, particularly where they may become territorial at nectar sources which can be manipulated (Wiens 1989). Competition between honeyeaters will depend on the nature of their nectar sources and the degree of territoriality exhibited by them.

1.3 The general ecology of honeyeaters.1 Honeyeaters (Meliphagidae) are distinguishable from other nectarivores in having a protrusible tongue formed into a single open channel that splits into four fimbriated brush-like segments at the tip. Their slender heads and long decurved bills with slit-like operculate nostrils set honeyeaters apart from other songbirds (Schodde & Mason 1999).

There are about 170 species world-wide (Schodde & Mason 1999) and 73 of these occur within (Higgins et al. 2001). They are endemic to the southwest Pacific, centred in Australia and New Guinea and are the largest family of Australian birds (Schodde & Mason 1999). In the absence of representatives of similar groups (eg. Nectariniidae, Oriolidae, Sylviinae, Pycnonotidae and Promeropidae), the evolving meliphagids were able to effectively diversify and acquire the adaptive zones of such groups (Keast 1 Only very general features of honeyeater ecology are discussed here as the more relevant aspects of their ecology are described and reviewed in greater detail in the introductions for data chapters 4,5,6 & 7.

9 1976). They occupy arboreal niches throughout the continent (Schodde & Mason 1999). Honeyeaters are also very important plant pollinators in Australia and New Zealand (Higgins et al. 2001).

There is much variation in diet between honeyeater species and this probably relates to morphological differences between them. They are basically nectarivores or insectivores, while most species combine both foraging behaviours to varying degrees (Pyke 1980, Keast 1985a). Long- billed honeyeaters are more nectarivorous than short-billed which are more insectivorous (Ford & Paton 1977). Large honeyeaters supplement their energy requirements by eating insects and fruit (Brown et al. 1978) and take nectar and insects mainly from eucalypts and (Ramsey 1989, Mac Nally 1995, Franklin 1997). Smaller honeyeaters are often denied access to rich nectar sources by aggression from larger honeyeaters and are sometimes limited to poorer nectar sources (Collins & McNee 1991). Variation in diet probably also relates to energy and nutritional requirements at different stages of life history. Nectar and alternate carbohydrates (manna, honeydew and lerp) contain negligible protein, but provide better rates of energy gain than do insects, which are collected to satisfy protein requirements (Recher & Abbott 1970, Paton 1982).

Many honeyeater species are highly mobile and this is a likely response to a heterogeneous flowering landscape (Keast 1968, Mac Nally & McGoldrick 1997, Paton et al. 2004). Their movements are not related to climatic differences between regions (Keast 1968, McGoldrick & Mac Nally 1998), but are strongly related to seasonal variation in flowering (Mac Nally 1995, 1997). Their densities have been correlated with nectar abundance in forests and woodland (Ford & Paton 1976, Ford 1983, Collins & Newland 1986, McFarland 1986, Ramsey 1989), while no clear correlations were found in heathlands (Paton 1979, Pyke 1983a, 1983b, 1985, Pyke & Recher 1988, Pyke

10 et al. 1993). However, caution is needed when comparing studies as nectar has been quantified in different ways, the scale at which it was measured varied between studies and because different honeyeaters (i.e. species dependent or not dependent on nectar) were studied. Correlations that have been observed, may occur at some scales but not others (Franklin & Noske 1999). Some studies suggest that large honeyeaters forage selectively where nectar is richest and leave the area when foraging at nectar is no longer profitable (eg Collins 1985).

1.4 Some comparisons with other nectarivores. Nectarivores from other regions of the world are similar to honeyeaters in that they are also not exclusively nectarivorous. Hummingbirds (Feinsinger 1976, Feinsinger & Colwell 1978), sunbirds (Gill & Wolf 1975) and honeycreepers (Carpenter 1978) have been recorded taking insects and carbohydrates, other than nectar.

Most models of foraging behaviour for nectarivores have been developed for hummingbirds. However, there are differences in foraging behaviour between hummingbirds and honeyeaters. Many hummingbird species have bill shapes that are closely matched to particular flower types (Feinsinger & Colwell 1978) and show a closer affinity to particular flowering plant species (Feinsinger 1978, Snow & Snow 1980, Cotton 1998). In contrast, honeyeater bill shapes are more generalised and not specifically related to particular flower types (Keast 1985a). However, honeyeaters that are largely insectivorous and glean insects from surfaces have generally shorter bills than more nectarivorous honeyeaters which also take smaller flying insects than gleaning insectivores of similar size (Recher 1981, Wooller 1984).

Many hummingbird species are territorial (i.e. individuals restrict the use of resources by others within an area either through song or aggression) and

11 do not flock (i.e. move through the landscape in groups), whereas honeyeaters are less territorial, often with overlapping feeding areas, and do flock (Wolf 1970, Pyke 1980, Pyke et al. 1996).

Hummingbirds and honeyeaters differ in size, foraging tactics and tongue structure, yet they collect nectar at similar rates (Paton & Collins 1989). Honeyeaters perch while foraging for nectar while hummingbirds hover. Honeyeaters are on average heavier birds and forage at clustered flowers rich in nectar while hummingbirds often visit single flowers that are more spaced and need to move quickly between flowers (Pyke 1981). The broader brush tongues of honeyeaters allows them to exploit nectar thinly spread over the relatively large surface area of their host flowers and honeydew from leaf and bark surfaces, while the narrow tube-like tongues of hummingbirds may be better adapted to the tubular flowers which they visit (Paton & Collins 1989).

Foraging models that describe hummingbird behaviour may not apply to honeyeaters for other reasons. Nectar production rates have been found to be greater within honeyeater habitats than in those of hummingbirds, and nectar may be the limiting resource for hummingbirds but not for honeyeaters when nectar is not a major component of their diet (Craig & MacMillen 1985).

If the Noisy Friarbird Philemon corniculatus (Latham) or the Red Wattlebird Anthochaera carunculata (Shaw) are typical honeyeaters they will not be reliant on nectar and this will not be a limiting resource for them. If this is the case, then the foraging models that describe territorial nectarivores at limiting resources will not be applicable to them and competition may not account for the foraging niche of either species.

12 1.5 The basis for selection of species for this study. The Noisy Friarbird and Red Wattlebird are both common and highly mobile bird species with complex movements related to flowering (Pizzey & Doyle 1980). They are able to use a wide diversity of food resources (Lea & Gray 1936, Barker & Vestjens 1990, Lepschi 1993, 1997). Thus they may be considered to be nomadic generalists. As such, they occupy several trophic levels and are likely to have important roles as pollinators, seed dispersers and insect population suppressors. These factors suggest they are also more likely to persist within a fragmented landscape compared with sedentary specialists and may have an important role in maintaining remnant patch quality.

They are found throughout a wide range of woodland and forest types and will use remnant vegetation patches, areas of revegetation and both native and exotic plantings in rural and urban landscapes (Pizzey & Doyle 1980). They may be found throughout the entire range of remnant size classes from continuous forest, through narrow native vegetation corridors to isolated trees in open pastureland (pers. obs.).

They often share food resources, but can also be found independent of one another within the sympatric part of their range (Saunders 1993). Movement patterns and niche overlap between these two honeyeaters appear to be highly variable and are likely to be strongly influenced by unpredictable resources over both space and time. Movements and food resource utilisation may also differ between these two species, and if so, they are responding to the landscape in different ways.

Declines in habitat quality through tree loss and loss of understorey vegetation, further fragmentation and the existence of narrow corridors favours edge-specialists such as the Manorina melanocephala

13 (Major et al. 2001). The presence of Noisy Miners has been negatively correlated with diversity of many woodland species (Loyn 1987). In contrast, both Noisy Friarbirds and Red Wattlebirds appear to be able to coexist with Noisy Miners at shared food resources and are able to use patches occupied by them (pers. obs.). Hence the ecological role of Noisy Friarbirds and Red Wattlebirds may be more significant in the present landscape than historically. As other pollinators, seed dispersers and insectivores decrease in the landscape, Noisy Friarbirds and Red Wattlebirds may have increased their share of these roles within degraded remnants.

An examination of the foraging ecology of these two species may provide an understanding of how they cope within the fragmented landscape and the role they may play in maintaining habitat patch quality. A comparison between their foraging ecologies may reveal the conditions controlling variation in niche overlap and how competition between them is affected by unpredictable food resources. Comparison with other honeyeaters and other nomadic fauna may also provide further insights as to the role of highly mobile species in fragmented landscapes.

1.6 The general ecology of the Noisy Friarbird and the Red Wattlebird. Many foraging ecology studies of honeyeaters in heathlands demonstrated that flowering and nectar production followed seasonal patterns that were reasonably consistent between years (e.g. Pyke 1983a, Pyke & Recher 1988, Pyke et al. 1993). However, many honeyeaters occupy forests and woodlands which have less consistent flowering and nectar production, and which show large fluctuations in abundance of honeyeaters (Collins 1985, Craig & MacMillen 1985, Ford & Paton 1985, Pyke 1985, McFarland 1986). It has been suggested that many forest and woodland honeyeaters populations fluctuate in response to unreliable food resources (Keast 1968, Mac Nally & McGoldrick 1997). The Noisy Friarbird and Red Wattlebird are

14 two such honeyeaters.

The Noisy Friarbird is found in the moist eastern part of Australia from just south of Cape York to eastern , while the Red Wattlebird is found across southern Australia from the southwest corner of to the New South Wales/ border in the east (Figure 1). They are the two largest honeyeaters in New South Wales (107g and 125g respectively, Ford et al. 1986). They have similar habitat preference, social and foraging behaviour, and within New South Wales there is a high overlap in distribution (Figure 1). They are considered to be common throughout most of their range. Most of the data on either species are contained within broader studies of woodland or forest bird communities (e.g. Lamm & Wilson 1966, Marchant 1979, Loyn 1980, McFarland 1984, Ford 1985, Pyke 1985 Recher & Holmes 1985, Slater 1995, McFarland 1996, Traill et al. 1996, Egan et al. 1997, Mac Nally 1997, Er et al. 1998), while few papers specifically examine or compare either species (Saunders 1993, Ford & Debus 1994, Ford 1999, Ford & Tremont 2000).

Both honeyeaters are described as gregarious, conspicuous, pugnacious and strongly arboreal (Pizzey & Doyle 1980, Longmore 1991, Higgins et al. 2001). They are generally found in eucalypt forest and woodland (Lyon 1980, Paton 1980, Ford et al. 1985, Osborne & Green 1992, Saunders 1993), but can also be found in various other woodlands comprising , Acacia, Callitris, Melaleuca and Callistemon species (Recher & Abbot 1970, Smith 1980, Pyke & Recher 1988, Turner 1992).

The movements of both species are complex and poorly understood, with much of the published material based on incomplete information and different classifications i.e. migration versus nomadism (Higgins et al. 2001). It is thought that flowering phenology of their food plants is an important

15 factor affecting their movement (Keast 1968, Lyon 1980, McFarland & Ford 1991).

Nectar, taken mainly from eucalypts (banksias, grevilleas and epacrids to a lesser extent), is a principal food source (Ford & Paton 1977, Paton 1986, Egan 1997). Invertebrates (mainly arthropods), fruit, seed, lerp, manna, honeydew are also taken (Paton 1980, Gosper 1999, Rose 1999). The Noisy Friarbird has also been recorded preying on eggs and nestlings of other small birds (Smith 1980).

The Red Wattlebird forages singly, in pairs or small groups (Higgins et al. 2001) although foraging flocks, sometimes numbering in excess of one hundred individuals, have also been recorded (Baxter 1989). Noisy Friarbirds forage mainly in larger flocks (McFarland 1984) and have been observed moving through the landscape in small flocks of up to 30 birds (pers. obs). The two species often forage together and with other nectarivores, including other honeyeaters (Saunders 1993, McFarland 1996). They forage mostly in the canopy, although they will also forage in the shrub layer, take insects while sallying and will occasionally forage on the ground (Higgins et al. 2001). Nectar is taken while flower probing and invertebrates are taken while probing bark, gleaning foliage or sallying through the air within or just over the canopy. To a lesser extent, manna, honeydew and lerp are gleaned from foliage and bark (Higgins et al. 2001).

Little is known of the social behaviour and organisation of either species (Higgins et al. 2001). The Noisy Friarbird is known to establish long term pair bonds (Ford 1999), but much of the published material on behaviour of either species relates to aggression (e.g. Davis & Recher 1993, Leonard 1995, Geering & French 1998, Oliver 1998). Both are described as aggressive birds, with much of the aggression directed at conspecifics, smaller honeyeaters at

16 shared food sources, or potential predators (Ford & Debus 1994).

They are a significant component of forest and woodland bird communities as attested by past studies (Lyon 1980, Ford et al. 1985, Pyke 1985, Smith 1989, Er & Tidemann 1996, Traill et al. 1996, Egan 1997, Shelly 1998). Although aggressive behaviour is often described as characteristic of these two honeyeaters (Higgins et al. 2001), few studies have quantified this behaviour and explored its relationship with food resource abundance (e.g. McFarland 1986, Ford & Debus 1994). Because they are aggressive and are likely to have strong competitive interactions with other birds, more needs to be known about the behaviour of these two large honeyeaters.

Limited data have been collected on seasonality of food sources or their reliability over consecutive years. Few attempts have been made to correlate these honeyeaters’ occurrence with large scale variation in food resources ( e.g. Mac Nally & McGoldrick 1997). There have been few studies over large spatial and temporal scales that investigate their foraging behaviour or quantify their food resources (e.g. Mac Nally & McGoldrick 1997). Hence the study of the foraging behaviour of Noisy Friarbirds and Red Wattlebirds over a broad region and over several years would help elucidate relationships between regional patterns of food availability and patterns of abundance and distribution in animals, and the study of the affects of resource patterns on foraging niche and competition.

Different types of food resources are available seasonally over their range. It was hypothesised that many of these food sources are seasonally unreliable and that they are unpredictable throughout their range. This may account for their little understood and complex movements. As a consequence, a small-scale ‘snapshot’ study was not perceived as sufficient in the study of the comparative ecology of these two honeyeaters. Patterns at

17 a larger regional scale and over several seasons and years were considered important. Food sources were identified and the abundance of these two honeyeaters was related to changing resource patterns at several sites from the coast, over the tablelands and onto the slopes in the middle of their sympatric range. This was done over two years to see if yearly patterns were consistent. Time budgets were collected for both species and the relationships between behaviour and site, season, and foods chosen were explored. Differences in behaviour of these two honeyeaters was accounted for, based on the morphological differences between them.

1.7 The questions for this study. The general question was to consider how regional patterns of food availability affect the patterns of abundance and distribution of birds, and how patterns of food availability affect niche overlap and competition between species.

The specific questions to be addressed in the data chapters are: 1. What food sources are Noisy Friarbirds and Red Wattlebirds utilising? 2. Is their occurrence and that of their food sources predictable over space and time? 3. Do Noisy Friarbirds and Red Wattlebirds respond to the landscape in the same way? 4. What degree of niche overlap exists between Noisy Friarbirds and Red Wattlebirds and what does this imply about competition between them? 5. Are Noisy Friarbirds and Red Wattlebirds really nomadic generalists? 6. What are the likely ecological roles of Noisy Friarbirds and Red Wattlebirds in the fragmented landscape?

The final discussion will be a synthesis of the findings in relation to these

18 questions and will compare the observed patterns of Noisy Friarbirds and Red Wattlebirds with those of other fauna relying on food resources with similar distribution patterns. It will also consider whether unpredictable food availability favours nomadic generalists over other foraging strategies in a fragmented landscape and the appropriate temporal and spatial scales for the study of nomadic generalists relying on unpredictable food resources.

The different aspects of their ecology are presented here as discrete studies since the contents of chapters four, six and seven have been published as papers in refereed journals. In each chapter’s introduction, the current status of research in each area is discussed. In the following chapter, chapter two, the location, topography and climate at each site is described. Floristics and vegetation structure are described in chapter three. The final chapter, chapter eight, provides a synthesis of the findings of each chapter and further examines the role of nomadic generalists in a fragmented landscape with unpredictable resources.

2. LOCATION, TOPOGRAPHY AND CLIMATE OF STUDY SITES

19 Noisy Friarbirds are distributed down the east coast of Australia from north Queensland to southeast Victoria, while Red Wattlebirds are distributed across southern Australia from northern New South Wales to southwestern Western Australia (Figure 2.1). Their distributions overlap in southeastern Australia taking in the coastal, tableland and inland slope regions. The centre of overlap in their distributions is in central eastern New South Wales.

Movement of Noisy Friarbirds and Red Wattlebirds between coastal, tableland and inland slope regions is more significant than latitudinal movements in New South Wales (Saunders 1993). Hence sites were selected to examine patterns of occurrence and food resource use across these areas. Sites were also selected that occur within the core of the sympatric range of these birds to facilitate comparisons between the species’ response to food resource patterns. To satisfy this criterion, sites were selected along an east- west transect from the Central Coast to the Central Western Slopes (Figure 2.1, Table 2.1).

The three topographic regions (coast, tablelands and western slopes) are distinguished by their climate (Table 2.1 and Figure 2.2, historical data, Climate Services, Bureau of Meteorology). The Central Coast receives most rainfall from the east and temperatures are moderate compared with the other two regions. The Central Tablelands experience higher rainfall due to an orographic effect. Temperatures are generally lower when compared with the other two regions. The Central Western Slopes receive rain from both east and west, but there is a strong rain shadow effect from the Central Tablelands and as a result the Slopes receive lower rainfall than the other two regions. Temperatures are more extreme compared with the other regions, although minimum temperatures are not as low as on the Central

20 DUBBO

Munghorn Gap N.R. Goobang N.P. Clandulla Back Yamma S.F. S.F. King’s Douglas Park COWRA Tableland SYDNEY Royal TABLELANDS N.P. WESTERN SLOPES COAST

0 500 1000m km

Clandulla Kings S.F. Tableland Munghorn Goobang Gap N.R. N.P. 500m

Back Yamma S.F. 0 100 200 km Douglas Park Royal N.P. W Central Western Slopes Central Tablelands Central E Coast Figure 2.1 Location map and cross-section of sites, topographical regions and range of Anthochaera carunculata (vertical strips), and Philemon corniculatus (horizontal stripes). (N.P. = National Park, N.R. = Nature Reserve, S.F. = State Forest.)

21 Table 2.1 Location, altitude, climate and transect lengths at study sites (a = historical data, b = 1992 to 1993, except for Goobang National Park which was sampled from 1993 to 1996, N.P. = national park, N.R. = nature reserve, S.F. = state forest).

Region Central Western Slopes Central Tablelands Central Coast

Site Back Yamma Goobang Munghorn Clandulla King’s Douglas Royal S.F. N.P. Gap N.R. S.F. Tableland Park N.P.

Latitude 33˚19’s 32˚49’s 32˚24’s 32˚54’s 33˚49’s 34˚12’s 34˚04’s

Longitude 148˚14’e 148˚21’e 149˚50’e 149˚55’e 150˚25’e 150˚42’e 151˚06’e

Altitude (m) 340 500 600 720 670 130 50

Mean Rainfall a 527.9 560.5 668.5 655.6 1401 856.4 1103 (mm) b 700.2 619.3 718.0 672.4 1147 836.0 919.4

Mean Max. Daily a 23.8 24.3 23.1 22.3 16.5 23.6 22.1 Temp. (˚C) b 23.2 23.6 22.4 21.8 16.4 22.4 22.1

Mean Min. Daily a 10.0 11.7 8.3 8.8 7.9 10.0 13.1 Temp. (˚C) b 9.8 11.5 8.8 8.5 7.5 9.8 13.9

Transect T1 600 2000 2000 2000 2000 1250 540 Length (m) T2 700 2000 2000 2000 2000 625 630 3000 30

20

2000 10 Daily Precipitation Temperature (mm) 0 (°C)

1000 -10 23

-20

0 -30 Back Yamma Goobang Munghorn Clandulla Kings Douglas Royal S.F. N.P. Gap N.R. S.F. Tableland Park N.P.

Rainfall - Mean All Years (mm) Mean Max. Daily Temperature (°C)

Rainfall - Mean Study Period (mm) Mean Min. Daily Temperature (°C)

Figure 2.2 Mean precipitation and mean minimum and maximum temperatures for study sites (historical data, Bureau of Meteorology) and mean precipitation for study period 1992 to 1993. (S.F. = state forest, N.P. = national park, N.R. = nature reserve) Tablelands.

The mean maximum and minimum daily temperatures for the study period were lower than means for all years. Rainfall means for the study period differed from means for all years at some sites (see Figure 2.2). The Royal National Park and Kings Tableland received less than average rainfall, while sites on the Central Western Slopes, particularly Back Yamma State Forest, received higher than average rainfall. Back Yamma State Forest experienced flooding during the last survey in spring 1993.

Initially six sites were selected for their accessibility, low level of human disturbance and the predicted occurrence of open forest, recognised as one of the preferred habitats of both honeyeaters (Saunders 1993). Two transects were established at each site with three plots along each transect as described in Section 3.1.

A site at Goobang National Park was added to the study in 1993, when it was recognised as a valuable site for testing the importance of nectar to these two honeyeaters, and data were collected here from 1993 to 1996. At all other sites data were collected from 1992 to 1993. In each year, sampling was undertaken during each season over a period of two to three weeks in December/January, April/May, June/July and September/October. Details of data collection are discussed in the methods of chapters three through six.

Much of the area from the Central Coast to the Central Western Slopes has a long history of forestry and agriculture, and finding sites that satisfied the above criteria meant that small deviations from a direct east-west transect were necessary. All sites had some evidence of past disturbance. Douglas Park and Back Yamma State Forest both experience occasional grazing of understorey grasses. Clandulla State Forest has been

24 logged previously. All other sites are in Nature Reserves or National Parks. The Royal National Park, Clandulla State Forest and Goobang National Park all contained evidence of recent fires. Sparse undergrowth and charring on tree trunks was evident at the Royal National Park and Goobang National Park, while epicormic regrowth in Clandulla State Forest related to a fire in the summer of the year prior to this study.

The three topographic regions (coast, tablelands and western slopes) are distinguished by their climate. Rainfall and temperatures for each site are shown in Table 2.1 and Figure 2.2 (historical data, Climate Services, Bureau of Meteorology). The Central Coast receives most rainfall from the east and temperatures are moderate compared with the other two regions. The Central Tablelands experience higher rainfall due to an orographic effect. Temperatures are generally lower when compared with the other two regions. The Central Western Slopes receive rain from both east and west, but there is a strong rain shadow effect from the Central Tablelands and as a result the Slopes receive lower rainfall than the other two regions. Temperatures are more extreme compared with the other regions, although minimum temperatures are not as low as on the Central Tablelands.

The mean maximum and minimum daily temperatures for the study period were lower than means for all years. Rainfall means for the study period differed from means for all years at some sites (see Figure 2.2). The Royal National Park and Kings Tableland sites received less than average rainfall, while sites on the Central Western Slopes, particularly Back Yamma State Forest, received higher than average rainfall. Back Yamma State Forest experienced flooding during the last survey in spring 1993.

25 3. FLORISTICS AND VEGETATION STRUCTURE AT SITES

3.1 Methods Two 2000 metre long transects were established along tracks at each of the seven sites, unless shorter transects were necessary because of patch size and the need to avoid ecotones between habitat types (see Table 2.1). In general, all trees along a transect within a 20 m wide band, with a diameter at breast height of ≥ 30 cm, and at least eight metres high, were counted and identified. This ensured that only mature trees were sampled and avoided bias towards species that produced prolific saplings. At Royal National Park and Back Yamma State Forest, where the transects were restricted in length, the band width was increased to 50 m. These initial tree counts were then used to calculate the total tree density per hectare and the tree species composition along each transect.

Where necessary for identification, samples of leaves, fruit, and flowers were collected, along with description of bark for each tree species. All trees were identified to species (Costermans 1981, Harden 1990, Tame 1992, Brooker & Kleinig 1999).

To determine tree heights and canopy cover, three plots, each 10 m x 50 m, were established with their long axis perpendicular to the transect at points 20%, 50% and 80% of the transect length from the start of each transect. Plots were placed in order to spread them along the transect and to avoid ecotones near the ends of transects at some sites. In each plot, nine trees were selected along the mid line of each plot as these were the same trees selected for estimating the canopy cover (see below). Tree heights were estimated visually, to the nearest whole metre, using a five metre long pole marked in alternating white and red one metre long sections placed at the base of the tree which was sighted through a transparent ruler hand held at

26 20 metres from the base of the tree to estimate the height of the portion of tree above the pole. Means of tree height for each plot were tested for significant differences using a three factor nested ANOVA for plots nested within transect and transect nested within site. Crown diameters were measured for each of the nine trees and its nearest neighbour, along with the crown separation between the trees in each pair. Measurements, to the nearest metre, were made with a tape measure from points on the ground immediately beneath the tree edge. If a tree crown deviated from circular then measures were made across the shortest and the longest crown width. The average width, to the nearest metre, was then used as the crown diameter. The crown cover for each plot was estimated by the method described by Walker and Hopkins (1990). Average crown cover for each transect was determined from plots and used to classify the structure of the forest at each transect (cf. Specht et al. 1974).

Basic soil types at each site were classified as clays, loams or sandy soils and underlying rock type was also recorded for each site.

Species richness (S), diversity (H) and evenness (J) were calculated for each transect. Species richness equals the number of species recorded along each transect. Diversity was calculated as H=∑pi(log pi) where pi equals the proportion of the total sample belonging to the ith species, and evenness was calculated as J=H/log S (Krebs 1989). Transects were also compared via correspondence analysis using tree species densities for each transect.

27 3.2 Results

3.2.1 Forest Structure Forest canopy structure for transects at each site are shown in Table 3.1. Percent crown cover varied from 48 to 94% between transects, although all but three transects had between 30 and 70% crown cover and thus were tall open forest (cf. Specht et al. 1974). Both transects at Royal National Park and one of the transects at Kings Tableland were tall closed forest.

Tree density is lowest at Coast sites and greatest on Central Tableland sites and then decreases onto the Western Slopes. There is a significant positive correlation between altitude and tree density for sites ( r=0.969, d.f. =5, P<0.01, Figure 3.1).

The mean tree heights for each plot are shown in Figure 3.2. Differences between plots within a site vary between sites; appearing low at Goobang National Park and high at Royal National Park, with no discernible pattern along an east-west transect through sites. The analysis of tree heights indicated that there were highly significant differences between plots within transects within sites, but no significant difference between transects within sites nor between sites (Table 3.2). There was also considerable variation between replicates within plots as attested by the estimated variance component (see Table 3.2).

The understorey structure varied considerably within transects, between transects and between sites. In general, at Royal National Park, Douglas Park and Back Yamma State Forest there was a very sparse shrub layer and ground cover was dominated by forbs and grasses, although there were some thick patches of saplings. At Kings Tableland and Goobang National Park the shrub layer was well developed and frequently difficult to

28 Table 3.1 Forest canopy structure and floristic associations of transects at each site. N.P. = National Park, N.R. = Nature Reserve, S.F. = State Forest, TOF = tall open forest, TCF = tall closed forest, S = species richness, H = diversity index, J = evenness index, canopy associations list first 3 most common tree species with % composition along transect shown in brackets (4 where 2 species have equal composition), genera are Al = Allocasuarina, An = , B = Banksia, C = Callitris and E = Eucalyptus.

Site Transect Structure Trees/ha Crown Canopy Associations S H J Cover(%) (%)

Royal 1 TCF 45 78 Al. casuarina(25), An. costata(23), E. punctata(17) 7 0.79 0.93 N.P. 2 TCF 63 94 E. punctata(27), An. costata(24), B. serrata(24) 9 0.77 0.81

Douglas 1 TOF 52 38 E. fibrosa(49), E. crebra(15), E. globoidea(14) 9 0.67 0.70 Park 2 TOF 42 48 E. fibrosa(55), E. moluccana(12), E. globoidea(12) 7 0.61 0.72

Kings 1 TCF 137 73 E. piperita(37), E. oblonga(23), E. gummifera(20) 7 0.67 0.79 Tableland 2 TOF 144 55 E. piperita(47), E. oblonga(25), E. gummifera(10), E. sieberi(10) 6 0.60 0.77

Clandulla 1 TOF 144 56 E. oblonga(63), E. punctata(19), E. rossii(12) 7 0.50 0.59 S.F. 2 TOF 133 61 E. oblonga(37), E. crebra(23), E. rossii(20), E. punctata(20) 4 0.59 0.98

Munghorn 1 TOF 131 48 C. endlicheri(30), E. punctata(21), E. rossii(15) 8 0.76 0.84 Gap N.R. 2 TOF 133 56 E. blakelyi(68), E. melliodora(9), An. floribunda(8) 9 0.56 0.59

Goobang 1 TOF 100 51 E. sideroxylon(53), E. blakelyi(20), E. fibrosa(13) 8 0.60 0.66 N.P. 2 TOF 92 67 E. sideroxylon(51), E. fibrosa(24), E. blakelyi(12) 6 0.58 0.75

Back 1 TOF 90 54 E. sideroxylon(57), E. microcarpa(17), C. glaucophylla(13) 7 0.59 0.70 Yamma S.F.2 TOF 56 66 C. glaucophylla(54), E. albens(23), E. microcarpa(15) 4 0.51 0.85 20

15 Height

10 Tree Mean

31 5

0 Back Goobang Munghorn Clandulla Kings Douglas Royal Yamma S.F. N.P. Gap N.R. S.F. Tableland Park N.P.

Figure 3.2 Mean tree heights ± S.E. for each plot (n = 9, N.P. = national park, N.R. = nature reserve, S.F. = state forest). Table 3.2 Nested ANOVA of tree heights for plots nested within transects, nested within sites (number of levels witnin each factor shown after each source of variation).

Source of Variation Numerator Denominator Mean Square F-value Prob(>F) Estimated Variance d.f. d.f. Component

Site (7) 6 7 88.709 2.294 >0.05 0.9266 Transect(Site) (2) 7 28 38.675 0.942 >0.05 -0.0890 Plot(Transect(Site)) (3) 28 336 41.077 6.012 <0.001 3.8048 Residuals (9) 336 6.833 6.8333 penetrate, while at Munghorn Gap Nature Reserve and Clandulla State Forest the shrub layer was more variable and there were often open areas with thick leaf litter.

Soils at most sites were generally poor and shallow. Soils were sandy at Royal National Park, Douglas Park, Kings Tableland, Munghorn Gap Nature Reserve, Clandulla State Forest, and Goobang National Park. The underlying rock type at the first four sites was sandstone while at the latter two sites it was quartzite. The soils at Back Yamma State Forest were clay loams based on hornfels, although there were some differences between transects. Transect 1 was on a stoney hornfels ridge with much quartz in a shallow clay soil, while transect 2 was on the lower flats with deep clay loam.

3.2.2 Floristic Associations The associations of tree species for each transect, along with species richness, diversity and evenness indices are shown in Table 3.1. Eucalypt species dominate at most sites. However, at Royal National Park the genera Angophora, Allocasuarina and Banksia make up a large portion of the canopy and at sites on the Western Slopes Callitris can be the dominant genus (see also Figure 3.3).

A closer inspection of eucalypt bark groups reveals that boxes and ironbarks dominate at Douglas Park and at sites on the Western Slopes. Other eucalypts such as Peppermints, Ashes, Stringybarks and Bloodwoods appear to dominate at Tableland sites, while Gums have no particular association with region (Figure 3.2).

When the eucalypts are divided into subgenera (Figure 3.4) some patterns emerge between subgenus and region. The subgenus Eucalyptus dominates sites on the Tablelands and was absent from most Western Slope sites,

33 100 80 60 Gums 40 20 0 100 80 60 Ironbarks 40 20 0 100 80 60 Boxes 40 20

Composition 0

% 100 80 60 Other Eucalypts 40 20 0 100 80 60 Non-Eucalypts 40 20 0 BSF GNP MG CSF KT DP RNP

Figure 3.3 Percent tree composition for tree groups at each site. Other eucalypts include Peppermints, Ashes, Stringybarks and Bloodwoods. Non-eucalypt species include the genera Acacia, Angophora, Banksia, Callitris, and Casuarina. (BSF = Back Yamma State Forest, GNP = Goobang National Park, MG = Munghorn Gap Nature Reserve, CSF = Clandulla State Forest, KT = Kings Tableland, DP = Douglas Park, and RNP = Royal National Park).

34 800

600

400 Altitude

200

0 0 50 100 150 Tree Density (Trees/ha)

Figure 3.1 Correlation between altitude and tree density across sites (n=7, d.f.=5, r=0.969, P<0.01)

30 100

75

50 Corymbia

25

0

100

75

50 Symphyomyrtus

25

0

100

75

50 Eucalyptus

25

0 BSF GNP MG CSF KT DP RNP

Figure 3.4 Percent of eucalypt trees in each subgenus at each site. (BSF = Back Yamma State Forest, GNP = Goobang National Park, MG = Munghorn Gap Nature Reserve, CSF = Clandulla State Forest, KT = Kings Tableland, DP = Douglas Park, RNP = Royal National Park).

35 except for Munghorn Gap Nature Reserve. Munghorn Gap is placed in the Central Tablelands in some classifications (e.g. Morris et al. 1981) while on the Western Slopes in other classifications (e.g. Harden 1990). It occurs on a western spur of the and is perhaps best considered to be an intermediate site between the two regions. The subgenus Eucalyptus is also less common at Douglas Park, making this site more like a Western Slopes site. The similarity of this site to those on the Western Slopes is also supported by the relatively high percentage of trees of the subgenus Symphyomyrtus at this site and on Western Slope sites. The subgenus Corymbia is only represented by one species and occurs at relatively low percentages at Kings Tableland on the Central Tablelands and at The Royal National Park on the Central Coast.

The species richness and canopy associations for each transect reveal that there is variation for tree species within a site as well as between sites (Table 3.1). There was a high overlap of tree species composition between transects within each site, except at Munghorn Gap Nature Reserve and Back Yamma State Forest. Altitude and local topography differed between transects at these two sites, but were very similar between transects at other sites. Diversity index (H) does not differ dramatically between transects within sites (range 0.02 to 0.09), except for transects at Munghorn Gap Nature Reserve (0.20). Difference between transect species’ evenness (J) are greater for Clandulla State Forest (0.39) and Munghorn Gap Nature Reserve (0.25) than between transects at all other sites (range 0.02 to 0.15). At Clandulla State Forest Eucalyptus oblonga is twice as abundant on one transect compared with the other and at Munghorn Gap Nature Reserve Callitris glaucophylla is three times more abundant on one transect compared with the other.

Correspondence analysis between tree species’ counts and transects

36 examines similarity between transects across all sites based on tree species composition (Figure 3.5). Constructed ellipses show three clusters and group transects according to region. Douglas Park could possibly be linked with the Western Slopes cluster as it contains both box and ironbark eucalypts. Eucalyptus moluccana and , both called Grey Box, are difficult to separate except by range. A mixed stand of both species was found in a patch at Munghorn Gap Nature Reserve and this was only noticed when one of the species flowered (pers. obs.). Eucalyptus fibrosa was found at Douglas Park and Goobang National Park. The presence of the stringybark Eucalyptus globoidea separates the Douglas Park site from Western Slopes sites.

One of the transects from Munghorn Gap Nature Reserve is in the cluster with Tableland sites while the other is found with Western Slopes sites. The tree species composition along transect one, along a spur of the range, was very similar to the transects in Clandulla State Forest, while that along transect two, on a valley floor, was more similar to sites on the Western Slopes.

The number of tree species recorded from plots, transects and sites are shown in Figure 3.6 along with the cumulative number of tree species for plots within each transect. At most sites, the tree species composition varies between plots along a transect, while there is less variation between transects within a site. Plot data underestimate species richness at each site, while transect data better estimate species richness at each site, except at Munghorn Gap Nature Reserve where there were greater differences in tree species between transects.

37 2 Callitris columellaris Eucalyptus microcarpa BSF T2

Eucalyptus blakelyi Eucalyptus sideroxylon BSF T1 Eucalyptus melliodora Eucalyptus moluccana GNP T1 1 GNP T2 MG T2 DP T2 Eucalyptus fibrosa WESTERN SLOPES Eucalyptus globoidea DP T1 DOUGLAS PARK

Angophora floribunda 0 Callitris endlicheri MG T1 Eucalyptus sclerophylla Eucalyptus rossii CSF T2 Eucalyptus punctata CSF T1 Eucalyptus oblonga RNP T2 & RNP T1 -1 TABLELANDS & KTT2 Angophora costata ROYAL NATIONAL KT T1 PARK Allocasuarina littoralis Eucalyptus sieberi Eucalyptus haemasatoma Eucalyptus piperita Eucalyptus gummifera

-2

-2 -1 0 1 2

Figure 3.5 Correspondence analysis between tree species’ densities (boxes) and transects (crosses). (BSF = Back Yamma State Forest, CSF = Clandulla State Forest, DP = Douglas Park, GNP = Goobang National Park, KT = Kings Tableland, MG = Munghorn Gap Nature Reserve, RNP = Royal National Park, T1 = transect 1, T2 = transect 2, ellipses constructed to show clusters about regions)

38 Royal N.P. Douglas Park 10

8

6 3 4 3 2 2 4 3 3 2 4 4 4 2 3

0

Kings Tableland Clandulla S.F. 10

8

6 4 4 4 5 4 3 3 3 3 4 4 2 Species 2 2 0

Tree Munghorn Gap N.R. Goobang N.P. 10 of 8

No. 6 3 4 5 3 3 3 3 3 4 4 2 2 2 2 0 P11 P12 P13 T1 P21 P22 P23 T2 SITE Back Yamma S.F. 10

8

6

4 4 2 2 3 3 3 2 0 P11 P12 P13 T1 P21 P22 P23 T2 SITE Figure 3.6 Number of tree species at plots, transects and sites. Cumulative number of tree species shown for plots and the number of tree species in each plot is shown below each plot. Number of tree species along each transect are shown after plots for each transect. Number of tree species at each site is based on cumulative numbers for transects at each site. For example, at Goobang N.P. the plots along transect 1 each had 2,4 and 3 tree species respectively while the cumulative number of species rose from 2 to 4 to 6 for plots. When all tree species along the transect were counted independently of plots, 8 species were recorded. Only 6 species were recorded along transect 2, but this did not change the total number of species for the site. (N.P. = national park, N.R. = nature reserve, S.F. = state forest).

39 4. THE IMPORTANCE OF EUCALYPT NECTAR IN THE DIETS OF LARGE HONEYEATERS

4.1 Introduction

Several studies have revealed a positive correlation between honeyeater numbers and flower or nectar abundance over both spatial and temporal scales (eg. Ford 1983, Collins, Briffa & Newland 1984, McFarland 1985, Newland & Wooller 1985), with the correlation being stronger for larger honeyeaters in some studies (e.g. Collins & Newland 1986)2. In studies comparing patches within a site at any one time, positive correlations were found between nectar availability and honeyeater numbers (Collins 1985, Ford & Paton 1985, Collins & Newland 1986, Paton 1986). Over time, honeyeater numbers diminished as nectar became scarce in patches that were initially nectar rich (Collins 1985). Patchy and asynchronous flowering of eucalypts is thought to drive nomadism in many honeyeaters (McGoldrick & Mac Nally 1998) and dispersal patterns are in the order of hundreds, rather than the thousands of kilometres for pronounced seasonal migrants (Mac Nally 1996). Thus nectar probably exerts considerable force on the dynamics of honeyeater communities.

In open forest, honeyeater density was positively correlated with nectar both seasonally and spatially, however, these correlations were weak (Pyke 1985). In heathlands no correlation was found between honeyeaters and nectar (Pyke 1983b, Pyke & Recher 1988, Armstrong 1992, Pyke et al. 1993), although these studies involved resident honeyeaters and correlations would not be expected.

2 There has been sufficient evidence to support this relationship, despite nectar supply having been measured in different ways in different studies. This often reflects the scale of the study i.e. what is possible at a patch or plot scale would be difficult at a regional scale. In addition, this suggests that the relationship may hold true at many different scales.

40 Honeyeaters are not exclusively nectarivorous and other foods may affect their densities over both spatial and temporal scales. Their diet includes insects and fruit (Brown et al. 1978). A seasonal shift in diet has also been observed when nectar is in short supply (Craig & MacMillen 1985). Alternative carbohydrates such as lerp, manna and honeydew become important components of the diet when nectar is scarce (Paton 1980, Ford & Paton 1985). Invertebrates are relatively more common dietary components during warmer months than at other times of the year (Collins & Newland 1986), and are important as a source of protein which nectar does not provide (Paton 1982).

In many studies, the lack of correlation between honeyeater numbers and nectar supply could also relate to other factors. Inappropriate scales of measurement, over-abundance of nectar locally or over a broad area and other social behavioural factors may contribute (Franklin & Noske 1999). Mac Nally and McGoldrick (1997) argued that it is difficult to interpret the dynamics of honeyeater communities as scale affects most studies. Insufficient data could explain the lack of correlation in some cases.

Large honeyeaters utilise nectar mainly from Eucalyptus and Banksia species (Keast 1968, Ford & Paton 1976, 1977, Franklin 1997). Paton and Ford (1977) suggested that this was because the flowers in these two genera are clumped, which may lead to improved foraging efficiency.

Several honeyeater species may share abundant nectar sources with little interspecific aggression (Keast 1968, Rasch & Craig 1988, Armstrong 1991). However, when nectar is less abundant, aggression may occur. Larger honeyeaters dominate such aggressive interactions (Ford 1979, McFarland 1986, Collins & Paton 1989, Ford & Debus 1994).

41 Both P. corniculatus and A. carunculata typify these patterns and have varied diets that consist mainly of nectar and arthropods but include some fruit and seeds (Blakers et al. 1984, Longmore 1991).

Nectar, when available, is an important food for both P. corniculatus and A. carunculata. In some studies, when an abundant nectar supply from Eucalyptus and/or Banksia species was available, one or both of these large honeyeaters were present at sites within their range (Keast 1968, Ford 1979, 1983, Newland & Wooller 1985, McFarland 1986). Both species are described as ‘blossom nomads’ whose appearance at many sites is tied to the flowering of appropriate nectar sources (Keast 1968). Between 36% and 65% of foraging observations for these species were spent at nectar sources (Pyke 1980, Recher & Holmes 1985). Ford et al. (1986) found that P. corniculatus and A. carunculata spent 53% and 47% of their foraging time at nectar respectively. Both species will exploit and defend rich patches of nectar (Ford 1981, Ford & Paton 1982, Collins 1985, Ford & Debus 1994).

Alternative carbohydrates have been recognised as important for A. carunculata when nectar is not available (Paton 1980). They will forage at manna, lerp and honeydew in the field (Recher & Holmes 1985) and have been shown experimentally to take lerp (Woinarski et al. 1989).

Few studies (e.g. Mac Nally & McGoldrick 1997, McGoldrick & Mac Nally 1998) have investigated honeyeater dynamics over broad spatial scales in forests. Most have attempted to correlate honeyeaters with their nectar sources at small spatial and temporal scales and mostly in heathlands. In this chapter correlations are examined at larger spatial and temporal scales in eucalypt forests over two years and at seven sites along a 350 kilometre east-west transect. In particular, it was investigated whether there were correlations between honeyeater density and flowering tree density at local

42 and regional scales, and whether these densities correlated with the proportions of time spent foraging at flowering trees and in aggression.

4.2 Methods 4.2.1 Study sites Two sites were selected in three regions within the sympatric range of P. corniculatus and A. carunculata in central eastern New South Wales (see Figure 2.1). Sites were selected in open eucalypt forest as this habitat is frequented by both species (Saunders1993).

Goobang National Park, on the western slopes, was selected in order to investigate correlation between flowering trees and large honeyeaters at a smaller spatial scale in Eucalyptus sideroxylon forest. Eucalyptus sideroxylon was an important source of nectar for honeyeaters in winter and spring on the western slopes. Although E. sideroxylon was present at Back Yamma State Forest, the patch was too small for setting-up independent, replicated plots.3

Within each site, two transects were established. These were along tracks and fire trails because some sites had dense leaf litter or a dense shrub layer. At such sites the noise created while conducting trial counts along randomly placed transects through the bush resulted in displacement of birds and reduced detectability, and hence an under-estimation of bird density. At other sites bird counts were not so affected. Off-track transects were considered to introduce unequal bias between sites. Sites were to be compared and relative densities were considered important. Using tracks also meant that a steady pace could be maintained and birds could be detected by call as far as 80 m from the transect line. Hanowski and Niemi (1995)4 observed that counts of bird species that forage in canopy that is 3 Transects counts were undertaken at each site, including Goobang National Park. The smaller scale study of plots was only undertaken at Goobang National Park. 4 This study was conducted in northern hemisphere forests.

43 continuous over tracks, did not differ significantly between transects placed on tracks and off tracks. Hence the use of tracks is unlikely to affect counts as both P. corniculatus and A. carunculata are mainly canopy foragers (Higgins et al. 2001). The availability of tracks at most of the sites was limited and the distribution of both honeyeaters appeared patchy, hence long transects were established to compensate. In a simulation of transect counts of randomly distributed birds, Engle-Wilson et al. (1981) found that more accurate estimates were achieved for longer transects.

4.2.2 Trees counts and species composition In general, all trees along a transect within a 20 m wide band, with a diameter at breast height of ≥ 30 cm, and at least eight metres high, were counted and identified. This ensured that only mature trees were sampled and avoided bias towards species that produced prolific saplings. Because of small patch size and the need to avoid ecotones between habitat types, shorter transects were necessary at some sites (see Table 2.1). At Royal National Park and Back Yamma State Forest, where the transects were restricted in length, the band width was increased to 50 m. These initial tree counts were then used to calculate the tree species density per hectare.

During each visit, the number of trees with flowers were counted for each species. These counts were made at the same transect width as the initial tree species composition surveys. In addition, in spring 1994 at Goobang National Park, counts were made along 41 fixed width transects 500 m long by 100 m wide, each separated by 50 m. The density of flowering trees was determined for each visit. Only those flowering tree species that were flower-probed by either species of honeyeater during the study are included in the analyses, and henceforth, counts of flowering trees refer only to those species.

44 4.2.3 Birds counts Data were collected at all sites in each year (1992 and 1993) during mid- season over a period of two to three weeks in January, April, July and October. Bird counts were started approximately one hour after sunrise, during still and sunny weather. Birds, when seen or heard, were recorded in each ten-metre band either side of the transect up to the 70-80 metre band width. Only counts up to fifty metres from the transect were used in the analysis as few data were collected beyond this range. In spring 1994 at Goobang National Park, counts of A. carunculata and P. corniculatus were made along the 41 fixed width transects.

Time budgets for individual honeyeaters were collected as follows: When a bird was located it was observed through binoculars. If the bird’s behaviour changed when it became aware of the recorder’s presence (i.e. it stopped feeding, preening or calling and watched the researcher) observations were not recorded until the bird resumed its previous behaviour or started another activity. Every 10 seconds, the behaviour that occupied the majority of the time unit was recorded. Observations were terminated after 5 minutes, or when the bird was lost from sight. Hence up to 30 observations may have been recorded for an individual. When both species were present, observations alternated between species. Generally, it was possible to keep track of the birds in a patch, so there was little risk of reselecting an individual.

The two behaviours of concern here are aggression and flower-probing. Aggression is defined as an approach of a bird which disrupts the behaviour of the bird under observation or when the observed bird disrupts another bird’s behaviour. No attempt was made to distinguish types of disruptive behaviour. The direction of aggression was not considered here (see Figure 5.7 for data pooled and appendices 10.14 through to 10.18 for a breakdown

45 by site). Flower-probing was used to describe the time spent at flowers. Birds at flowers may be taking nectar, insects or both, but the actual food is difficult to identify (McFarland 1984).

4.2.4 Analysis of bird count data Count data are Poisson distributed rather than normally distributed, hence Generalised Linear Models were used to model bird counts with a Poisson distribution specified for the error term (McCullagh & Nelder 1989). Estimation of the regression coefficients was by maximum likelihood, using iteratively reweighted least-squares. All models were initially assessed by comparing the residual deviances of competing models and further validated by examining plots of regression diagnostics.

Counts of honeyeaters from different locations were often made over differing transect lengths so an adjustment was made by including an offset in the Poisson regression models for the logarithm of the area surveyed (Venables & Ripley 1997).

The deviance from a Generalised Linear Model is a measure of the variability in the data. A large reduction in the deviance of a model, achieved by fitting an additional term, indicates that the term is important in explaining the variability of the bird counts (McCullagh & Nelder 1989).

Poisson variables often exhibit over-dispersion which is indicated when the residual deviance exceeds the number of degrees of freedom for the model. The variance of a Poisson distributed variable Y is Var (Y) = fm where m represents the mean of Y and the dispersion parameter f is fixed at 1 (i.e. a Poisson distribution assumes the variance equals the sample mean). Moderate over-dispersion can be accounted for by relaxing this constraint and estimating the value of f via quasi-likelihood estimation. This produces

46 more realistic estimates of the standard errors of the regression coefficients (McCullagh & Nelder 1989).

As P. corniculatus was absent from Royal National Park, this site was not incorporated in the analysis for this bird species. Similarly only a single A. carunculata was counted on only one survey at Douglas Park and at Kings Tableland, so these sites were not included in the analysis for this bird species. Counts were pooled over transects for each season within each site to investigate site, season, year and counts of flowering trees as explanatory variables for bird counts.

4.2.5 Analyses of time budget data Variables were tested for normality using a Shapiro-Wilk W Test (Shapiro et al. 1968). All correlations were calculated as Spearman Rank Correlations as only time spent in aggression was normally distributed.

Time budgets were collected on a per site visit basis. Transect counts were treated as sub-samples and were summed for correlations with other variables for each site visit5 . Across all sites there was the potential to collect time budget samples for each species from 64 surveys. However, one or both species were often absent from some sites and during some seasons, or in such low densities that collecting time budget data was impractical. Only 22 samples were collected for P. corniculatus and 35 samples for A. carunculata. The means of percent time spent in aggression during each survey for each honeyeater were compared using a student’s t-test.

5 Time budget data was collected on a per site basis so the count data from both transects at each site were summed before correlations with these other variables.

47 4.3 Results 4.3.1 Flowering trees The tree densities of the canopy species for each site are presented in Figure 4.1.6 The potential importance of sites for honeyeaters is reflected in the density of those tree species used as a nectar source. The three most frequently flower-probed coastal and tableland species were Eucalyptus gummifera, E. crebra and Banksia serrata, while on the western slopes E. sideroxylon, E. albens and E. melliodora were most frequently visited7 . On the western slopes, flower-probed species represent an average of 43% of the forests’ trees, while they averaged 20% of the forests’ trees at all other sites to the east. Goobang National Park and Back Yamma State Forest (both on the western slopes) have the highest densities of flower-probed trees and E. sideroxylon was a dominant species at these sites. Flower probing consumed most of the birds’ time on the western slopes sites, often more than 85% and as much as 100% of foraging time (details of time budgets will be considered in Chapter 5). Both flora and fauna of Munghorn Gap Nature Reserve have similarities with Clandulla State Forest and Goobang National Park, and this site is probably best considered as an intergrade between the tablelands and the western slopes, having properties of both (see Chapter 3.3.2).

4.3.2 Honeyeater counts at the regional scale Densities of P. corniculatus and A. carunculata varied considerably between sites, years and seasons (Figures 4.2, 4.3, Table 4.1). Their relationship to flowering trees also varied between sites. Flowering trees did not appear to determine the presence of either honeyeater at coastal and tableland sites, whereas they did appear to be important at sites on the western slopes.

6 Includes all trees i.e. both those fower-probed and those not flower-probed. 7 The percentage of flower-probing at flowering trees was E. gummifera 10%, E. crebra 72%, B. serrata 12% for coastal and tableland sites, and E. sideroxylon 62%, E. albens 22%, E. melliodora 11% at western slope sites.

48 150

E. sideroxylon

E. albens

E. melliodora

E. crebra

E. gummifera

E. agglomerata

B. serrata

100 Others

NFPT Trees/ha

50

0

Back Yamma Munghorn Gap King's Tableland Royal N.P. S.F. N.R. Goobang N.P. Clandulla S.F. Douglas Park

Figure 4.1 Density of trees at each site. (NFPT = trees not flower-probed by either P. corniculatus or A. carunculata. )

49 15 1 Royal N.P. 2.1 10 0.5 5

0 0 15 1 Douglas Park 10 0.5 5

0 0 15 1 Kings Tableland 10 0.5 5 Trees/ha 0 0 15 1

Clandulla S.F. Birds/ha 10

Flowering 0.5 5

0 0 15 1 Munghorn Gap N.R. 10 0.5 5

0 0 15 1

Back Yamma S.F. 57.6 16.1 10 0.5 5

0 0 Summer'92 Autumn'92 Winter'92 Spring'92 Summer'93 Autumn'93 Winter'93 Spring'93

Flowering Trees P. corniculatus A. carunculata

Figure 4.2 Seasonal densities of flowering trees, P. corniculatus and A. carunculata at sites from 1992 to 1993. (N.P.= national park, N.R.= nature reserve, S.F.= state forest)

50 80 2.5

Flowering Trees 2 P. corniculatus 60 A. carunculata 1.5 Trees/ha 40

1 Birds/ha 51 Flowering 20 0.5

0 0

Spring'93 Spring'94 Spring'95 Spring'96 Summer'93 Autumn'93 Winter'93 Summer'94 Autumn'94 Winter'94 Summer'95 Autumn'95 Winter'95 Summer'96 Autumn'96 Winter'96

Figure 4.3 Densities of flowering trees, P. corniculatus and A. carunculata for each season at Goobang National Park for 1993 to 1996. Table 4.1 Total counts of P. corniculatus and A. carunculata by site and season (years and transects pooled).

Site Summer Autumn Winter Spring

P. corniculatus

Royal National Park 0 0 0 0 Douglas park 0 0 0 1 Kings Tableland 24 0 0 0 Clandulla State Forest 32 2 0 28 Munghorn Gap Nature Reserve 56 17 18 36 Goobang National Park 0 0 34 136 Back Yamma State Forest 0 0 6 3

A. carunculata

Royal National Park 0 14 31 7 Douglas park 1 0 0 0 Kings Tableland 0 1 0 0 Clandulla State Forest 11 21 6 21 Munghorn Gap Nature Reserve 25 34 29 27 Goobang National Park 16 64 66 143 Back Yamma State Forest 0 2 24 1

52 For P. corniculatus coastal sites did not appear to be important, as this species was absent from Royal National Park and was seen in very low numbers at Douglas Park. Counts of P. corniculatus were higher on the tableland sites of Kings Tableland and Clandulla State Forest in spring and summer, but at these sites flowering trees were not important during these seasons. Goobang National Park and Back Yamma State Forest, on the western slopes, appeared to be more important for the winter to spring period, and P. corniculatus was only present when flowering trees were present. At Munghorn Gap P. corniculatus was not present in winter unless trees were flowering, but it was present in other seasons when trees were not flowering. In terms of patterns in counts of honeyeaters and flowering trees, Munghorn Gap exhibits characteristics of both tablelands and western slopes.

At the coastal site of Royal National Park counts of A. carunculata were high but these high counts did not coincide with the flowering periods of trees. At Clandulla State Forest there was little seasonal variation and no relationship with flowering trees. At sites on the western slopes A. carunculata often persisted all year-round even when no trees were in flower, but were more common when numbers of flowering trees increased.

For both species of honeyeater site, year, season, flowering trees and the site x year, site x season and year x season interactions were all highly significant in explaining the variation in bird counts (Table 4.2). For A. carunculata the site x year interaction could not be fitted to the model as there were insufficient degrees of freedom. The change in deviance was very large for the number of flowering trees (log-transformed) and indicates that flowering trees was still a very important explanatory variable after adjusting for temporal and spatial variability in honeyeater numbers.

53 Table 4.2 Analysis of deviance summary for counts of P. corniculatus and of A. carunculata for all sites, times and seasons. (Terms were added sequentially from first to last as a series of nested models, P =Pearson Chi-square probability.)

Term Change in d.f. Change in deviance Residual d.f. Residual deviance P(>c2)

P. corniculatus null 43 482.1 site 5 139.3 38 342.8 <0.001 year 1 8.3 37 334.4 <0.01 season 3 78.0 34 256.4 <0.001 log(flowering trees + 1) 1 23.2 33 233.2 <0.001 site x year 4 20.5 29 212.7 <0.001 site x season 15 151.2 14 61.6 <0.001 year x season 3 17.9 11 43.7 <0.001

A. carunculata null 47 545.1 site 4 58.8 43 486.4 <0.001 year 4 95.5 39 390.8 <0.001 season 3 96.3 36 294.5 <0.001 log(flowering trees + 1) 1 26.7 35 267.8 <0.001 site x season 12 110.8 23 157.0 <0.001 year x season 12 118.3 11 38.7 <0.001 4.3.3 Honeyeater counts at the local scale In spring 1994 at Goobang National Park bird counts were strongly correlated with counts of flowering trees (log scale) for both species of honeyeater (Figure 4.4). Poisson regression models of counts against log(flowering trees + 1) produced a better fit than models of flowering trees on an untransformed scale as indicated by the smaller residual deviance for the former model (Table 4.3). Philemon corniculatus was absent on plots without flowering trees and from many plots where the density of flowering trees was as high as 19 trees/plot, whereas A. carunculata was occasionally present when no trees were in flower. The maximum density of flowering trees on a plot without A. carunculata was seven trees/plot.

4.3.4 Correlations of bird behaviour with densities of flowering trees All correlations were significant except for the proportion of time spent foraging versus time spent in aggression for A. carunculata (Table 4.4). The correlations do not appear to produce clusters based on sites (Figure 4.5) and sample size per site was too small to do separate site analyses.

The means and standard errors of percent time spent in aggression during each survey by P. corniculatus (2.03±0.486) and A. carunculata (1.46±0.491) did not differ significantly (t=0.799, df=70, P>0.05).

4.4 Discussion

4.4.1 Correlations between flowering trees and honeyeaters Previous studies (Ford 1983, Collins, Briffa & Newland 1984, McFarland 1985, Newland & Wooller 1985, Pyke 1985) have reported a positive correlation between flowering and honeyeater density. In this study it was observed that the numbers of A. carunculata and P. corniculatus were strongly influenced by the density of flowering trees at both regional and local scales.

55 25

20

15 corniculatus

P. 10 of

Number 5

0

0 1 2 3 4 Ln(Number of flowering trees + 1)

20

15

10 A. carunculata

5 Number of

0

0 1 2 3 4 Ln(Number of flowering trees + 1)

Figure 4.4 Poisson regression models for counts of P. corniculatus (Y1) and A. carunculata (Y2) against flowering trees(X) at Goobang National Park in spring 1994 (n=41) with 95% confidence limits. Log (Y1) = -3.85(0.709) + 1.85(0.220)log[X + 1], f=2.01, P<0.001, and Log (Y2) = -0.93(0.359) + 0.95(0.127)log[X + 1], f=1.59, P<0.001, (Standard errors are as given in parentheses).

56 Table 4.3 Analysis of deviance summary for alternative models for honeyeater counts at Goobang National Park, spring 1994 (n=41).

Model d.f. Residual deviance P. corniculatus A. carunculata null 40 304.5 178.5 flowering trees 39 116.4 88.6 log(flowering trees + 1) 39 78.9 68.6

57 Table 4.4 Spearman Rank correlations between honeyeaters, flowering trees and behaviours as described in the text. ( *p<0.05, **P<0.01, ***P<0.001, n.s. = not significant)

Correlation A. carunculata P. corniculatus (n=35) (n=22)

Honeyeater density x Proportion of time in aggression 0.49 ** 0.55 * Proportion of time at flowers x Proportion of time in aggression 0.46 ** 0.70 ** Proportion of time foraging x Proportion of time in aggression -0.22 n.s. 0.57 ** Flowering tree density x Proportion of time at flowers 0.87 *** 0.69 ** A. carunculata P. corniculatus 0.12

0.09

0.06

Aggression 0.03

0 0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 HE/ha HE/ha

0.12

0.09

0.06

Aggression 0.03

0.00 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 Flower Probing Flower Probing

0.12

0.09

0.06

Aggression 0.03

0.00 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 Foraging Foraging

1.00

0.75

0.50

0.25 Flower Probing

0.00 0 10 20 30 40 50 60 0 10 20 30 40 50 60 FT/ha FT/ha

Figure 4.5 Correlations between honeyeaters, flowering trees and the behaviours described in the text. (HE/ha = honeyeaters per hectare, FT/ha = flowering trees per hectare, Aggression = proportion of time spent in aggression, Flower Probing = proportion of time spent probing flowers, Foraging = proportion of time spent foraging, = Royal National Park, = Clandulla State Forest, = Munghorn Gap Nature Reserve, + = Goobang National Park, = Back Yamma State Forest.)

59 High density of honeyeaters and trees in flower were observed on the western slopes during winter and spring. A similar pattern was noted by Mac Nally & McGoldrick (1997) where A. carunculata densities were greater on the northern slopes of the Great Divide in winter and spring when E. sideroxylon was in flower. However, both honeyeaters were also common on tableland sites in spring but they were not feeding at flowers. On the tablelands, P. corniculatus were common in spring and summer when few trees were in flower. At these times other foods such as insects, seed and fruit may be important dietary components (Chapter 5.3.2, Collins & Newland 1986).

Anthochaera carunculata has been observed to persist at sites during seasons when insects, nectar, seeds and fruit were scarce and are able to switch to alternate carbohydrates when other foods are depleted (Ford & Paton 1985). This pattern was observed at sites on the coast and tablelands where A. carunculata gleaned foliage in winter when P. corniculatus was absent. Numbers of P. corniculatus did not always match flowering tree density and the density of birds was often lower than expected. There are several possible explanations for the apparent lack of birds. The density of flowering trees may not be an accurate measure of the quality of the nectar supply, flowering may have been better at alternative sites (Ford & Paton 1985), patch size may have been below some minimum requirement, birds may have migrated to the tablelands to breed (pers. obs.) or honeyeater populations may be declining due to habitat degradation.

When both honeyeaters were present at sites where appropriate trees were flowering, they may spend considerable time foraging at nectar, even at sites not on the western slopes and when other foods were available (e.g. Clandulla State Forest in the summer of 1993). In winter, insects and fruit may be in short supply, temperatures are relatively low and energy rich

60 foods, such as nectar, may be required to enable birds to survive and build energy reserves for the breeding season. When nectar is available in winter, both A. carunculata and P. corniculatus exploit this resource8 . This would explain their “nomadic or migratory” tendencies. Comparison of the densities of each honeyeater with respect to density of flowering trees (see Figure 4.5) revealed that P. corniculatus may be absent when appropriate trees are in flower, whereas A. carunculata may be present when trees are not flowering. In winter, P. corniculatus was only present at sites with flowering trees (see Figures 4.3, 4.4). Philemon corniculatus possibly requires higher flowering tree density than A. carunculata and it may vacate a patch that falls below a certain threshold. By switching to alternative carbohydrates sources, A. carunculata may not experience the same pressure.

4.4.2 Aggression Movement between foraging bouts involves a loss of energy and it is thus less profitable to feed at patches of low resource density (Collins 1985). Rich patches would require less movement while foraging and enable honeyeaters to acquire more energy per unit time (Collins 1985, Collins & Paton 1989). Both A. carunculata and P. corniculatus spend time defending rich nectar sources (Ford & Paton 1985, Newland & Wooller 1985, McFarland 1996). Where this occurs, the energy spent in aggression may be less than that gained when competitors are excluded.

There were highly significant positive correlations between proportion of time foraging at flowers and time spent in aggression for either honeyeater. This suggests that nectar is an important resource worth defending. When both species were spending upwards of 90% of their foraging time at flowers, aggression was often relatively high (10% of time budget). However, overall time spent in aggression was low (≤ 2%) for both species, which was similar to the results of others (McFarland 1986, Ford & Debus

8 As revealed by time budget data on flower-probing.

61 1994). Nectar from eucalypt blossom is a localised, energy-rich and replenishable resource, and hence worth defending. Establishing exclusive use of this resource through aggression, would lead to a predictable food supply that can be harvested efficiently (Gill 1978). In this study and others (eg. Ford & Debus 1994) aggression away from flowers was much lower than at flowers, and often non-existent. This is further supported by the lack of correlation between time spent in aggression and time spent foraging in general for A. carunculata. Anthochaera carunculata could be found foraging at other food sources at sites and seasons when P. corniculatus was absent. Other food sources, such as fruit and insects, are not self-replenishing in the short term and may not be worth defending.

Highly significant correlations were observed between the density of flowering trees and the proportion of foraging time spent at flowers for both honeyeaters, thus when nectar was available, much of the foraging time was spent flower-probing. However, measures of nectar supply may not necessarily give a direct measure of how important the resource is to honeyeaters. With a decrease in available nectar, time spent feeding at flowers may increase because more time would have to be spent at each flower and more flowers would need to be visited in a given time to obtain sufficient nectar (Gill 1978, McFarland 1986, Armstrong 1992). Conversely, when nectar is abundant, fewer flowers need to be visited to provide an equivalent amount of nectar.

If nectar is super-abundant there would be no need to aggressively defend the supply. However, if nectar is a scarce or patchy resource, the cost of defending it may be prohibitive and it would not pay to aggressively defend the resource. For aggression to be worthwhile the nectar supply must lie somewhere between these two extremes. Several studies have found that aggression was highest at moderate nectar levels and lower when nectar

62 was poor or very rich (Carpenter & McMillen 1976, McFarland 1986, 1996). Carpenter and McMillen (1976) proposed a model predicting that territorial exclusiveness will occur between a lower and upper threshold of nectar supply. The data presented here do not support this. However, this may reflect the scale in this study. Perhaps such relationships are only relevant at a scale matching the size of feeding territories.

Measures of bird behaviour are a good measure of nectar availability (Gill 1978). Hutto (1990) also advocated measuring bird behaviour to confirm measures of food availability and suggested that doing so avoids the problems associated with the bird’s perception of the food, scale-of- measurement, and renewal rates. Since birds spend much of their time feeding at flowers, this resource must therefore be valuable to them. Correlation between flower-probing and time spent in aggression was demonstrated in this study, suggesting that flowers are worth defending. Hence both time spent at flowers and aggressive behaviour are probably good measures of the importance of nectar as a resource.

The data presented here support the contention that nectar from specific species of eucalypt is an important resource for both P. corniculatus and A. carunculata, particularly in the winter/spring period on the western slopes. The presence of P. corniculatus at some sites only in winter in association with preferred eucalypts in flower suggests that this species may not be able to switch to alternative carbohydrates as readily as A. carunculata. The availability of nectar in winter and early spring on the western slopes may therefore be important for the survival and reproductive success of P. corniculatus. Conservation of forest remnants on the western slopes may be critical to sustaining viable populations of these two honeyeaters, particularly P. corniculatus.

63 5. COMPARATIVE FORAGING BEHAVIOUR OF TWO LARGE HONEYEATERS: A LANDSCAPE PERSPECTIVE

5.1 Introduction

Studies that examine and compare niche have contributed much to our understanding of community ecology and knowledge of habitat requirements for individual species, and studies of foraging behaviour of birds are numerous (e.g. Paton 1982, Chan 1990; Plumpton & Anderson 1997). How species coexist is a foremost question for many ecologists and studies examining resource utilisation and niche overlap are also numerous for a wide range of groups, including molluscs (Creese & Underwood 1982), insects (Palestrini et al. 1998), fish (De Pirro et al. 1999), reptiles (Luiselli et al. 1999), birds (Cody 1974, Kossenko & Fry 1998) and mammals (Wahungu 1998). Niche comparison studies are also required when assessing impacts of species on others and may be of a general nature (Runciman 1996; Edwards et al. 1998), or relate to economically important domestic species (Hester et al. 1999), impacts from (Byorth & Magee 1998, Kirchoff & Larsen 1998), and suspected impacts on threatened species (Ford et al. 1993, Moysey 1997, Halley & Gjershaug 1998, Oliver 1998, van der Wal et al. 1998).

Collection of time budgets for niche studies is a useful way of quantifying bird behaviour and to establish the relative importance of different resources to a species or group of species (e.g. see Franklin 1997, Sun & Moermond 1997, Weathers & Seymour 1998, Iason et al. 1999). This approach allows for the collection of data on behaviour, spatial and temporal changes, foraging efficiency and competitive interactions to be quantified. The scale of bird behavioural studies can vary greatly. Dependent upon their purpose; they may be based on data collected over several days (eg. Hixon et al. 1983), months (Franklin 1997) or years (Ford et al. 1990; Caceres 1998). Many studies have examined seasonal variation in a single year (e.g. Jackson 1998,

64 Barreto & Herrera 1998). Some (e.g. Schmitt & Coyer 1983) compared behaviour in sympatry and allopatry. However, although these studies were undertaken over various temporal scales, they were limited in spatial scale. Comparative studies based upon natural experiments that contrast behaviours, spatially and temporally, continue to make a significant contribution to describing patterns of variation in foraging ecology (Recher & Gebski 1990).

Most of the data on A. carunculata and P. corniculatus are contained within broader studies of woodland or forest bird communities and thus provide only limited information on these honeyeaters (e.g. Lamm & Wilson 1966, Ford & Paton 1977, Marchant 1979, Loyn 1980, McFarland 1984, Ford 1985, Pyke 1985 Recher & Holmes 1985, McFarland 1986, Osbourne & Green 1992, Ford et al. 1993, Slater 1995, McFarland 1996, Traill et al. 1996, Egan et al. 1997, Mac Nally 1997, Er et al. 1998), while few papers specifically examine or compare either species (Saunders 1993, Ford & Debus 1994, Ford 1999, Ford & Tremont 2000). Other studies examine only prominent behaviour, such as aggression, as it impacts on other species (Davis & Recher 1993, Leonard 1995, Geering & French 1998, Oliver 1998). Much of what is understood about the ecology of these two honeyeaters was discussed in section 1.6 of the introduction. The lack of understanding of the movement patterns of these two species makes detailed behavioural studies difficult (Saunders 1993).

The aim here was to investigate the foraging ecology of A. carunculata and P. corniculatus over a broad scale to consider how patterns of abundance and distribution of these two honeyeaters related to regional patterns of food resource utilisation, and how patterns of food resource utilisation related to niche overlap and competition between these two species.

65 The specific questions considered here are: 1. Which foods and/or substrates are selected by A. carunculata and P. corniculatus? 2. What similarities or differences exist between them in resource selection and how can any differences be accounted for? 3. How much time does each species spend foraging and what may account for any differences? 4. How does food or substrate selection vary between sites at the landscape scale? 5. How does food or substrate selection vary over time within sites? 6. How does niche overlap vary at spatial and temporal scales and are there patterns in relation to food or substrates selected? 7. What proportion of time is spent in aggression, and to which birds is this aggression directed?

8. Is there a relationship between aggression and niche overlap?

5.2 Methods

5.2.1 Data collection Time budget data were collected along two 2000 m x 200 m transects at each site. No data were collected near nests to avoid possible bias between the behaviour of nesting and non-nesting birds. The sites at Royal National Park, Douglas Park, Kings Tableland, Clandulla State Forest, Munghorn Gap Nature Reserve and Back Yamma State Forest were visited during the middle of each season during 1992 and 1993. Goobang National Park was visited each season from 1993 to 1996. There was insufficient data collected at Douglas Park and Kings Tableland and so these sites were excluded from the analysis. Time budgets were collected for each species at each site as follows: bird behaviour was observed through binoculars and if behaviour changed due to the recorder’s presence (i.e. it stopped feeding, preening or

66 calling and watched the researcher) no observations were recorded until it reverted to ‘normal’ behaviour. Each bird’s behaviour was recorded within four variables: activity, foraging substrate type, substrate species and foraging height (see Table 5.1). They were observed for a maximum of five minutes, or until lost from sight. Within this period the behaviour that occupied the majority of each 10 second bout was recorded. When both species were present, observations alternated between species. Generally it was possible to distinguish individuals in a patch, so that the risk of double sampling an individual was minimised.

5.2.2 Statistical analysis The data constituted a series of repeated observations on individual birds. It is possible that this could introduce correlation among observations taken on the same bird. The consequence of this is that the P-values for statistical tests performed on the data may be smaller than if the observations were statistically independent. However, many birds were sampled which should negate this substantially. In addition, if the P-values for statistical tests are very highly significant then the repeated observations are unlikely to affect the conclusions of the tests.

Chi-square tests were used to analyse bout frequencies for single categories and two-way frequency tables (Zar 1984). Three-way frequency tables were analysed by fitting loglinear models to the counts (Agresti 1996). Loglinear models for contingency tables are a class of Generalized Linear Models (GLMs) whereby the logarithms of the expected cell frequencies can be expressed as a linear function of the row and column variables for two-way tables, or multiple variables for higher dimension tables (Bishop et al. 1975, Agresti 1996). Loglinear models for contingency tables are fitted as a nested hierarchy of models (Agresti 1996). The results of fitting a nested series of GLMs are often displayed in Analysis of Deviance tables. For loglinear

67 Table 5.1 Categories used for time-budgeting behaviours of Anthochaera carunculata and Philemon corniculatus at each site. (Categories defined where necessary. Greater than signs (>) indicate direction of aggression. Ac = Anthochaera carunculata, Pc = Philemon corniculatus, OH = other honeyeater)

Activity Substrate Foraging Height 0m 0-2m 2-5m 5-10m 10-15m 15-20m >20m Foraging Air Flying out from a perch to catch prey mid-air and returning to perch with prey. (includes searching and Flowers Probing flowers for nectar and/or invertebrates. prey handling) Foliage Gleaning invertebrates, lerp, manna or honeydew from leaves. Bark Gleaning invertebrates or sap from trunk or branches. Seeds Ingestion of seeds. Fruit Ingestion of friut or fruit pedicel. Ground Ingestion of any food item taken while on the ground. Sitting Resting while perching and while engaging in no other activity. Preening Care of while sitting. Calling Includes only calling while sitting. Flying All flight except during aggressive encounters. Aggression Approach which disrupts another bird’s behaviour. Direction of Aggression Ac>Ac Ac>Pc Ac>OH OH>Ac Pc>Pc Pc>Ac Pc>OH OH>Pc models, the deviance is equivalent to the Likelihood-Ratio statistic. Successive, nested models can be compared by examining the changes in this statistic. A significant reduction, assessed against the chi-square distribution, signifies an improvement in the model. Models without any interaction terms were compared with the null model; that is, a table in which the row and column structure is ignored. Models containing two-way associations terms were compared with the independence model and the saturated model was compared with the two-way association model.

5.2.3 Niche breadth and overlap Niche breadth was calculated as the Shannon-Wiener diversity index

H = ∑|pi log pi| where pi is the proportion of observations in category i, for all categories. Where all foraging is limited to one category H = 0 and when diverse H approaches one. Niche overlap was calculated as Oij = 1-0.5∑|pih - pjh| where pih equals the proportion of observations for species i in category h and pjh equals the proportion of observations for species j in category h, for all categories (Manly 1990). Oij = 0 indicates no overlap and Oij = 1 indicates complete overlap. Niche breadth and overlap were calculated for foraging height and substrate combined (see Table 5.1 for breakdown of categories used for each variable). Plant species were not used in the calculation of niche because in most cases only one plant species was chosen by birds within each substrate category. For example, birds flower probed at the one plant species in flower at the time of the survey. Foliage gleaning did occur at several plant species at many sites during most surveys but this is described in more detail elsewhere (see Chapter 6).

5.3 Results

5.3.1 General patterns in foraging substrate selection

Both honeyeaters spent most of their time foraging at flowers in the canopy (i.e. > 10 metres), although, they also fed at flowers, seed and fruit in the

69 understorey (5 to 10 metres) and in the shrub layer (< 5 metres) when food resources were there (Figure 5.1, Table 5.2 and 5.3). Anthochaera carunculata also fed at foliage in the canopy (Figure 5.1). The proportion of foraging in the canopy was very highly significantly different from foraging in other height classes for both honeyeaters (A. carunculata c2 = 4394.7, df = 3,

P < 0.001, P. corniculatus c2 = 2064.8, df = 3, P < 0.001). However, there were also very highly significant differences in the pattern of use of foraging height classes between the two honeyeaters (c2 = 96.7, df = 3, P < 0.001). The greatest difference between the two honeyeaters occurs between the two canopy layers (10 - 15 metres) and (>15 metres), with A. carunculata foraging more often between 10 and 15 metres in the canopy, while P. corniculatus spent nearly equal time in both height classes (Table 5.2).

Foraging substrate selection by each honeyeater was very highly significant (A. carunculata c2 = 11707.8, df = 4, P < 0.001, P. corniculatus c2 = 10535.2, df = 4, P < 0.001) with both honeyeaters spending more time at flowers and then foliage than other substrate types (Table 5.3). The proportion of time spent foraging on different substrates was dependent on the honeyeater species (c2 = 1329.6, df = 4, P < 0.001). Anthochaera carunculata spent proportionately more time foraging on bark and foliage and less time on fruit and seed than P. corniculatus. While both spent most of their time at flowers, P. corniculatus spent 71% at flowers while A. carunculata spent 54% at flowers (Table 5.3). The utilisation of different foraging substrates by the two honeyeaters is complex (see Table 5.4). While there is significant selection among foraging substrates, the proportional use not only varies with foraging height but also with foraging species (Table 5.5). All three variables are needed to explain bout frequencies. The largest reductions in deviance (i.e. improvements in model) were due to the inclusion of foraging substrate and substrate by species associations.

70 100 Anthochaera carunculata 80

60 n = 10114 Hi = 0.94 40

20 > 15m 10-15m 5-10m 0 < 5m Air Bark Fruit Seeds Ground Foliage Flowers

Oij = 0.66

100 Philemon corniculatus 80

60 n = 6014 Hj = 0.81 40

20 > 15m 10-15m 5-10m 0 < 5m Air Bark Fruit Seeds Ground Foliage Flowers

Figure 5.1 Percent of foraging bouts at various substrates by Anthochaera carunculata and Philemon corniculatus (data pooled: n = no. of 10 second observation periods, H i and H j = niche breadths and O ij = niche overlap).

71 Table 5.2 Number of bouts within each height class for each honeyeater. (Ac = Anthochaera carunculata, Pc = Philemon corniculatus, percentages show proportion of foraging for each species)

Height Class Ac Pc > 15 m 3735 (37%) 2315 (40%) 10 - 15 m 4551 (45%) 2278 (40%) 5 - 10 m 1093 (11%) 563 (10%) 0 - 5 m 674 ( 7%) 587 (10%)

Table 5.3 Number of bouts within each substrate selected by each honeyeater for sites pooled where both species were present. (Ac = Anthochaera carunculata, Pc = Philemon corniculatus, percentages show proportion of foraging for each species)

Substrate Ac Pc Bark 674 ( 8%) 89 ( 2%) Flowers 4660 (54%) 3703 (71%) Foliage 3080 (35%) 791 (15%) Fruit 250 ( 3%) 423 ( 8%) Seeds 41 (0.5%) 246 ( 5%)

Table 5.4 Number of bouts for foraging substrates selected in each height class for each honeyeater.

0-5m 5-10m 10-15m >15m Anthochaera carunculata Bark 83 117 319 155 Flowers 115 236 2094 2215 Foliage 123 408 1363 1186 Fruit 250 0 0 0 Seeds 0 41 0 0 Philemon corniculatus Bark 0 8 61 20 Flowers 44 170 1758 1731 Foliage 57 174 359 201 Fruit 423 0 0 0 Seeds 0 195 0 0

72 Table 5.5 Summary of analysis of deviance for hierarchical log-linear models for species, substrate and height interactions. All changes in deviance were significant at P < 0.001.

Deviance DF Change in Deviance Change in DF Species l Null 29506.2 38 594.5 1 Height l Null 24006.6 36 6094.1 3 Substrate l Null 14472.6 35 15628.1 4 Substrate + Height l Null 8080.5 32 22020.2 7 #1 Substrate + Height + Species l Null 7188.0 31 22912.7 8 Substrate + Height + Species + Height:Species l #1 7121.9 28 66.1 3 Substrate + Height + Species + Substrate:Species l #1 5862.3 27 1325.7 4 Substrate + Height + Species + Substrate:Height l #1 1453.9 19 5734.1 11 #2 Sub + Ht + Sp + Sub:Ht + Sub:Sp + Ht:Sp l #1 104.8 12 7083.2 19 Saturated l #2 0.0 0 104.8 12 The proportion of bouts spent in foraging and in non-foraging activities is shown in Figure 5.2. Anthochaera carunculata spent significantly more time foraging than P. corniculatus (c2 = 394.3, df = 1, P < 0.001), with 60% and 47% respectively (Table 5.6).

5.3.2 Spatial and temporal variation in substrate selection

Foraging substrate selection by each honeyeater for sites where both species were present is shown in Table 5.7. Bark, flowers and foliage comprise most of the substrates selected at most sites, while seeds and fruit were selected only at Munghorn Gap Nature Reserve. Flowers were selected predominantly at all sites except Clandulla State Forest where most of the foraging was at foliage. There was a very highly significant interaction between species, site and substrate (Table 5.8). Seeds and fruit were not considered in the analysis as they were only selected at one site and were not present at other sites. Even without their inclusion, the substrate:site association was the strongest two-way associations among factors, indicating that the proportion of bouts on different foraging substrates varies significantly among sites.

Foraging substrate selection by each honeyeater for each site within each year is shown in Table 5.9 and Table 5.10. The data contained in these tables is very difficult to formally analyse as there are many structural and sampling zeros9. Many of the differences are clear-cut and any loglinear analysis would require a saturated model i.e. a model containing all associations between variables to account for variation in foraging substrate selection between years. However, a chi-square test of frequency of foraging bouts on flowers among years was found to be very highly significant (c2 = 1461.4, df = 2, P < 0.001). Foraging at flowers was highly

9 Structural zeros result when no frequency counts are expected e.g. no foraging because the species is absent, while sampling zeros occur when no frequency counts were made in one or more categories.

74 Foraging

Sitting

Preening

Calling n=17155 n=12739

Flying

Aggression

100 75 50 25 0 0 25 50 75 100 % of Bouts % of Bouts Anthochaera carunculata Philemon corniculatus

Figure 5.2 Activity time budgets for Anthochaera carunculata and Philemon corniculatus (data pooled: n = no. of bouts).

75 Table 5.6 Number of foraging and non-foraging bouts for Anthochaera carunculata and for Philemon corniculatus (percentages are for bouts for each species).

Anthochaera carunculata Philemon corniculatus Foraging 10385 (60%) 6041 (47%) Non-foraging 7236 6698

Table 5.7 Number of bouts at each substrate selected for sites where both Anthochaera carunculata and Philemon corniculatus were found.

A. carunculata P. corniculatus Clandulla State Forest Bark 337 17 Flowers 617 117 Foliage 1564 280 Fruit 0 0 Seeds 0 0

Munghorn Gap Nature Reserve Bark 133 50 Flowers 1045 1015 Foliage 986 280 Fruit 250 463 Seeds 41 423

Goobang National Park Bark 185 11 Flowers 2595 2067 Foliage 530 30 Fruit 0 0 Seeds 0 0

Back Yamma State Forest Bark 19 11 Flowers 403 504 Foliage 0 18 Fruit 0 0 Seeds 0 0

76 Table 5.8 Summary of analysis of deviance for hierarchical log-linear models for species, substrate and site interactions for sites where both A. carunclulata and P. corniculatus were present (seed and fruit categories removed from analysis). All changes in deviance were significant at P < 0.001.

Deviance DF Change in Deviance Change in DF Species | null 16584.9 22 1146.1 1 Substrate | null 10251.9 21 7479.1 2 Site | null 14188.5 20 3542.5 3 #1 Substrate + Species + Site | null 5563.2 17 8625.3 3 Substrate + Species + Site + Substrate:Species | #1 4645.4 15 917.8 2 Substrate + Species + Site + Site:Species | #1 4640.3 14 922.9 3 Substrate + Species + Site + Substrate:Site | #1 1627.6 11 3935.6 6 #2 Sub + Sp + Site +Sub:Sp + Site:Sp + Sub:Site | #1 262.2 6 5301.0 11 Saturated | #2 0.0 0 262.2 6 Table 5.9 Number of bouts observed for A. carunculata and P. corniculatus on different substrates by year and site.

A. carunculata 1992 1993 RNP CSF MG GNP BSF RNP CSF MG GNP BSF Bark 104 174 92 - 0 112 163 41 41 19 Flowers 0 110 749 - 303 0 507 296 1007 100 Foliage 571 818 322 - 0 561 746 664 113 0 Fruit 0 0 0 - 0 0 0 250 0 0 Seeds 0 0 41 - 0 0 0 0 0 0

P. corniculatus 1992 1993 RNP CSF MG GNP BSF RNP CSF MG GNP BSF Bark - 0 5 0 0 - 17 45 2 11 Flowers - 0 677 117 471 - 117 338 467 33 Foliage - 70 233 0 0 - 210 230 30 18 Fruit - 0 0 0 0 - 0 423 0 0 Seeds - 0 246 0 0 - 0 0 0 0 variable between years at most sites. Fruit and seeds were only selected at Munghorn Gap Nature Reserve, but in alternate years. At the Royal National Park only bark and foliage were selected by A. carunculata and there was no significant difference in substrate selectivity between years (c2 = 0.382, df = 1, P >0.05).

Table 5.10 Number of bouts observed for A. carunculata and P. corniculatus on different substrates for each year at Goobang National Park (seed and fruit were not recorded as substrates at this site).

A. carunculata 1992 1993 1994 1995 Bark - 41 85 59 Flowers - 1007 1547 41 Foliage - 113 85 211

P. corniculatus 1992 1993 1994 1995 Bark 0 2 9 - Flowers 117 467 1483 - Foliage 0 30 0 -

5.3.3 Niche Overlap and aggression Niche breadth and niche overlap between species for all sites, years and seasons are shown in Table 5.11. Niche breadth varied over sites, years and seasons and was more often narrow when flowers were the main substrate selected, for example at Back Yamma State Forest in spring 1992. Niche breadth varied from 0 to 0.79 for A. carunculata (mean = 0.440, S.E. = 0.038, n = 35) and varied between 0 and 0.77 for P. corniculatus (mean = 0.428, S.E. = 0.049, n = 17). Niche overlap varied between 0.31 and 1.00 (mean = 0.661, S.E. = 0.048, n = 17). High niche overlap (Oij > 0.8) occurred at some sites and seasons, for example at Goobang National Park in winter 1993 and at Back Yamma State Forest in spring 1992, where they shared the same flowering

79 Table 5.11 Niche breadth based on foraging heights and substrate type for Anthochaera carunculata and Philemon corniculatus (Hi and Hj respectively) and niche overlap (Oij). (- indicates one or both species absent)

Site Royal Clandulla Munghorn Gap Goobang Back Yamma National Park State Forest Nature Reserve National Park State Forest

Indices Hi Oij Hj Hi Oij Hj Hi Oij Hj Hi Oij Hj Hi Oij Hj 1992 Summer - - - 0.15 0.61 0.38 0.45 0.46 0.67 ------Autumn 0.08 - - 0.42 - - 0.67 0.83 0.60 ------Winter 0.49 - - 0.79 - - 0.77 0.66 0.42 - - - 0.29 - - Spring 0.31 - - 0.75 0.47 0.33 0.38 0.83 0.30 - - - 0.16 1.00 0.16 1993 Summer - - - 0.72 0.77 0.77 0.46 0.84 0.56 ------Autumn 0.33 - - 0.65 0.40 0 0.29 0.41 0.27 - - - 0 - - Winter 0.76 - - 0.41 - - 0.52 - - 0.41 0.89 0.39 0.53 0.31 0.51 Spring 0.39 - - 0.64 0.67 0.55 0.64 0.62 0.70 0.53 0.77 0.30 - - - 1994 Summer 0 - - Autumn 0.78 - - Winter 0.14 - - Spring 0.30 0.69 0.37 1995 Summer - - - Autumn 0.37 - - Winter 0.46 - - Spring 0.36 - - substrates. Variations in niche overlap between sites, between years and between seasons are shown in Figure 5.3. The analysis of variance for spatial and temporal variation of niche overlap revealed no significant variation patterns (Table 5.12). Interactions between site, year and season were not included as there was insufficient replication to allow for an interaction model.

The proportion of activity bouts where aggression occurred was very low at around two percent (see Chapter 4.3.4). Correlations between time spent in aggression and counts of honeyeaters and flowering trees were discussed in Chapter 4. The breakdown of direction of aggression is shown in Figure 5.4. Aggression was mostly intraspecific (A. carunculata 34%, P. corniculatus 24%), while interspecific aggression between them accounted for 20% of bouts of aggression. The remaining 22% was between these two honeyeaters and other honeyeaters. Interspecific aggression between A. carunculata and P. corniculatus was found to be positively correlated with niche overlap after arcsine and square root transformation of both variables (Figure 5.5, r = 0.729, t = 4.12, df = 15, P < 0.001).

Table 5.12 Analysis of variance for spatial and temporal variation in niche overlap.

Source of Variation df Sum of Sq Mean Sq F Value P (F) Site 1 0.019 0.019 0.46 > 0.05 Year 1 0.0004 0.0004 0.01 > 0.05 Season 3 0.036 0.012 0.29 > 0.05 Residuals 6 0.247 0.041

81 1.1 1 0.9 0.8 0.7 0.6 0.5 Niche Overlap 0.4 0.3 CSF MG GNP BSF

1.1 1 0.9 0.8 0.7 0.6 0.5 Niche Overlap 0.4 0.3 1992 1993 1994

1.1 1 0.9 0.8 0.7

0.6 0.5 Niche Overlap 0.4 0.3 Summer Autumn Winter Spring

Figure 5.3 Spatial and temporal variation in niche overlap.

82 Anthochaera carunculata

34 %

8 % 12 %

6 % Philemon corniculatus 7 %

24 %

1 % 8 %

Other Honeyeaters

Figure 5.4 Direction of aggression between honeyeaters shown as percentage of bouts spent in aggression. Arrows indicate the direction of successful aggressive attacks defined as an encounter that disrupts the behaviour of another bird. Horizontal arrows indicate aggression between conspecifics. Sample size was 973 ten-second observation periods; data pooled for all sites and surveys.

83 1.6

1.4

1.2

1

0.8

arcsin(sqrt(niche overlap)) 0.6

0.4 0 0.05 0.1 0.15 arcsin(sqrt(interspecific aggression))

Figure 5.5 Plot of transformed niche overlap against transformed interspecific aggression (r=0.729, t=4.12, df=15, P<0.001).

84 5.4 Discussion

5.4.1 Foraging patterns and relationship to morphological differences It was observed that these honeyeaters are mainly canopy foragers, although they will forage within the shrub layer when appropriate food sources are locally abundant. Flower probing and foliage gleaning in the canopy make up the bulk of their foraging substrate selection, while nectar, fruit and seeds are taken from the subcanopy and shrub layer when available. Previous studies have also shown that these two honeyeaters are mainly canopy foragers that probe flowers and glean foliage, mainly in eucalypts (Ford & Paton 1977, Ford et al. 1986, McFarland 1986) and that the flowering phenology of eucalypts has considerable influence over the occurrence of these birds (Ford et al. 1986). This is typical of birds that tend to be opportunistic since they often make ready use of super abundant resources (Szaro et al. 1990).

Differences between foraging behaviour for A. carunculata and P. corniculatus are probably related to their morphologies. Anthochaera carunculata spent more time foraging than P. corniculatus. Hugh Ford also found that A. carunculata spent more time foraging than P. corniculatus at Newholme in 1990 (unpublished data). Anthochaera carunculata also spent more time at foliage and bark substrates than P. corniculatus, and was less able to ingest large prey items (see Chapter 7). Anthochaera carunculata (125g) is heavier than P. corniculatus (107g) (Ford et al. 1986). However, the bill of A. carunculata (gape width 10.12 ± 0.13 mm) is smaller than for P. corniculatus (12.56 ± 0.18 mm; see Chapter 7.3). When nectar from flowering eucalypts was abundant both honeyeaters spent a considerable percentage of their foraging time flower probing. Anthochaera carunculata spent more time foliage gleaning in areas vacated by P. corniculatus, when eucalypts were not flowering during cooler months. This is hypothesised to be due to the smaller bill of A. carunculata that enables them to glean foliage

85 more efficiently and thus they are able to persist for longer periods at such sites. In contrast, P. corniculatus appears to be more dependent on flowering eucalypts in the cooler months (see Chapter 4.4.1). Because A. carunculata are limited to smaller prey items by their smaller gape and because they are larger than P. corniculatus, they need to spend more time foraging, but are probably able to persist at lower resource densities than P. corniculatus.

5.4.2 Spatial and temporal variation in foraging patterns Substrate selection was found to vary considerably between sites. Bark, flowers and foliage were the most often selected substrates, but their relative selection varied considerably between sites. Fruit and seeds were only recorded at one site, i.e. Munghorn Gap Nature Reserve. Substrate selection was also found to vary between years within sites and between seasons within a year (see appendices for a complete breakdown of foraging substrate selectivity for sites, years and seasons) and is likely to reflect variation in resource availability between years and during any year e.g. the density of flowering trees between seasons and between years (Chapter 4). The phenology of several food plant species utilised was highly variable. For example flowering of E. sideroxylon at Goobang National Park was generally over winter and spring, but varied from 4% to 62% of trees between years (1993-1996: see Figure 4.3). Timing of seed and fruit set and eruption of cicadas was also highly variable between years at Munghorn Gap Nature Reserve. Despite this, such foods were observed to be important during the warmer months when they were available. Many of the other foods utilised by these two honeyeaters are likely to vary in similar ways. The variation in substrate selectivity between seasons within a year at all sites, between years within a site, and between sites suggest that resource availability is highly variable over both space and time, at both patch and landscape scales. With such variable food sources, the ability to switch between resources and/or sites would play a critical role in survival. The patterns of resource

86 availability reflect the distribution patterns shown by A. carunculata and P. corniculatus which also vary at the landscape scale.

5.4.3 Niche variation and relationship to aggression No patterns were found in the spatial and temporal pattern of niche overlap, but this may be due to the small number of sampling periods where both honeyeaters were present. Niche breadth was found to be narrow at times when particular food sources were abundant, and often at these times, niche overlap was found to be greater. This suggests that both species gather to feed at abundant food sources. Niche overlap can be expected to increase at such resources (Wiens 1989). However, high niche overlap does not mean that competition is strong between species sharing resources as these may be abundant and non-limiting (Underwood 1986, Wiens 1989). If competition exists between these two honeyeaters, then it would be expected that they would defend food sources through aggression.

Although aggression is an obvious behaviour and probably affects inter- and intraspecific interactions, it consumed only a small portion of these birds’ time. Ford and Debus (1994) also reported similar levels of aggression (A. carunculata 1.8% and P. corniculatus 1.7% of total time budget) and that much was conspecific (24% and 37% respectively). When A. carunculata foraged without P. corniculatus aggression varied between 0.4 and 2.1% and 77% was conspecific (McFarland 1986). In this study it was found that most aggression was intraspecific. However, the interspecific aggression that did occur was positively correlated with niche overlap. It appears that when resource sharing increased so did their need to defend those resources. Thus, quantification of interspecific aggression may be a good measure of interspecific competition. However, this aggression would have to be solely for resource defence and not related to, for example breeding territory defence. This study quantified aggression at food sources only and time-

87 budgeting near nests was avoided. Intraspecific aggression is a good measure of competition between individuals of the same species. Underwood (1986) suggested that foraging overlap would be greater between individuals of the same species than between individuals of different species as they are more likely to overlap in morphologies and food preference, and hence intraspecific competition is likely to be greater than interspecific competition. The relative proportions of interspecific and intraspecific aggression found in this study support this suggestion.

5.4.4 Landscape perspectives of foraging studies

Studies of the behavioural interactions of A. carunculata and P. corniculatus have indicated seasonal differences (Ford & Paton 1977). However, few studies have revealed the extent of this variation because they were of relatively short duration or data were pooled over several years (eg. Ford et al. 1986), and because there were a limited number of sites or there were multiple sites in close proximity (Ford & Paton 1977, Ford et al. 1986).

This study was designed to examine patterns at local and landscape scales (seven sites over a 350 km long east-west transect across the central sympatric range of these species) as well as within and between years (seasonally over two years). This study demonstrated great variation at both site and landscape scales between seasons and consecutive years, and found no regular patterns. From this it may be inferred that a lack of regular pattern would be expected elsewhere through the range of these two honeyeaters.

Time budgeting animal behaviour is an important tool for identifying and quantifying the relative importance of food sources, quantifying interactions, habitat use and behaviour (see Caithamer et al. 1995, Plumpton & Andersen 1997, Jones et al. 1998, Iason et al. 1999, Oliver 2000). Proportion of time

88 spent at various foods allows estimation of the importance of such resources (Hutto 1990). It is particularly useful where animals utilise a diversity of food resources that are difficult to quantify or manipulate. In this study, there was much switching between diverse food resources which varied unpredictably at both local and landscape scales. Thus time budgeting was the appropriate tool for quantifying the behaviour of these two honeyeaters. Studies collecting time budget data over broader scales need to be interpreted cautiously when heterogeneous data on foraging behaviour are pooled. When temporal and/or spatial variation exists, such pooling can lead to misinterpretations about community organisation, competitive interactions and foraging ecology (Hejl & Verner 1990, Szaro et al. 1990). The behavioural response at one time and place will depend on conditions and, for adaptable species foraging in variable environments, behavioural data will likely be heterogeneous. This study found that site, year and season were all important variables in determining the foraging substrates selected, and that the data were heterogeneous over both space and time.

It follows that data collected over a short time period in one location do not necessarily represent the foraging ecology of a species, especially if the species is likely to be a generalist forager living in a variable and unpredictable environment. This can be illustrated by the variation that was found between years and sites in this study. For example, at Munghorn Gap Nature Reserve, both honeyeaters took Acacia seed in summer 1992 and Exocarpus fruit in summer 1993. These foods were not available together during both years. Great variation was also found in the flowering of Eucalyptus sideroxylon and populations of A. carunculata and P. corniculatus between season and years at Goobang National Park (see Chapter 4). The relative importance of particular food species was also found to vary greatly between sites. The seeds and fruit utilised at Munghorn Gap Nature Reserve were not found at other sites. Flowering tree species selected for flower

89 probing were also found to be different between sites (see Chapter 4). Pooling the data across sites or years would have hidden important information about the variability of food sources over space and time, and would not have provided any explanation for the distribution of these two honeyeaters. Thus, for nomadic generalist foragers, such as A. carunculata and P. corniculatus, data collected over broader spatial and temporal scales will provide a more complete picture of behaviour provided the data is not pooled. The interpretation of data needs to account for consistent patterns as well as heterogeneity.

When looking for patterns in behaviour of species, long term studies that match the scale of environmental variation within the target species range are required. Short studies may provide a “snapshot” in space and time that is not representative (Wiens 1989). Szaro et al. (1990) found that the activity patterns for nine ponderosa pine forest bird species differed significantly between years in their study, and noted that only seven of 150 papers they reviewed examined differences between years. They concluded that when data were collected over a single year or when data were pooled over several years, misinterpretations were likely. Examination of temporal variation may identify mechanisms for such variability, such as resource availability, climatic conditions, predation and phenology. The results in this study supported these findings. The foraging patterns of the two honeyeaters demonstrated that there was considerable variation between sites, years and seasons, that probably reflected variable resource availability, which may in turn have related to variable phenology and climatic variation. The unpredictability of resources in time and space indicate that to manage these species, and presumably other wide ranging opportunistic species, requires a better knowledge of their distribution, resource use and their habitat variability to ensure the long term viability of populations.

90 6. SELECTIVE FOLIAGE FORAGING BY RED WATTLEBIRDS Anthochaera carunculata AND NOISY FRIARBIRDS Philemon corniculatus

6.1 Introduction Foliage gleaning in eucalypts and nectar collection are the most common foraging behaviours observed in honeyeaters (Keast 1985a, Recher et al. 1985). For example, the Noisy Miner Manorina melanocephala and Black- headed Honeyeater Melithreptus affinus may spend a substantial amount of their foraging time foliage gleaning (Ma. melanocephala 55-75%, Dow 1977, Me. affinus 93%,Thomas 1980). In contrast, other honeyeaters appear to be more dependent on bark and/or flowers (Thomas 1980). For many honeyeaters, lerp, manna and honeydew gleaned from eucalypt foliage are important sources of carbohydrates when nectar is not available (Paton 1980) and may be critical for winter survival (Recher et al. 1985).

Many foliage gleaning birds select specific tree species (Holmes & Robinson 1981, Recher et al. 1985) and food density has been observed to be higher in such species (Holmes & Robinson 1981). In addition, tree species with more lerp and higher density of preferred invertebrates, also have higher levels of nutrients in their foliage than non-target species (Recher et al. 1991). However, foliage structure may also determine the ease with which birds can effectively forage and hence species’ choice (Robinson & Holmes 1984). Despite these observations, caution is necessary when assessing whether birds actively choose tree species because they may simply forage on the common tree species (Ford et al. 1986).

Selective foraging on the foliage of particular eucalypts has been observed for many Australian birds, including honeyeaters (Meliphagidae) (Wooller

91 1984, Keast 1985b, Recher et al. 1985, Ford et al. 1986, Franklin 1997), and thornbills (Pardalotidae) (Bell 1985, Woinarski 1985, Recher et al. 1991, Recher 1989, Recher & Gebski 1990); the main canopy foliage foragers. Preference may be shown for specific eucalypts (Recher et al. 1985). However, preference may vary locally and depend upon the tree species present, for example Bell (1985) observed that Striated Thornbills Acanthiza lineata choose stringybarks and box over gums, while Recher (1989) observed that they preferred ironbark to box, although there were no ironbarks at Bell’s site and no stringybarks at Recher’s site.

Lerp, produced by psyllids, often accounts for the bulk of the invertebrate biomass on eucalypt foliage and are an important food source for foliage gleaners (Woinarski 1985). For example, in enclosures 22 out of 29 bird species collected lerp (Woinarski et al. 1989). Lerp are an important food source for honeyeaters (Paton 1980). Such carbohydrate-dependent birds forage preferentially in trees with abundant lerp during cooler months, when invertebrates are less abundant (Recher et al. 1991).

Large honeyeaters, Anthochaera and Philemon spp., are mostly nectarivorous (McFarland & Ford 1991), but have also been described as opportunistic generalists (Wooller 1984). Throughout the year, foliage gleaning was observed to be the second most common foraging behaviour after flower- probing in woodland (Ford et al. 1986). Gleaning for lerp, especially when nectar is scarce, comprises a large proportion of foraging time for the Xanthomyza phrygia, a medium-sized honeyeater (Oliver 2000). In this chapter the extent to which P. corniculatus and A. carunculata select foliage of particular eucalypt species is investigated. These birds have high niche and range overlaps, yet little data have been published on the ecology of either species.

92 6.2 Methods

Data were collected in open forest along an east-west transect within the sympatric range of Noisy Friarbirds and Red Wattlebirds in central eastern

New South Wales, in the middle of each season for two to three weeks in

January, April, July and October from 1992 to 1996.

Generally, trees were counted and identified along two transects 20 m wide and 2 km long at each site. At Royal National Park and Back Yamma State

Forest transects were 50 m wide but shorter to avoid ecotones with other habitats.

Time budgets for individual honeyeaters were collected as follows: when a bird was located it was observed through binoculars. If the bird’s behaviour ceased when it became aware of the recorder’s presence (ie it stopped feeding, preening or calling and watched the researcher) observations were discontinued until the bird resumed its previous behaviour or moved to some other form of activity. Each bird’s behaviour was recorded for a maximum of 5 minutes. Every 10 seconds, the behaviour that occupied the majority of the time unit was recorded. Hence up to 30 observations may have been recorded for an individual. Each 10 second period became the unit of observation. Observation was terminated after 5 minutes, or when the bird was lost from sight. When both species were present observations alternated between species. Generally, it was possible to keep track of the birds in a patch, so that the risk of reselecting an individual was minimised.

93 The time spent foliage gleaning on each tree species was compared with tree species abundance. For each bird species, data were collected as frequency of foliage gleaning in each tree species at each site. Expected frequencies of foliage use are estimated as the proportion of the number of trees of each species in the forest at each site. Invertebrate and Lerp densities and foliar nutrients were not determined due to time restraints.

Observed and expected frequencies are compared for significant differences using log-likelihood ratio tests. Only the four to six more common Eucalyptus and Angophora tree species are considered in the analysis, as inclusion of other tree species produced zeros in the contingency tables. These other species were removed as a prerequisite for the analysis. Where there were sufficient data, sites and seasons were analysed separately and where a similar pattern for seasons and years within a site was found, data were pooled for comparison. Changes in preference for tree species over years and seasons for each site were tested for significant differences using loglinear analysis.

The method of continuous sampling suffers from a lack of independence of data. However, it records visits to tree species of short duration, which can be missed by single-point observations (Recher & Gebski 1990). The large number of observations made should overcome the lack of independence, and as a precaution, differences between frequencies and associations between effects were not considered significant unless p < 0.01.

6.3 Results Selection of tree species was significantly different from that expected from

94 the species’ frequencies in all seasons and at all sites (A. carunculata, Figure 6.1 and P. corniculatus, Figure 6.2). Most foliage gleaning occurred at Eucalyptus and Angophora trees (A. carunculata 95%, n=4200, P. corniculatus 96%, n=838). Acacia, Allocasuarina and Callitris tree species accounted for the remaining foliage gleaning.

In general, A. carunculata and P. corniculatus showed significant preference for E. punctata foliage when data were pooled for each site where this tree species was present (P < 0.001 for both species at each site). However, in Winter 1992 there was a significant preference by A. carunculata for E. crebra in Clandulla State Forest (c2 = 439.2, n = 192, df = 2, P < 0.001), and at

Munghorn Gap Nature Reserve both birds preferred to foliage glean in A. floribunda, although sample sizes were too small to test for significance. In Goobang National Park, A. carunculata foraged over a wider range of tree species. There was a significant shift in preferred tree species between season and years. Eucalyptus fibrosa was strongly selected in spring 1993 (c2 = 35.89, n = 101, df = 3, P < 0.001) and autumn 1994 (c2 = 159.1, n = 144, df = 2, P < 0.001), whereas E. blakelyi was strongly selected in summer 1994 (c2 = 70.17, n = 28, df = 1, P < 0.001), winter 1995 (100 %, n = 51) and spring 1995 (c2 = 27.62, n = 29, df = 1, P < 0.001). Eucalyptus sideroxylon was strongly selected in spring 1994 (100 %, n = 14) and winter 1995 (100 %, n = 131). Philemon corniculatus visited Goobang National Park during winter and spring where they foraged at E. melliodora and E. sideroxylon, although there was insufficient data for analysis. The more widely preferred species for foliage gleaning, E. punctata, was not present at the site.

Although A. carunculata consistently selected foliage in E. punctata in all seasons and both years at the Royal National Park, there was a significant difference between seasons (c2 = 9.2543, df = 2, P < 0.01) and between years (c2 = 7.1637, df = 1, P < 0.01). Results for Munghorn Gap Nature Reserve

95 100 A

75

50 Percent

25

0 A. costata E. punctata E. gummifera E. haemastoma E. piperita 100 B

75

50 Percent

25

0 E. oblonga E. punctata E. rossii E. crebra E. melliodora 100 C 75

50 Percent

25

0 E. blakelyi E. punctata E. rossii A. floribunda E. crebra E. oblonga

Figure 6.1 Expected and observed use of foliage-gleaned trees by Anthochaera carunculata , 1992-1993, at a) Royal National Park b) Clandulla State Forest and c) Munghorn Gap Nature Reserve (n = number of 10 second observation periods). Eucalyptus punctata was strongly selected overall. (Royal National Park c2 = 1427.88, n = 1044, df = 3, p < 0.001) (Clandulla State Forest c2 = 1703.52, n = 968, df = 3, p < 0.001) (Munghorn Gap Nature Reserve c2 = 2144.93, n = 947, df = 4, p < 0.001)

96 100 A

75

50 Percent

25

0 E. oblonga E. punctata E. rossii E. crebra E. melliodora

100 B

75

50 Percent

25

0 E. blakelyi E. punctata E. rossii A. floribunda E. crebra E. oblonga

Figure 6.2 Expected and observed use of foliage-gleaned trees by Philemon corniculatus , 1992-1993, at a) Clandulla State Forest and b) Munghorn Gap Nature Reserve (n = number of 10 second observation periods). Eucalyptus punctata was strongly selected overall. (Clandulla State Forest c2 = 453.95, n = 233, df = 3, p < 0.001) (Munghorn Gap Nature Reserve c2 = 601.12, n = 380, df = 4, p < 0.001)

97 were similar (seasons: c2 = 60.689, df = 3, P < 0.001; years: c2 = 905.11, df = 3, P < 0.001). Although there were significant differences between seasons at Clandulla State Forest (c2 = 275.48, df = 3, P < 0.001), the two sampling years were not significantly different (c2 = 1.054, df = 1, P > 0.05). For P. corniculatus, at Clandulla State Forest, there were no significant differences between seasons (c2 = 0.048, df = 1, P > 0.05) or years (c2 = 3.9861, df = 1, P > 0.01) in their selection of E. punctata foliage. There were, however, significant differences between seasons (c2 = 13.341, df = 3, P < 0.01) and years (c2 = 7.6961, df = 1, P < 0.01) at Munghorn Gap Nature Reserve.

6.4 Discussion In general, A. carunculata and P. corniculatus preferred E. punctata to other species. However, A. carunculata spent substantially more time in winter gleaning the foliage of E. crebra at Clandulla State Forest, and at several other species when E. punctata was not present.

The association with E. punctata was strong. At Clandulla State Forest, several other species of honeyeater also selected E. punctata for foraging and ignored other species. These preferences suggested that E. punctata provided a superior food source to other species. At all sites, eucalypt species preferentially selected belonged to the subgenus Symphyomyrtus. This subgenus was represented only by E. punctata at the Royal National Park, by E. crebra, E. melliodora and E. punctata at Clandulla State Forest, and by E. crebra, E. blakelyi and E. punctata at Munghorn Gap Nature Reserve. All species used at Goobang National Park belonged to this subgenus. The majority of other eucalypt species present belonged to the subgenus Monocalyptus, except E. gummifera, which belonged to Corymbia. Eucalyptus oblonga (subgenus Monocalyptus), the most common species at Clandulla State Forest, was seldom selected. This pattern of selectivity has been previously observed at other sites where arthropod density in the subgenus

98 Symphyomyrtus was higher than in Monocalyptus (Woinarski 1985). Symphyomyrtus species have also been found to have significantly greater Psyllid densities on foliage compared with Monocalyptus species (Majer et al. 1997). No data were found on psyllids or other invertebrates on E. punctata.

Anthochaera carunculata and P. corniculatus removed lerp from foliage of E. albens. At Munghorn Gap Nature Reserve a small group of trees on a slope had foliage that was close to the ground and the Lerp were large enough to confirm that the birds were removing it. Lerp, manna and honeydew have been described as important alternative carbohydrates to nectar for honeyeaters (Paton 1980). Although P. corniculatus was not present at Clandulla State Forest in winter, A. carunculata remained and their predominant behaviour was foliage gleaning. In contrast, P. corniculatus was more common on the Western Slopes during winter when nectar was available (see Chapter 4.3.2). On the Western Slopes, both species foliage gleaned less frequently than they flower-probed when nectar was abundant. In contrast, at the Royal National Park, A. carunculata mostly gleaned foliage, while P. corniculatus was at alternative food sources elsewhere within the Park.

Anthochaera carunculata spent substantially more time foliage gleaning than P. corniculatus (42 % and 14 % of total foraging time respectively, see Chapter 5.3.1), and have shorter bills (25 mm and 39 mm respectively, Longmore 1991). It was hypothesised that because of their longer bills, P. corniculatus forage less efficiently in foliage than A. carunculata, and hence leave the Great Divide in the winter to seek food elsewhere.

99 7. GAPE WIDTH AND PREY SELECTIVITY IN Philemon corniculatus AND Anthochaera carunculata

7.1 Introduction Optimal foraging theory predicts that predators will select those prey that maximise either energy gain per prey item or per unit foraging time, except when the prey do not satisfy the predator’s nutritional requirements (Goss- Custard 1977, Krebs et al. 1977, Avery et al. 1993).

Factors limiting maximum prey size are the gape width of a predator (Zaret 1980, Wheelwright 1985), prey handling time (Recher & Recher 1968, Goss- Custard 1977, Krebs et al. 1977, Sherry & McDade 1982, Avery et al. 1993) and prey defence (Webb & Shine 1993). The relative importance of each of these factors varies considerably for each predator and its prey. Birds will often spend considerable time bashing prey to improve its palpability, however, swallowing prey whole is a feature of many birds (Lack & Owen 1955, Goss-Custard 1977). Food items are often swallowed whole by frugivorous birds (Wheelwright 1985, White & Stiles 1991, Avery et al. 1993,) and insectivorous birds (Lack & Owen 1955). Search time is dependent upon prey availability, and prey handling time by birds is generally determined by prey size rather than prey defences (Sherry & McDade 1982). The bill of a bird affords a greater advantage over other predatory tools in this respect as prey can be killed at a safer distance from the face. Therefore, the main factor determining maximum prey size in birds is gape width.

Anthochaera carunculata and P. corniculatus are sympatric in south-eastern Australia and occupy similar ecological niches (Blakers et al. 1984, Saunders 1993). Both are generalised feeders, consuming nectar, fruit, seeds, insects and nestlings (Blakers et al. 1984, Longmore 1991). However, there is little

100 published information on the insectivorous component of their diets. Barker and Vestjens (1990) found cicadas in the diet of both species, but Lea and Gray (1936) recorded them only in that of P. corniculatus; neither of these studies identified the cicada species. I observed P. corniculatus prey upon insects, for example scarab beetles, cicadas, stick insects and grasshoppers, that were larger than those normally taken by A. carunculata. Nearly all insects were ingested whole and head- first by both species. Redeye Cicadas, , and Black Prince Cicadas, Psaltoda plaga (Hemiptera: Cicadidae), are favourite food items of P. corniculatus during spring/summer. Philemon corniculatus took these cicadas while foraging alongside A. carunculata, which ignore them.

To determine if the differences in prey preferences were related to gape- limitation, a study of gape widths of both honeyeaters and maximum widths of both cicadas was undertaken.

7.2 Methods All New South Wales specimens of the A. carunculata and P. corniculatus in the collection of The Australian Museum were measured for gape width, while their specimens of Ps. moerens and Ps. plaga (Moulds,1990) were measured for prey width. Measurements to the nearest 0.05 mm were made with Vernier Callipers. The width of a cicada’s head was measured at the widest point when viewed dorsally, i.e. the distance between the outer edges of the eyes. The bill gape width of birds was measured at commissural points of the gape. Comparisons with skulls of these two honeyeaters revealed that the commissural points are very close to the junction of the bill with the skull and that shrinkage is minimal at this point. A more flexible gape in a live bird is still limited by skull dimensions.

There is no evidence of geographical variation in size of these honeyeaters

101 within New South Wales and the nature of their movements within this state is poorly understood (Saunders,1993). The sizes of both cicadas can vary considerably; generally they are larger in wetter seasons, rather than showing geographical variation (M. Moulds pers. comm.). Data were therefore pooled for each cicada and bird species. Honeyeater adults and juveniles were measured, provided the tail of the juveniles was at least two thirds adult length because juvenile P. corniculatus of this size in the field were observed to be fed whole cicadas.

Data were tested for normality using a Kolmogorov-Smirnov Goodness of Fit and found to be neither normal nor log-normal, therefore, the Mann- Whitney U-test was used to compare the medians of the samples (Zar, 1984) using Statview Student®.

7.3 Results Gape and prey widths are presented in Figure 7.1. The mean (± standard error) gape widths of A. carunculata and P. corniculatus were 10.12 ± 0.13 and 12.56 ± 0.18 mm, respectively. The mean prey width of Ps. moerens was 13.35 ± 0.08 mm compared with 12.73 ± 0.08 mm for Ps. plaga. The mean gape width of A. carunculata was significantly smaller than that of P. corniculatus and the mean widths of each cicada species (all P<0.001). The mean gape width of P. corniculatus was also significantly smaller than the mean width of Ps. moerens (P<0.001) but not Ps. plaga (P>0.05). However, only two Ps. moerens specimens (two percent of measured specimens) were larger than the maximum P. corniculatus gape width.

When the sizes of prey species were combined, only two percent were larger than the maximum gape width of the P. corniculatus sample, whereas 68% were larger than the maximum gape width of the A. carunculata sample. Nine percent of gapes of P. corniculatus were too

102 14 12 10 Anthochaera carunculata (n = 56) 8 6 4 2 0

10

8 Philemon corniculatus 6 (n = 38)

4

2

0

20

Frequencies 16 Psaltoda moerens 12 (n = 53)

8

4

0

20

15 Psaltoda plaga (n = 71) 10

5

0 7 8 9 10 11 12 13 14 15 16 17 Width (mm)

Figure 7.1. Gape widths of Anthochaera carunculata and Philemon corniculata. Prey widths of Psaltoda moerens and Psaltoda plaga.

103 small for any of the sample population of cicadas, whereas 80% of gapes of A. carunculata were too small.

7.4 Discussion Even though there is some sample overlap between prey size and A. carunculata gape width they were not observed to forage for these cicadas. This could be because A. carunculata learn that they are not suitable prey due to their size. They may attempt to ingest large cicadas but quickly learn that they are unable to swallow them. Hence they would soon learn to avoid cicadas because they would be non-profitable prey. This pattern of behaviour has been observed between many predators and potential prey in many birds (Recher & Recher 1968, Sherry & McDade 1982).

In most cases prey handling time and search time appeared to be similar whether the target was small hymenoptera or large cicadas. However, in some cases P. corniculatus was seen to make several attempts to swallow Ps. moerens before they were successful, but they were not observed to experience difficulty ingesting Ps. plaga. This suggests that Ps. moerens may be approaching the maximum prey size that P. corniculatus can ingest whole. They were not observed to discard large cicadas, although sometimes insects were observed to escape or out-manoeuvre the pursuing birds during sallies in the warmer part of the day.

During seasons when these cicadas are common, other smaller prey items are also abundant and are taken by P. corniculatus. Since search time is much longer than prey handling time for all insect prey items, it would not be profitable to ignore smaller insects and concentrate only on large prey (Herrera 1978, Thompson 1978, Sherry & McDade 1982, Webb & Shine 1993). However, when feeding dependent young it is more profitable to

104 feed them optimally sized prey10 since each food packet would contain more energy and nutrients, and incur equal transport costs; thus promoting higher survivorship of the young. Breeding success of P. corniculatus may be closely correlated with cicada density (H.A. Ford pers. comm.). For example during the drought of the summer of 1993/94 cicadas failed to emerge at Norton’s Basin, Wallacia and at Blue Mountains National Park. Philemon corniculatus did not breed during this period and generally did not remain at these sites, although they had bred there in previous summers. With appropriate rains and the emergence of cicadas in abundance during the summer of 1994/95, P. corniculatus density was much higher and breeding was recorded again at these sites (Saunders, unpublished data).

According to optimal foraging theory, it is most profitable in terms of energy intake, to take the largest prey items that can be consumed (Lack & Owen 1955, Goss-Custard 1977, Krebs et al. 1977, Zach & Falls 1978, Zaret 1980, Perrins & Birkhead 1983, White & Stiles 1991). When common, the search time for cicadas is not substantially different from that of other prey and since the insects are swallowed whole, handling times are short compared with search time. Although not quantified, observations indicate that search time and prey handling time of cicadas and smaller prey classes are equivalent. Thus the most likely explanation for differences in prey selectivity between A. carunculata and P. corniculatus is gape width as a limitation on the maximum prey size that can be swallowed.

10 i.e. the maximum size of prey that can be ingested by dependent young rather than all size classes of prey up to the maximum.

105 8. DISCUSSION

8.1 The nature of food resources selected by Noisy Friarbirds and Red Wattlebirds in the landscape and their effects on distribution and foraging behaviour. The distribution of A. carunculata and P. corniculatus was found to vary between sites, years and seasons and reflected the selection of food sources by each honeyeater (see Chapters 4 & 5). Philemon corniculatus was not recorded at coastal sites except in occasional low numbers, while the Royal National Park was an important site for A. carunculata in all seasons except summer. The pattern of abundance across the other three seasons was not the same between 1992 and 1993, peaking in winter in 1992 and in autumn in 1993 (see Figure 4.2). Philemon corniculatus was found at tableland sites mainly during spring and summer. However, there was variation in densities between years (see Figure 4.2). Anthochaera carunculata was seldom detected during the study at Kings Tableland, but was found in all seasons for both years at Clandulla State Forest. At sites on the western slopes densities of both honeyeaters fluctuated considerably with no regular patterns, but densities were positively correlated with flowering trees (see Chapter 4). The emerging pattern from this study is that these two honeyeaters tend to be nomadic over most sites, although A. carunculata is a regular visitor to Royal National Park and is largely resident on the tablelands, where P. corniculatus is a regular spring-summer visitor. However, the numbers of each honeyeater vary considerably even at sites where they are regular visitors.

Anthochaera carunculata has been found to be resident at national (Blakers et al. 1984, Barrett et al. 2003), state (Emison et al. 1987) and regional levels (Saunders 1993), with regular altitudinal movement from high altitude on the tablelands to lower altitudes during autumn and winter (Taylor 1992, Er

106 et al. 1998). It is a summer migrant to southern coastal New South Wales (Recher and Holmes 1985, Gregory-Smith 1991), while in other coastal areas is considered resident (Morris 1986, Kutt 1996, Saunders 2002) or nomadic (Loyn 1980, Smith 1989, Leishmann 1994, Egan 1997, Egan et al. 1997). At many sites on the tablelands it was found to be resident (Balfour 1980, McFarland 1986, Taylor 1987, Ford and Debus 1994) or nomadic (Baldwin 1973, French et al. 2003). On the western slopes A. carunculata is nomadic and this relates to eucalypt flowering (Morris 1984, Schrader 1987, Traill et al. 1996, Mac Nally and McGoldrick 1997). Nomadism on the coast and tablelands is also considered to relate to flowering of eucalypts (Smith 1989, Mac Nally 1996). In the Hunter region they are considered to be resident but with local movements following flowering eucalypts (Morris 1975). At the national level, the range of P. corniculatus shrinks towards the southeast during summer, with lower reporting rates in inland Queensland and higher reporting rates in eastern Victoria (Barrett et al. 2003). On the coast they are more common in the north during winter and in the south during summer (Smith 1989, Catterall et al. 1991, Gregory-Smith 1991, Saunders 1993, Slater 1995, Martin and Catterall 2001, Farmer et al. 2004). On the tablelands they are considered to be summer migrants (Emison et al. 1987, Taylor 1992, Saunders 1993, Ley et al. 1997, Ford 1998, Lindenmayer et al. 2002). On the western slopes P. corniculatus is considered to be nomadic in relation to eucalypt flowering (Morris 1984, Schrader 1987). They are also nomadic at some sites on the tablelands (McFarland 1985, Osbourne and Green 1992, French et al. 2003) and on the coast (Hindwood 1944, Morris 1986, Smith 1989, Egan et al. 1997, Saunders 2002).

Reports of the distribution patterns for each honeyeater were found to vary within regions. These were generally derived from single site studies (£ 100 ha) that have been undertaken for short periods (£ 1 year), e.g. during only breeding seasons. Studies in adjacent areas and studies at different times in

107 the same area have shown different patterns of occurrence for these honeyeaters. This is likely to be an artefact of the sampling strategy and probably represents a ‘snapshot’ in time and/or space of their behaviour where they have a variable response in distribution over time and space. Morris (1975) describes A. carunculata as resident in the Hunter region (an area of 606575 ha and many different habitat types) with local movements in response to flowering eucalypts. This means that A. carunculata was found somewhere within the County at any one time, but that the actual location within the County depended on where eucalypts were in flower. Thus at a regional scale this honeyeater would be considered to be resident while at the local scale it would be considered nomadic. It is difficult to describe the distribution pattern for a species from small scale studies (single sites £ 100 ha and £ 1 year) as they do not record patterns outside the study site or at different times. However, broader studies have revealed variation between years (e.g. Balfour 1979, Smith 1989, Leishmann 1994, Egan 1997) and between sites (e.g. Saunders 1993, Mac Nally and McGoldrick 1997, Martin and Catterall 2001). These studies confirm that movements are often irregular in response to variable flowering of nectar sources.

Most previous studies use presence or absence of these two honeyeaters when describing occurrence without knowledge of their occurrence, or movements outside the study area. Although many of these studies relate occurrence to the presence of flowering eucalypts and other flowering shrubs, it is only when each is viewed in context with other studies that they begin to elucidate movement patterns for these two honeyeaters. However, patterns from one area cannot be extended to other areas unless there are sufficient studies from these other areas that reveal similar patterns. The broader scale at which this study was undertaken has meant that both spatial and temporal patterns of distribution were explored simultaneously. Many of the patterns observed in the current study matched patterns found in

108 other studies from outside the area of central overlap of the range of these two species. Thus it may be argued that the patterns found in this study may apply over much of the overlap range of these two honeyeaters. Some studies in a particular area (e.g. Central Coast)describe these two honeyeaters as resident (e.g. Morris 1975), while others in the same area describe them as migratory (e.g. Hindwood 1944) or even nomadic (e.g. Egan 1997). It can now be seen that these different descriptions of their distribution do not conflict and that when they are viewed together produce a similar model of movement patterns to the one found in this study. While the honeyeaters may be resident at a regional scale they may also be nomadic at the local scale. This study confirmed that there was movement in relation to flowering eucalypts at both regional and local scales, although other food sources were also important as they were also found foraging at sites where nectar was not available (see Chapter 5). Data from the three western slope sites (i.e. Munghorn Gap Nature Reserve, Goobang National Park and Back Yamma State Forest) add to our knowledge of the movement of these two honeyeaters in this region as previous studies from this region are few. The data suggested that both honeyeaters are highly nomadic within this region and that their occurrence is closely related to tracking flowering eucalypts. Along a north-south transect across central Victoria, Mac Nally and McGoldrick (1997) found a similar trend in movement patterns between tableland and inland slope sites in relation to flowering eucalypts. In this study and others, food resource variation was seen to strongly influence the distribution of these two honeyeaters.

Anthochaera carunculata and P. corniculatus were found to be mainly canopy foragers spending most of their foraging time at flowers, particularly at eucalypts, where flower probing occupied as much as 85% of their time (Chapters 4 & 5). Numbers of both honeyeaters were strongly influenced by the density of flowering trees at both local and regional scales and

109 positive correlation between flower-probing and aggression suggested that nectar was an important food source (Chapter 4). This was not found for other food sources. Foliage gleaning was the next most common foraging activity and both preferred the foliage of Eucalyptus punctata when this was available (Chapter 6). Other foods selected by both honeyeaters included fruit, insects and seeds (Chapters 5 & 7). However, there were some differences in food selection between the two species. Philemon corniculatus spent more time foraging at flowering trees than A. carunculata, which spent more time foraging at foliage and bark than P. corniculatus (Chapters 4 & 5). Philemon corniculatus was also able to take larger insect prey (Chapter 7). Anthochaera carunculata spent a greater proportion of their time foraging than did P. corniculatus (Chapter 5).

Similar foraging patterns were found in other studies. Anthochaera carunculata was found to spend more time foraging than P. corniculatus (Hugh Ford pers. com.). Flower probing was the most common foraging behaviour for both honeyeaters, but was more important for P. corniculatus (Ford et al. 1986). Gleaning foliage, followed by gleaning bark were the next two most common foraging behaviours recorded in that study. Nectar from eucalypts was found to be an important food for both honeyeaters (Hindwood 1944, Barker & Vestjens 1990, Lepschi 1993, 1997) and taking nectar from E. sideroxylon has been recorded (Shelly 1998). Nectar from Banksia (Hindwood 1944, McFarland 1986, Barker & Vestjens 1990) and from Grevillea (Osbourne and Green 1992, Lepschi 1997) was also taken by both honeyeaters. Philemon corniculatus has also been found to take Lerp (McCulloch 1997), large insects (Rose 1973, Gregory-Smith 1991, Rose 1999) and in particular small cicadas (Lea & Gray 1936, McCulloch 1990, Lepschi 1993), as well as seeds (Lea & Gray 1936) and fruit from Exocarpus sp. (Barker & Vestjens 1990, Rose 1999).

110 Differences between this and other studies were also found. Anthochaera carunculata was found to forage higher than P. corniculatus (Ford et al. 1986). Anthochaera carunculata was found to feed on mistletoe fruit (Lea & Gray 1936), nectar from Hakea and Callistemon (Lepschi 1997), fruit from Ficus (Rose 1999) and Psaltoda moerens (the cicada species discussed in Chapter 7, Lepschi 1999). Philemon corniculatus was also found to feed on nectar from Hakea and Callistemon also (Lepschi 1999). Hakea and Callistemon were not detected at any of the sites in this study, but the other foods listed above were available. Anthochaera carunculata was not found to take seeds or fruit from Exocarpus sp. in any of these other studies and P. corniculatus was not recorded taking seeds from Acacia, whereas both these seeds and fruit were taken at Munghorn Gap in this study by both honeyeaters.

A total of 64 surveys were undertaken across seven sites, two years (four years at Goobang National Park) and four seasons in each year. These sites were along a 350 km transect from the coast, through the tablelands and onto the western slopes. Because of the area and time period covered in this study, it is not surprising that most of the foods recorded from other studies were also recorded here. Additional foods were added to the list of items taken during each new survey at many of the sites and it is likely that more would have been added if the survey period had been extended. In addition, foods not available from sites in this study were available outside the area. These data indicate that both honeyeaters have broad diets and are able to switch between available foods.

Differences in the morphologies between Anthochaera carunculata and P. corniculatus may account for the differences observed in their food selection. Philemon corniculatus has a larger bill which enables it to ingest larger prey items, but which may restrict its ability to glean foliage for lerp and manna. In contrast, A. carunculata was found to foliage glean at sites which P.

111 corniculatus had vacated. Anthochaera carunculata is larger than P. corniculatus (Longmore 1991), and if it is confined to forage on smaller items it would need to spend more time foraging than P. corniculatus to gain sufficient food to satisfy energy and nutritional requirements.

Foods selected by these two honeyeaters were mostly different between seasons and years at a site and between sites (Chapter 5). The phenology of flowering and subsequent fruiting and seed production were highly variable between years and often between sites for the same species of plants. There was also variation between densities of flowering species within transects and between transects within a site (Chapter 4). This may be due in part to the variability in tree species composition between plots along a transect (Chapter 3). The species composition and phenology of plant species appeared to vary at the patch, site and landscape scales. The seasonal availability of many suspected foods was also highly unpredictable. Several tree species were expected to be in flower during particular seasons at many sites, but were often found not to do so. For example, E. sideroxylon flowers mainly in winter and spring on the western slopes, but at Goobang National Park and Back Yamma State Forest flowering was found to be highly variable between years. At Goobang, flowering was around 60 trees per hectare in winter and spring in 1993, three and 25 trees per hectare in 1994, no flowering in 1995 and an increase to 30 and 40 trees per hectare in 1996 in each season respectively. Similar irregularities have been found for flowering eucalypts used by these honeyeaters (Mac Nally 1997, Lindenmayer et al. 2002). Thus many of the foods used were highly unpredictable over both space and time.

The unpredictable phenology of many flowering plants and, as a consequence the foods they provide, has been noted for many Australian species (e.g. eucalypt flowering, Law et al. 2000, fruit, Stanley & Lill 2002).

112 Emergence of cicadas is also unpredictable between years (Moulds 1990). In these studies good flowering and fruiting, and emergence of cicadas was found to follow periods of good rainfall. Even in areas where rainfall is seasonal, the amount received can be highly variable between years and is likely to affect the availability of many food sources. The lack of predictability of food sources in Australia is likely to have consequences for foragers dependent on such foods (Mac Nally 2000). They need to be mobile and able to detect resource rich areas as they become available. Generalist foragers that are able to switch between foods are well adapted to cope with this resource pattern.

There was no single site where P. corniculatus was detected during every visit (Chapter 4). At many sites it was absent for particular seasons during both years. There were only two sites where A. carunculata was recorded during all surveys and numbers fluctuated considerably between surveys. Numbers of both honeyeaters fluctuated with numbers of flowering trees, especially at sites on the western slopes. Although A. carunculata would still persist in low numbers at such sites when flowering was poor, P. corniculatus would often vacate these sites. Anthochaera carunculata was recorded gleaning for insects in all seasons although they may not have been present at all sites during any particular season. However, P. corniculatus was only present at non-flowering sites during warm seasons when insects were likely to be more abundant. If flowering was poor at a site during winter they were likely to be absent. Thus food patterns appear to influence the occurrence of these two honeyeaters, and the influence appears to be stronger for P. corniculatus . In other studies of these two honeyeaters it was found that their occurrence was determined by the flowering of eucalypts (e.g. Mac Nally 1997, Mac Nally & McGoldrick 1997, Lindenmayer et al. 2002). Honeyeaters were observed to avoid sites when eucalypt nectar was not available. The unpredictability of their food sources means that regular

113 migration between sites for preferred foods is unlikely, and that a nomadic existence is more in keeping with their food resource patterns.

Overall, in response to the temporal and spatial variation of their food resources, Anthochaera carunculata and P. corniculatus are best described as nomadic generalists.

8.2 Niche overlap and competition between A. carunculata and P. corniculatus . Niche overlap was highly variable between sites, seasons and years with little discernible pattern (Chapter 5). Niche breadth was narrow with high overlap at sites with high flowering tree density and when foraging was concentrated on a few flowering trees at sites on the tablelands and on the western slopes during winter and spring. This suggested that when particular resources were abundant these honeyeaters concentrated on this resource. Theory predicts that when a food source becomes abundant, foragers will concentrate at the resource with a corresponding narrowing of niche and greater niche overlap (Ford 1989, Schoener 1986). This study provided some support for this relationship at flowering trees. Niche overlap was found to vary between 31% and 100% with a mean value of 66% in this study. Ford et al. (1986) found that foraging niche overlap between A. carunculata and P. corniculatus was 83% based on foraging height, foraging method, substrate type and species. However, this was calculated from pooled data from two close sites on the northern tablelands and from data collected over three years at these sites. No comparisons were made between sites or sampling periods. No other measures of niche overlap were found in the literature.

Aggression between these two honeyeaters was less than two percent of total time and much of this is intraspecific (Chapter 5). Intraspecific

114 competition is more likely to limit populations as there is a greater similarity of requirements within a species than between species. This may facilitate coexistence between species (Underwood 1986). In this study, the proportion of aggression that was intraspecific, suggested that intraspecific competition may be stronger than interspecific competition. This would be expected as niche overlap between individuals within a species is likely to be greater than between species. Interspecific aggression was found to positively correlate with niche overlap, which indicated that these two honeyeaters compete for, and defend, shared resources. Some interspecific competition therefore exists between A. carunculata and P. corniculatus at shared resources.

When resources were abundant (e.g. flowering trees), A. carunculata and P. corniculatus often coexisted at many of the sites. However, they appear to be seldom together at any one site and using similar food sources for more than a few months. The large fluctuations in numbers of each bird species and frequent changes in foraging resources selected, suggested that equilibrium between resources and populations of these two honeyeaters is unlikely to occur. Interspecific competition implies that a competitor is either excluded, suffers a reduction of its population or shifts its ecological niche so that coexistence is possible (Cody 1974, Krebs 1978, Begon et al. 1996), but also requires that environments are stable and not fluctuating (Krebs 1994), and that potential competitors are together for sufficient time for niche differentiation to occur (Mac Nally 2000). None of these conditions were met by either honeyeater or their food resources and hence competition between these two honeyeaters is unlikely to shape foraging niche. Instead it is more likely to be shaped by the temporal and spatial variability of their food resources.

8.3 The ecological role of Noisy Friarbirds and Red Wattlebirds in the

115 landscape. The long-term maintenance of some plant communities in Australia is considered to be dependent on maintaining populations of bird pollinators (Paton et al. 2004). A decline in honeyeater numbers has lead to declines in seed set in some Australian plant communities (Paton et al. 2004). Anthochaera carunculata and P. corniculatus spend much of their foraging time at flowers and are likely to be important pollinators. As they also include fruit and seeds in their diets, they are also likely to aid seed dispersal. When they spend time foraging for insects, which are taken during foliage and bark gleaning and during sallies from perches, they may help reduce the impacts of insects on vegetation. For example, both honeyeaters will take Plague Locusts Chortoicetes terminifera in open forest and woodland areas (pers. obs.). Hence they are likely to have an important ecological role in the landscape.

They will forage in continuous forest, open forest, woodland, pastureland with scattered trees, orchards, parks and gardens (Pizzey & Doyle 1980) and in areas of revegetation (Paton et al. 2004). They have also been observed to use remnant trees, occurring singly or in twos or threes, several kilometres from the nearest other tree and have been observed crossing large open areas between trees that separate habitat patches by as much as 10-20 kilometres (pers. obs.). When survey work was undertaken between sites on the tablelands and the western slopes, both honeyeaters were encountered in these areas at different times. That is, when they occurred on the tablelands they were often absent on the western slopes, and when they were present at sites on the western slopes they were often absent from the tablelands. These sites were often separated by as much as 200 kilometres with no encounters with either species in the intervening area, despite the occurrence of suitable habitat. Banding records of P. corniculatus have found that it may move as much as 510 kilometres between banding and recapture

116 sites (Anon 1980) and A. carunculata has been recorded up to 200 kilometres from where it was banded (Purchase 1969). It is likely that movements on this scale are frequent, but recapturing banded birds of these two species would be infrequent as they spend a high proportion of their time foraging in the canopy. Hence, these birds are highly mobile and will move through the landscape matrix to reach suitable habitat patches hundreds of kilometres away from their last feeding patch. Movements on this scale, in relation to flowering eucalypts, have been suggested from other studies (e.g. Mac Nally & McGoldrick 1997). Paton et al. (2004) recorded movements of A. carunculata of approximately 100 kilometres when studying the effects of habitat fragmentation on the foraging of honeyeaters at nectar sources.

Habitat fragmentation and understorey thinning favour edge specialists such as Manorina melanocephala (Loyn 1987, Major et al. 2001, Piper & Catterall 2003). This bird species will aggressively exclude other woodland birds. particularly smaller insectivorous birds (Dow 1977, Loyn 1987, Major et al. 2001, Parsons 2001, Piper & Catterall 2003). They are largely sedentary and territorial (Dow 1977, Higgins et al. 2001). As they occupy fragmented habitat and displace other bird species within small habitat patches and narrow corridors (Major et al. 2001, Piper & Catterall 2003) they are likely to cause change in community structure and habitat patch quality across the landscape. Loyn (1987) found that small heavily grazed patches of woodland contained M. melanocephala but few other canopy insectivores and had signs of dieback due to insect damage. Thus, it is unlikely that they fill all the niche space vacated by the displaced birds. In addition, because they are a sedentary species, they are less likely to fill the pollination and seed dispersal role of some of the species they replace. Anthochaera carunculata and P. corniculatus can defend themselves against aggressive and territorial M. melanocephala, and are able to coexist with them in habitat patches (pers. obs.).

117 Because Anthochaera carunculata and P. corniculatus are highly mobile, nomadic generalists that appear to cope with a high degree of habitat fragmentation and habitat modification they are likely to have an important role as pollinators, seed dispersers and insect population limitation in the landscape. Because they are able to coexist with M. melanocephala, while others are not, they may also account for an increased proportion of this role since clearing began.

8.4 The relevance of scale to studies of nomadic generalists in unpredictable environments. Food resource selection by these two honeyeaters was found to vary considerably and often unpredictably over the spatio-temporal scale of this study. Although variation in foods selected and occurrence within a site over two years or more could support the nomadic generalist view of the behaviour of these two honeyeaters, evidence from several sites increases the strength of such assertions. In several studies undertaken at small sites over several years it has been found that A. carunculata and P. corniculatus are possibly nomadic, as suggested by irregular visits to such sites (e.g. Smith 1989, Mac Nally 1996) but larger scale studies provide stronger evidence as several sites over a large regional were monitored simultaneously (e.g. Mac Nally and McGoldrick 1997, Paton et al. 2004). With smaller scale studies it is difficult to determine where the absent birds have gone and to estimate how far they are likely to have travelled. The spatial scale of this study provided strong evidence for nomadism in these two honeyeaters.

Previous studies of bird behaviour have been undertaken at scales ranging from small plots or territories (e.g. Holmes & Robinson 1981, Lundquist &

118 Manuwal 1990, Greenberg et al. 1993) to sites over several kilometres (e.g. Brotons et al. 1998). The scales in these studies were chosen to match the questions being considered. For example, small scale studies found relationships between density of honeyeaters and flowering trees at small sites (e.g. Smith 1989, Mac Nally 1996) while regional scale studies looked for patterns of occurrence of honeyeaters through the landscape in relation to spatial variation in nectar sources at the landscape scale (e.g. Mac Nally and McGoldrick 1997, Paton et al. 2004). These latter studies are better suited to the investigation of the scales at which the environmental processes which determine distribution of nectar sources are likely to operate. Hence, broad scale studies are also more likely to be relevant to many questions about bird behaviour as they need to match the scales of the environmental processes that affect them.

In this study the spatial scale was very large and consisted of seven sites spread over a 350 kilometre east-west transect, each containing two 30 to 50 hectare transects. The temporal scale was four seasons over two years at six of the sites, and over four years at Goobang National Park.

The scales used in many previous studies of bird behaviour were generally small and often undertaken at only one site (e.g. Beaver & Baldwin 1975, Hejl & Verner 1990, Sun & Moermond 1997), although they sometimes consisted of several plots. Sites have ranged from 10 hectares to about 1200 hectares, which in some cases were as much as a forty kilometres apart (e.g. Szaro et al. 1990). The duration of these studies ranged from days, through months in a year, to several years, although the latter was much less common. Often, multiple year studies were undertaken during breeding or non- breeding seasons within those years (e.g. Goss-Custard 1977, Airola & Barrett 1985, Hejl & Verner 1990, Kellner et al. 1990), particularly in northern hemisphere sites where seasons are more distinct. Often the data were

119 pooled across years and no inter-year comparisons were made (e.g. Goss- Custard 1977, Holmes & Robinson 1981, Airola & Barrett 1985).

Australian studies have been undertaken at similar spatial and temporal scales to elsewhere. Where two or more sites were involved, often data were pooled (e.g. Ford et al. 1986, Er 1997) and if data were collected during breeding seasons over two or more years, again data were often pooled (e.g. Rice 1978, Robinson & Holmes 1982). However, there were also many studies undertaken over a single year where data were collected into months or seasons, and where variability over this time scale was examined (e.g. Recher 1989, Cale 1994). Many of the longer studies collected continuous data over the study period in Australia (e.g. all seasons over three years: Bell 1985, every month for three years: Bell & Ford 1990, monthly for three years: Ford et al. 1990), whereas many northern hemisphere studies were discontinuous and data were collected in one part of the annual cycle over the years of the study (e.g. two consecutive summers: Beaver & Baldwin 1975, two consecutive breeding seasons: Airola & Barrett 1985, and Hejl & Verner 1990). Discontinuous studies have been undertaken in Australia, but most of these relate to breeding (e.g. Ford 1999) and not foraging behaviour, except, for example Mac Nally (1996), which examined foraging by honeyeaters in three consecutive winters. These proportions of continuous and discontinuous studies reflect that seasonal variation throughout much of Australia is not as sharp as experienced in continental northern hemisphere sites, and that northern hemisphere birds may show more distinctive seasonal behaviour.

Behavioural studies involving nectarivorous birds outside Australia have been mostly carried out at smaller spatial scales (of the order between 0.1 to 5 ha) and have related mainly to feeding territories (e.g. Wolf 1970, Carpenter & MacMillen 1976, Gill & Wolf 1977). Temporal scales were also

120 much smaller and were sometimes only a few days (e.g. Wolf 1970, Gill & Wolf 1977). Some of these studies investigated diurnal variation in foraging behaviour (Gill & Wolf 1977). Where longer time frames were employed the data were often pooled and variation over the study period was not considered (e.g. Wolf 1970, Wolf & Hainsworth 1971, Yamgishi & Eguchi 1996). Only a few studies examined variation over months, seasons or between years (e.g. Carpenter & MacMillen 1976, Feinsinger & Swarm 1982).

Australian studies involving honeyeaters have employed similar scales to those applied to nectarivorous birds overseas (e.g. Carpenter 1978, Armstrong 1992), although, there have been several studies at sites ranging from 15 to 100 kilometres apart (e.g. Ford 1979, Chan 1990). Many studies were carried-out within one year and considered monthly or seasonal variation over the year (Pyke 1983a, Newland & Wooller 1985, Rasch & Craig 1988). Long term studies over several years were few and only some of these considered inter-year variation (e.g. Vaughton 1990, Slater 1994). Many of the studies dealt with communities of honeyeaters that consisted of a mixture of residents, nomads and migratory species (e.g. Collins et al. 1984, Newland & Wooller 1985). Some made reference to tracking resources at the local scale within a site (e.g. Collins et al. 1984). Specific behavioural studies of P. corniculatus and A. carunculata are few. The scales used varied from one tree (Ford 1981) to a site 240 hectare in area (Ford & Tremont 2000) and from a few weeks to a few months. In one study (Ford & Tremont 2000), breeding was investigated over nine years, although no inter-year comparisons were made. Some studies from other parts of the ranges of these two honeyeaters have found similar patterns of behaviour to those found in this study for P. corniculatus and A. carunculata while others have found different patterns. For example, P. corniculatus was found to be a summer migrant in the tablelands (Balfour 1980, Taylor 1992, Chan 1995, Ford and Debus 1994, Ford 1998, Saunders 1993) as well as in this study. In

121 other studies within the tablelands they were found to be nomadic (McFarland 1986, Osboune and Green 1992, French et al. 2003). This disparity between studies within the same region reflects the scales used. Landscape scale studies are more likely to account for movements or occurrence in the surrounding area while longer studies are more able to accurately discern temporal patterns of occurrence. However, there have been few studies involving either species, where the scales used are comparable with this study (e.g. Mac Nally & McGoldrick 1997, Paton et al. 2004). These latter studies found that occurrence of honeyeaters was often determined by their nectar sources and that such resources vary at a regional scale.

Most studies were not designed to explore large scale spatial-temporal variation, and hence could not be expected to represent behaviour of a species over their whole range nor over longer time periods. Several authors, both overseas and in Australia, point out that between year variation of resources exists and that studies need to expand the temporal range of investigation (Carpenter & McMillen 1976, Carpenter 1978). Variation within and between seasons, and between years suggested that caution is required when drawing inferences that are based on data from a single year or from data pooled over several years. Hejl & Verner (1990) found variation in species selected for foraging for each month within a season which would have been lost if data were pooled for season. Kelly & Wood (1996) found both diurnal and intra-seasonal variation in foraging behaviour that would have been masked by pooling for seasons. Miles (1990) and Szaro et al. (1990) both found variation in foraging behaviours between seasons and between years and concluded that aspects of community organisation would have been lost if data were pooled between years in the study of seasonal variation. In this study, year was a significant variable accounting for much of the variation in foraging substrates selected

122 (see Chapter 5). For example, at Munghorn Gap Nature Reserve, seeds and fruit were taken by both honeyeaters, but in alternate years. If both years had been pooled differences between the same seasons in alternate years would have been lost. It would have appeared that these honeyeaters took both seeds and fruit in summer without showing that taking seeds or fruit did not occur in all years. As many studies have dealt with resident or migratory birds, caution is also required when applying the same methods and scales to behaviour of nomadic species (Mac Nally 1995). Resident and migratory species have predictable occurrence in the landscape and so they are easier to target when studying foraging behaviour, while sites can be setup within a nomads range without them being present during any part of the study. Hence larger scales need to be considered when designing studies of nomads.

Differences in foraging between honeyeaters and other nectarivores also have implications for the scale of behavioural studies. Alternative foods seem to be more important for honeyeaters than for hummingbirds, and many hummingbirds show greater specialisation for different nectar sources and are more sensitive to temporal changes in abundance (Ford & Paton 1985). Hence models based on the assumption that nectar is the major component of diet may not hold for many honeyeaters (Craig & MacMillen 1985). The two honeyeaters in this study were seen to utilise insects, fruit, seed and other carbohydrates when nectar was not available (see Chapter 5). Australian habitats have also been shown to be richer in nectar than other habitats producing about twice as much energy per flower (Craig & MacMillen 1985). Furthermore, as many honeyeaters are highly mobile, investigations in one small area may ignore reliance on neighbouring habitat and not give a representative view of behaviour (Craig & MacMillen 1985, Rasch & Craig 1988). These mobile honeyeaters may acquire resources outside the study area and hence their real resource use is unknown. Studies

123 involving nomadic honeyeaters need to be on a much larger scale that matches the scale of movements and temporal variability of resources. Ford (1998) has suggested that the general pattern of dispersal and philopatry for non-sedentary honeyeaters was poorly understood and that there was a need for long-term studies of migratory or nomadic Australian .

There is a need to measure patterns of resources and birds over several scales simultaneously, because measurement at only one scale cannot be extrapolated to other scales and because important processes can be missed completely (Levin 1992, Huhta et al. 1998). Franklin & Noske (1999) argued that when exploring correlations between nectarivores and nectar the question should not be whether correlations occur but at what scale they occur. In this study, correlations between flowering trees and honeyeaters were investigated at local and regional scales to test whether flowering trees affected honeyeater density both across sites and between patches within a site. Positive correlations were found at both scales. There was a positive correlation between numbers of flowering trees and numbers of both honeyeaters in the forty one plots counted in spring 1994 at Goobang National Park as well as across all sites (Chapter 6). These correlations may be expected for individual sites where flowering trees are selected for foraging, but would not necessarily be expected across sites if flowering trees were seldom selected at all sites when available. This would occur if other resources were more important at some sites. In other words, the existence of correlations at the local scale does not imply that the same correlations apply at a regional scale. These need to be tested separately.

Studies of bird behaviour need to be undertaken at much broader spatial and temporal scales that match temporal scales with environmental variation over time, and spatial scale to the scales of variation in resources and bird

124 distribution patterns (Wiens 1989).

The importance of scale in such studies cannot be over-emphasised. Plot- based research may yield relatively little about the ecology of nectarivorous bird communities (Mac Nally & McGoldrick 1997) because this scale is not likely to match the scale over which birds are ‘experiencing’ resources (McGoldrick & Mac Nally 1998). The scale of studies needs to match the scale of the processes under examination or incorrect conclusions may be reached (Wiens 1981, 1989, Mac Nally & Quinn 1998). An example of this is provided by a comparison of studies investigating the relationship between occurrence of Noisy Miners and the diversity of woodland birds. As noted previously, Noisy Miners have been found to negatively impact on the diversity of woodland species, particularly small insectivorous birds (Dow 1977, Loyn 1987, Major et al. 2001, Parsons 2001, Piper & Catterall 2003). However, in a study of bird atlas data collected at the ten-minute grid scale, reporting rates for Noisy Miners were not negatively correlated with the reporting rates for most other woodland birds (Reid 2001). Study plots were of similar size to Noisy Miner territories (of the order of hectares to tens of hectares) in the former studies while the ten-minute grids (16 x 19 kilometres) used in Reid’s study, are very much larger than Noisy Miner territories. Although Noisy Miners will aggressively exclude smaller birds from their territories, there are likely to be many areas within a ten-minute grid that do not contain Noisy Miner territories and in which bird diversity would not be affected. The negative impacts of aggressive exclusion would operate at the scale of Noisy Miner territories and would be detected at this or smaller scales, and not at the landscape scale.

It was found that the large scales used in this study were necessary to establish the lack of regular patterns in foraging substrate selection by P. corniculatus and A. carunculata over sites and between years. An example of

125 differences between years has been discussed previously. Site was also found to account for much of the variation in foraging substrate selectivity. For example, fruit and seeds were only selected at Munghorn Gap Nature Reserve. If this site had been excluded from the study then no selection of fruit or seeds would have been detected. If this study had been confined to either one site over several years, or to all sites over one year, the spatial and/or temporal patterns would have appeared different.

The use of appropriate scales in ecological studies of birds is also important for other reasons. Resources were found to be seasonally unreliable with much between year variation and this needs to be considered when assessing habitat requirements. Conservation of habitat is critical for maintenance of bird community structure and so a deeper understanding is required of birds’ needs and the spatial and temporal scales at which they vary. More studies need to use scales that match the scales relevant to nomadic species, at least to the scales of this study.

8.5 Movement of birds in response to the distribution of food resources. The complex relationship between movement of birds and predictability of food resources was introduced in Chapter 1 (Section 1.1). If food resources are locally abundant and available over long periods of time, or if species can forage at a broad range of local food sources then foragers may be resident. If food resources are seasonally reliable but patchily distributed at a regional or continental scale, then foragers are more likely to be migratory and move in regular patterns between seasonally available resource rich patches. If however, food resources become locally abundant for short periods and are unpredictable in both space and time, then foragers dependent on such food resources are more likely to be nomadic. Thus bird movements at regional, continental or global scales are generally related to the patterns of their food resources in the landscape. Birds provide many examples of these

126 relationships between movement patterns and distribution of food resources.

The Platycercus elegans mainly inhabits forests, but can also be found in heathland, open woodland and partially cleared farmlands along the east coast of Australia. It is mostly sedentary, with local movements in winter (Forshaw and Cooper 1981). Seeds are taken from a wide variety of trees, for example, over twenty species of Eucalyptus, three species of Acacia, and species of Angophora, Allocasuarina, and Callitris have been recorded, and from thirty or more shrub species (pers. obs.). This species is thus able to utilise a wide range of seed sources. Even if the seed sources are seasonally unreliable they appear able to readily switch to alternate food sources. There is no need for this species to move great distances in search for food, and this would explain why only local movements have been recorded (Forshaw and Cooper 1981). The Superb Fairy-wren Malurus cyaneus and the Striated Thornbill Acanthiza lineata are resident for much of the year with local movement outside their breeding seasons (Serventy 1982). The habitat of M. cyaneus is dense shrub understorey across most woodland types, including weedy understoreys of Lantana camara, as well as gardens and parks, while A. lineata inhabits a wide range of eucalypt forest and woodland canopies (Serventy 1982). Both feed on insects and other small arthropods, taken from the ground or shrub foliage in M. cyaneus or from tree foliage in A. lineata (Serventy 1982). They are able to find food all year round within a local area and do not need to travel large distances in search of food. These three species utilise locally abundant and/or diverse local food resources and can be found in a range of habitat types including modified habitats. They do not need to cope with unpredictable or dispersed seasonal food resources, and hence do not need to migrate or exhibit nomadic behaviour. Their food selection allows them to remain in an area and hence would explain why they are considered residents.

127 When suitable foods are not available locally, birds must move in search of food. The Dunlin Calidris alpina is a regular annual migrant from High Arctic tundra to European tidal mudflats after breeding (van de Kam 2004). While in the High Arctic they take a variety of insects, spiders, snails and worms, and time their breeding so that chicks hatch during the peak abundance of terrestrial insect larvae as the chicks feed themselves on hatching. When the short summer is over they migrate to more southern mudflats where they feed on worms, small shellfish and crustaceans (van de Kam 2004). Food availability varies dramatically in their breeding grounds and requires that they migrate regularly and consequently switch between food types. The Leaden Flycatcher Myiagra rubecula is a regular migrant from northern Queensland and New Guinea to southern Australia in the summer where it is found in forests and coastal scrublands (Officer 1969). It spends much of its foraging time chasing flying insects (Boles 1988) and takes a broad range of these (Blakers et al. 1984), but flies predominate in the diet(Officer 1969). Pyke (1985) found that larger flying insects were more common in warmer months between October and March than at other times of the year near Sydney and that they consisted mostly of Diptera. The summer occurrence of M. rubecula in southern Australia coincides with this increase in flying insects. The Cuculus pallidus is also a regular summer migrant to the southern areas of Australia preferring forests, woodlands and partially cleared areas (Shields 1994). It feeds on insects and their larvae, particularly hairy caterpillars which are more common in warmer months (Shields 1994). Their summer occurrence may also be due to their hosts feeding the young mainly on insects at this time of year. The movements of these three migrants would thus be predictable in response to predictable food sources that show a marked seasonal variation in availability.

128 If, however, foods are not predictable in space or time regular migration may not direct foragers to areas of abundant resources. The Banded Stilt Cladorhynchus leucocephalus has complex and erratic movements with no consistent pattern (Lane and Davies 1987). It generally occupies highly saline shallow wetlands (Lane and Davies 1987) and moves from coastal and sub- coastal areas into inland areas after rain or flooding (Marchant and Higgins 1993). When inland salt lakes flood it feeds on abundant small crustacea, water beetles and molluscs and times its’ breeding to coincide with these events (Jones 1945). The coastal and sub-coastal wetlands comprise the non- breeding distribution for this species. As flooding of inland lakes is highly irregular and unpredictable, so is the occurrence of C. leucocephalus. Irregular rainfall can also determine the distribution of some terrestrial birds. The numbers of Stubble Quail Coturnix pectoralis will vary in most areas with the amount of rainfall (Marchant and Higgins 1993). Its’ food consists mainly of grass seeds and it prefers tall dense ground cover, particularly in natural and improved grasslands (Marchant and Higgins 1993). Such conditions are only provided when suitable rain falls in an area so that grasses grow and seeds form. This species has been observed to move into an area a month or two after good rains have promoted strong grass growth where they were absent for many years previously (pers. obs.). Both these species are nomadic in response to unpredictable food resources and are habitat specialists. This requires that they be able to track resources in some way and move long distances between resource rich areas, often over areas of unsuitable habitat in the landscape matrix.

In this study the occurrence of A. carunculata and P. corniculatus was found to respond to unpredictable food resources, particularly nectar from eucalypts. They were found to be typical nomadic generalists. Because they are able to switch between many different food types and forage in most types of woodland, partially cleared farmland, parks and gardens, they are more

129 likely to cope with fragmented habitat than many other bird species. They may use resources within the landscape matrix between patches of abundant food sources. Resident bird species restricted to remnant habitat patches will be limited by the size of the patch and the quality of resources within the patch. It is unlikely that such species will be able to move between patches in a highly fragmented landscape and cope with further habitat fragmentation. They are dependent on a patch providing for all their needs. Regular migrants are likely to suffer from habitat fragmentation if there is a net loss of suitable habitat between destinations when migrants use these as ‘stepping stones’ for resting or feeding, and if habitat is reduced at their destinations. Because nomadic species need to adapt to unpredictable resources they may be able to change their movement responses when further habitat fragmentation occurs or when suitable habitat is removed from the matrix. However, habitat patches need to be sufficiently close to reach and large enough to find for both migrants and nomads, and large enough to support all three types of birds even if only for short periods for migrants and nomads.

There appears to be no pattern between taxonomic groups or foraging guilds as to whether they can cope well in a fragmented landscape i.e. move between patches readily. Species life histories and habitat preferences must play an important role. Habitat generalists are more likely to cope than specialists (Krebs 1994). Many Australian birds can cope with a fragmented landscape because they had already evolved to move between suitable habitat patches in the landscape where the intervening matrix did not provide resources (Ford & Barrett 1995, Reid 2000). However, it would be expected that the nature of the new intervening matrix is likely to affect some species differently from the matrices in which they became adapted. Insectivorous bats have been found to use scattered trees within the rural landscape and appear to have a higher tolerance of fragmentation than

130 many other groups (Lumsden et al. 1995, Lumsden & Bennett 2005). Not all species seem to cope with habitat fragmentation as well. Some forest bird species were inhibited by forest gaps, but would use narrow corridors (St. Clair et al. 1998). Less mobile species, dependent on understorey vegetation, will use corridors with shrub cover but will not cross clearings to reach other patches (Du Guesclin et al. 1995).

If further habitat fragmentation occurs the effects on the dispersal of species are likely to become more pronounced. Initially, habitat fragmentation affects populations through loss of habitat, but as fragmentation continues patch isolation becomes more important. Habitat fragmentation may separate resource patches to the point where migrants or nomads will not be able to reach them, or patches may become too small to provide sufficient resources to warrant visitation, and so disrupt metapopulation dynamics (Wiens 1997). For example, nectarivorous birds are now forced to move large distances in order to obtain year-round resources (Paton et al. 2004).

Another aspect of fragmentation is the effect it may have when resources are also unpredictable. If resources in a patch fluctuate erratically it does not mean that resources at the landscape scale are also unstable (O’Neill and King 1998). This may have been the natural pattern of resources in the landscape before post-European fragmentation took place and thus the local populations would have already adapted to such patterns. In an erratic climate with unpredictable resources, migration and nomadism are important (Krebs 1994) and such adaptations may have pre adapted them to habitat fragmentation (Ford & Barrett 1995, Reid 2000). Such species can cope with modified landscapes, provided habitat heterogeneity is maintained (Ford & Barrett 1995). Although, species such as those studied here, may be adapted to resource fluctuations between patches that cause them to disperse through the landscape in search of other patches, declines of some

131 species may still occur because some patch types are cleared disproportionately e,g, fertile soil areas with White Box Eucalyptus albens (Paton 2000, Ford et al. 2001). Nectarivores appear better able to cope with fragmentation because they are mobile and are adapted to spatially and temporarily variable resources (Katten et al. 1994, Renjifo 1999).

Past studies have concentrated on the effects of habitat fragmentation on declining resident species (e.g. Reid 2000, Brooker and Brooker 2003) and little attention has been paid to population persistence and dispersion with common nomadic generalists such as A. carunculata and P. corniculatus. The focus has also been on the role of corridors in connecting patches of habitat (Saunders & Hobbs 1991) and there is concern that maintaining large patches should be the focus of conservation efforts because small patches and narrow corridors favour edge specialists that impact on the diversity of other species (Major et al. 2001). Large patches of habitat interspersed within a heterogeneous matrix are more likely to provide for both sedentary habitat specialists and nomadic generalists. For habitat generalists, mobile non-sedentary species and species that are wide ranging and using resources opportunistically, large areas at the landscape scale rather than reserves, need to be managed to maintain metapopulation dynamics (Wiens 1997, Hobbs 1998, Brooker and Brooker 2003) and they need to include better quality areas that have been disproportionately cleared (Paton et al. 2004). However, even patches that are as small as ten hectares have conservation value (Ford and Barrett 1995).

Nomadic generalists may have important roles in fragmented landscape as they may be important invertebrate and herbivore population suppressors, pollinators and seed dispersers that can cover relatively large distances between habitat patches which may be too small or too isolated to support residents having the same roles. These patches may only be able to support

132 species for short periods and hence it is important that species be able to move-on when resources are depleted. Nomadic and migratory species are adapted to such patches. Resident species are more likely to become locally extinct when this happens. In this study, A. carunculata and P. corniculatus were found to take nectar, fruit, seeds and insects and were able to cover relatively large distances between habitat patches. Thus they are likely to have important roles in the fragmented landscape. Although they appear to be coping in the fragmented landscape, greater knowledge of their metapopulation dynamics is required in order to conserve such species in the longer term. Loss or indeed decline of such species from the landscape is likely to accelerate the decline of habitat patch quality and have consequences for a wide range of other species.

133 10. APPENDICES

171 1992 1993

100 80 n=0 n=0 60

40 >15m Summer 20 10-15m 5-10m 0 <5m

100 80 n=79 n=140 60 Hi =0.08 Hi =0.33 40 >15m Autumn 20 10-15m 5-10m 0 <5m

100 80 n=499 n=520 60 Hi =0.49 Hi =0.76 40 >15m Winter 20 10-15m 5-10m 0 <5m

100 80 n=114 n=44 60 Hi =0.31 Hi =0.39

40 >15m Spring 20 10-15m 5-10m 0 <5m Air Air Bark Fruit Bark Fruit Seeds Seeds Ground Ground Foliage Flowers Foliage Flowers

Figure 10.1 Percent foraging bouts spent at various substrates by Anthochaera carunculata at Royal National Park in 1992 and 1993 (n = no. of 10 second observation periods, H i = niche breadth).

172 Oij =0.61

100 80 n=155 n=33 60 Hi =0.15 Hj =0.38

40 >15m 10-15m

Summer 20 5-10m 0 <5m

100 80 n=542 n=0 60 Hi =0.42

40 >15m

Autumn 20 10-15m 5-10m 0 <5m

100 n=310 80 n=0 Hi =0.79 60

40 >15m Winter 20 10-15m 5-10m 0 <5m

Oij =0.47

100 80 n=128 n=104 60 Hi =0.75 Hj =0.33

40 >15m Spring 20 10-15m 5-10m 0 <5m Air Air Bark Fruit Bark Fruit Seeds Seeds Ground Ground Foliage Foliage Flowers Flowers A. carunculata P. corniculatus

Figure 10.2 Percent foraging bouts spent at various substrates by Anthochaera carunculata and Philemon corniculatus at Clandulla State Forest in 1992 (n = no. of 10 second observation periods, H i and H j = niche breadths and O ij = niche overlap).

173 Oij =0.77 100 80 n=362 n=113 60 Hi =0.72 Hj =0.77 40 >15m 20 10-15m Summer 5-10m 0 <5m

Oij =0.40 100 80 n=820 n=75 60 Hi =0.65 Hi =0 40 >15m

Autumn 20 10-15m 5-10m 0 <5m

100 80 n=113 n=0 Hi =0.41 60 40 >15m Winter 20 10-15m 5-10m 0 <5m

Oij =0.67

100 80 n=216 n=180 60 Hi =0.64 Hj =0.55 40 >15m Spring 20 10-15m 5-10m 0 <5m Air Air Bark Fruit Bark Fruit Seeds Seeds Ground Ground Foliage Foliage Flowers Flowers A. carunculata P. corniculatus

Figure 10.3 Percent foraging bouts spent at various substrates by Anthochaera carunculata and Philemon corniculatus at Clandulla State Forest in 1993 (n = no. of bouts, H i and H j = niche breadths and O ij = niche overlap).

174 Oij = 0.46

100 80 n = 144 n = 396 60 Hi = 0.45 Hj = 0.67 40 >15m

Summer 10-15m 20 5-10m 0 <5m

Oij = 0.83 100 80 n = 327 n = 244 60 Hi = 0.67 Hj = 0.60 40 >15m

Autumn 20 10-15m 5-10m 0 <5m

Oij = 0.66 100 80 n = 314 n = 156 60 Hi = 0.77 Hj = 0.42 40 >15m

Winter 20 10-15m 5-10m 0 <5m

Oij = 0.83

100 80 n = 436 n = 395 60 Hi = 0.38 Hj = 0.30 40 >15m Spring 20 10-15m 5-10m 0 <5m Air Air Bark Fruit Bark Fruit Seeds Seeds Ground Ground Foliage Flowers Foliage Flowers Anthochaera carunculata Philemon corniculatus

Figure 10.4 Percent foraging time spent at various substrates by Anthochaera carunculata and Philemon corniculatus at Munghorn Gap Nature Reserve in 1992 (n = no. of bouts, H i and H j = niche breadths and O ij = niche overlap).

175 Oij = 0.84

100 80 n = 412 n = 660 60 Hi = 0.46 Hj = 0.56

40 >15m Summer 20 10-15m 5-10m 0 <5m

Oij = 0.41 100 80 n = 51 n = 40 60 Hi = 0.29 Hj = 0.27 40 >15m

Autumn 20 10-15m 5-10m 0 <5m

100 80 n = 333 n = 0 60 Hi = 0.52 40 >15m 20 Winter 10-15m 0 5-10m <5m

Oij = 0.62 100 80 n = 507 60 n = 417 Hi = 0.64 Hj = 0.70 40 >15m Spring 20 10-15m 5-10m 0 <5m Air Air Bark Fruit Bark Fruit Seeds Seeds Ground Ground Foliage Foliage Flowers Flowers Anthochaera carunculata Philemon corniculatus

Figure 10.5 Percent foraging time spent at various substrates by Anthochaera carunculata and Philemon corniculatus at Munghorn Gap Nature Reserve in 1993 (n = no. of bouts, H i and H j = niche breadths and O ij = niche overlap).

176 100

80 n=0 n=0 60

40 >15m

Summer 10-15m 20 5-10m 0 <5m

100 80 n=0 n=0 60

40 >15m Autumn 20 10-15m 5-10m 0 <5m

Oij =0.89

100 80 n=408 n=356 Hi =0.41 Hj =0.39 60

40 >15m Winter 20 10-15m 5-10m 0 <5m

Oij =0.77

100 80 n=830 n=150 60 Hi =0.53 Hj =0.30

Spring 40 >15m 20 10-15m 5-10m 0 <5m Air Air Bark Fruit Bark Fruit Seeds Seeds Ground Ground Foliage Foliage Flowers Flowers A. carunculata P. corniculatus

Figure 10.6 Percent foraging bouts spent at various substrates by Anthochaera carunculata and Philemon corniculatus at Goobang National Park in 1993 (n = no. of bouts, H i and H j = niche breadths and O ij = niche overlap).

177 100 80 n=28 n=0 60 Hi =0 40 >15m

Summer 20 10-15m 5-10m 0 <5m

100 80 n=281 n=0 60 Hi =0.78 40 >15m

Autumn 20 10-15m 5-10m 0 <5m

100

80 n=208 n=0 60 Hi =0.14

40 >15m

Winter 20 10-15m 5-10m 0 <5m

Oij =0.69 100 80 n=1326 n=1492 60 Hi =0.30 Hj =0.37 40 >15m Spring 20 10-15m 5-10m 0 <5m Air Air Bark Fruit Bark Fruit Seeds Seeds Ground Ground Foliage Flowers Foliage Flowers A. carunculata P. corniculatus

Figure 10.7 Percent foraging bouts spent at various substrates by Anthochaera carunculata and Philemon corniculatus at Goobang National Park in 1994 (n = no. of bouts, H i and H j = niche breadths and O ij = niche overlap).

178 100

80 n=0 n=0 60

40 >15m

Summer 10-15m 20 5-10m 0 <5m

100

80 n=162 n=0 60 Hi =0.37

40 >15m Autumn 20 10-15m 5-10m 0 <5m

100 80 n=82 n=0 60 Hi =0.46

40 >15m Winter 20 10-15m 5-10m 0 <5m

100 80 n=74 n=0 60 Hi =0.36 40 >15m Spring 20 10-15m 5-10m 0 <5m Air Air Bark Fruit Bark Fruit Seeds Seeds Ground Ground Foliage Foliage Flowers Flowers A. carunculata P. corniculatus

Figure 10.8 Percent foraging bouts spent at various substrates by Anthochaera carunculata and Philemon corniculatus at Goobang National Park in 1995 (n = no. of bouts, H i and H j = niche breadths and O ij = niche overlap).

179 100

80 n=0 n=0 60

40 >15m

Summer 10-15m 20 5-10m 0 <5m

100

80 n=0 n=0 60

40 >15m Autumn 20 10-15m 5-10m 0 <5m

100

80 n=74 n=0 60 Hi =0.29

40 >15m Winter 20 10-15m 5-10m 0 <5m

Oij =1.00 100

80 n=229 n=472 60 Hi =0.16 Hi =0.16

40 >15m Spring 20 10-15m 5-10m 0 <5m Air Air Bark Fruit Bark Fruit Seeds Seeds Ground Ground Foliage Foliage Flowers Flowers A. carunculata P. corniculatus

Figure 10.9 Percent foraging bouts spent at various substrates by Anthochaera carunculata and Philemon corniculatus at Back Yamma State Forest in 1992 (n = no. of bouts, H i and H j = niche breadths and O ij = niche overlap).

180 100

80 n=0 n=0 60

40 >15m

Summer 10-15m 20 5-10m 0 <5m

100

80 n=28 n=0 60 Hi =0

40 >15m Autumn 20 10-15m 5-10m 0 <5m

Oij =0.31 100

80 n=91 n=134 60 Hi =0.53 Hi =0.51

40 >15m Winter 20 10-15m 5-10m 0 <5m

100

80 n=0 n=0 60

40 >15m Spring 20 10-15m 5-10m 0 <5m Air Air Bark Fruit Bark Fruit Seeds Seeds Ground Ground Foliage Foliage Flowers Flowers A. carunculata P. corniculatus

Figure 10.10 Percent foraging bouts spent at various substrates by Anthochaera carunculata and Philemon corniculatus at Back Yamma State Forest in 1993 (n = no. of bouts, H i and H j = niche breadths and O ij = niche overlap).

181 1992 1993 Forage

Sit

Preen Summer Call

Fly n=0 Aggression n=0

Forage

Sit

Preen Autumn Call

Fly n=322 Aggression n=331

Forage

Sit

Preen Winter Call

Fly n=800 Aggression n=733

Forage

Sit

Preen Spring Call

Fly n=148 Aggression n=60

100 50 0 0 50 100 % Bouts % Bouts

Figure 10.11 Percent bouts at various activities for Anthochaera carunculata at Royal National Park (n = no. of bouts).

182 1992 1993

Forage Forage

Sit Sit

Preen Preen

Call Call Summer Fly Fly

n=186 Aggression n=153 n=701 Aggression n=476

Forage Forage

Sit Sit

Preen Preen

Call Call Autumn Fly Fly

n=684 Aggression n=0 n=1439 Aggression n=151

Forage Forage

Sit Sit

Preen Preen

Call Call Winter Fly Fly

n=612 Aggression n=0 n=214 Aggression n=0

Forage Forage

Sit Sit

Preen Preen

Call Call Spring Fly Fly

n=232 Aggression n=556 n=467 Aggression n=494

100 50 0 0 50 100 100 50 0 0 50 100 % Bouts % Bouts % Bouts % Bouts A. carunculata P. corniculatus

Figure 10.12 Percent bouts at various activities for Anthochaera carunculata and Philemon corniculatus at Clandulla State Forest (n = no. of bouts).

183 1993

Forage

Sit

Preen

Summer Call

Fly n=0 Aggression n=0

Forage Sit

Preen

Call Autumn Fly

n=0 Aggression n=0

Forage

Sit

Preen

Call Winter Fly n=688 Aggression n=755

Forage

Sit

Preen

Call Spring Fly

n=1412 Aggression n=243

100 50 0 0 50 100 % Bouts % Bouts A. carunculata P. corniculatus

Figure 10.13 Percent bouts at various activities for Anthochaera carunculata and Philemon corniculatus at Goobang National Park in 1993 (n = no. of bouts).

184 1994 1995

Forage Forage

Sit Sit

Preen Preen

Call Call Summer Fly Fly

n=39 Aggression n=0 n=0 Aggression n=0

Forage Forage

Sit Sit

Preen Preen

Call Call Autumn Fly Fly

n=463 Aggression n=0 n=209 Aggression n=0

Forage Forage

Sit Sit

Preen Preen

Call Call Winter Fly Fly

n=307 Aggression n=10 n=131 Aggression n=0

Forage Forage

Sit Sit

Preen Preen

Call Call Spring Fly Fly

n=1901 Aggression n=2371 n=134 Aggression n=4

100 50 0 0 50 100 100 50 0 0 50 100 % Bouts % Bouts % Bouts % Bouts A. carunculata P. corniculatus

Figure 10.14 Percent bouts at various activities for Anthochaera carunculata and Philemon corniculatus at Goobang National Park in 1994 and 1995 (n = no. of bouts). 185 1992 1993

Forage Forage

Sit Sit

Preen Preen

Call Call Summer Fly Fly n=0 Aggression n=0 n=0 Aggression n=0

Forage Forage

Sit Sit

Preen Preen

Call Call Autumn Fly Fly

n=0 Aggression n=0 n=48 Aggression n=0

Forage Forage

Sit Sit

Preen Preen

Call Call Winter Fly Fly

n=161 Aggression n=0 n=285 Aggression n=447

Forage Forage

Sit Sit

Preen Preen

Call Call Spring Fly Fly

n=282 Aggression n=633 n=0 Aggression n=0

100 50 0 0 50 100 100 50 0 0 50 100 % Bouts % Bouts % Bouts % Bouts A. carunculata P. corniculatus

Figure 10.15 Percent bouts at various activities for Anthochaera carunculata and Philemon corniculatus at Back Yamma State Forest (n = no. of bouts).

186 1992 1993

RW - RW

NF - NF

RW - NF

Summer RW - OHE

(n=0) NF - OHE (n=0)

RW - RW

NF - NF

RW - NF

Autumn RW - OHE

(4%,n=322) NF - OHE (10%,n=331)

RW - RW

NF - NF

RW - NF

Winter RW - OHE

(2%,n=800) NF - OHE (5%,n=733)

RW - RW

NF - NF

RW - NF

Spring RW - OHE

(0%,n=148) NF - OHE (n=0)

100 50 0 0 50 100

A. carunculata P. corniculatus Other Honeyeaters

Figure 10.16 Percent bouts spent in aggression between P. corniculatus , A. carunculata and other honeyeaters at Royal National Park. Bars indicate proportion of aggression and shading indicates dominant honeyeater in aggressive encounters. Quantities in brackets show percent of total bouts spent in aggression and total number of bouts in each survey period.

187 1992 1993

RW - RW

NF - NF

RW - NF

Summer (1%,n=459) RW - OHE (0.7%,n=1107) NF - OHE

RW - RW

NF - NF

RW - NF

Autumn RW - OHE

(0%,n=754) NF - OHE (0.3%,n=1590)

RW - RW

NF - NF

RW - NF Winter RW - OHE

(3%,n=612) NF - OHE (0%,n=214)

RW - RW

NF - NF

RW - NF

Spring RW - OHE

(0.3%,n=668) NF - OHE (2%,n=961)

100 50 0 0 50 100

A. carunculata P. corniculatus Other Honeyeaters

Figure 10.17 Percent bouts spent in aggression between P. corniculatus , A. carunculata and other honeyeaters at Clandulla State Forest. Bars indicate proportion of aggression and shading indicates dominant honeyeater in aggressive encounters. Quantities in brackets show percent of total bouts spent in aggression and total number of bouts in each survey period.

188 1992 1993

RW - RW

NF - NF

RW - NF

Summer (1%,n=459) RW - OHE (0.7%,n=1107) NF - OHE

RW - RW

NF - NF

RW - NF

Autumn RW - OHE

(0%,n=754) NF - OHE (0.3%,n=1590)

RW - RW

NF - NF

RW - NF Winter RW - OHE

(3%,n=612) NF - OHE (0%,n=214)

RW - RW

NF - NF

RW - NF

Spring RW - OHE

(0.3%,n=668) NF - OHE (2%,n=961)

100 50 0 0 50 100

A. carunculata P. corniculatus Other Honeyeaters

Figure 10.17 Percent bouts spent in aggression between P. corniculatus , A. carunculata and other honeyeaters at Clandulla State Forest. Bars indicate proportion of aggression and shading indicates dominant honeyeater in aggressive encounters. Quantities in brackets show percent of total bouts spent in aggression and total number of bouts in each survey period.

188 1992 1993

RW - RW

NF - NF

RW - NF

Summer RW - OHE

(0.2%,n=998) NF - OHE (0.8%,n=2031)

RW - RW

NF - NF

RW - NF Autumn RW - OHE

(1%,n=1104) NF - OHE (0%,n=219)

RW - RW

NF - NF

RW - NF Winter RW - OHE (8%,n=890) (3%,n=913) NF - OHE

RW - RW

NF - NF

RW - NF Spring RW - OHE

(5%,n=1306) NF - OHE (6%,n=1821)

100 50 0 0 50 100

A. carunculata P. corniculatus Other Honeyeaters

Figure 10.18 Percent bouts spent in aggression between P. corniculatus , A. carunculata and other honeyeaters at Munghorn Gap Nature Reserve. Bars indicate proportion of aggression and shading indicates dominant honeyeater in aggressive encounters. Quantities in brackets show percent of total bouts spent in aggression and total number of bouts in each survey period.

189 1994 1995

RW - RW

NF - NF

RW - NF Summer RW - OHE

(0%,n=39) NF - OHE (n=0)

RW - RW

NF - NF

RW - NF

Autumn RW - OHE

(0.2%,n=463) NF - OHE (0%,n=209)

RW - RW

NF - NF

RW - NF Winter RW - OHE

(2%,n=317) NF - OHE (0%,n=131)

RW - RW

NF - NF

RW - NF Spring RW - OHE

(6%,n=4272) NF - OHE (0.7%,n=138)

100 50 0 0 50 100

A. carunculata P. corniculatus Other Honeyeaters

Figure 10.19 Percent bouts spent in aggression between P. corniculatus , A. carunculata and other honeyeaters at Goobang National Park. Bars indicate proportion of aggression and shading indicates dominant honeyeater in aggressive encounters. Quantities in brackets show percent of total bouts spent in aggression and total number of bouts in each survey period.

190 1992 1993

RW - RW

NF - NF

RW - NF

Summer RW - OHE

(n=0) NF - OHE (n=0)

RW - RW

NF - NF

RW - NF

Autumn RW - OHE

(n=0) NF - OHE (0%,n=48)

RW - RW

NF - NF

RW - NF Winter RW - OHE

(11%,n=161) NF - OHE (4%,n=732)

RW - RW

NF - NF

RW - NF

Spring RW - OHE

(10%,n=915) NF - OHE (n=0)

100 50 0 0 50 100 A. carunculata P. corniculatus Other Honeyeaters

Figure 10.20 Percent bouts spent in aggression between P. corniculatus , A. carunculata and other honeyeaters at Back Yamma State Forest. Bars indicate proportion of aggression and shading indicates dominant honeyeater in aggressive encounters. Quantities in brackets show percent of total bouts spent in aggression and total number of bouts in each survey period.

191 9. REFERENCES

Abramsky, Z., Bowers, M.A. & Rosenzweig, M.L. 1986 Detecting interspecific competition in the field: testing the regression method. Oikos 47: 199-204

Agresti, A. 1996 An Introduction to Categorical Data Analysis. John Wiley & Sons, Inc. New York.

Aho, T., Kuitunen, M., Suhonen, J., Jantti, A. & Hakkari, T 1997 Behaviour responses of Eurasian Treecreepers, Certhia famillaris, to competition with ants. Animal Behaviour 54: 1283-1290

Airola, D.A. & Barrett, R.H. 1985 Foraging and habitat relationships of insect-gleaning birds in a Sierra Nevada mixed-conifer forest. Condor 87: 205-216

Alcock, J. 1989 Animal Behaviour 4th ed. Sinauer, Massachusetts

Armstrong, D.P. 1991 Aggressiveness of breeding territorial honeyeaters corresponds to seasonal changes in nectar availability. Behavioural Ecology and Sociobiology 29: 103-111

Armstrong, D.P. 1992 Correlation between nectar supply and aggression in territorial honeyeaters: causation or coincidence? Behavioural Ecology and Sociobiology 30: 95-102

Atkinson R.P.D., Rhodes C.J., Macdonald D.W. & Anderson R.M. 2002 Scale- free dynamics in the movement patterns of jackals. Oikos 98: 134-140

134 Avery, M.L., Goocher, K.J. & Cone, M.A. 1993 Handling efficiency and berry size preferences of Cedar Waxwings. Wilson Bull. 105: 604-611

Baldwin, M. 1973 Nests of the Red Wattlebird Sunbird 4: 10-12

Balfour, D. 1979 Honeyeater distribution Canberra Bird Notes 4: 15-16

Balfour, D. 1980 Birds of Mount Ainslie Canberra Bird Notes 5: 11-16

Barker, R.D. & Vestjens, W.J.M. 1990 The Food of Australian Birds: 2 Passerines. CSIRO, Melbourne.

Barreto, G.R. & Herrera, E.A. 1998 Foraging patterns of Capybaras in a seasonally flooded savanna of Venezuela. Journal of Tropical Ecology 14: 87-98

Barrett, G., Silcocks, A., Barry, S. Cunningham, R. & Poulter, R. 2003 The New Atlas of Australian Birds Royal Australasian Ornithologists Union

Baxter, C.I. 1989 An Annotated List of the Birds of South Australian National Parks and Wildlife Service, Kingscote, Kangaroo Island,

Beaver, D.L. & Baldwin, P.H. 1975 Ecological overlap and the problem of competition and sympatry in the Western and Hammond’s Flycatchers. Condor 77: 1-13

Begon, M., Mortimer, M. & Thompson, D.J. 1996 Population ecology: a unified study of animals and plants. Blackwell Scientific Publications, Cambridge.

Bell, H.L. 1985 The social organisation and foraging behaviour of three

135 syntopic Thornbills Acanthiza spp. In: Birds of Forests and Woodland: Ecology, Conservation, Management Keast, A., Recher, H.F., Ford, H. & Saunders, D. eds. Pp. 151-163 Surrey Beatty & Sons,Chipping Norton

Bell, H.L. & Ford, H.A. 1990 The influence of food shortage on interspecific niche overlap and foraging behaviour of three species of Australian Warblers (). Studies in Avian Biology 13: 381-388

Bennetts, R.E. & Kitchens, W.M. 2000 Factors influencing movement probabilities of a nomadic food specialist: proximate foraging benefits or ultimate gains from exploration? Oikos 91: 459-467

Bishop, Y.V.V., Fienberg, S.E. & Holland, P.W. 1975 Discrete Multivariate Analysis MIT Press, Cambridge, MA

Blakers, M., Davies, S.J.J.F. & Reilly, P.N. 1984 The Atlas of Australian Birds RAOU & Melbourne University Press, Melbourne

Boles, W.E. 1988 The Robins and Flycatchers of Australia Angus and Robertson, Sydney

Bradley, D.W. 1985 The effects of visibility bias on time-budget estimates of niche breadth and overlap Auk 102: 493-499

Brooker, M.I.H. & Kleinig, D.A. 1999 (2nd ed.) Field Guide to Eucalypts: South-eastern Australia Bloomings Books, Hawthorn

Brooker, L. & Brooker, M. 2003 Local distribution, metapopulation viability and conservation of the Blue-breasted Fairy-wren in fragmented habitat in the Western Australian wheatbelt Emu 103: 185-198

136 Brotons, L., Magrans, M., Ferus, L. & Nadal, J. 1998 Direct and indirect effects of pollution on the foraging behaviour of forest passerines during the breeding season. Canadian Journal of Zoology 76: 556-565

Brown, E.D. & Hopkins, M.J.G. 1996 How New Guinea rainforest resources vary in time and space: implications for nectarivorous birds. Australian Journal of Ecology 21: 363-378

Brown, J.H., Calder III, W.A. & Kodric-Brown, A. 1978 Correlates and consequences of body size in nectar-feeding birds. American Zoologist 18: 687-700

Byorth, P.A. & Magee, J.P. 1998 Competitive interactions between Artic grayling and brook trout in the Big Hole River drainage, Montana. Transactions of the American Fisheries Society 127: 921-931

Cáceres, C.E. 1998 Seasonal dynamics and interspecific competition in Oneida Lake Daphnia. Oecologia 115: 233-244

Caithamer, D.F., Gates, R.J. & Tacha, T.C. 1995 A comparison of diurnal time budgets from paired Interior Canada Geese with and without offspring. Journal of Field Ornithology 67: 105-113

Cale, P. 1994 Temporal changes in the foraging behaviour of insectivorous birds in a forest in . Emu 94: 116-126

Carpenter, F.L. 1978 A spectrum of nectar-eater communities. American Zoologist 18: 809-819

137 Carpenter, F.L. 1979 Competition between hummingbirds and insects for nectar. American Zoologist 19: 1105-1114.

Carpenter, F.L. 1987 Food abundance and territoriality: to defend or not to defend? American Zoologist 27: 387-399

Carpenter, F.L. & MacMillen, R.E. 1976 Threshold model of feeding territoriality and test with a Hawaiian honeycreeper. Science 194: 639-642

Catterall, C.P., Green, R.J. & Jones, D.N. 1991 Habitat use by birds across a forest-suburb interface in Brisbane: implications for corridors Pp 247-258 In: Saunders, D.A. & Hobbs, R.J. (eds.) Nature Conservation: the Role of Corridors Surrey Beattie & Sons

Chan, K. 1990 Habitat selection in the White-plumed Honeyeater and the Fuscous Honeyeater at an area of sympatry. Australian Journal of Ecology 15: 207-217

Chan, K. 1995 Bird community patterns in fragmented vegetation zones around streambeds of the Northern Tablelands, New South Wales Australian Bird Watcher 16: 11-20

Cody, M.L. 1974 Competition and The Structure of Bird Communities. Monographs in Population Biology No. 7 Princeton University Press, Princeton, New Jersey

Collins, B.G. 1985 Energetics of foraging and resource selection by honeyeaters in forest and woodland habitats of Western Australia. New Zealand Journal of Zoology 12: 577-587

138 Collins, B.G.,Briffa, P. & Newland, C. 1984 Temporal changes in abundance and resource utilisation by honeyeaters at Wongamine Nature Reserve. Emu 84: 159-166

Collins, B.G. & McNee, S. 1991 Resource partitioning within Australian nectarivorous communities. Acta XX Congress of International Ornithology pp. 1166-1174. New Zealand Ornithological Trust Board, Wellington, New Zealand.

Collins, B.G. & Newland, C. 1986 Honeyeater population changes in relation to food availability in the jarrah forest of Western Australia. Australian Journal of Ecology 11: 63-76

Collins, B.G. & Paton, D.C. 1989 Consequences of differences in body mass, wing-length and leg morphology for nectar-feeding birds. Australian Journal of Ecology 14: 269-289

Connell, J.H. 1983 On the prevalence and relative importance of interspecific competition: evidence from field experiments. American Naturalist 122: 661-696.

Costermans, L.F. 1981 Native trees and shrubs of South-eastern Australia Rigby Publishers

Cotton, P.A. 1998 The hummingbird community of a lowland Amazonian rainforest. Ibis 140: 512-521

Craig, J.L. & MacMillen, R.E. 1985 Honeyeater ecology: an introduction. New Zealand Journal of Zoology 12: 565-568

139 Creese, R.G. & Underwood, A.J. 1982 Analysis of inter- and intraspecific competition amongst intertidal limpets with different methods of feeding. Oecologia 53: 337-346

Dann, P. 1981 Resource allocation in three congeneric species of sandpiper Stilt 1: 3

Davis, W.E. & Recher, H.F. 1993 Notes on the breeding biology of the Regent Honeyeater. Corella 17: 1-4

Dayton, P.K. 1973 Two cases of resource partitioning in an intertidal community: making the right prediction for the wrong reason. American Naturalist 107: 662-670

De Pirro, M., Marchetti, G.M. & Chelazzi, G. 1999 Foraging interactions among three benthic fish in a Posidonia oceanica reef lagoon along the Tyrrhenian Coast. Journal of Fish Biology 54: 1300-1309

Dow, D.D. 1977 Indiscriminate interspecific aggression leading to almost sole occupancy of space by a single species of bird. Emu 77: 115-121

Du Guesclin, P, Smith, S., O’Shea, B. & Dennis, C. 1995 “Brushing for bristles”: habitat corridors for the Rufous Bristlebird Pp 163-165 In: Bennett, A., Backhouse, G., Clark, T. (eds.) People and Nature Conservation: Perspectives on Private Land Use and Endangered Species Recovery Royal Zoological Society of New South Wales, Sydney

Egan, K.H. 1997 Seasonal changes in the pollen sampled from nectarivorous birds visiting an open forest at Menai, New South Wales. Corella 21: 83-87

140 East, M. 1982 Time-budgeting of European Robins Erithacus rubecula: inter- and intrasexual comparisons during autumn, winter, and early spring. Ornis Scandinavica 13: 85-93

Edwards, J.W., Heckel, D.G. & Guynn, D.C. 1998 Niche overlap in sympatric populations of Fox and Grey Squirrels. Journal of Wildlife Management 62: 354-363

Egan, K.H. 1997 Seasonal changes in the pollen sampled from nectarivorous birds visiting an open forest at Menai, New South Wales. Corella 21: 83-87

Emison, W.B., Beardsell, C.M., Norman, F.I., Loyn, R.H. & Bennett, S.C. 1987 Atlas of Victorian Birds Department of Conservation, Forests and Lands and Royal Australasian Ornithologists Union

Engel-Wilson, R.W., Kurt Webb, A., Rosenberg, K.V., Ohmart, R.D. & Anderson, B.W. 1981 Avian censusing with the strip method: a computer simulation. Studies in Avian Biology 6: 445-449

Enoksson B., Angelstam P. & Larsson K. 1995 Deciduous forest and resident birds: the problem of fragmentation within a coniferous forest landscape. Landscape Ecology 10: 267-275

Er, K.B.H. 1997 Effects of eucalypt dieback on bird species diversity in remnants of native woodland. Corella 21: 101-111

Er, K.B.H. & Tidemann, C.R. 1996 Importance of Yellow Box-Blakely’s Red Gum woodland remnants in maintaining bird species diversity: inferences from seasonal data. Corella 20: 117-128

141 Er, K.B.R., Wong, T.H. & Tidemann, C.R. 1998 An analysis of the occurrence of terrestrial bird species in the lowland Yellow Box-Blakely’s red Gum woodland remnants of the Australian Capital Territory. Australian Bird Watcher 17: 370-382

Farmer, D., Catterall, C.P. & Piper, S.D. 2004 Abundance patterns across months and locations, and their differences between migrant and resident landbirds in lowland subtropical eucalypt forest. Emu 104: 283-296

Feinsinger, P. 1976 Organisation of a tropical guild of nectarivorous birds. Ecological Monographs 46: 257-291

Feinsinger, P. 1978 Ecological interactions between plants and hummingbirds in a successional tropical community. Ecological Monographs 48: 269-287

Feinsinger, P. & Colwell, R.K. 1978 Community organisation among neotropical nectar-feeding birds. American Zoologist 18: 779-95

Feinsinger, P & Swarm, L.A. 1982 “Ecological release”, seasonal variation in food supply and the hummingbird Amazilia tobaci on Trinidad and Tobago. Ecology 63: 1574-1587

Fleay, D. 1968 Nightwatchmen of Bush and Plain Jacaranda Press, Brisbane

Ford, H.A. 1979 Interspecific competition in Australian honeyeaters- depletion of common resources. Australian Journal of Ecology 4: 145-164

Ford, H.A. 1981 Territorial behaviour in an Australian nectar feeding bird. Australian Journal of Ecology 6: 131-134

142 Ford, H.A. 1983 Relation between number of honeyeaters and intensity of flowering near Adelaide,South Australia. Corella 7: 25-31

Ford, H.A. 1985 A synthesis of the foraging ecology and behaviour of birds in eucalypt forests and woodlands. Pp 333-340 In: Birds of Eucalypt Forests and Woodland: Ecology, Conservation and Management RAOU & Surrey Beatty and Sons, Sydney

Ford, H.A. 1989 Ecology of Birds Surrey Beatty & Sons, Chipping Norton

Ford, H.A. 1998 Faithfulness to breeding site and birthplace in Noisy Friarbirds Philemon corniculatus. Emu 98: 269-275

Ford, H.A. 1999 Nest site selection and breeding success in large Australian honeyeaters: are there benefits from being different? Emu 99: 91-99

Ford, H.A. & Barrett, G. 1995 The role of birds and their conservation in agricultural systems. Pp 128-135 In: Bennett, A., Backhouse, G., Clark, T. (eds.) People and Nature Conservation: Perspectives on Private Land Use and Endangered Species Recovery. Royal Zoological Society of New South Wales, Sydney

Ford, H.A., Barrett, G.W., Saunders, D.A. & Recher, H.F. 2001 Why have birds in the woodlands of South Australia declined? Biological Conservation 97: 71-88

Ford, H.A., Bridges, L. & Noske, S. 1985 Density of birds in eucalypt woodland near Armidale,north-eastern New South Wales. Corella 9: 78-107

143 Ford, H.A. & Debus, S. 1994 Aggressive behaviour of Red Wattlebirds Anthochaera carunculata and Noisy Friarbirds Philemon corniculatus. Corella 18: 141-147

Ford, H.A., Huddy, L. & Bell, H.L. 1990 Seasonal changes in the foraging behaviour of three species of passerines in Australian eucalypt woodland. Studies in Avian Biology 13: 245-253

Ford, H.A., Noske, S. & Bridges, L. 1986 Foraging of birds in eucalypt woodlands in north east New South Wales. Emu 86: 168-179

Ford, H.A. & Paton, D.C. 1976 Resource partitioning and competition in honeyeaters of the genus Meliphaga. Australian Journal of Ecology 1: 281-287

Ford, H.A. & Paton, D.C. 1977 The comparative ecology of ten species of honeyeaters in South Australia. Australian Journal of Ecology 2: 399-407

Ford, H.A. & Paton, D.C. 1982 Partitioning of nectar sources in an Australian honeyeater community. Australian Journal of Ecology 7: 149-159

Ford, H.A. & Paton, D.C. 1985 Habitat selection in Australian Honeyeaters,with special reference to nectar productivity. Pp 367-388 In: Habitat Selection in Birds Cody, M.C. (ed.) Academic Press, New York.

Ford, H.A., & Trémont, S. 2000 Life history characteristics of two Australian honeyeaters (Meliphagidae). Australian Journal of Zoology 48: 21-32

Ford, H.A., Davis, W.E., Debus, S., Ley, A., Recher,H. & Williams,B. 1993 Foraging and aggressive behaviour of the Regent Honeyeater Xanthomyza phrygia in northern New South Wales. Emu 93: 277-281

144 Forde, N. 1986 Relationships between birds and fruits in temperate Australia. Pp 42-58 Pp 42-58 In: The Dynamic Partnership: The Birds and Plants of Southern Australia Ford, H.A. & Paton, D.C. (eds.) J. Woolman Government Printers, South Australia

Forman, R.T.T. & Godron, M. 1986 Landscape Ecology Wiley, New York.

Forshaw, J.M. & Cooper, W.T. 1981 Australian Parrots (2nd. ed.) Lansdowne Editions, Sydney

Franklin, D.C. 1997 The foraging behaviour of avian nectarivores in a monsoonal Australian woodland over a six-month period. Corella 21: 48-54

Franklin, D.C. & Noske, R.A. 1999 Birds and nectar in a monsoonal woodland: correlations at three spatio-temporal scales. Emu 99: 15-28

French, K., Paterson, I., Miller, J. & Turner, R.J. 2003 Nectarivorous bird assemblages in Box-Ironbark woodlands in the Capertee Valley, New South Wales. Emu 103: 345-356

Geering, D. & French, K. 1998 Breeding biology of the Regent Honeyeater Xanthomyza phrygia in the Capertee Valley, New South Wales. Emu 98: 104- 116

Gill, F.B. 1978 Proximate costs of competition for nectar. American Zoologist 18: 753-763

Gill, F.B. & Wolf, L.L. 1975 Economics of feeding territoriality in the Golden- winged Sunbird. Ecology 56: 333-345.

145 Gill, F.B. & Wolf, L.L. 1977 Non-random foraging by sunbirds in a patchy environment. Ecology 58: 1284-1296

Gosper, C.R. 1999 Plant food resources of birds in coastal dune communities in New South Wales. Corella 23: 53-62

Goss-Custard, J.D. 1977 Optimal foraging and the size selection of worms by redshank (Tringa totanus), in the field. Animal Behaviour 25: 10-29

Greenberg, R., Caballero, C.M. & Bichier, P. 1993 Defence of Homopteran honeydew by birds in the Mexican highlands and other warm temperate forests. Oikis 68: 519-524

Gregory-Smith, R. 1991 Birds of North Araluen, New South Wales Canberra Bird Notes 16: 65-74

Halley, D.J. & Gjershaug, J.O. 1998 Inter- and intraspecific dominance relationships and feeding behaviour of Golden Eagles Aquila chrysaetos and Sea-Eagles Haliaeetus albicilla at carcasses. Ibis 140: 295-301

Hanowski, J.M. & Niemi, G.J. 1995 A comparison of on- and off-road bird counts: do you need to go off road to count birds accurately? Journal of Field Ornithology 66: 469-483

Harden, G.J. 1990 Flora of New South Wales. New South Wales University Press, Kensington

Hejl, S.J. & Verner, J. 1990 Within-season and yearly variations in avian foraging locations. Studies in Avian Biology 13: 202-209

146 Herrera, C.M. 1978 Individual dietary differences associated with morphological variation in robins, Erithacus rubecula. Ibis 120: 542-545

Hester, A.J., Gordon, I.J., Baillie, G.J. & Tappin, E. 1999 Foraging behaviour of sheep and Red Deer within natural heather grass mosaics. Journal of Applied Ecology 36: 133-146

Higgins, P.J., Peter, J.M. & Steele, W.K. (eds.) 2001 Handbook of Australian, New Zealand and Antarctic Birds. Volume 5: Tyrant-flycatchers to Chats. Oxford University Press, Melbourne

Hindwood, K.A. 1944 Honeyeaters of the Sydney district (County of Cumberland) New South Wales. Australian Zoologist 10: 231-251

Hixon, M.A., Carpenter, F.L. & Paton, D.C. 1983 Territory area,flower density and time budgeting in hummingbirds: an experimental and theoretical analysis. American Naturalist 122: 366-391

Hobbs, R.J. 199 8 Managing ecological systems and processes Pp 459-484 In: Peterson, D.L. & Parker, V.T. (eds.) Ecological Scale: Theory and Applications Columbia University Press, New York

Hobbs, R.J. 1999 Clark Kent or Superman: where is the phone booth for landscape ecology? Pp 11-23 In: Klopatek, J.M. & Gardner, R.H. (eds.) Landscape Ecological Analysis Springer, New York

Holmes, R.T. & Robinson, S.K. 1981 Tree species preferences in foraging insectivorous birds in a northern hardwoods forest. Oecologia 48: 31-35

147 Holt, R.D., Pacala, S.W., Smith, T.W. & Liu, J. 1995 Linking contemporary vegetation models with spatially explicit animal population models. Ecological Applications 5: 20-27

Huhta, E., Jokimaki, J. & Rahko, P. 1998 Distribution and reproductive success of the Pied Flycatcher Ficedula hypoleuca in relation to forest patch size and vegetation characteristics; the effect of scale. Ibis 140: 214-222

Hutto, R.L. 1990 Measuring the availability of food resources. Studies in Avian Biology 13: 20-28

Iason, G.R., Mantecon, A.R., Sim, D.A., Gonzalez, J., Foreman, E., Bermudez, F.F. & Elston, D.A. 1999 Can grazing sheep compensate for a daily foraging time constraint? Journal of Animal Ecology 68: 87-93

Jackson, T.P. 1998 Seasonal variation in the diet and foraging behaviour of Brants Whistling Rat, Parotomys bratsil, in northern Namaqualand. South African Journal of Zoology 33: 37-41

James, K.E.S. & Poulin, R. 1998 The effects of perceived competition and parasitism on the foraging behaviour of the Upland Bully (Eleotridae). Journal of Fish Biology 53: 827-834

Johnson, N.K. 1966 Bill size and the question of competition in allopatric and sympatric populations of Dusky and Gray Flycatchers. Systematic Zoology 15: 70-87

Jones, J. 1945 The Banded Stilt Emu 45: 1-36

Jones, M.J., Lace, L.A., Harrison, E.C. & Stevens-Wood, B. 1998 Territorial

148 behaviour in the Speckled Wood Butterflies Pararge xiphia and P. aegeria of Madeira- a mechanism for interspecific competition. Ecography 21: 297-305

Kattan, G.R., Alvarez-Lopez, H. & Giraldo, M. 1994 Forest fragmentation and bird extinctions: San Antonio eighty years later Conservation Biology 8: 138-146

Keast, A. 1968 Seasonal movements in the Australian honeyeater (Meliphagidae) and their ecological significance. Emu 67: 159-209.

Keast, A. 1976 The origins of adaptive zone utilisations and adaptive radiations, as illustrated by the Australian Meliphagidae. Pp 71-82 In: Proceedings of the 16th International Ornithologists Congress

Keast, A. 1985a An introductory ecological biogeography of the Australo- Pacific Meliphagidae. New Zealand Journal of Zoology 12: 605-622

Keast, A. 1985b Bird community structure in southern forests and northern woodlands. Pp. 97-116 In: Birds of Forests and Woodland: Ecology, Conservation, Management Keast, A., Recher, H.F., Ford, H. & Saunders, D. (eds.) Surrey Beatty & Sons, Chipping Norton

Kellner, C.J., Smith, K.G., Wilkinson, N.C. & James, D.A. 1990 Influence of periodic cicadas on foraging behaviour of insectivorous birds in an Ozark forest. Studies in Avian Biology 13: 375-380

Kelly, J.P. & Wood C. 1996 Diurnal, intraseasonal, and intersexual variation in foraging behaviour of the Common Yellowthroat. The Condor 98: 491-500

Kie, J.G., Bowyer, R.T., Nicholson, M.C., Boroski, B.B. & Loft, E.R. 2002

149 Landscape heterogeneity at different scales: effects on spatial distribution of Mule Deer. Ecology 83: 530-544.

Kirchhoff, M.D. & Larsen, D.N. 1998 Dietary overlap between native Sitka Black-tailed Deer and introduced Elk in southeast Alaska. Journal of Wildlife Management 62: 236-242

Kossenko, S.M. & Fry, C.H. 1998 Competition and coexistence of the European Bee-eater apiaster and the Blue-cheeked Bee-eater Merops persicus in Asia. Ibis 140: 2-13

Krebs, C.J. 1972 Ecology: The Experimental Analysis of Distribution and Abundance Harper and Row, New York.

Krebs, C.J. 1989 Ecological Methodology Harper & Row, New York

Krebs, C.J. 1994 The Experimental Analysis of Distribution and Abundance 4th ed. Harper Collins, New York

Krebs, J.R., Ericksen, T., Webber, M.I. & Charnov, E.L. 1977 Optimal prey selection in the Great Tit (Parus major). Animal Behaviour 25: 30-38

Kutt, A.S. 1996 Bird populations density in thinned, unthinned and old lowland regrowth forest, East Gippsland, Victoria. Emu 96: 280-284

Lack, D. & Owen, D.F. 1955 The food of the swift. Journal of Animal Ecology 24: 120-136

Lamm, D.W. & Wilson, S.J. 1966 Seasonal fluctuations of birds in the Brindabella Ranges, ACT. Emu 65: 183-207

150 Lane, B.A. & Davies, J.N. 1987 Shorebirds in Australia Nelson Publishers, Melbourne

Law B., Mackowski C., Schoer L. & Tweedie T. 2000 Flowering phenology of myrtaceous trees and their relation to climatic, environmental and disturbance variables in northern New South Wales. Austral Ecology 25: 160- 178

Lea, A.M. & Gray, J.T. 1935 The food of Australian birds. Part I Emu 34: 275- 292

Lea, A.M. & Gray, J.T. 1936 The food of Australian birds. Part IV Emu 35: 251-280

Leishman, A.J. 1994 The birds of Humewood/Beulah Forest, Campbelltown NSW Australian Birds 28: 14-26

Leonard, J. 1995 The interaction of a solitary Regent honeyeater with a group of Red Wattlebirds. Canberra Bird Notes 20: 81-83

Lepschi, B.J. 1993 Food of some birds in eastern New South Wales: additions to Barker & Vestjens. Emu 93: 195-199

Lepschi, B.J. 1997 Food of some birds in southern Australia: additions to Barker & Vestjens, Part 2. Emu 97: 84-87

Levin, S.A. 1992 The problem of pattern and scale in ecology. Ecology 73: 1943-1967

151 Leviten, P.J. 1978 Resource partitioning by predatory gastropods of the genus Conus on subtidal Indo-Pacific coral reefs: the significance of prey size. Ecology 59: 614-631

Ley, A.J., Oliver, D.L., & Williams, M.B. 1997 Theft of nesting material involving Honeyeaters (Meliphagidae). Corella 21: 119-123

Lindenmayer, D.B., Cunningham, R.B., Donnelly, C.F., Nix, H. & Lindenmayer, B.D. 2002 Effects of forest fragmentation on bird assemblages in a novel landscape context. Ecological Monographs 72: 1-18

Longmore N.W. 1991 Honeyeaters and Their Allies of Australia. The National Photographic Index of Australia. Angus & Robertson, Sydney

Loyn, R.H. 1980 Bird populations in mixed eucalypt forest used for production of wood in Gippsland, Victoria. Emu 80: 145-156

Loyn, R.H. 1987 Effects of patch area and habitat on bird abundances, species numbers and tree health in fragmented Victorian forests. pp 65-77 in:- Saunders,D.A., Arnold,G.W., Burbidge,A.A. & Hopkins,A.J.M. (eds.) Nature Conservation: the Role of Remnants of Natural Vegetation. Surrey Beatty, Sydney

Luiselli, L., Akani, G.C. & Capizzi, D. 1999 Is their interspecific competition between Dwarf Crocodiles (Osteolaemus tetraspis) and Nile Monitors (Varanus niloticus ornatus) in the swamps of central Africa? A study from southeastern Nigeria. Journal of Zoology 247: 127-131

Lumsden, L.F., Bennett, A.F., Krasna, S.P. & Silins, J.E. 1995 The conservation of insectivorous bats in rural landscapes of northern Victoria.

152 Pp 142-148 In: Bennett, A., Backhouse, G., Clark, T. (eds.) People and Nature Conservation: Perspectives on Private Land Use and Endangered Species Recovery Royal Zoological Society of New South Wales, Sydney

Lumsden, L.F. & Bennett, A.F. 2005 Scattered trees in rural landscapes: foraging habitat for insectivorous bats in south-eastern Australia. Biological Conservation 122: 205-222

Lundquist, R.W. & Manuwal, D.A. 1990 Seasonal differences in foraging habitat of cavity-nesting birds in the southern Washington Cascades. Studies in Avian Biology 13: 218-225

MacArthur, R.H. 1972 The machinery of competition and predation. Pp 21- 32. In: Geographical ecology: patterns in the distribution of species Harper and Row, New York

MacArthur, R.H. & Levins, R. 1967 The limiting similarity,convergence and divergence of coexisting species. American Naturalist 101: 377-385.

Mac Nally, R.C. 1990a An analysis of density responses of forest and woodland birds to composite physiognomic variables. Australian Journal of Ecology 15: 267-275.

Mac Nally, R.C. 1990b The roles of floristics and physiognomy in avian community composition. Australian Journal of Ecology 15: 321-327

Mac Nally, R.C. 1995 A protocol for classifying regional dynamics, exemplified by using woodland birds in south-eastern Australia. Australian Journal of Ecology 20: 442-454

153 Mac Nally, R.C. 1996 A winter’s tale-among-year variation in bird community structure in a south-eastern Australian forest. Australian Journal of Ecology 21: 280-291.

Mac Nally, R.C. 1997 Population densities in a bird community of a wet sclerophyllous Victorian forest. Emu 97: 253-258

Mac Nally, R. 2000 Coexistence of a locally undifferentiated foraging guild: avian snatchers in a southeastern Australian forest. Austral Ecology 25: 69-82.

Mac Nally, R.C. & McGoldrick, J.M. 1997 Landscape dynamics of bird communities in relation to mass flowering in some eucalypt forests of central Victoria, Australia. Journal of Avian Biology 28: 171-183

Mac Nally, R.C. & Quinn, G.P. 1998 The importance of scale in ecology. Australian Journal of Ecology 23: 1-7

Majer, J.D., Recher, H.F., Wellington, A.B., Woinarski, J.C.Z. & Yen, A.L. 1997 Invertebrates of eucalypt formations. Pp 278-302 In: Eucalypt ecology: individuals to ecosystems. Williams, J.E. & Woinarski, J.C.Z. (eds.) Cambridge University Press, Cambridge, New York

Major, R.E., Christie, F.J. & Gowing, G. 2001 Influence of remnant and landscape attributes on Australian woodland bird communities. Biological Conservation 102: 47-66.

Malizia, L.R. 2001 Seasonal fluctuations of birds, fruits, and flowers in a subtropical forest of Argentina The Condor 103: 45-61

Manly, B.F. 1990 On the statistical analysis of niche overlap data. Canadian

154 Journal of Zoology 68: 1420-1422

Marchant, S. 1979 The birds of forests and woodland near Moruya, New South Wales, Australia. Australian Birds 13: 59-68

Marchant, S. & Higgins, P.J. (eds.) 1993 Handbook of Australian, New Zealand and Antarctic Birds, Volume 2 , Raptors to Lapwings Oxford University Press, Melbourne

Martin, T.G. & Catterall, C.P. 2001 Do fragmented coastal heathlands have habitat value to birds in eastern Australia? Wildlife Research 28: 17-31

McCullage, P. & Nelder, J.A. 1989 Generalized Linear Models 2nd ed. Chapman and Hall, London

McCulloch, E.M. 1997 Lure of the lerp. The Bird Observer 778: 12

McCulloch, E.M. 1990 Parental feeding behaviour of the Noisy Friarbird Philemon corniculatus. Australian Bird Watcher 13: 226-230.

McFarland, D.C. 1984 Insects in flowers: a potential source of protein for honeyeaters. Australian Birds 18: 73-78

McFarland, D.C. 1985 Diurnal and seasonal changes in aggression in a honeyeater community. Corella 9: 22-25

McFarland, D.C. 1986 The organisation of a honeyeater community in an unpredictable environment. Australian Journal of Ecology 11: 107-120

McFarland, D.C. 1996 Aggression and nectar use in territorial non-breeding

155 New Holland Honeyeaters Phylidonyris novaehollandiae in eastern Australia. Emu 96: 181-188

McFarland, D.C. & Ford, H.A. 1991 The relationship between foraging ecology and social behaviour in Australian honeyeaters. Pp. 1141-1155 Acta XX Congress of International Ornithology Bell, D.B. (ed.) New Zealand Ornithological Trust Board, Wellington, New Zealand.

McGoldrick, J.M. & Mac Nally, R. 1998 Impact of flowering on bird community dynamics in some central Victorian eucalypt forests. Ecological Research 13: 125-39.

McKnight, S.K. & Hepp, G.R. 1998 Foraging-niche dynamics of Gadwells and American Coots in Winter. Auk 115: 670-683

Meagher, D. 1991 The MacMillan Dictionary of the Australian Environment MacMillan, Crows Nest

Miles, D.B. 1990 The importance and consequences of temporal variation in avian foraging behaviour. Studies in Avian Biology 13: 210-217

Morris, A.K. 1975 The birds of Gosford,Wyong and Newcastle (County of Northumberland) Australian Birds 9: 37-76

Morris, A.K. 1984 Birds Pp 46-54 In: The Dubbo Region: A Natural History Macquarie Publications

Morris, A.K. 1986 The birds of Sydney Harbour National Park, New South Wales. Australian Birds 20: 65-81

156 Morris, A.K., McGill, A.R. & Holmes, G. 1981 Handlist of Birds in New South Wales. New South Wales Field Ornithologists Club, Sydney

Morris, W.J. & Wooller, R.D. 2001 The structure and dynamics of an assemblage of small birds in a semi-arid eucalypt woodland in south-western Australia Emu 101: 7-12.

Moulds, M.S. 1990 Australian Cicadas. New South Wales University Press, Kensington.

Moysey, E.D. 1997 A study of resource partitioning within the Helmeted Honeyeater Lichenostomus melanops cassidix during non-breeding season. Emu 97: 207-219

Newland, C.E. & Wooller, R.D. 1985 Seasonal changes in a honeyeater assemblage in Banksia woodland near Perth, Western Australia. New Zealand Journal of Zoology 12: 631-636

Norris K. & Johnstone I. 1998 Interference competition and the functional response of Oystercatchers searching for Cockles by touch. Animal Behaviour 56: 639-650

Nour N., Suhonen J., Vandamme R., Matthysen E. & Dhondt A.A. 1997 Does the dominant Nuthatch Sitta europaea affect the foraging behaviour of the subordinate treecreeper Certhia brachydactyla in small forest fragments. Ardea 85: 259-267

Officer, H.R. 1969 Australian Flycatchers The Bird Observers Club of Australia, Melbourne

157 Oliver, D.L. 1998 The breeding behaviour of the endangered Regent Honeyeater, Xanthomyza phrygia, near Armidale, New South Wales. Australian Journal of Zoology 46: 153-170

Oliver, D.L. 2000 Foraging behaviour and resource selection of the Regent Honeyeater Xanthomyza phrygia in northern New South Wales. Emu 100: 12-30

O’Neill, R.V. & King, A.W. 1998 Homage to St. Michael: or why are there so many books on scale? Pp 3-16 In: Peterson, D.L. & Parker, V.T. (eds.) Ecological Scale: Theory and Applications Columbia University Press, New York

Osbourne, W.S. & Green, K. 1992 Seasonal changes in composition, abundance and foraging behaviour of birds in the Snowy Mountains. Emu 92: 93-105

Pacala, S. & Roughgarden, J. 1982 Resource partitioning and interspecific competition in two-species insular Anolis lizard communities. Science 217: 444-446

Pacala, S.W. & Roughgarden, J. 1985 Population experiments with the Anolis lizards of St. Maarten and St. Eustatius. Ecology 66: 129-141

Palestrini, C., Barbero, E. & Rolando, A. 1998 Intra- and interspecific aggregation among dung beetles (Coleoptera, Scarabaeoidea) in an alpine pasture. Journal of Zoology 245: 101-109

Parsons, H.M. 2001 Factors influencing urban bird assemblages of the Greater Sydney region. Honours thesis, University of Wollongong

158 Paton, D.C. 1979 The behaviour and feeding ecology of the Phylidonyris novaehollandiae in Victoria. Ph.D. Thesis, Monash University, Melbourne

Paton,D.C. 1980 The importance of manna,honeydew and lerp in the diets of honeyeaters. Emu 80: 213-226

Paton, D.C. 1982 The diet of the New Holland honeyeater, Phylidoyris novaehollandiae. Australian Journal of Ecology 7: 279-298

Paton, D.C. 1986 Honeyeaters and their plants in south-eastern Australia. Pp 9-19 In: The Dynamic Partnership: Birds & Plants in Southern Australia. Ford, H.A. & Paton, D.C. (eds.) Government Printers, South Australia.

Paton, D.C. 2000 Disruption of bird-plant pollination systems in South Australia. Conservation Biology 14: 1232-1234

Paton, D.C., Rogers, D.J. & Harris, W. 2004 Birdscaping the environment: restoring the woodland systems of the Mt Lofty region, South Australia. Pp 331-358 In: Conservation of Australia’s Forest Fauna (2nd ed.), edited by Lunney, D. Royal Zoological Society of New South Wales, Mosman, NSW, Australia.

Paton, D.C. & Collins, B.G. 1989 Bills and tongues of nectar-feeding birds: A review of morphology, function and performance, with intercontinental comparisons. Australian Journal of Ecology 14: 473-506

Paton, D.C. & Ford, H.A. 1977 Pollination by birds of native plants in South Australia. Emu 77: 73-85

159 Perrins, C.M. & Birkhead, T.R. 1983 Pp 159-183 In: Avian Ecology. Blackie & Son, Glasgow

Petit, L.J., Petit, D.R. & Smith, K.G. 1990 Precision, confidence, and sample size in the quantification of avian foraging behaviour. Studies in Avian Biology 13: 193-198

Piper S.D. & Catterall C.P. 2003 A particular case and a general pattern: hyperaggressive behaviour by one species may mediate avifaunal decreases in fragmented Australian forests. Oikos 101: 602-614.

Pizzey, G. & Doyle, R. 1980 A Field Guide to the . Collins, Sydney.

Plumpton, D.L. & Andersen, D.E. 1997 Habitat use and time budgeting by wintering Ferruginous Hawks. The Condor 99: 888-893

Pyke, G.H. 1980 The foraging behaviour of Australian honeyeaters: a review and some comparisons with hummingbirds. Australian Journal of Ecology 5: 343-369

Pyke, G.H. 1981 Why hummingbirds hover and honeyeaters perch. Animal Behaviour 29: 861-867

Pyke, G.H. 1983a Seasonal pattern of abundance of honeyeaters and their resources in heathland areas near Sydney. Australian Journal of Ecology 8: 217- 233.

Pyke, G.H. 1983b Relationships between honeyeater numbers and nectar

160 production in heathlands in Sydney. Australian Journal of Ecology 8: 217-233

Pyke, G.H. 1985 The relationships between abundance of honeyeaters and their food resources in open forest areas near Sydney. Pp. 65-77 In: Birds of Forests and Woodland: Ecology, Conservation, Management Keast, A., Recher, H.F., Ford, H. & Saunders, D. (eds.) Surrey Beatty & Sons, Chipping Norton

Pyke, G.H., Christy, M. & Major, R.E. 1996 Territoriality in honeyeaters: reviewing the concept and evaluating available information. Australian Journal of Zoology 44: 297-317

Pyke, G.H., O’Connor, P.J. & Recher, H.F. 1993 Relationship between nectar production and yearly and spatial variation in density and nesting of resident honeyeaters in heathland near Sydney. Australian Journal of Ecology 18: 221- 229

Pyke, G.H. & Recher, H.F. 1988 Seasonal changes of capture rate and resource abundance for honeyeaters and Silvereyes in heathland near Sydney. Emu 88: 33-42

Ramsey, M.W. 1989 The seasonal abundance and foraging behaviour of honeyeaters and their potential role in the pollination of . Australian Journal of Ecology 14: 33-40

Rasch, G. & Craig, J.L. 1988 Partitioning of nectar resources by New Zealand honeyeaters. New Zealand Journal of Zoology 15: 185-190

Recher H.F. 1981 Nectar-feeding and its evolution among Australian vertebrates. Pp 1637-1648 In:- Ecological Biogeography of Australia (ed. A.

161 Keast), Dr. W. Junk, The Hague

Recher, H.F. 1989 Foraging segregation of Australian Warblers (Acanthizidae) in open forests near Sydney, New South Wales. Emu 89: 204-215

Recher, H.F. 1990 Specialist or generalist: Avian response to spatial and temporal changes in resources. Studies in Avian Biology 13: 333-336

Recher, H.F. & Abbott, I.J. 1970 The possible significance of hawking by honeyeaters and its relation to nectar feeding. Emu 70: 90

Recher, H.F. & Gebski, V. 1990 Analysis of the foraging ecology of eucalypt forest birds: sequential versus single-point observations. Studies in Avian Biology 13: 174-180

Recher, H.F. & Holmes, R.T. 1985 Foraging ecology and seasonal patterns of abundance in a forest avifauna. Pp. 79-96 In: Birds of Forests and Woodland: Ecology, Conservation, Management Keast, A., Recher, H.F., Ford, H. & Saunders, D. (eds.) Surrey Beatty & Sons, Chipping Norton

Recher, H.F., Holmes, R.T., Shultz, M., Shields, J. & Kavanagh, R. 1985 Foraging patterns of breeding birds in eucalypt forest and woodland of southeastern Australia. Australian Journal of Ecology 10: 399-420

Recher, H.F., Majer, J.D. & Ford, H.A. 1991 Temporal and spatial variation in the abundance of eucalypt canopy invertebrates: response of forest birds. Pp 1568-1775 In: XX Congressus Internationalis Ornithologoci, Christchurch. Bell, D.B. et al. (eds.)New Zealand Ornithological Trust Board.

162 Recher, H.F. & Recher, J.A. 1968 Comments on the escape of prey from avian predators. Ecology 49: 560-562

Reid, J.R.W. 2000 Threatened and declining birds in the New South Wales Sheep- Wheat Belt: II. Landscape relationships - modelling bird atlas data against vegetation cover. Consultancy report to NSW National Parks and Wildlife Service CSIRO Sustainable Ecosystems, Canberra

Renjifo, L.M. 1999 Compositional changes in a subandean avifauna after long-term forest fragmentation Conservation Biology 13: 1124-1139

Rice, J. 1978 Ecological relationships of two interspecific territorial vireos. Ecology 59: 526-538

Robinson, S.K. & Holmes, R.T. 1982 Foraging behaviour of forest birds: the relationship among search tactics, diet and habitat structure. Ecology 63: 1918-1931

Robinson, S.K. & Holmes, R.T. 1984 Effects of plant species and foliage structure on the foraging behaviour of forest birds. Auk 101: 672-684

Rose, A.B. 1973 Food of some Australian birds. Emu 73: 177-183.

Rose, A.B. 1999 Notes on non-nectar foods of some honeyeaters in eastern New South Wales. Australian Bird Watcher 18: 26-34

Roughgarden, J. 1986 A comparison of food-limiting and space-limiting animal competition communities. Pp 492-516. In: Diamond,J & Case,T.J. (eds) Community Ecology Harper and Row, New York

163 Runciman, D. 1996 Activity budget of non-breeding Helmeted Honeyeaters. Emu 96: 62-65

Saunders, A.S.J. 1993 Seasonal variation in the distribution of the Noisy Friarbird Philemon corniculatus and the Red Wattlebird Anthochaera carunculata in eastern New South Wales. Australian Bird Watcher 15: 49-59

Saunders, A.S.J. 2002 Red Wattlebirds and Noisy Friarbirds in the Cumberland County. Cumberland Bird Observers Club Newsletter 23(6): 4-5

Saunders, D.A. & Hobbs, R.J. 1991 Nature Conservation: The Role of Corridors Surrey Beatty and Sons, Sydney

Schmitt, R.J. & Coyer, J.A. 1983 Variation in surfperch diets between allopatry and sympatry: circumstantial evidence for competition. Oecologia 58: 402-410

Schodde, R. & Mason, I.J. 1999 The Directory of Australian Birds: Passerines CSIRO Publishing, Collingwood

Schoener, T.W. 1983 Field experiments on interspecific competition. American Naturalist 122: 240-285.

Schoener, T.W. 1986 Resource partitioning. Pp. 91-126. In:- Kikkawa,J. & Anderson,D.J. Community Ecology: Patterns and Process Blackwell Scientific Publications, Melbourne.

Schrader, N. 1987 The Flora and Fauna of the Parkes Shire Parkes Naturalist Group, Parkes

164 Serventy, V.N. (ed.) 1982 The Wrens and Warblers of Australia Angus and Robertson, Sydney

Shapiro, S.S., Wilk, M.B. & Chen, H.J. 1968 A comparative study of various tests for normality. Journal of the American Statistics Association 63: 1343-1372

Shields, J. 1994 Pallid Cuckoo Cuculus pallidus Pp. 3-7 In: Strahan, R. (ed.) Cuckoos, Nightbirds and Kingfishers of Australia Angus and Robertson, Sydney

Shelly, D. 1998 Survey of vertebrate fauna and habitats in a cypress pine- ironbark forest in Central-West New South Wales. Australian Zoologist 30: 426-436

Sherry, T.W. & McDade, L.A. 1982 Prey selection and handling in two neotropical hover-gleaning birds. Ecology 63: 1016-1028

Slater, P.J. 1994 Niche overlap between three sympatric, short-billed honeyeaters in Tasmania. Emu 94: 186-192

Slater, P.J. 1995 The interaction of bird communities with vegetation and season in Brisbane Forest Park. Emu 95: 194-207

Smith, N. 1980 Noisy Friarbird attacks Black-fronted Plover chick. Australian Birds 14: 78

Smith, P. 1989 Changes in a forest community during a period of fire and drought near Bega, New South Wales. Australian Journal of Ecology 14: 41-54

Snow, D.W. & Snow, B.K. 1980 Relationships between hummingbirds and flowers in the Andes of Columbia. Bulletin of the British Museum of Natural

165 History: Zoology 38: 105-139

Specht, R.L., Roe, E.M. & Boughton, V.H. (eds.) 1974 Conservation of major plant communities in Australia and Papua New Guinea. Australian Journal of Botany Supplement 7: 667

Stanley M.C. & Lill A. 2002 Avian fruit consumption and seed dispersal in a temperate Australian woodland Austral Ecology 27: 137-148

St, Clair, C.C., Belisle, M., Desrochers, A. & Hannon, S. 1998 Winter responses of forest birds to habitat corridors and gaps. Conservation Ecology 2: http://www.consecol.org/vol2/iss2/art13

Sun, C. & Moermond, T.C. 1997 Foraging ecology of three sympatric Turacos in a montane forest in Rwanda. The Auk 114: 396-404

Swart J.M., Richardson P.R.K. & Ferguson J.W.H. 1999 Ecological factors affecting the feeding behaviour of Pangolins (Manis temminckii). Journal of Zoology 247: 281-292

Szaro, R.C., Brawn, J.D. & Balda, R.P. 1990 Yearly variation in resource-use behaviour by Ponderosa Pine forest birds. Studies in Avain Biology 13: 226-236

Tame, T. 1992 Acacias of southeast Australia. Kangaroo Press, Kensington

Taylor, I.M. 1987 Murrumbidgee River corridor bird surveys: summary of results. Canberra Bird Notes 12: 110-131

Taylor, M. 1992 Birds of the Australian Capital Territory: An Atlas

166 Canberra Ornithologists Group and National Capital Planning Authority

Thomas, D.G. 1980 Foraging of honeyeaters in an area of Tasmanian sclerophyll forest. Emu 80: 55-58

Thompson, D.J. 1978 Prey size selection by larvae of the damsel-fly Ischnura elegans (Odonata). Journal of Animal Ecology 47: 769-785

Traill, B.J., Collins, E., Peake, P. & Jessup, S. 1996 Current and past status of the birds of Chiltern- A box-ironbark forest in north-eastern Victoria. Australian Bird Watcher 16: 309-326

Turner, R.J. 1992 Effect of wildfire on birds at Weddin Mountains, New South Wales. Corella 16: 65-74

Underwood, A.J. 1986 The analysis of competition by field experiments. Pp. 240-268. in :- Kikkawa,J. & Anderson,D.J. (eds.) Community Ecology: Pattern and Process Blackwell Scientific Publications, Melbourne. van de Kam, J., Ens, B., Piersma, T. & Zwarts, L. 2004 Shorebirds: An Illustrated Behavioural Ecology KNNV Publishers, The Netherlands van der Wal, R., Kunst, P. & Drent, R. 1998 Interactions between hare and Brent Goose in a salt marsh system: evidence for food competition? Oecologia 117: 227-234

Vaughton, G. 1990 Seasonal variation in honeyeater foraging behaviour, inflorescence abundance and fruit set in (Protacae). Australian Journal of Ecology 15: 109-116

167 Venables, W.N. & Ripley, B.D. 1997 Modern Applied Statistics with S-Plus 2nd ed. Springer, New York

Wahungu, G.M. 1998 Diet and habitat overlap in two sympatric primate species, the Tana Crested Mangabey Cercocebus galeritus and Yellow Baboon Papio cynocephalus. African Journal of Ecology 36: 159-173

Ward, P. 1965 Feeding ecology of the black-faced Dioch Quelea quelea in Nigeria. Ibis 107: 173-214

Walker, J. & Hopkins, M.S. 1990 Vegetation. Pp 58-77 In: Australian Soil and Land Survey: Field Handbook McDonald R.C., Isbell, R.F., Speight, J.G., Walker, J. & Hopkins, M.S. 2nd ed. Goanna Print, Canberra

Walker, S.R., Novaro, A.J. & Branch, L.C. 2003 Effects of patch attributes, barriers, and distance between patches on the distribution of a rock-dwelling rodent (Lagidium viscacia) Landscape Ecology 18: 187-194

Weathers, W.W., Paton, D.C. & Seymour, R.S. 1996 Field metabolic rate and water flux of nectarivorous honeyeaters. Australian Journal of Zoology 44: 445-460

Weathers, W.W. & Seymour, R.S. 1998 Behaviour and time-activity budgets of Mallefowl Leipoa ocellata in South Australia. Emu 98: 288-296

Webb, J.K. & Shine, R. 1993 Prey-size selection, gape limitation and predator vulnerability in Australian blindsnakes (Typhlopidae). Animal Behaviour 45: 1117-1126

Westphal, M.I., Field, S.A., Tyre, A.J., Paton, D. & Possingham, H.P. 2003

168 Effects of landscape pattern on bird species distribution in the Mt. Lofty Ranges, South Australia Landscape Ecology 18: 413-426

Wheelwright, N.T. 1985 Fruit size, gape width, and the diets of fruit-eating birds. Ecology 66: 808-818

White, D.W. & Stiles, E.W. 1991 Fruit harvesting by American Robins: influence of fruit size. The Wilson Bulletin 103: 690-692

Wiens, J.A. 1981 Scale problems in avian censusing. Studies in Avian Biology 6: 513-521

Wiens, J.A. 1989 The Ecology of Bird Communities Vols. I & II. Cambridge University Press

Wiens, J.A. 1997 Metapopulation dynamics and landscape ecology Pp 43-62 In: Hanski, I.A. & Gilpin, M.E. (eds.) Metapopulation Biology: Ecology, Genetics and Evolution Academic Press, San Diego

Wiens, J.A., Rotenberry, J.T. & Van Horne, B. 1986 A lesson in the limitations of field experiments: shrubsteppe birds and habitat alteration. Ecology 67: 365-376

Woinarski, J. 1985 Foliage-gleaners of the tree-tops, the Pardalotes. Pp. 165- 175 In: Birds of Forests and Woodland: Ecology, Conservation, Management Keast, A., Recher, H.F., Ford, H. & Saunders, D. (eds.) Surrey Beatty & Sons, Chipping Norton

Woinarski, J.C.Z., Cullen, J.M., Hull, C. & Nayudu, R. 1989 Lerp-feeding in birds: A smorgasbord experiment. Australian Journal of Ecology 14: 227-234

169 Wolf, L.L. 1970 The impact of seasonal flowering on the biology of some tropical hummingbirds. Condor 72: 1-14

Wolf, L.L. & Hainsworth, F.R. 1971 Time and energy budgets of territorial hummingbirds. Ecology 52: 980-988

Wooller, R.D. 1984 Bill shape and size in honeyeaters and other small insectivorous birds in Western Australia. Australian Journal of Zoology 32: 657- 661

Yamagishi, S. & Eguchi, K. 1996 Comparative foraging ecology of Madagascar vangids (Vangidae). Ibis 138: 283-290

Zach, R. & Falls, J.B. 1978 Prey selection by captive ovenbirds (Aves: Parulidae). Journal of Animal Ecology 47: 929-943

Zar, J.H. 1996 Biostatistical Analysis 3rd. ed. Prentice-Hall Incorporated, Englewood Cliffs, New York

Zaret, T.M. 1980 Predation and Freshwater Communities. Yale University Press, New Haven

170