University of Wollongong Research Online

University of Wollongong Thesis Collection University of Wollongong Thesis Collections

2011 Variation in the health of tropical finches in relation to conservation status, season and land tenure Kimberly Lynn Maute University of Wollongong, [email protected]

Recommended Citation Maute, Kimberly Lynn, Variation in the health of tropical finches in relation to conservation status, season and land tenure, Doctor of Philosophy thesis, Department of Biological Sciences, Faculty of Science, University of Wollongong, 2011. http://ro.uow.edu.au/ theses/3390

Research Online is the open access institutional repository for the University of Wollongong. For further information contact Manager Repository Services: [email protected].

Variation in the health of tropical finches in relation to conservation status,

season and land tenure

A thesis submitted in fulfilment of the requirements for the award of the degree

Doctor of Philosophy

from

University of Wollongong

by

Kimberly Lynn Maute, BSc (Hons)

Department of Biological Sciences

2011

i

Thesis Certification

I, Kimberly Lynn Maute, declare that this thesis, submitted in fulfilment of the requirements for the award of Doctor of Philosophy, in the Department of Biological

Sciences, University of Wollongong, is wholly my own work unless otherwise referenced or acknowledged. The document has not been submitted for qualifications at any other academic institution. The data chapters (2 – 5) have been written and formatted for submission to academic journals. The role of potential co-authors in these chapters has been strictly supervisory. This thesis does not contain data that was substantially collected and analyzed by anyone other than myself.

Kimberly Lynn Maute

22 March 2011

ii Acknowledgements

I thank my supervisors, Sarah Legge (Australian Wildlife Conservancy), Lee

Astheimer and William Buttemer (Deakin University), Kris French and Jack Baker

(University of Wollongong) for helping me to take advantage of the opportunity to complete a PhD project. Their advice was indispensable in the completion of this research and the writing of my thesis. William Buttemer blind tested several of my finch plasma samples for osmolality, assisting in the validation of my assumptions that birds were not dehydrated (Chapter 1). Stephen Murphy and Douglas

Schaefer, while working for the Australian Wildlife Conservancy, sexed several of my bird blood samples. Numerous AWC staff at the Mornington Wildlife Sanctuary were kind enough to look after key field gear and red-blood samples while I was in the laboratory in Wollongong. I thank them for their caretaking. Lee Astheimer,

Amanda Guy and Colin Cortie organized my laboratory space and trained me in the basics of laboratory work, giving me the background I needed to perform and troubleshoot stress hormone (Corticosterone (CORT) and Corticosterone-binding- globulin (CBG)) assays using bird plasma.

I thank the many people who went to great lengths to ensure I was given access to funds, equipment, and study sites. I thank the Kowanyama and Pormpuraaw

Community Councils for the privilege of visiting their land and studying their wildlife, as well as the key people in the Lands offices who helped initiate contact

(Viv Sinnamon, Arvid, Kath, Colin and Rangers in Kowanyama; Rob Morris, John

Clark and Rangers in Pormpuraaw). The Jawoyn Association was also kind iii enough to allow me to study birds on their traditional lands in the Yinberri Hills, NT.

I would like to thank Mick Peirce and Ian McConnell for liaising with me for access to Jawoyn land. Vista Gold Mining also assisted with access to the Yinberri Hills area. I thank Brad Bell, Robbie Friel and the mine staff for their help and patience with my requests for access at such early hours of the morning. I thank the

Department of Defence officials who provided me access to Bradshaw Field

Training Area and helped me locate finches and waterholes (Kylie Browne, Range

Control Officers Sargent Pop Dawes, John McCrystal, Peter Watts, and the wonderful caretakers Peter and Jenny from Wildman). I would have been a bit lost without the help of countless QPWS and NT rangers and staff who guided me through the country and helped find me shelter (Andrew Hartwig and Steve

Anderson in QLD, Anne Walters, David Hooper, and Edith Falls Rangers in the

NT). The warm hospitality of pastoralists Shane Johnstone, James and Jenel,

Mark and Linda, Tom and Sue Shephard, and Don and Wendy was gratefully appreciated. Mark and Linda from Delamere were immensely helpful in providing me advice and support. My field work was made so much more pleasant by the amazing birders and researchers who helped us find finches (Michael Todd, Sue and Gary at Lotus Bird Lodge). I thank the Black-throated Finch Recovery Team and friends (Jo Wieneke, Tony Grice, Bill Holmes, Rosemary Payet, Marnie

McCullough) for keeping me informed of finch movements near Townsville. In particular, the Venables family were an invaluable source of moral and physical encouragement. I thoroughly enjoyed the company and support of so many remarkable individuals, organizations and communities.

iv Funding for this project was supported though a Linkage grant from the Australian

Research Council (ARC), the Australian Wildlife Conservancy (AWC), and the NT

Research and Innovation Board. I also thank Birds Australia, who generously awarded me three years of funding through the Stuart Leslie Bird Research award

(2007 and 2009) and Professor Allen Keast Award (2008). Other partners included

NT Parks and Wildlife, and Charles Darwin University (CDU) who both provided in- kind support. Dr. Stephen Garnett and CDU staff helped with field equipment maintenance and storage and I thank them for the time and effort required to administer the ARC Linkage Grant to my project and others affiliated with AWC.

I could not have survived my field seasons without all of the help and fun company of my friends and field assistants; Emily Haber, Jill Gautreaux and the ever-positive force of Jessica Griffiths! I would not have finished this thesis without the moral support of my friends and family. I cannot thank my mother, sister, and father,

James, Jacob, Karen and Olya enough for the endless hours of comforting words and advice that have helped me through this important process.

v Abstract

Despite the fact that Australia’s savanna is still a highly connected landscape dominated by native vegetation, over a quarter of Australia’s granivorous birds have declined in abundance or experienced range contractions over the last fifty years. It is widely assumed that cattle grazing and changed fire regimes have lowered the quality of savanna habitats for grass-seed eating species by lowering the productivity of important grasses. However, the relative impacts of changed land use on granivores such as finches remains poorly defined because of difficulties in monitoring the abundance of these semi-nomadic birds in remote areas. Consequently, my study compares the health measures of finch populations to determine if the timing and severity of changes in health indicators coincide with differing land management. The relative number of unhealthy animals can indicate a population’s probability of decline, as sick or chronically stressed individuals are less likely to survive and breed.

I monitored a variety of finch species in order to describe relative granivore responses to land management. Declining species are characterized by specialized diets, prolonged breeding seasons and strictly timed moulting periods, while non-declining species have opportunistic feeding, breeding and moult strategies which could aid them in responding to the seasonal savanna environment and recent habitat changes in Northern Australia. This led me to hypothesize that health patterns would differ between these species seasonally.

Therefore I monitored finches during the peak breeding season, when food vi resources are plentiful just after the wet growing season, and again in the non- breeding season, when grass seeds become scarce, and most finches are moulting. Based on the theory that grazing and burning lowers the quality of finch habitat, I also hypothesized that finches would show differences in health between populations living in areas of differing land management. Finches were monitored for two years in three study areas in the Northern Territory and three in

Queensland; one site per state was conservation managed (without grazing and fire), one was pastoral (grazed but without fire), and one was aboriginal (with frequent burning, but without high grazing). The health of declining Gouldian finches and non-declining Long-tailed and Masked finches were monitored in the

Northern Territory while declining Star and Black-throated finches were monitored in Queensland. Health indices included body condition measures (body mass, fat stores and muscle contour), haematocrit and stress response to capture using plasma corticosterone (CORT) and binding globulin levels. Finch habitat was surveyed in the Northern Territory in order to describe any differences in the grass layer in response to grazing or fire and among land tenures.

Health monitoring results suggest that all finch species have better body condition during breeding seasons and declining species have a possible pattern of chronic stress during non-breeding seasons. All finch species had higher body mass and muscle scores and lower fat stores during breeding compared to non-breeding.

Measures of CORT showed that Gouldian, Star, and Black-throated finches have stress response patterns opposite to that seen in most passerine studies; low stress responses during breeding and high stress during non-breeding. This vii pattern suggests that declining species are experiencing chronic stress during non- breeding, possibly due to low food resource availability during this time of year.

Non-declining Long-tailed and Masked finches had a normal stress response pattern, opposite to declining species.

High stress, low body condition and unusually high haematocrit were found in

Gouldian finches monitored on pastoral land during a non-breeding season.

Gouldian finches surveyed on aboriginal land during this season also had lowered body condition and unusually high haematocrit compared to birds sampled on conservation land. Habitat surveys showed that annual grass cover was low on pastoral land and perennial grass cover and productivity was low on aboriginal land compared to conservation managed land. These low grass measures coincided with the poor health of Gouldian finches on pastoral and aboriginal land, suggesting lowered grass abundance as a likely cause of poor finch health.

Though Black-throated finch body condition and haematocrit were similar between land tenures, stress responses were higher on pastoral land compared to conservation managed land during the non-breeding season in 2007. This suggests that Black-throated finches may be sensitive to grazing effects during some years. Star Finches monitored in open grasslands tended to have higher haematocrit levels, suggesting a lifestyle with higher energy demands. No consistent significant differences were found between the health measures of

Long-tailed or Masked finches sampled on different land tenures.

viii The results of this study suggest that tropical finch species have seasonal health variation which implies poorer condition during non-breeding compared to breeding seasons and a possible pattern of chronic stress in three declining species. This supports my hypothesis that declining species are more sensitive to seasonal changes. Coinciding differences in finch health, land tenure and grass cover estimates suggest that non-conservation management adversely affects finch health and grasses important to finches. This was particularly evident for the declining Gouldian finch and indicates a probability of future finch population declines on non-conservation managed land. Thus this research provides evidence of the threat of intensification of pastoral practices and continued frequent burning to declining savanna species. This study also reveals the utility of using health indices to determine the sensitivity of birds to environmental perturbations.

ix Table of Contents

Title Page ...... i

Thesis Certification ...... ii

Acknowledgements ...... iii

Abstract ...... vi

Table of Contents ...... x

List of Figures ...... xiv

1.0 Chapter One: Introduction and review of literature ...... 1

1.1 Australian savanna finches: History of declines and life history ...... 2

1.2 Changing land use may explain patterns of species declines ...... 5

1.3 Variation in finch species susceptibility to decline ...... 10

1.4 Biogeography of surviving grass-finch populations ...... 13

1.5 Finch health indices as bioindicators of population susceptibility to decline ...... 15 Measures of finch health predict survival and breeding success ...... 15 Factors affecting bird body condition ...... 17 Factors affecting bird haematocrit and haemoglobin ...... 21 Factors affecting stress in bird populations ...... 27

1.6 Research aims and hypotheses ...... 35

1.7 Experimental design and thesis structure ...... 37

2.0 Chapter Two: Seasonal patterns of health measures in three Australian grass-finches ...... 42

2.1 Introduction ...... 42

2.2 Methods ...... 49 Study site characteristics and survey seasons ...... 49 Finch capture and handling ...... 50 Measurement of finch body and haematological condition in the field ...... 52 Measurement of health indices from plasma ...... 53

x Validation of the assumption that health index variation was not due to dehydration ...... 58 Statistical analysis ...... 59

2.3 Results ...... 60 Residual CORT levels and CBG capacity ...... 60 Body condition (muscle score, fat score and body mass) ...... 67 Haematocrit and haemoglobin ...... 74 Interactions between health indices ...... 80

2.4 Discussion ...... 82 Seasonal stress response variation ...... 82 Seasonal body condition variation ...... 89 Correlations between health indices ...... 91 Variation in health indices due to age and sex ...... 93 Utility of health indices ...... 95 Conclusion ...... 100

3.0 Chapter Three: The effects of land management on grass cover in tropical savanna, northern Australia...... 102

3.1 Introduction ...... 102

3.2 Methods ...... 105 Study site descriptions ...... 105 Study plots and monitoring ...... 107 Estimation of grazing and burning levels ...... 108 Rainfall data ...... 109 Estimation of percent ground cover ...... 110 Measures of grass productivity ...... 110 Statistical analysis ...... 111

3.3 Results ...... 112 Site specific grazing, burning and rainfall levels ...... 112 Cover patterns by year and land tenure ...... 115 Cover patterns within tenures: effects of grazing and burning ...... 118 Annual grass cover ...... 119 Perennial grass cover and productivity...... 121

3.4 Discussion ...... 126 General ground cover differences between properties ...... 126 General ground cover characteristics within properties ...... 126

xi Differences in specific ground cover types between properties ...... 128 Implications for savanna conservation management ...... 132

4.0 Chapter Four: Variation in the health of declining and non-declining Australian savanna finch populations in relation to land tenure ...... 134

4.1 Introduction ...... 134

4.2 Methods ...... 138 Study species ...... 138 Study site and survey seasons ...... 139 Estimation of grazing and burning levels ...... 141 Finch capture and handling ...... 142 Measurement of finch condition in the field ...... 143 Measurement of health indices from plasma ...... 144 Statistical analysis ...... 144

4.3 Results ...... 145 Grazing and burning impacts ...... 145 Gouldian finch health indices ...... 148 Long-tailed finch health indices ...... 154 Masked finch health indices ...... 159

4.4 Discussion ...... 163 Health variation between tenures in a declining species ...... 163 Health variation between tenures in a non-declining species ...... 164 Health variation between individuals ...... 168 Conservation implications ...... 171

5.0 Chapter Five: Does variation in Gouldian finch health indices predict variation in Star finch and Black-throated finch health indices? ...... 174

5.1 Introduction ...... 174

5.2 Methods ...... 177 Study species populations...... 177 Study site characteristics and survey seasons ...... 178 Finch capture and handling ...... 179 Measurement of finch condition in the field ...... 180 Measurement of health indices from plasma ...... 180

5.3 Results ...... 183

xii Star finches ...... 183 Black-throated finches ...... 191

5.4 Discussion ...... 197 Seasonal variation in stress ...... 197 Seasonal variation in finch health indices ...... 198 Site differences finch health indices ...... 200 Inter-annual patterns in finch health indices...... 201 Conclusion ...... 202

6.0 Chapter Six: Conclusion and General Discussion ...... 204

6.1 Savanna finch health variation offers insight into conservation priorities ...... 204 Background and summary of findings ...... 204 Seasonal heath changes in finches and conservation implications ...... 205 Heath differences between populations on different land tenures ...... 209 Poor finch health coincides with lower grass cover and productivity ...... 215

6.2 Improving the use of physiological measures for gauging health status of free-living birds ...... 220

6.3 Concluding summary ...... 227

7.0 References ...... 229

xiii List of Figures

Chapter One

Figure 1.1: Map of Australia outlining the location of regions of the Northern

Territory and Queensland where this study was conducted. 39

Chapter Two

Figure 2.1: Graphs on the left are regression plots of a) Gouldian finch, b) Long- tailed finch, and c) Masked finch total plasma CORT levels (ng/mL) to handling time (minutes). Resulting residuals for each species are shown on graphs on the right. 56

Figure 2.2: Mean residual CORT levels (ng/mL) for Gouldian, Long-tailed, and

Masked finches during two non-breeding seasons (2007 and 2008) and two breeding seasons (2008 and 2009). Symbols represent means with standard error bars. Error values are small, and bars are not often visible. 62

Figure 2.3: Mean residual CORT levels (ng/mL) for Gouldian finches sampled at three sites during two non-breeding seasons (2007 and 2008) and two breeding seasons (2008 and 2009). Symbols represent means with standard error bars.

The direction of significant differences between sites are shown using abbreviations for each site (Tukey – Kramer HSD). 63 xiv

Figure 2.4: Mean CBG Capacity (nM) for adult Gouldian and Long-tailed finches during two non-breeding seasons (2007 and 2008) and two breeding seasons

(2008 and 2009). Symbols are means with standard error bars. Error values are small, and bars are not often visible. 64

Figure 2.5: Mean residual CORT levels (ng/mL) for Long-tailed finches sampled at three sites during two non-breeding seasons (2007 and 2008) and two breeding seasons (2008 and 2009). Symbols represent means with standard error bars.

The direction of significant differences between sites are shown using abbreviations for each site (Tukey – Kramer HSD). 66

Figure 2.6: Mean Gouldian finch residual body mass (a), muscle scores (b) and fat scores (c). Plots depict means for two non-breeding seasons (2007 and 2008) and two breeding seasons (2008 and 2009) and include values for differing ages and sexes. Standard error bars are shown when variation is large enough to be visible.

69

Figure 2.7: Mean Long-tailed finch residual body mass (a), muscle scores (b) and fat scores (c). Plots depict means for two non-breeding seasons (2007 and 2008) and two breeding seasons (2008 and 2009) and include values for differing ages and sexes. Standard error bars are shown when values are large enough to be visible. 71

xv Figure 2.8: Mean Masked finch residual body mass (a), muscle scores (b) and fat scores (c). Plots depict means for two non-breeding seasons (2007 and 2008) and two breeding seasons (2008 and 2009) and include values for differing ages and sexes. Standard error bars are shown where values are large enough to be visible. 73

Figure 2.9: Mean Gouldian finch haemoglobin (top graph) and haematocrit for two non-breeding seasons (2007 and 2008) and two breeding seasons (2008 and

2009). Plots depict means for different ages and sexes. Standard error bars are shown where values are large enough to be visible. 75

Figure 2.10: Mean Long-tailed finch haemoglobin (top graph) and haematocrit for two non-breeding seasons (2007 and 2008) and two breeding seasons (2008 and

2009). Plots depict means for different ages and sexes. Standard error bars are shown where values are large enough to be visible. 77

Figure 2.11: Mean Masked finch haemoglobin (top graph) and haematocrit for two non-breeding seasons (2007 and 2008) and two breeding seasons (2008 and

2009). Plots depict means for different ages and sexes. Standard error bars are shown where values are large enough to be visible. 79

xvi Chapter Three

Figure 3.1: Grazing and burning measure levels on each property during 2007,

2008 and 2009. Bars are means +/- SE, and significant differences between properties with values above 0 are signified with abbreviations for sites (P < 0.05).

114

Figure 3.2: Mean percent cover of grasses on plots by property and year. Bars are means +1 standard error, significant differences between properties are designated with abbreviations for sites (P < 0.05). 117

Figure 3.3: Mean height of grasses on plots by property and year. Means were calculated using estimation of grass height to the nearest 0.5m for each plot, which were then pooled and averaged for the property each year. Bars are means +1 standard error, significant differences between properties are designated with abbreviations for sites (P < 0.05). 117

Figure 3.4: Mean percent forb cover of plots by property and year. Bars are means +1 standard error, significant differences between properties are designated with abbreviations for sites (P < 0.05). 118

Figure 3.5: Mean percent annual grass cover on plots by property and year. Bars are means +1 standard error, significant differences between properties are designated with abbreviations for sites (P < 0.05). 120 xvii

Figure 3.6: Mean percent perennial grass cover on plots by property and year.

Bars are means +1 standard error, significant differences between properties are designated with abbreviations for sites (P < 0.05). 123

Figure 3.7: Mean perennial grass productivity (mean crown diameter in cm) within plots by property and year. Graphs represent the productivity of a) Triodia spp., b)

Chrysopogon fallax, c) Eriachne spp., d) Heteropogon spp., and e) Themeda triandra. Bars are means +1 standard error, significant differences between properties are designated with abbreviations for sites (P < 0.05). 125

Chapter Four

Figure 4.1: Grazing and burning levels on each property during 2007 and 2008.

Bars are means +/- SE, and significant differences between land tenures are signified with abbreviations for each site (P < 0.05). 147

Figure 4.2: Body condition measures for Gouldian finches sampled during non- breeding seasons 2007 and 2008. Graphs refer to measures of finch a) body mass, b) pectoral muscle score, and c) furcular fat score. Measures are means +/-

SE. Significant differences between land tenures are signified with abbreviations for each site (Tukey – Kramer HSD, P < 0.05). 150

xviii Figure 4.3: Haematocrit levels for Gouldian finches sampled during non-breeding seasons 2007 and 2008. Measures are means +/- SE. Significant differences between tenures are signified with abbreviations for each site (Tukey – Kramer

HSD, P < 0.05). 151

Figure 4.4: Residual CORT levels (ng/mL) for Gouldian finches sampled during non-breeding seasons 2007 and 2008. Measures are means +/- SE. Significant differences between land tenures are signified with abbreviations for each site

(Tukey – Kramer HSD, P < 0.05). 152

Figure 4.5: Long-tailed finch a) body mass and b) furcular fat score sampled during non-breeding seasons 2007 and 2008. Measures are means +/- SE.

Significant differences between land tenures are signified with abbreviations for each site (Tukey – Kramer HSD, P < 0.05). 155

Figure 4.6: Haematocrit levels for Long-tailed finches during non-breeding seasons 2007 and 2008. Measures are means +/- SE. Significant differences between land tenures are signified with abbreviations for each site (Tukey –

Kramer HSD, P < 0.05). 156

Figure 4.7: Residual CORT levels (ng/mL) for Long-tailed finches sampled during non-breeding seasons 2007 and 2008. Measures are means +/- SE. Significant differences between land tenures are signified with abbreviations for each site

(Tukey – Kramer HSD, P < 0.05). 157 xix

Figure 4.8: Body mass for Masked finches sampled during non-breeding seasons

2007 and 2008. Measures are means +/- SE. Significant differences between land tenures are signified with abbreviations for each site (Tukey – Kramer HSD, P

< 0.05). 160

Figure 4.9: Residual CORT levels (ng/mL) for Masked finches sampled during non-breeding seasons 2007 and 2008. Measures are means +/- SE. There were no significant differences between land tenures (Tukey – Kramer HSD, P > 0.05).

161

Chapter Five

Figure 5.1: Graphs on the left are regression plots of a) Star finch and b) Black- throated finch CORT levels (ng/mL) to handling time (minutes). Resulting residuals for each species are shown on graphs on the right. 182

Figure 5.2: Body condition measures for Star finches sampled during the non- breeding seasons 2007 and 2008, and the breeding seasons 2008 and 2009.

Graphs refer to measures of finch a) body mass, b) pectoral muscle score, and c) furcular fat score. Bars are means +/- SE. Significant differences between properties for each season determined by Tukey – Kramer HSD are designated with abbreviations for each site (P < 0.05). 185

xx Figure 5.3: Haematocrit levels for Star finches sampled during the non-breeding seasons 2007 and 2008, and the breeding seasons 2008 and 2009. Bars are means +/- SE. Significant differences between properties determined by Tukey –

Kramer HSD are designated with abbreviations for each site (P < 0.05). 186

Figure 5.4: Residual CORT levels (ng/mL) for Star finches sampled during the non-breeding seasons 2007 and 2008, and the breeding seasons 2008 and 2009.

Bars are means +/- SE. No significant differences between properties. 187

Figure 5.5: CBG capacity (nM) for Star finches sampled during the non-breeding seasons 2007 and 2008, and the breeding seasons 2008 and 2009. Bars are means +/- SE. No significant differences between properties were found. 188

Figure 5.6: Body condition measures for Black-throated finches sampled during the non-breeding seasons 2007 and 2008, and the breeding seasons 2008 and

2009. Graphs refer to measures of finch a) body mass, b) pectoral muscle score, and c) furcular fat score. Bars are means +/- SE. Significant differences between properties for each season determined by Tukey – Kramer HSD are designated with abbreviations for each site (P < 0.05). 193

Figure 5.7: Haematocrit levels for Black-throated finches sampled during the non- breeding seasons 2007 and 2008, and the breeding seasons 2008 and 2009. Bars are means +/- SE. A significant difference between properties determined by

Tukey – Kramer HSD is designated with abbreviations for sites (P < 0.05). 194 xxi

Figure 5.8: Residual CORT levels (ng/mL) for Black-throated finches sampled during the non-breeding seasons 2007 and 2008, and the breeding seasons 2008 and 2009. Bars are means +/- SE. Significant differences between properties determined by Tukey – Kramer HSD are designated with abbreviations for sites (P

< 0.05). 195

Chapter Six

Figure 6.1: Seasonal changes in mean body mass and muscle scores of finches

(Taken from data used in Chapters 2 and 5). Graphs represent, a) Gouldian,

Long-tailed, and Masked finch body mass, b) Gouldian, Long-tailed, and Masked finch mucle scores, c) Star and Black-throated finch body mass, d) Star and Black- throated finch muscle scores. Measures are means +/- SE. 206

Figure 6.2: Seasonal changes in mean stress response levels of finches (Taken from data used in Chapters 2 and 5). Graphs represent stress response levels for a) Gouldian finches, b) Long-tailed finches, c) Masked finches d) Star finches, and e) Black-throated finches. Measures are means +/- SE. Note the much larger range of stress response scores in d) and e) which necessitate the larger range of values on the x-axis of these graphs. 207

Figure 6.3: Body condition measures for Gouldian finches sampled during the non- breeding seasons 2007 and 2008 (Reproduced from Figure 4.2). Graphs refer to xxii measures of finch a) body mass, b) pectoral muscle score, and c) furcular fat score. Measures are means +/- SE. Significant differences between land management types are signified with abbreviations for sites (Tukey – Kramer HSD,

P < 0.05). 211

Figure 6.4: Haematocrit levels for Gouldian finches sampled during the non- breeding seasons 2007 and 2008 (Reproduced from Figure 4.3). Measures are means +/- SE. Significant differences between land management types are signified with abbreviations for sites (Tukey – Kramer HSD, P < 0.05). 212

Figure 6.5: Residual CORT levels (ng/mL) for Gouldian finches sampled during the non-breeding seasons 2007 and 2008 (reproduction of Figure 4.4). Measures are means +/- SE. Significant differences between land tenures are signified with abbreviations for sites (Tukey – Kramer HSD, P < 0.05). 213

Figure 6.6: Mean percent cover of grasses on plots by land management type and year (taken from Figure 3.2). Bars are means +1 standard error, significant differences between properties are designated with abbreviations for sites (P <

0.05). 216

Figure 6.7: Mean height of grasses on plots by land management type and year

(taken from Figure 3.3). Height is a general measure of grass productivity. Bars are means +1 standard error, significant differences between land types are designated with abbreviations for sites (P < 0.05). 216 xxiii List of Tables

Chapter Two

Table 2.1: Basic patterns of CORT and CBG levels found in birds during breeding compared to other times of the year. The five patterns listed are based on those found in the literature (cited in text) and in this study. 83

Chapter Three

Table 3.1: Rainfall totals and length of wet season based on data for properties from the Australian Bureau of Meteorology. 115

Table 3.2: Characteristic annual and perennial grass species as identified by similarity analysis. Average percent cover is mean cover of each species out of the total grass cover for each property and all years (Total grass cover Figure 3.2).

Cattle and wildlife use of grasses is based on published studies (Garnett & Crowley

1994; Ash et al. 1997; Crowley & Garnett 2001; Dostine & Franklin 2002). 124

Chapter Four

Table 4.1: Direction of significant differences in Gouldian finch health indices between tenures during the two years of the study (determined using Tukey –

Kramer HSD). 153 xxiv

Table 4.2: Direction of significant differences in Long-tailed finch health indices between tenures during the two years of the study (determined using Tukey –

Kramer HSD). 158

Table 4.3: Direction of significant differences in Masked finch health indices between tenures during the two years of the study (determined using Tukey –

Kramer HSD). 162

Chapter Five

Table 5.1: Direction of significant differences in Star finch health indices between sites (Lakefield National Park = L; Pormpuraaw = P) during the two years of the study (determined using Tukey – Kramer HSD). The abbreviation ‘ns’ refers to health indices found to have no significant differences between sites. 190

Table 5.2: Direction of significant differences in Black-throated finch health indices between sites (Lakefield National Park = Lake; Laudham = Laud) during the two years of the study (using Tukey – Kramer HSD). The abbreviation ‘ns’ refers to health indices found to have no significant differences between sites. 196

xxv 1.0 Chapter One: Introduction and review of literature

Habitat destruction, fragmentation and disturbance are some of the greatest threats to biodiversity worldwide, as shown by recent declines in species that are largely resident in areas facing widespread habitat loss, fragmentation or impacted by the encroachment of exotic species introductions (Brooks et al. 2002). Habitat specialists and poor dispersers constitute a significant proportion of these declining species (Travis 2003; Munday 2004). Travis (2003) and Munday (2004) hypothesized that specialist species are impacted by habitat degradation before generalist species due to their inability to adapt to environmental change or disperse across fragmented landscapes. In regions long occupied and altered by humans, generalist species, including many introduced species which are able to adapt to modified habitat are increasing while specialist species continue to decline

(Gibbs & Stanton 2001; Devictor et al. 2008). As a result of the intensification of agriculture and continued anthropogenically induced habitat fragmentation there is evidence of a new wave of avian declines in these areas of the world (Newton

2004).

In contrast, the tropical savanna of northern Australia is a relatively intact landscape subject to few human impacts other than low-intensity pastoralism

(Woinarski & Braithwaite 1990; Dyer et al. 2001; Woinarski et al. 2007). However, a suite of specialist species endemic to northern Australia seem to be on the brink of an extinction crisis similar to the mammal extinction event in the continent’s arid interior soon after the advent of European colonization (Burbidge & McKenzie 1 1989; Recher & Lim 1990; Woinarski et al. 2001; Woinarski & Fisher 2003).

Australian savanna-specialist plant and animal species follow the world-wide pattern of susceptibility to population decline, with many fire-sensitive species or those dependent on ground and understorey habitats showing the most pronounced declines in abundance and distribution (Woinarski & Ash 2002;

Andersen et al. 2005).

1.1 Australian savanna finches: History of declines and life history

The tropical northern savannas of Australia cover 2 million hectares of land and encompass a wide range of habitat types, most of which are characterized by grassy understorey with sparse tree cover (Woinarski et al. 2007). Northern

Australia has a tropical monsoonal climate with distinct wet and dry seasons.

Unlike most of the world’s savanna, Australia’s savanna landscape is largely intact and dominated by native vegetation. However, recent studies have documented the decline of a number of ground foraging granivorous birds across the tropical savannas of northern Australia (Franklin 1999; Franklin et al. 2005). Not only has the abundance of certain species changed, but there has been range contractions in 11 species of granivores with nearly half of these being small Estrildid finches

(Franklin 1999). Similar to changes observed in avian assemblages due to woodland clearing in south-eastern Australia (Ford et al. 2001; Martin &

Possingham 2005), there seems to be a decline in smaller, specialist species of birds, while larger species with generalist diets are increasing. The decrease in abundance of some granivores, and changes in regional granivorous bird 2 assemblages suggest the habitat changes in Australia’s tropical savannas are detrimental to small finches.

The first observed declines in granivorous birds occurred in the eastern savannas, following the advance of pastoralism across the continent (Woinarski 1992b;

Franklin 1999; Franklin et al. 2000; Woinarski & Ash 2002). The range contraction of the southern subspecies of the Black-throated finch (Poephila cincta cincta)1 is thought to have started as early as the 1950s in the central and south-eastern savanna of Queensland (Garnett 1992; Franklin 1999; Barrett et al. 2003). The southern subspecies of the Black-throated finch is now only found in a few key locations surrounding Townsville, Queensland. The disappearance of the critically endangered and possibly extinct south-eastern subspecies of the Star finch

(Neochmia ruficauda ruficauda) has followed a similar pattern of decline to the

Black-throated finch (Garnett 1992; Barrett et al. 2003). This subspecies was previously found throughout south-eastern Queensland but has vanished from this area in the past 60 years. In one of the most dramatic instances of granivore decline, the Gouldian finch (Erythura gouldiae) was seen and captured in the thousands in almost all tropical Australian savanna until the early 1980s, and is now thought to have been extirpated east of the Gulf of Carpentaria. Today, only small populations are found scattered across the Northern Territory and Western

Australia (Garnett 1992; Tidemann 1996; Franklin 1999; O'Malley 2006a). These and similar declines in other ground-foraging granivores coincide with the long

1 Avian nomenclature follows the current taxonomy of Australian birds (Christidis & Boles 2008). 3 history of grazing in southern and central Queensland (Ash et al. 1997; Woinarski

& Catterall 2004; Franklin et al. 2005).

Northern Australia’s 14 species of Estrildid finches (grass-finches) share numerous biological and ecological traits. These birds are relatively strict granivores as adults, only take insects occasionally and even feed their young grass seeds

(Higgins et al. 2006). Finches tend to rely on abundant annual grass seeds during the dry season and late wet season, and then switch to perennial grass seed sources when they become ripe in the early to late wet season. Because grass seed abundance changes seasonally, many grass-finches have large home-ranges

(2 - 20 km radius), and are predisposed to abrupt regional movements (Legge

2009). Most grass-finches either build dome nests in trees, grass or understorey shrubs, or nest in small cavities in trees. Grass-finches are highly productive breeders. Their breeding season is often long (up to 11 months) and they can produce 2 - 4 broods a season with averages of 4 - 6 young per brood. Most grass-finches are small (9 - 17cm) and short-lived, with an average life span of only

1 - 4 years (Higgins et al. 2006). These traits suggest grass-finches could be prone to large fluctuations in population size in response to interannual variation in environmental conditions that affect grass seed production and thus, breeding success.

Grass-finches undergo a complete moult of all body and flight feathers annually.

Some species have abbreviated moult timed to when particular short-term high protein seed sources are available, while others have a longer moult period, most 4 likely due to their ability to exploit non-seed protein sources during energetically demanding periods (Tidemann & Woinarski 1994; Franklin et al. 1998; Higgins et al. 2006). For example, the Long-tailed and Crimson finches (Neohmia phaeton) are known to have a higher proportion of insects in their diet during breeding and moult, which could help these birds support their longer moult period and opportunistic breeding seasons (Tidemann & Woinarski 1994; Todd et al. 2003;

Higgins et al. 2006). In contrast, Gouldian finches have a consistently low proportion of insects in their diet, and have extended breeding seasons and abbreviated moult periods that coincide with either sprouted seed availability or newly ripened seed, both of which are high in dietary protein (Allen & Hume 1997;

Higgins et al. 2006).

Grass-finches are found throughout most of Australia, but because of their grass seed diet and nesting requirements, there is a higher diversity and abundance of grass-finches found in northern Australia’s savannas. Nine of the 19 species found in Australia are confined to the tropics and another 5 are found in both the savannas, arid zones, or along the east coast. Thus the northern tropical savannas are considered to be highly important habitat for the conservation of these finches.

1.2 Changing land use may explain patterns of species declines

Changing land use across northern Australia is strongly correlated with population declines in many savanna flora and fauna, including granivorous birds. One of the 5 more obvious changes to Australia’s tropical savannas is the increase in native woody shrub and tree cover due to the combination of overgrazing by livestock and feral herbivores, fire suppression and decreased rainfall during the 1970s (Crowley

& Garnett 2000; Edwards et al. 2003; Fensham & Fairfax 2003; Sharp & Whittaker

2003; Werner et al. 2006). Though this increase in woody vegetation may be restricted to certain vegetation communities (Sharp & Bowman 2004), it has implications for species reliant on grasses for food or shelter, as a thicker understorey shrub layer can suppress growth in the grass layer (Crowley & Garnett

1998; Liedloff et al. 2001). Research has demonstrated a link between the absence of some avian species in areas that have experienced an increase in woody plants (Tassicker et al. 2006). This pattern of change has had a direct impact on the endangered northern subspecies of the Star finch (Neochima ruficauda clarescens) in Cape York Peninsula, as it is reliant on open grassland habitat that has been notably reduced by the increased density and abundance of woody shrubs (Crowley & Garnett 2000; Todd et al. 2003; Garnett et al. 2005).

Because all grass-finches have similar reliance on the grass layer for food and/or nest sites, it is very likely that shrub encroachment in the understorey of savanna could adversely affect finch species in other areas as well.

Grazing and changes in fire frequency have also been linked to changes in the savanna grass layer. Grazing decreases the productivity, diversity and abundance of perennial grass species, changes being apparently in some instances irreversibly (Ash et al. 1997; Friedel 1997; Crowley & Garnett 1998; Fensham &

Skull 1999; Landsberg et al. 1999; Bastin et al. 2003). Overgrazing can so 6 severely decrease fuel loads in the understorey that the consequent reduction in fire intensity is insufficient to kill woody vegetation, thus increasing growth of woody plant and suppression of the grass layer through shading and competition. The decline of perennial grass species is typically due to reduced seed production and seedling recruitment that can be caused by any type of defoliation (grazing or fire) or competition with woody plants. These conditions are also correlated with increases in abundance of annual grasses and forbs (McIvor & Gardner 1991;

O'Connor & Pickett 1992; Walker et al. 1997b; Ash & McIvor 1998; Crowley &

Garnett 2001; Bastin et al. 2003). Because fire is a form of defoliation, increased fire frequency and intensity in western savannas has also caused shifts in grass- layer productivity and composition, which include reduction or absence of seed production in perennial and fire-sensitive species (Mott & Andrew 1985; Lonsdale et al. 1998; Fisher et al. 2003; Russell-Smith et al. 2003a; Yates & Russell-Smith

2003; Andersen et al. 2005). Most granivorous birds in the tropical savannas depend on perennial grass seed during the early wet season, when annual seeds begin to sprout and are no longer available as a food source (Crowley & Garnett

1999). Increased grazing intensity and fire frequency, with concomitant declines in key grass seed food resources, could have serious implications for the survival of granivore populations in affected areas.

Though the direct effects of pastoral practices on grassland species are sometimes contested, the changes wrought by cattle and fire on riparian systems are well known worldwide. Overgrazing and frequent fire have both been shown to reduce the complexity of understorey structure and recruitment of trees, while increasing 7 erosion and silting of water bodies in wetlands and riparian habitats (Meeson et al.

2002; McIntyre et al. 2003). This has resulted in the decline in terrestrial and aquatic biota dependent on these areas in Australia and around the world. Once exotic plants are found in significant numbers, local extinctions of the native flora and fauna often occur in the affected areas (Kauffman & Krueger 1984; Jansen &

Healey 2003; Andersen et al. 2005; Martin et al. 2006). Most granivores are dependent on riparian systems for some part of their life and riparian degradation has been implicated in the decline of two northern Australian finches; the Crimson finch and the Star finch in the Cape York Peninsula (Todd et al. 2003; Garnett et al.

2005).

Pastoral practices have also had important impacts on Australia’s tropical savannas through the introduction of exotic plants. Though much of the savannas are still dominated by native pasture, surveys suggest that cattle and feral herbivores transport and help establish exotic grasses (McIntyre et al. 2003).

Once exotic plants are found in significant numbers, the native flora and fauna often respond with local extinctions in the areas affected (Rossiter et al. 2003;

Ferdinands et al. 2005). Though it is in the interest of pastoralists and conservationists to address the issue of exotic plant spread, few eradication programs are in place, and some rangeland scientists still favour the use of sowed non-native pasture in certain areas (Ash et al. 1997; Bortolussi et al. 2005). Thus the perturbing effects of exotic plants are continuing to contribute to changes in the grass layer that are negatively affecting a suite of sensitive native species.

8 Recent increases in fire frequency and the intensification of pastoralism are known to substantially reduce the abundance of fauna dependent on the savanna grass layer for food or shelter (Woinarski et al. 2006). Invertebrates such as ants, spiders and grasshoppers are intolerant of areas experiencing frequent fires, preferring ungrazed patches unburned for at least a year (Andersen 1991;

Landsberg et al. 1997; James et al. 1999; Woinarski et al. 2002; Hoffmann 2003;

Churchhill & Ludwig 2004). Similarly, surveys and experiments identify a range of vertebrates, mostly diurnal reptiles, ground-foraging birds and small mammals, that are sensitive to grazing and increased fire frequency (Woinarski & Fisher 1995;

Woinarski et al. 1999; Woinarski & Ash 2002; Fraser et al. 2003; James 2003;

Woinarski et al. 2004b; Andersen et al. 2005). Though most researchers attribute declines in these sensitive species to changes in the landscape due to pastoralism, few studies have successfully linked vertebrate abundance to actual habitat traits defined by pastoral management (vegetation or soil characteristics), due either to too fine or too course a scale of measurement (Woinarski & Ash 2002; Layme et al.

2004; Woinarski et al. 2004a; Price et al. 2005b). One suggestion to gain more insight into the processes associated with habitat/species interaction is to monitor specific species and the characteristics of their habitats as they change by using experimental grazing and fire manipulation of landscapes (Whitehead et al. 2005).

The increase in the number of artificial watering points has intensified the impact of cattle and feral herbivores on savanna and riparian systems. Distance from water has been shown to be a major influence on the distribution of grazing-sensitive species and changes in the ground vegetation attributed to overgrazing are 9 concentrated near water (Landsberg et al. 1997; James et al. 1999; Jansen &

Robertson 2001; Ash et al. 2004). As pastoralists increase the number of water points across the savanna, not only are they increasing the area available to cattle, but to exotic animals as well. Horses, donkeys, goats, buffalo and pigs have detrimental effects on native plants and animals when present in substantial numbers (Bayliss & Yeomans 1989; Landsberg et al. 1997; Werner et al. 2006).

These rangeland biologists also theorize that artificial water points may concentrate the activities of exotic and native predators. This has serious implications for prey groups such as small mammals, reptiles and ground foraging birds.

Thus changes to savanna vegetation, particularly the grass layer and riparian habitat, as well as increases in exotic plants, herbivores and predators are concurrent with declines in native biodiversity in northern Australian. These changes help explain the current extinction crisis facing specialist species endemic to northern Australia, despite the low intensity of agricultural use of savannas and abundance of native vegetation (Burbidge & McKenzie 1989; Recher & Lim 1990;

Woinarski et al. 2001; Woinarski & Fisher 2003).

1.3 Variation in finch species susceptibility to decline

Many species of grass-finches may be naturally susceptible to population decline.

It is notable that half of all tropical Australian species of grass-finches are declining, while others show no indication of changed abundance or distribution (Garnett 10 1992; Franklin 1999; Higgins et al. 2006). Several bodies of research suggest this may be due to a combination of effects including the specialized diet of some species, seasonal changes in food supply, and differences in life history traits among grass-finches.

Despite the fact that all Estrilldids are grass seed specialists, the diets of individual species vary considerably, even when found in the same habitats (Higgins et al.

2006). Masked and Long-tailed finches tend to take seeds from a diverse range of grass species and supplement their diet with insects, while endangered Gouldian finches focus on seeds of a few key species (Dostine & Franklin 2002). Adding protein sources such as insects to avian diets has positive effects on survival and breeding success for many species, including finches (Birkhead et al. 1999;

Schoech & Bowman 2003). If changes to the grass layer due to pastoral practices remove preferred grass species or reduce their productivity, it is likely that finch species with specialized diets will show more dramatic population declines or range contractions than more generalist species.

Australia’s finches may naturally encounter a period of low seed availability when seeds of annuals sprout during the early wet season before most perennial grasses have set seed (Tidemann 1996; Crowley & Garnett 2000). Declining species such as the Gouldian finch and Pictorella mannikin (Heteromunia pectoralis) are known to switch their diet to perennial grass seed species such as

Triodia spp. and Alloteropsis spp., both of which are sensitive to grazing and fire

(Crowley & Garnett 2001; Dostine et al. 2001; Higgins et al. 2006). In areas prone 11 to yearly burning, grass species important to finch diets such as Triodia spp. and

Alloteropsis spp. will have lowered seed output or completely fail to seed, causing a food resource bottleneck in the early wet season. Crowley and Garnett (2001) have shown that defoliation due to grazing or frequent fire substantially reduces the amount of seed produced and increases the time and rainfall needed for these plants to set seed. They hypothesize that finches reliant on perennial grasses may be disadvantaged in areas prone to overgrazing and frequent fire. The food resource bottleneck in the early wet season could be less stressful for finch species able to switch to a wider variety of seed sources and/or insect prey.

Differences in life history strategies of grass-finches may also cause different species to be more susceptible to population fluctuations. All grass-finches are typically short-lived and highly productive breeders but declining species like the

Gouldian finch tend to have more continuous and prolonged seasonal breeding attempts, shorter periods of moult and possibly shorter life spans than similar common finches (Tidemann & Woinarski 1994; Franklin et al. 1998; Dostine et al.

2001; Higgins et al. 2006). Thus, the Gouldian finch is likely to be more sensitive to annual fluctuations in food supply and breeding site availability and therefore require a higher rate of energy intake that may make these events more stressful to finches living in substandard habitats (Tidemann & Woinarski 1994). This would be especially true during the period of Gouldian finch moult, which can overlap with the period of food bottleneck in the early wet season. If declining species like the

Gouldian finch are naturally shorter-lived than other species, this trait would cause them to be less resilient to long periods of poor environmental conditions or 12 unusual disturbances. Less is known about the lives of other declining species, but research currently underway may determine whether they are under similar pressures to the Gouldian finch.

1.4 Biogeography of surviving grass-finch populations

Regions of higher rainfall, topography and shorter histories of grazing show less variation in granivore species abundance and distribution (Woinarski 1999;

Woinarski & Catterall 2004; Franklin et al. 2005; Woinarski et al. 2005). These trends imply that places with these qualities could be acting as refugia for grazing- or fire-sensitive species of plants and animals (Williams et al. 2002). For example, fire sensitive plant species have become rarer across the open savannas, and are now typically found in areas with high topographic relief (Bowman et al. 2001;

Fisher et al. 2003; Yates & Russell-Smith 2003).

Higher rainfall areas can be more resilient to grazing and fire due to the predictability and abundance of water for grass layer regeneration (Ludwig et al.

2001). Studies on fire frequency show that the majority of fires each year occur in high rainfall coastal and sub-coastal regions, due to both higher fuel loads in these areas and a high incidence of fire near coastal communities (Williams et al. 2002;

Russell-Smith et al. 2003b). However, the greatest extent of grass layer disturbance is associated with more arid regions (Landsberg et al. 1999). This could be due to the ability of perennial grasses to recover from fire events quickly after high rainfall wet seasons (Walker et al. 1997b) and to the less predictable 13 nature of arid regions, which amplifies the effect of defoliation in dry years (Friedel

1997; Bastin et al. 2003). Once drought-affected areas are subject to overgrazing or widespread fires, and rainfall does not return, the changes wrought by extreme defoliation may be irreversible (Sharp & Whittaker 2003).

High topographic variability tends to exclude cattle and widespread fire from some areas, creating a mosaic of microhabitats that are useful to granivores (Gill &

Bradstock 1995; Gill et al. 2000; Price et al. 2005a). Surveys have shown that fire- sensitive plant species have become rarer and are now typically found in areas with high topographic relief (Bowman et al. 2001; Fisher et al. 2003; Yates &

Russell-Smith 2003). Finches that rely on these plants during the wet season may more easily survive in patches of habitat left unburnt and ungrazed for several years. Gouldian finches, Painted finches (Emblema pictum) and Pictorella

Manikins are known to feed predominately on Triodia spp. (spinifex) seed during the early wet season in areas where the seed is available and Painted finches and

Pictorella Manikins use older tussocks of Triodia spp. for nesting (Higgins et al.

2006). Triodia spp. grasses tend to seed more regularly if left unburnt for 2 - 3 years, and often take longer to become suitable finch nest substrates, therefore unburnt habitat is important to these finches (O'Malley 2006b). Though reliant finch species do survive in areas where few patches of Triodia spp. are left unburnt for more than two years, it is likely that this is suboptimal habitat and birds are switching to other perennial grasses that may not be suitable for the long-term survival of these populations (O'Malley 2006a).

14 1.5 Finch health indices as bioindicators of population susceptibility to decline

Measures of finch health predict survival and breeding success

Land management changes have coincided with the decline of some granivorous bird species but actual fluctuations in populations due to increases in grazing or fire frequency have not been causally substantiated. Australian finch populations are difficult to monitor due to their semi-nomadic nature and thus, the existence of either continued declines or evidence of recovery of sensitive species has not been found. Despite the fact that late dry season waterhole counts of Gouldian finches indicate dramatic annual variation in population size, no clear trend in finch numbers has emerged in the last 10 years of intensive monitoring (Price et al. In preparation). The failure of current methods of population monitoring to predict the sustainability of finch populations prompts a call for new and more sensitive approaches for assessing population stability or change.

One such method involves using biomarkers of health and condition of individuals within a population as an indication of the population’s susceptibility to decline.

Recent studies of a variety of vertebrates highlight the usefulness of physiological measures in demonstrating that individuals under higher stress or in poor condition are less likely to breed and survive in the wild (Siegel 1980; Romero & Wikelski

2001; Stevenson 2005; Tracy et al. 2006).

15 Healthy wild populations are able to sustain population numbers through the balance of recruitment and mortality. Many biologists have demonstrated that avian reproductive success is affected by overall health of parent birds and that a wide range of environmental conditions can affect health. Diet quality and quantity and predator pressure are all examples of known factors found to affect avian breeding success (Clinchy et al. 2004; Cyr & Romero 2006; Zanette et al. 2006;

O'Dwyer et al. 2007). Because the health of individuals is a consequence of the integration of all environmental challenges confronting them, measures of an individual’s health status through characterization of certain biomarkers of stress and condition is likely to be highly and dynamically correlated with overall habitat quality for a given species. Such measures will also identify the proportion of individuals within an area that are likely to breed or survive, which will predict the likelihood of population decline. This is because birds in poor health have lower survival as a result of being less successful in avoiding predation, finding food and producing young (Dubiec & Cichon 2001; Kilgas et al. 2006b). If measures of stress and physiological condition can predict bird survival and breeding success, it follows that these measures will be useful in comparing the relative health of finch populations and providing a basis for understanding finch population trends.

Measures of individual condition have been used successfully to predict the likelihood of bird survival and breeding success, as well as explaining the physiological consequences of habitat disturbance (Koivula et al. 1996; Suorsa et al. 2004; Kilgas et al. 2006b; Nadolski et al. 2006; Stevenson & Woods 2006;

Bonier et al. 2007). Ecologists have documented that geographic, seasonal, 16 habitat, age and sex status interact in affecting overall health, including body condition, hormonal, metabolic, immunological and haematological status of birds

(see discussion below). Consequently, measurements of selected physiological variables can serve as biomarkers that aid in identifying overall avian population health.

Factors affecting bird body condition

Most studies examining health of free-living birds use combinations of body mass, body size, and estimates of visible fat depots to formulate condition indices. These measures are relatively easy to obtain in the field and the resultant body condition indices are generally good predictors of survival (Lundberg 1985; Magrath 1991;

Houston & McNamara 1993; Brown 1996; Merino et al. 2001; Stevenson & Woods

2006). In many small to medium-sized temperate species, heavier birds and those carrying more fat typically have a higher probability of over-winter survival. This could be because birds with higher fat scores are able to fast longer and maintain higher body temperatures at night (Stuebe & Ketterson 1982), but this could also represent those with better foraging abilities. It is well known that migratory birds increase fat and protein storage to fuel long-distance flights and the birds with larger stores are those most likely to reach their destination (Blem 1976; Lindstrom

& Piersma 1993; Seaman et al. 2006). The availability of suitable food affects avian foraging success which is in turn strongly linked to body mass and fat stores

(Leary et al. 1999; Strong & Sherry 2000; Eeva et al. 2003; Sanchez-Guzman et al.

2004). Therefore body mass and fat indices can be indicative of the foraging 17 quality of the environment. It is not surprising then that bird body mass and fat are typically reliable indicators of breeding success (Kitaysky et al. 1999; Eeva et al.

2003; O'Dwyer et al. 2007).

However, higher mass, muscle and fat levels are not always adaptive. Research has revealed that factors other than diet quality can affect how wild birds regulate body mass. For example, Great tits (Parus major) reduce their mass, but not their pectoral muscle volume, in response to the reduced wing surface area during moult (Senar et al. 2002). This means that moulting birds faced with lowered wing surface area have optimized mass without compromised muscle ability which maintains flight efficiency. Other studies show that birds in habitats with high predation risks tend to decrease their mass which then increases agility and/or lowers the energy requirements of carrying more weight (Lima 1986; Ekman &

Hake 1990; Houston & McNamara 1993; Metcalfe & Ure 1995; Brodin 2001).

Increased agility and decreased energy demands in predator dense habitats may increase these birds probability of survival. Therefore, these background effects should be taken into account when evaluating mass variation in birds. Though predation risk is difficult to control for or quantify, researchers avoid some biases by standardizing the time of day that birds are measured for fat and mass and recording the moult status of birds. Because mass increases proportionally with bird size, mass must be evaluated with an overall estimate of bird volume based on skeletal measures (Buttemer 1992).

18 While fat storage benefits birds with sporadic food availability or long-term fasting needs, it comes at a cost by increasing non-muscle mass, reducing agility and leaving animals more prone to predation. Birds experiencing periods of fasting and/ or living in habitats with unpredictable food resources are more likely to store fat than birds with access to predictable and abundant food sources (Rogers 1987;

Ekman & Hake 1990; Cuthill et al. 2000; Cresswell 2003). This has doubly negative implications for birds living in habitats with higher predation risk and lower or patchy food resources that require birds to be more agile and fly further distances (Lima 1986; Witter & Cuthill 1993). Social status is also correlated with variation in bird condition. Dominant individuals usually have better access to food or advanced food handling efficiency, making them less reliant on stored fat which, in turn, optimizes their body mass for flight efficiency. These traits are associated with greater individual survival in temperate zone non-breeding seasons (Lundberg

1985; Piper & Wiley 1990; Ekman & Lilliendahl 1993; Gosler 1996; Koivula et al.

1996; Carrascal et al. 1998; Leary et al. 1999). The overall evidence suggests that fat accumulation is chiefly related to energy requirements, which are higher for birds exposed to short days and cold nights, or during migration. In summary, increased fat storage in temperate zone birds is generally thought to relate to the unpredictability of food resources caused by lower social status and/or food shortages in winter.

In lower latitudes and altitudes, and in the southern hemisphere, birds may not need to carry fat over winter due to higher overnight temperatures and the lower energy demands of their non-migratory lifestyle. They consequently show less 19 annual variation in body fat levels or even opposite trends in mass variation compared with temperate zone conspecifics (King & Farner 1966; Blem 1976;

Rogers et al. 1993; Breuer et al. 1995). In both hemispheres, winter tends to be a season with limited food resources. However, it has been suggested that the lower body fat content of Zebra finches (Taeniopygia guttata) during winter aids in lowering their flight energy expenditure and improves their flight agility (Rozman et al. 2003). This would reduce their overall cost of making more frequent foraging flights during this season but also make them less vulnerable to predators. The literature describes two general scenarios of fat storage: (i) Arctic; one in which birds use fat as a critical source of fuel during periods when foraging time is limited; during long and/or cold nights or during migratory flights (King & Farner 1966; Blem

1976). These are typical adaptations of northern temperate and arctic species allowing them to exploit temporarily resource rich habitats. (ii) Alternatively, resident species in milder and less demanding climates, like Australian finches, are less likely to require fat storage (Breuer et al. 1995; Rozman et al. 2003). In fact, fat may actually disadvantage birds facing higher predation rates, due to the impact of higher mass on the flight agility of birds. However, during seasonal resource bottlenecks, moult and the early stages of courtship and breeding, small finches in the savannas of northern Australia may need extra energy stores and will therefore carry some extra fat despite the disadvantage to flight efficiency.

Many factors may affect the relative advantage of carrying fat, thus changes or differences in the fat levels of bird populations must be interpreted with care.

20 Factors affecting bird haematocrit and haemoglobin

Haematological measures also provide useful indicators of physiological status and have long been used as indicators of animal metabolic activity, disease, stress and nutritional status. The ratio of red blood cells to the total volume of plasma is termed haematocrit. Red blood cells contain haemoglobin, an oxygen-binding protein that is essential for oxygen transport throughout the body. Generally, low levels are indicative of unhealthy or less active individuals. Very high haematocrit levels increase blood viscosity, which decreases blood flow and heat loss (Sturkie

1986). Haematocrit and haemoglobin increase in response to higher energy demands, extreme temperature, dehydration and elevation, but can decrease due to blood parasites, chronic fasting and disease (Carpenter 1975; Averbeck 1992;

Carmi et al. 1994; Morton 1994; Horak et al. 1998b; Soler et al. 1999; Booth &

Elliott 2002). Dehydration and high ambient or animal body temperature affect haematological values through reduction in blood plasma volume. Few studies have detected changes in plasma volume caused by dehydration, due to the highly efficient mechanisms for conservation of plasma volume in birds (Goldstein &

Zahedi 1990; Carmi et al. 1993). Changes due to dehydration have mainly been observed in experiments under high ambient temperature and during forced fasting and removal of drinking water for extended periods (Vleck & Priedkalns 1985;

Carmi et al. 1994; Vleck et al. 2000). During experimental dehydration, avian haemoglobin levels decreased early due to erythrocyte shrinkage, while haematocrit did not decrease due to this dehydration effect. Eventually, the relative number and size of blood cells remained constant, while plasma volume 21 decreased due to dehydration, causing an increase in haematocrit. Under less extreme temperatures and water loss, plasma volume was protected (and thus blood flow critical for metabolism is preserved) due to thermal expansion of water, capillary permeability changes and the shift of body water from interstitial fluid and extracellular fluid to the blood (Dawson et al. 1983; Arad et al. 1989).

Despite their high surface area to volume ratio and potential for evaporative water loss, small granivorous birds are likely to be similarly resistant to water stress compared to larger arid-adapted birds. This results from their high-carbohydrate diet and the concomitant efficient production and use of metabolic water, high body temperatures and feather protection against evaporative water loss (MacMillen

1990; MacMillen & Baudinette 1993; Williams 1996). However, because

Australia’s tropical savannas are seasonally water restricted, it is still possible that some finches are susceptible to dehydration effects on plasma volume and the resultant increased haematocrit and haemoglobin readings. With warming climate trends, this may be an increasing symptom during summer.

Because of their potential to vary with physiological status, haematocrit and haemoglobin concentration can be effective markers for the study of health variation in birds. Clinically ill birds have below average haematocrit and haemoglobin levels, a symptom of anaemia and nutritional stress (Sturkie 1986;

Averbeck 1992). Haematocrit and haemoglobin levels have been linked to individual survival (Dubiec & Cichon 2001) and breeding success (Andersson &

Gustafsson 1995), and higher levels have commonly been correlated with superior 22 body condition, elevated testosterone, immunological health and lack of ectoparasites (Dawson & Bortolotti 1997a; Potti et al. 1999; O'Brien et al. 2001;

Sergent et al. 2004; Dudaniec et al. 2006; Cuervo et al. 2007). Sources of variation in bird fitness such as diet, food availability and body condition have also been commonly associated with differences in haematological state. Birds subjected to periods of fasting or poor diet often show declines in condition and lowered haematocrit and haemoglobin levels, and normally take a week or longer to recover levels once diet is improved (Piersma et al. 2000a; Hoi-Leitner et al.

2001; Sanchez-Guzman et al. 2004). Lack of nutrients, (notably sources of iron) can lead to inhibition of erythropoiesis (red blood cell production) and anaemia, which would cause an initial drop in haematocrit. The relatively long recovery period can be explained by the typical life span of bird erythrocytes (less than a month) and the time it takes birds to produce enough new cells to increase haematocrit (e.g. 13 days in fasted sandpipers) (Sturkie 1986; Piersma et al.

2000a). Foraging proficiency may also affect bird haematocrit; juvenile birds with higher foraging success had higher body condition and haematocrit levels

(Donnelly & Sullivan 1998). Studies of fat reserves in Blue Tits (Parus caeruleus) found that birds with lower reserves had lower haematocrit (Merila & Svensson

1997). However, other studies have not found correlations between haematocrit and nutritional reserves, despite clear differences in body condition among individuals and populations (Gavett & Wakeley 1986; Dawson & Bortolotti 1997a;

Ouillfeldt et al. 2004). Thus haematocrit alone is not always a reliable indicator of health in birds.

23 There is a common perception that haematocrit increases with increased exercise.

Strong fliers and those species more reliant on flight for survival had higher haematocrit levels than less active birds when natural haematocrit levels are compared between species of different flight capabilities (Carpenter 1975). In some situations, individuals that are experiencing higher energy demands due to higher levels of exercise display higher haematocrits. For example, migratory birds preparing for long-distance flights are known to increase their red blood cell count beyond the levels predicted for their size (Piersma & Everaarts 1996). Several studies have used the possible link between haematological indices and energy expenditure to explain the negative correlation between haematocrit levels and other condition indices. Breeding birds with high haematocrits were found to have other health measures indicating increased stress and lowered body and immune condition, including increased heterophil/lymphocyte ratios and lowered body condition (Horak et al. 1998a; Mazerolle & Hobson 2002; Bleeker et al. 2005).

Though haematocrit is generally elevated above normal levels in response to, or anticipation of, intense activity, research has found that individuals actually slightly decrease haematocrit during endurance flights and periods of increased exercise

(Soler et al. 1999; Jenni et al. 2006). Part of this response to increased exercise can be explained by the idea that larger birds need more oxygen for flight, and as birds lose mass during extended bouts of exercise, their total oxygen needs are reduced despite mass specific oxygen requirements remaining stable or possibly increasing (Landys-Ciannelli et al. 2002). This response was found to occur without any change in total body water. Captive birds subjected to slightly increased exercise and lower food availability also experienced lowered 24 haematocrit levels, suggesting predictability of food is a more important regulator of haematocrit in some studies, or in captive birds not likely to experience significant changes in energy demands (Birkhead et al. 1998). If the food shortage experienced by the captive birds resulted in poor nutrition, it is possible that RBC production was compromised, resulting in dying cells not being replaced, and an eventual lowering of haematocrit.

Varied relationships between blood parasite infections and haematocrit levels have been found in birds, with higher or lower haematocrit in response to parasite infection depending on the intensity of infection, time since infection and type of parasite (Fair et al. 2007). The response varies considerably as some bird species develop anaemia due to infection while others recover from initial infection quickly, and still others increase haematocrit, possibly a result of erythropoeisis to replace infected erythrocytes (Dawson & Bortolotti 1997a; Booth & Elliott 2002). Thus we cannot generalize about the interaction between haematocrit and blood parasites without more research on the physiological responses of birds to blood parasites.

Bird haematocrit and/or haemoglobin levels vary seasonally: lower levels are typically found late in the breeding season or during moult; higher values occur in the winter, early breeding season and during migration when energy requirements may be higher (Morton 1994; Box et al. 2002; Gayathri & Hegde 2006; Owen &

Moore 2006). Part of this variation is explained by the normally dramatic, but usually temporary, drop in haematocrit during egg-laying in female birds (Williams et al. 2004; Gayathri & Hegde 2006). In this case, it is thought that blood plasma 25 volume is increased to allow for higher concentrations of yolk precursors and that higher levels of estrogens associated with breeding may inhibit erythropoiesis. In contrast, androgens tend to stimulate erythropoiesis and males with higher testosterone levels during the breeding season typically have higher haematocrit values (Sturkie 1986). Because passerine chicks have low RBC levels at hatching, nestling and fledgling birds may have low haematocrits compared to adults until they increase erthyropoiesis to adult levels (Fair et al. 2007). Thus, factors such as foraging efficiency, season, sex, age and breeding stage are important sources of variation to consider if haematocrit is to be used as a measure of condition.

Haematological status has been successfully used in combination with other condition measures to document bird health differences in contrasting environments. Despite their preference for deciduous breeding habitat, Great tits nesting in coniferous forest were found to have better condition and higher haematocrits than those nesting in preferred habitats (Kilgas et al. 2006a). Ots et. al. (1998) found lower haematocrit levels for birds breeding in urban versus rural habitats, while other studies documented no differences in birds breeding in native versus altered habitats (Gavett & Wakeley 1986; Owen et al. 2005). Male

Ovenbirds (Seiurus aurocapillus) nesting in continuous forest habitat known to support better breeding success than fragmented forest had higher mass and haematocrit and normal immune system activity, while males in fragmented forests were lighter, had lower haematocrits, and may have had depressed immune responses (Mazerolle & Hobson 2002). The haematological differences found between bird populations in many of these studies support the use of 26 haematological indices to aid in the determination of bird health responses to habitat quality. When combined with multiple health indices, haematological condition is relatively simple to measure, is a robust measure of relative health among individuals and populations, and can be used as an indicator of bird breeding potential and survival.

Factors affecting stress in bird populations

External stimuli perceived as a threat to survival and endogenous metabolic stimuli indicating an energy deficit can induce stress responses in birds through activation of the hypothalamo-pituitary adrenal (HPA) axis. This response is triggered by the release of corticotrophin-releasing factor (CRF) from the hypothalamus, which in turn causes the pituitary gland to release adrenocorticotropic hormone (ACTH).

ACTH is then transported to the adrenal glands, which quickly (within 5 minutes) release glucocorticoids, including the main avian stress hormone, corticosterone

(CORT). The glucocorticoids initiate a suite of physiological and behavioural effects that aim to reduce the state of stress, and begin self-suppression of the stress response through negative feedback to the hypothalamus and anterior pituitary (Harvey et al. 1984; Sturkie 1986). This feedback to the hypothalamus is important in limiting the duration and intensity of the stress response. At the cellular level, the stress response involves the up-regulation of heat shock proteins

(HSPs), which provide protection against protein degradation or improper protein folding due to stressors (Lindquist & Parsell 1993).

27 Acute stress causes a rapid increase of CORT that may remain elevated for long periods or may return to baseline levels rapidly, depending on the duration, source and intensity of the stressor. The rapid release of CORT in response to acute stressors (stimuli of minutes to hours) results in the cessation of feeding, mobilization of energy reserves and ‘flight or fight’ behaviour to aid in survival

(Wingfield et al. 1983; Wingfield & Silverin 1986). Individuals experiencing chronic stress (days to weeks) and more prolonged and constant release of CORT, may exhaust or suppress the stress response, over time resulting in lower baseline and stress induced levels of CORT (Rich & Romero 2005). As glucocorticoids activate gluconeogenesis, increased heart rate and muscle activity, prolonged stress can lead to hyperglycaemia, exhaustion of energy reserves, immunosuppression, suppression of reproduction and decreased territorial behaviour (Gray et al. 1990;

Sapolsky et al. 2000; Ilmonen et al. 2003). Because the long-term effects of elevated CORT can be disadvantageous to breeding, general health and survival, suppression of elevated stress hormone levels can benefit individuals by allowing a return to normal feeding, moult or breeding status and behaviour (Romero &

Wikelski 2002; Cyr & Romero 2006; Cyr et al. 2007).

The level of circulating CORT is influenced by the binding capacity and abundance of the specific plasma carrier protein, corticosteroid binding globulin (CBG), in the plasma (Fleshner et al. 1995; Breuner & Orchinik 2002a). Glucocorticoids bind reversibly to CBGs resulting in a homeostatic and dynamic plasma pool of free and

CBG-bound CORT. By binding to and reducing tissue availability of CORT, CBG

28 levels are thought to mitigate the behavioural effects of CORT (Breuner & Orchinik

2002a). This largely accepted ‘free hormone hypothesis’ suggests that it is the unbound CORT in plasma that is the biologically active component of the stress response. However, there is some evidence that CBGs may serve as a mechanism to transport bound CORT to certain tissues or a means of interacting with signalling pathways in cell membranes (Malisch & Breuner 2010). Thus the relative importance of free and CBG bound CORT in stress physiology is still a matter of contention. Many avian stress studies now measure CORT as well as

CBGs in order to calculate and analyse levels of free CORT, however all researchers still favour the inclusion of both CBG affinity and total CORT in the results of all studies (Romero 2004; Malisch & Breuner 2010).

New methods of measuring CBG levels in small plasma samples of individual birds

(Breuner & Orchinik 2002a) now allow the use of this measure to quantify the impact of stressors. A recent review indicates that CORT and CBGs are known to vary in basal concentrations, after acute stress, or in response to chronic stressors

(Malisch & Breuner 2010). Corticosterone binding globulins decrease during an acute stress response in some species (within 1 - 24 hours), allowing elevated free

CORT levels over extended periods (Fleshner et al. 1995; Breuner et al. 2006).

This prolonged and higher stress response may be an important adaptation during certain life stages or in different environments. More study is needed to fully understand the dynamics of CBGs and CORT; thus, it is crucial that publications include both measures of CORT and CBGs (Romero 2004; Malisch & Breuner

29 2010). This will allow the importance of different patterns of CORT and CBG levels to be re-examined and clarified as more knowledge about their biological significance is gained.

The function of the HPA axis response is age-dependant. Nestlings do not respond to the stress of capture by humans (handling stress) for the first few days after hatching due to the suppression of free CORT by higher CBG levels and develop stress responses similar to adults in stages (Sims & Holberton 2000;

Wada et al. 2006a). Once young birds have developed the ability to produce a stress response, they can still differ from their parents and other adults in their reaction to stressors. For example, nestling Magellanic penguins (Spheniscus magellanicus) show elevated stress responses in colonies exposed to tourism, while their parents did not (Walker et al. 2005). Nestling Wilson's storm-petrels

(Oceanites oceanicus) only display a stress response when in poor condition, while adults always elevate CORT in response to being handled (Ouillfeldt et al. 2004).

Gender (male) or higher dominance position is sometimes correlated with lower

CORT levels and higher breeding success or survival in species with social structures that affect habitat choice and feeding efficiency (Alexander & Irvine

1998; Jennings et al. 2000; Rubenstein 2007). Females and subordinate American redstarts (Setophaga ruticilla) were found to have higher stress at over-wintering sites, and possibly lower survival (Marra & Holberton 1998). Male redstarts measured in the above study and Mountain chickadees (Poecile gambeli) measured by Pravosudov and others benefit from their dominant feeding status 30 with lower stress and higher body condition (Pravosudov et al. 2001).

Testosterone has been implicated in this pattern, as it shares high affinity to CBGs with CORT (Deviche et al. 2001). Testosterone release and aggressive behaviour in Song sparrows (Melopisa melodia) and tree lizards (Urosaurus ornatus) are suppressed by chronically elevated CORT, supporting the view that subordinate individuals and females typically have higher stress levels in general (Wingfield &

Silverin 1986; Moore et al. 1991; Knapp & Moore 1997). Some male animals may lose feedback control of the stress response when testosterone levels are extremely elevated, possibly because testosterone can displace CORT on the

CBG molecule causing continuously high CORT levels (McDonald et al. 1986;

Astheimer et al. 1994; Klukowski et al. 1997). This response can help explain highly elevated CORT levels found in breeding males but not necessarily non- breeding males (Kern et al. 2005; Bonier et al. 2007). Birds with experimentally enhanced testosterone levels developed higher baseline and stress induced CORT and sometimes higher CBG levels, which were then related to increased aggressive behaviour characteristic of dominant breeding or foraging birds

(Klukowski et al. 1997; Creel 2001; Zysling et al. 2006). Though subordinate and female animals may suffer more acute stress from feeding inefficiency and harassment by more aggressive individuals, dominant male birds are more likely to suffer from increased stress levels due to their higher testosterone levels. The advantage to maintaining dominance should be greater when these individuals are able to quickly return to a lower stress state after aggressive encounters and if this aggressive behaviour provides them with a higher feeding efficiency.

31 Stress hormones have an important influence on metabolism and behaviour. Birds experiencing high stress due to cognitive stressors initially display symptoms of increased energy use and modify their behaviour to maximize survival (fight or flight), often at the expense of other activities such as breeding and moult (Siegel

1980). High levels of CORT are known to cause breeding birds to abandon breeding behaviour and territories during extreme weather and after receiving artificial implants of CORT (Wingfield et al. 1983; Silverin 1986; Wingfield & Silverin

1986; Astheimer et al. 1995). Birds experiencing higher stress levels in response to acute stress often have lowered breeding success (Love et al. 2005; Bonier et al. 2007). There is some evidence suggesting successful breeders are simply better at suppressing stress responses (Wingfield et al. 1995). A competing theory suggests that down-regulation of CORT is a sign of chronic stress and harms reproductive potential. In accordance with this theory, birds that were experimentally subjected to chronic stress were more likely to have decreased stress hormone levels and lowered breeding success or survival (Cyr et al. 2007).

Both scenarios are possible and we can differentiate between the two when a combination of stress and condition indices is used to determine overall bird health.

The patterns of stress in birds can vary depending on more general factors such as season and geography. In most birds, stress levels have been found to vary seasonally, with higher CORT levels found during the onset of breeding seasons and lower levels during post-breeding and moult (Dawson & Howe 1983;

Astheimer et al. 1994; Romero & Wingfield 1998, 1999; Deviche et al. 2000). This seasonal pattern differs between species depending on latitude and seasonality of 32 climate. The increased ability of high latitude birds to suppress free CORT levels compared to temperate birds suggests that the amount of stress that affects breeding success varies geographically (Wingfield et al. 1994; Breuner et al. 2002).

Snow buntings (Plectrophenax nivalis) breeding in the Arctic do not have elevated free CORT levels during breeding as in most temperate species, due to the simultaneous increase in both total CORT and CBG binding capacity during this season (Romero et al. 1998). Birds at very high latitudes are often restricted to one brood or breeding attempt per year due to the longer and colder winters, and any stress induced interference with breeding behaviour would have more detrimental effects on reproductive success. Thus, suppression of the stress response could be a great advantage to birds breeding in northern latitudes. The theory of the increased importance of stress hormones with latitude is supported by the finding that tropical birds appear to show low seasonal variation in free CORT levels and tend to express lower baseline and acute stress induced CORT levels than related temperate species (Martin et al. 2005; Wada et al. 2006b).

Physiological ecologists have found that differences between basal and stress- induced levels of CORT can be reliable predictors of survival for individuals and populations (Romero & Wikelski 2001). As expected, changes in stress levels or response often track changes in body condition (Kitaysky et al. 1999; Ouillfeldt et al. 2004; Kern et al. 2005). Higher free CORT levels are associated with fasting or poor foraging success, which in turn result in declines in body condition

(Pravosudov et al. 2001; Lynn et al. 2003a). Poor nutrition and declining energy stores can trigger a physiological stress response that will promote changes that 33 will help birds regain body condition. There is a significant correlation between increases in CORT and an increase in catabolism of fats (lipolysis) and eventually proteins (proteolysis) in birds subject to food supply limitations or during migration

(Cherel et al. 1988b; Sapolsky et al. 2000). In these cases, the increased CORT levels also cause changes in behaviour, such as increased foraging rates and local migration, which will assist birds in gaining condition and alleviating stress

(Schwabl et al. 1991; Astheimer et al. 1992; Marra & Holberton 1998; Holberton

1999). However, if behavioural adjustments are not effective in restoring condition due to added stressors such as predation pressure and severe food limitations, birds will continue to elicit a stress response that can lead to the down-regulation of

CORT levels and suppression of stress responses (Buchanan 2000; Jenni et al.

2000; Clinchy et al. 2004; Rich & Romero 2005). Thus symptoms of chronic stress or stress response suppression can indicate an individual or population that is not living under optimal conditions for breeding and survival.

Differences in the basal and stress-induced CORT and CBG levels in bird populations have identified the disadvantages of breeding in poor habitat or areas of increased disturbance. Individuals overwintering in poor habitat showed symptoms of declining fitness with lowered body condition, higher baseline CORT levels and lowered response to acute stress as the season progressed (Marra &

Holberton 1998; Kitaysky et al. 1999). European blackbirds (Turdus merula) living in urban areas had lowered response to acute stress compared to birds living in natural habitats, suggesting that urban birds may be habituated to urban stressors, and could be suppressing the stress response to handling by humans (Partecke et 34 al. 2006). Magellanic penguins exposed to higher levels of human disturbance produce young that have elevated responses to acute stress compared to those breeding in undisturbed areas (Walker et al. 2005). Eurasian treecreepers (Certhia familiaris) breeding in poor habitats were found to produce young with poor body condition and higher CORT levels (Suorsa et al. 2003). These effects can be detrimental to breeding populations of birds, as young animals exposed to elevated stress early in life may experience lowered breeding success or survival later in life

(Blount et al. 2003). Therefore many studies have found that birds living in poor quality habitat show symptoms of chronic stress.

Measures of stress are useful indicators of bird health, despite the tendency of patterns of stress levels and responses to change depending on a large number of variables (Romero 2004). Stress hormone and CBG levels are more easily interpreted once background information is collected about the seasonal, sex, age and regional patterns of CORT and CBG secretions for a species.

1.6 Research aims and hypotheses

In summary, it is still unknown what processes are causing declines in finch occurrence in Australian savanna habitats. The historic and geographic pattern of finch range contractions and the lack of other changes in savanna ecology suggest pastoral practices and changes in fire frequencies may be the root cause. We know the basic effects of grazing and fire on the savanna grass layer and assume this affects finches, which rely on the grass layer for food and shelter. However, 35 there is little information on how these changes in the grass layer affect finch productivity and survival. The literature reviewed in the first part of this chapter shows that measures of bird health changes over seasons and in different environments may be useful tools for determining when and where finches are experiencing suboptimal conditions for breeding and survival. A number of physiological indices can be used in concert to identify health differences among populations living in different habitats and during different seasons. In order to conserve granivorous birds in Australia’s tropical savannas, we need a better understanding of the processes contributing to granivore population declines to inform the management of these areas. More generally, this research will add to the knowledge of how physiological measures of bird health co-vary and change in relation to environmental conditions. In this thesis there are two main hypotheses:

1. High levels of grazing and burning are detrimental to the health of finches. It is

predicted that finches living in areas with high grazing pressure or frequent fire

will show patterns of health that suggest they are in poorer condition compared

to those living in areas with lower fire and grazing pressure.

2. Declining finches are more sensitive to seasonal changes in habitat quality and

the environmental effects of grazing and fire. It is predicted that declining

species will show more evidence of poorer health during stressful seasons and

on grazing or fire prone land compared to non-declining species.

36 1.7 Experimental design and thesis structure

To test these hypotheses, the habitat and health of multiple finch species (declining and non-declining) were measured during breeding and non-breeding seasons at sites with differing levels of cattle grazing and fire frequency. Finch health and habitat quality were monitored at each of the six study sites during the non- breeding and breeding season for two years. Data chapters two through five are written and presented in the format of multiple-authored papers for submission to academic journals, thus there will be some repetition of material between chapters.

Chapter 2

This chapter aims to test whether declining species show more sensitivity to seasonal environmental changes than non-declining finches (the first part of hypothesis 2) irrespective of the possible effects of differing land management.

Here I describe significant differences between the seasonal patterns found in the stress response, body condition, and haematological measures of the declining

Gouldian finch and non-declining Long-tailed and Masked finches using mean health measures which are pooled across three properties.

37 Chapter 3

Chapter Three investigates land tenure differences in grass layer components thought to be important to finch diets. Sites were selected in the Northern Territory that were (i) under high grazing pressure (pastoral), (ii) had a high frequency of fire across a large area (aboriginal) or (iii) under conservation-oriented management with low grazing pressure and small infrequent fires (Defence; Figure 1.1a). The role of this chapter is to provide evidence of differences in grass seed resources between sites with differing land management (grazing and fire patterns). This evidence is used in the conclusion (chapter 6) to provide possible explanations for differences found between the health measures of finch populations monitored on different land tenures.

38 a) Northern Territory Study Sites b) Queensland Study Sites

Figure 1.1: Map of Australia outlining the location of regions of the Northern

Territory and Queensland where this study was conducted. Inserts refer to: a) Northern Territory: 1. Defence managed: Bradshaw Field Training Area

2. Pastoral managed: Delamere Station

3. Aboriginal managed: Yinberri Hills b) Queensland: 1. Conservation managed: Lakefield National Park

2. Aboriginal managed: Pormpuraaw

3. Pastoral managed: Laudham Park Station

39 Chapter 4

In Chapter Four I present evidence of the threat of grazing and frequent intense fire to Australian tropical finch populations at sites in the Northern Territory (Figure

1.1a). The data for this chapter is similar to that used in Chapter Two, but in this case I examined mean health measures between three sites to assess the effect of different land management regimes on finch health (hypothesis 1). Here I also aim to examine whether a declining finch species, the Gouldian finch, shows more evidence of poor health on non-conservation managed land compared to non- declining species such as the Long-tailed finch (the second part of hypothesis 2).

Chapter 5

In Chapter Five, I describe similarities between Gouldian finch seasonal and site related health index variation and the variation found in Star finch and Black- throated finch health indices. The last two species are both declining in

Queensland, thus populations of these finches were monitored at two sites for each species: one site managed for conservation and one site managed for either pastoral or aboriginal interests (Figure 1.1b). Data on health measures of Star and

Black-throated finches is examined seasonally (pooled across two sites for each species) as well as between land tenures (not-pooled). This chapter aims to address both hypotheses: 1) whether Star and Black-throated finches show evidence of poorer health on non-conservation managed land and during stressful seasons, and 2) whether the variation in the health indices of these species are 40 more similar to the declining Gouldian finch, or the non-declining Long-tailed or

Masked finch.

Chapter 6

Chapter Six has two sections. First I propose links between finch health and grass layer changes at Northern Territory sites with differing land management. The implications of the research to the conservation of all five finch species (in the

Northern Territory and Queensland) are discussed. The second section of this chapter discusses possible improvements to the use of physiological measures for gauging the health status of free-living birds based on my review of the literature and my own research.

41 2.0 Chapter Two: Seasonal patterns of health measures in three Australian grass-finches

2.1 Introduction

Studies of seasonal variation in avian stress responses and body condition predominantly describe patterns commonly observed in northern hemisphere north-temperate and arctic birds. Because these birds live in seasonally resource rich environments, they tend to have highly seasonal lifestyles characterized by short and intense breeding seasons and long non-breeding and migratory periods.

The annual cycle of northern hemisphere birds shows variation in their health profile including increased stress levels during the peak of the breeding season; this is sometimes accompanied by lowered body condition measures (mass, muscle or fat) compared to non-breeding seasons (Wingfield et al. 1994; Kitaysky et al. 1999; Breuner & Orchinik 2002b).

Basal and stress-induced plasma levels of the key avian stress hormone corticosterone (CORT) tend to be highest in breeding birds and lower in non- breeding and moulting birds (Astheimer et al. 1994; Romero & Wingfield 1998;

Romero & Remage-Healey 2000; Breuner & Orchinik 2001). Circulating corticosterone binding proteins (CBGs), which are thought to limit the availability of plasma stress hormone to tissues by sequestering CORT in plasma, are also found in higher levels in breeding birds (Breuner & Orchinik 2002a). This results in

42 similarly low free or unbound CORT levels year round. Because only free CORT that is unbound to CBGs is thought to have biological activity, avoidance of high levels of active CORT can be achieved by either increasing CBG capacity and/or by releasing less CORT in response to a stressor. Such containment may be advantageous to facilitate successful breeding, as continually elevated CORT levels have been shown to coincide with interruptions to breeding behaviour and lowered body condition (Wingfield et al. 1983; Silverin 1986; Wingfield & Silverin

1986; Gray et al. 1990; Cyr & Romero 2006; Cyr et al. 2007). Corticosterone binding globulins may also be important in helping control CORT levels within tissues or movement across cell membranes (Malisch & Breuner 2010). Because levels of CORT bound to CBGs are higher during breeding, this type of action may be more useful during this season; however, why this occurs is still a matter of conjecture. Because of the lack of consensus on the importance of the dynamics between these two components of the stress response, studies examining stress in animals typically report on the patterns seen in both measures of CORT and CBGs

(Romero 2004; Malisch & Breuner 2010).

The typical avian response to a single acute stressor regardless of season is characterized by a rapid rise in total CORT levels (within 5 minutes), which overwhelm CBG capacity and result in increased circulating free CORT (Dawson &

Howe 1983; Fleshner et al. 1995; Lynn et al. 2003a; Breuner et al. 2006). These and other authors have proposed that acute stress responses also promote self- preserving behaviours, including foraging, following exposure to climatic or nutritional stressors that jeopardize an individual’s survival. Individuals with higher 43 baseline or stress-induced CORT levels often have poor body condition compared to birds with lower CORT levels; this is particularly apparent among birds exposed to high energetic demands, inclement weather or food limitations (Rees et al. 1985;

Astheimer et al. 1995; Marra & Holberton 1998; Kitaysky et al. 1999; Jenni et al.

2000). The few studies that address how exposure to stress affects haematological measures show varied results. For example, some arctic passerines show a correlation between increased CORT and lowered haematocrit

(percent of red blood cells to plasma; (Lynn et al. 2003b) while fasted penguins had increased haematocrits and high CORT levels (Vleck et al. 2000; Lynn et al.

2003b). Thus, it is unclear whether haematological measures vary independently of CORT levels, in relation to stress in a species specific manner or in relation to environmental conditions.

In general, many bird species modulate their response to stress by regulating the amount of circulating free CORT, but this modulation is strongly affected by overall body condition at the time of stress exposure. The pattern of CORT release immediately following stress exposure has been examined in more than 64 species over various stages of their annual cycle, although examination of seasonal variation in CORT profiles of native tropical southern hemisphere birds has received little attention (Astheimer & Buttemer 2002). Studies contrasting the stress responses of arctic versus temperate breeding bird populations found that species with shorter breeding seasons were more likely to suppress the stress response by reducing CORT secretion (Wingfield et al. 1994; Breuner et al. 2003).

Similar to temperate birds, the few tropical birds studied had similar increases in 44 CORT and CBG capacity during breeding but sometimes had lower baseline and stress induced CORT than temperate conspecifics (Martin et al. 2005; Wada et al.

2006b). The species studied in these comparisons, the Rufus-collared sparrow

(Zonotrichia capensis) and the House sparrow (Passer domesticus) are common birds, with relatively stable or increasing population sizes. It is possible that tropical birds, especially species declining in abundance due to environmental stressors, may modulate stress and condition differently from these common exemplars.

Seasonality in body condition and haematological state may also differ between birds found at different latitudes. Some tropical birds show less annual variation in body fat levels or even opposite trends in mass variation compared with temperate zone conspecifics (King & Farner 1966; Blem 1976; Rogers et al. 1993; Breuer et al. 1995). Bird haematocrit levels vary seasonally in temperate species: lower levels are typically found late in the breeding season or during moult; higher values occur in the winter, early breeding season and during migration when energy requirements may be higher (Morton 1994; Box et al. 2002; Gayathri & Hegde

2006; Owen & Moore 2006). More research is needed to fully describe this variation, as few studies have looked at possible seasonal differences in haematocrit in tropical species (Fair et al. 2007).

This study characterized physiological traits that reflect overall health status in common and rare tropical grass-finches. This was accomplished by evaluating the stress response, body condition and haematological status of three closely related 45 tropical species during breeding and non-breeding seasons. This study focuses on

Australian endemic Estrildid finches; all are small (less than 16g) passerines, which mainly feed on grass seeds.

Of the 15 species of grass-finch inhabiting northern Australia, 4 species have suffered large range contractions and decreased abundance in savanna within the past 50 years (Franklin 1999). Increased grazing by cattle and more frequent, extensive and intense fires are likely to have decreased the suitability of savanna habitat to granivores by reducing grass layer productivity and therefore have been proposed as the main cause of finch declines (Franklin et al. 2005). Despite grass- finch species often occupying the same habitat, the basic life history and conservation status of these finches differ, which, in turn, may reflect underlying differences in the physiological and morphological traits we examined.

Gouldian finch (Erythrura gouldiae) life history is characterized by a short life span, a long uninterrupted breeding season, a short moult period and a specialized diet, including a small number of preferred species of grass seeds. Based on banding and recapture studies, typical free-living Gouldian finches live less than two years

(Woinarski & Tidemann 1992; Legge 2009). Gouldian finches breed January to

September, with some pairs raising up to 3 broods in a season (Tidemann &

Woinarski 1994). Gouldian finches are strict cavity nesters and have an exclusive preference for branch hollows of Eucalyptus brevifolia or E. tintinnans. All

Gouldian finches undergo a complete flight-feather and body moult during one month of the late dry season, between October and late December (Tidemann & 46 Woinarski 1994; Franklin et al. 1998). Gouldian finches are grass seed specialists, rarely supplementing their diet with other food resources (Dostine & Franklin 2002).

Possibly due to their diet restrictions, this species is highly nomadic. Individuals are often recorded flying over 17 km a day to visit selected food, water and breeding sites (Dostine et al. 2001; Legge 2009). While this nomadic lifestyle may help the Gouldian finch access naturally dispersed resources, the highly energetic lifestyle and diet specialization of the Gouldian finch could limit this species ability to adapt to environmental changes, especially if resources become less available or more widely dispersed. Thus, these traits may explain the current endangered status and recent dramatic declines of the Gouldian finch in response to increased grazing, fire and resulting grass layer changes in Australia’s savannas.

In contrast, both Long-tailed (Poephila acuticauda) and Masked finches (P. personata) appear to have a longer lifespan, more opportunistic moult schedules, reduced annual fecundity, and more diversified diets. Long-tailed and Masked finches are commonly recaptured in banding studies, with 2 year old birds regularly resighted or recaptured near their original capture site (Maute, personal observation). This anecdotal evidence suggests that these species are not only

50% longer-lived than the Gouldian finch, but less nomadic as well. Long-tailed and Masked finches appear to be more opportunistic breeders, with a smaller proportion of individuals breeding during the late wet and dry season, compared to

Gouldian finches (Woinarski & Tidemann 1992; Tidemann & Woinarski 1994).

Captive breeding of Long-tailed and Masked finches requires protein supplementation (i.e. insect food) to encourage pairs to breed, with birds 47 interrupting breeding attempts in the absence of dietary enhancement (Evans &

Fidler 1986). This suggests a flexible approach to breeding, with an ability to delay or interrupt a breeding attempt in response to environmental conditions. In contrast, captive Gouldian finches tend to breed regardless of diet, as long as nesting sites and substrates are provided. Long-tailed and Masked finches are also less specialized in their choice of nesting substrates, with Long-tailed finches recorded building nests in tree canopies, hollows and human built structures

(Immelman 1982; Higgins et al. 2006). Masked finches build nests near or on the ground in grass tussocks (Higgins et al. 2006). Both species also have a relatively varied diet focused on a wider variety of grass seed species and regularly supplemented with other types of plant seeds, insects, lerps and nectar (Garnett &

Crowley 1994; Dostine & Franklin 2002). Moult timing in Long-tailed and Masked finches is irregular and appears to be extended, with individuals recorded moulting body and flight feathers during all months of the year (Tidemann & Woinarski 1994;

Franklin et al. 1998). The opportunistic feeding, breeding and moult strategies of these two species suggest that they are well adapted to dealing with recent changes in environmental conditions in northern Australia, while Gouldian finches may be disadvantaged by their more strictly timed and specialized lifestyle.

This study aims to characterize seasonal differences in the stress response, body condition and haematological status of common and rare tropical grass-finches.

The different ecology of Gouldian, Long-tailed and Masked finches leads to the prediction that their health profiles will also differ in response to annual and seasonal environmental variation. Due to their specialized diet and strictly timed 48 breeding and moulting schedules, Gouldian finches are more likely to show higher levels of stress and poor condition than the less specialized and more opportunistic

Long-tailed and Masked finches.

2.2 Methods

Study site characteristics and survey seasons

Study sites were selected where all three study species occurred in the vicinity.

Three properties in the northern Territory Australia were sampled twice a year:

Bradshaw Field Training Area (latitude -15.5S, longitude 130.4E), a Department of

Defence property, Delamere Station (latitude -15.7S, longitude 131.6E), a pastoral property, and Yinberri Hills (latitude -14.1S, longitude 132.0E), aboriginal land

(Figure 1.1a). The habitat on all properties was characterized by grassy understorey of perennial grasses such as Triodia spp., Heteropogon contortus, H. triticeus, Themeda triandra, Sehima nervosum, Chrysopogon fallax, and annual grasses such as Sarga spp. and Schizachyrium spp. At Bradshaw FTA and

Delamere Station the sparse tree layer was dominated by smooth barked gums

(Eucalyptus brevifolia) and box (E. tectifica), while the Yinberri Hills was dominated by salmon gum (E. tintinnans) and Corymbia conifertiflora. Shrubs and small trees make up a very small proportion of the vegetation, except in drainage areas and in patchy Calytrix spp. thickets. Soils in high rocky country tended to be well-drained sandy soils, while floodplains were dominated by red clay soils. Cracking clay or

49 black soil plains were present in most areas; however, finches were not found to use these areas during the monitoring sessions.

Populations of all three finch species were located within one area of 5 - 10 km2 on each property each season. The search areas were relatively large due to the semi-nomadic behaviour of these species. Individuals from each population were captured, measured and released during both the late dry/early wet season

(September – early December) when birds are not breeding but most are moulting and during the early dry season (May-June) which is the peak of the finch breeding season. All properties were visited in the non-breeding (late dry/early wet) season during 2007 and 2008 (first year of the study) and in the breeding (early dry) season in 2008 and 2009 (second year). During each visit, the goal was to capture, band and measure 30 - 70 individuals of each species, thereby sampling a range of ages and sexes. In total, 448 Gouldian finches, 439 Long-tailed finches and 360 Masked finches were included in the results of this study. No recaptured individuals from this or other studies were included in analysis and therefore all birds included in this study were naïve to the capture process.

Finch capture and handling

Birds were captured during fine weather and within 4 hours after sunrise using mist nets at sites used by finches for drinking or foraging. Sites were between 5 to 20 km from any sealed road or human settlement, thus the possibility of human caused disturbance (other than our trapping effort) was minimal. The length of 50 time between when each bird was captured (when the bird hit the mist net) and held before blood sampling was measured to the nearest minute to account for the effect of capture stress on CORT. All birds were removed from nets within 10 minutes of capture, and held in clean cloth bags until they could be measured and blood samples taken. Blood samples were taken by venipuncture of the left brachial vein with a 26 - gauge needle and the time of blood sampling was recorded. A blood volume of 50 - 140 μL was collected into heparinised haematocrit tubes and 5 μL placed in a HemoCue® microcuvette for immediate determination of blood haemoglobin levels. The haematocrit blood samples were kept on ice until they could be spun down in a centrifuge, measured, and have the plasma removed and frozen at -20C (within 6 hours of sampling). Each bird was banded with a uniquely numbered aluminium band and released at the capture site within 90 minutes of capture. Birds were measured for head to bill and tarsus lengths using vernier callipers and the flattened wing cord of the right wing was measured with a wing ruler. Each bird was assessed for the level of body and flight feather moult, and the presence of any external reproductive characters (e.g. brood patch, egg in oviduct). Adults of all species were often found to show evidence of reproductive status (body characters and nesting behaviour) during breeding season sampling, confirming that finches were actually breeding during this time of year. All finches were aged by either presence of juvenile plumage, bill gape or colour. Many finches retained some juvenile plumage until the late dry or non-breeding season, while very few were still young enough to show bill gape or dark bills later in the year. Thus it was assumed that most juveniles were older during the non-breeding season compared to the breeding season. Adult Gouldian 51 finches were sexed by plumage, while other species are not reliably sexed by physical characteristics. Following the methods described by Griffiths et. al.

(1998), a proportion of adult Long-tailed and Masked finches were sexed by

Australian Wildlife Conservancy staff using genetic material in blood samples.

Measurement of finch body and haematological condition in the field

Finch body condition was measured using three indices. (i) Fat score was determined using an estimation of the volume of fat stored in each bird’s furculum using a scale of 0 - 5 (Helms & Drury 1960). A fat score of 0 indicated no fat was seen in the furculum, 1 indicated 1 – 10% of the furculum full of fat, 2 indicated 10

– 60% full, 3 indicated 60 – 90% full (concave surface of fat in furculum), 4 indicated greater than 90% full (fat bulging over rim of furculum), and 5 indicated fat overflowing from furculum and connecting with other body fat stores. (ii) Muscle score was estimated based on the shape and relative volume of the pectoral muscles using a scale of 0 - 3 (Brown 1996; Lindstrom et al. 2000) (iii) Mass was measured in grams (± 0.5 g) using a Pesola scale, which was calibrated using a known weight once a week. Two haematological indices were measured in the field: (i) The blood concentration of haemoglobin (g/mL) using a HemoCue® 201+ spectrophotometric reader and (ii) haematocrit measured by spinning haematocrit tubes for 7 minutes at a force of 16 060 G in a Hettich Haematokrit 210 centrifuge.

The tube was measured for percent volume of red-blood cells using a ruler or haematocrit reader. Plasma was then drawn out of the tube with a fixed-tip syringe

(Houston et al. 1995) and transferred to a plastic cryovial and frozen at -20C in an 52 portable Engle freezer until samples could be transported and stored at -80C prior to measurements for plasma constituents (see below).

Measurement of health indices from plasma

Total CORT and CBG capacity were determined from field-collected plasma samples. Due to the difficulty of safely bleeding small finches, not all birds had blood samples with a sufficient volume of plasma to test for CORT and CBG. Thus only 287 Gouldian, 275 Long-tailed and 224 Masked finches were tested for

CORT, and 212 Gouldian and 136 Long-tailed finches were tested for CBG capacity. 2 Plasma CORT was measured for 5 μL aliquots of whole plasma from each bird in an EIA Kit ACETM Competitive Enzyme Immunoassay for corticosterone (Cayman Chemical Co. Ann Arbor, Michigan, USA). This assay uses competitive binding between CORT contained in the plasma aliquot and a fixed amount of a corticosterone-acetylcholinesterase (AChE) conjugate for a limited number of antiserum binding sites in the well of the assay plate to determine plasma CORT levels. The antiserum-CORT complex is bound to an immunoglobulin and combined with a reagent that contains the substrate AChE to produce an enzymatic reaction with a yellow colour that can be read by a spectrometer to estimate CORT levels (Pradelles et al. 1985; Maclouf et al. 1987).

Two standard CORT solutions (one high concentration and one low) were tested

2 Masked finch CBG capacity was not measured due to the time and budget constraints of this preliminary study. 53 along with finch plasma on all plates, which had an intra-assay variability of < 5% and an inter-assay variability of < 5%.

The capture and bleeding protocol used in other studies measuring CORT typically involves the collection of serial blood samples at 0, 15, and 45 minutes after capture from each individual bird (Romero 2004). These serial samples are then used to plot the release of CORT over time since capture, and therefore characterize the stress response to capture of each individual bird for use in comparisons to other birds. This protocol was not feasible for use when capturing large numbers of endangered birds with limited staff. Because finches are very social birds, we were unable to capture birds singly and due to the low number of bird handlers available in the field (one to two maximum), birds were typically captured and processed in groups of ten to thirty as soon as possible after capture.

It was also not feasible to bleed small endangered Gouldian finches more than once during capture, thus all finches were only bled once. Because each finch in this study was only bled once over a range of 5 – 90 minutes, we created a relative measure of CORT response to handling stress over time by regressing all CORT results from all seasons against handling time (the time measured between capture and blood sampling) for each species. Regressions explained significant amounts

2 of variation for each species (Gouldian r = 0.20 F1, 277 = 15.78, P < 0.0001; Long-

2 2 tailed r = 0.48 F1, 275 = 80.34, P < 0.0001; Masked r = 0.38 F1, 213 = 36.35, P <

0.0001), therefore we used the residual values from these regressions as a measure of stress response (relative CORT release in response to time since capture) in order to control for differences in handling times between samples 54 (regression plots and residuals shown in Figure 2.1). Residual values for each species (residual CORT) represent the release of CORT irrespective of handling time and may be either positive or negative. Residual CORT levels were then used in further analyses to compare ages, sexes, seasons and years.

55 a) Gouldian finch

80 60 70 50 40 60 30 50 20 40 10

CORT(ng/ 30 Residual ml) CORT(ng/mL) 0 20 -10 10 -20

10 20 30 40 50 60 70 80 90 18 20 22 24 26 28 30 32 34 handle time (min) CORT (ng/ml) Predicted b) Long-tailed finch

80 50 70 40 60 30 50 20 40 30 10

CORT(ng/

ml) Residual ml)

CORT(ng/mL) 20 0 10 -10 0 -20 10 20 30 40 50 60 70 80 90 12 16 20 24 28 32 36 40 44 handle time (min) CORT (ng/ml) Predicted c) Masked finch 80 70 50 60 40 50 30 40 20 30 10

CORT(ng/

ml) Residual ml) CORT(ng/mL) 20 0 10 -10 0 -20 10 20 30 40 50 60 70 80 90 12 16 20 24 28 32 36 40 44 handle time (minutes) CORT (ng/ml) Predicted

Figure 2.1: Graphs on the left are regression plots of a) Gouldian finch, b) Long- tailed finch, and c) Masked finch total plasma CORT levels (ng/mL) to handling time (minutes). Resulting residuals for each species are shown on graphs on the right.

56 Gouldian and Long-tailed finch plasma CBG capacity was measured using a radioligand-binding assay with tritiated corticosterone (described in Lynn, 2003).

This assay measures the amount of tritiated CORT that can be bound by the CBGs active and present in a finch plasma sample, and is therefore quoted as nM CBG capacity, an indicator of the CORT binding capacity of the CBGs present. Raw plasma was stripped of steroids during a 30 - minute incubation period in a dextran-coated charcoal solution. Plasma from all finch species was diluted 1:200 and all samples were incubated in triplicate with both total binding (known concentration of tritiated corticosterone, H3-CORT) and non-specific binding

(known concentration of unlabelled standard CORT) solutions for 2 hours at 4C, before being rinsed with 9 mL buffer through 25 mM Tris buffer and 0.3% polyethyleneimine soaked glass filters using vacuum filtration in a Brandel cell harvester. The amount of radiolabelled CORT bound to filters was measured using standard liquid scintillation spectroscopy (Ultima Gold scintillation fluid by brand;

Tricarb Liquid Scintillation Analyzer 2800TR 2004, Perkin Elmer, Inc.). To calculate nM concentration of CBG, the disintegrations per minute (DPM) were multiplied by plasma dilution (200) and K. K is the correction factor determined by label-specific activity (SA) and volume assayed (litres) using the formula:

Pooled plasma from captive Japanese quail (Coturnix coturnix japonica) served as a reference standard for all assays of CBG capacity and was tested along with

57 individual finch plasma samples on all filters. This revealed average intra-assay variability to be < 5% and inter-assay variability to be < 10%.

Validation of the assumption that health index variation was not due to dehydration

A subset of Gouldian and Long-tailed finch plasma collected at all three sites used in this study during the late dry season (when water is scarce, and birds are more likely to be dehydrated) was tested for differences in osmolality as an indicator of possible dehydration in individuals. Osmolality of avian plasma is known to increase up to 30% above normal in experimentally dehydrated birds (Arad et al.

1989; Zhou et al. 1999). It was important to determine whether finches in this study were significantly dehydrated, as reduced plasma volume due to dehydration has the potential to increase health measure readings, especially haematocrit,

CORT and CBG capacity (Vleck & Priedkalns 1985; Carmi et al. 1994).

Measurements were made using 10 uL finch plasma in a Wescor Vapor Pressure

Osmometer in duplicate and mean osmolality values (which ranged from 286–

401mmol/kg) were compared with regression against haematocrit and CORT readings for those individual finches. Regression showed no correlation between osmolality of plasma and health indices of Gouldian (haematocrit F1, 47 = 0.008, P =

0.93; CORT F1, 11 = 0.17, P = 0.69) or Long-tailed finches (haematocrit F1, 28 =

0.27, P = 0.61, CORT not compared in Long-tailed finches), thus differences in mean health indices are likely to be independent of bird hydration status, and more a reflection of other factors such as stress and nutritional status. 58

Statistical analysis

All health measurements (body mass, muscle score, fat score, haematocrit, haemoglobin and stress indices) were sampled from naïve individuals each season and found to be normally distributed. Thus, three-factor ANOVA models were used to analyse finch health measure response to the factors season (breeding and non- breeding), year (first year = non-breeding 2007 and breeding 2008, and second year = non-breeding 2008 and breeding 2009) and site (Bradshaw, Delamere and

Yinberri) (SAS institute JMP5.1 software). Separate three-factor ANOVA was used to analyse finch health measure variation between age and sex for each finch species, and whether this differed between seasons. Where factors were found to have an effect on health measures, Tukey – Kramer HSD tests were used to determine the direction of significant differences. Body mass was corrected for individual structural size using head-bill measurements due to the strong

2 correlation between mass and head-bill (Gouldian finch r = 0.10 F1, 445 = 51.00, P

2 < 0.0001, Long-tailed finch r = 0.11, F1, 443 = 56.23, P < 0.0001 and Masked finch

2 r = 0.16, F1, 356 = 69.96, P > 0.0001). Body mass, fat and muscle scores were analysed separately in order to fully describe seasonal and annual physiological variation in finches. Interactions between health measures were analysed separately for each species using regression analysis (SAS institute JMP5.1 software). Mean body size differences between sexes were analysed using one- way ANOVA by using the head-bill measurements and tarsus length as factors.

59 Wing chord, although measured in the field, was not used, as mean measures completely overlapped between known sexed birds of all three finch species.

2.3 Results

Residual CORT levels and CBG capacity

Gouldian finch

For Gouldian finch results, season had a significant effect on residual CORT levels and CBG capacity (CORT F1, 277 = 138.05, P < 0.0001, CBG capacity F1, 210 = 18.0,

P < 0.0001), with this species of finch showing significantly higher residual CORT and CBG capacity during non-breeding seasons compared to breeding seasons

(Tukey – Kramer HSD, Figures 2.2, 2.4). This seasonal effect on CORT was dependent on year and site (F2, 277 = 12.53, P < 0.0001), with CORT levels not differing between seasons at Yinberri during the first year (Figure 2.3). Mean CBG capacity was significantly higher during the second year of the study (F1, 210 = 15.5,

P < 0.0001), however, the CBG levels were still significantly higher during breeding seasons compared to non-breeding seasons during both years (Tukey – Kramer

HSD, Figure 2.4). Seasonal changes in CBG levels were not dependant on site

(interaction F2, 197 = 1.27, P = 0.28). Several differences were observed between adult and juvenile stress responses. The effect of age on stress response was dependent on season (F1, 285 = 12.58, P < 0.0001), with Gouldian finch juveniles having higher residual CORT compared to adults during non-breeding seasons 60 (adult residual CORT = 5.58 ng/mL ± 1.55 SE; juvenile residual CORT = 11.59 ng/mL ± 1.43 SE), and significantly lower levels during breeding seasons (adult residual CORT = -5.42 ng/mL ± 1.20 SE; juvenile residual CORT = -9.56 ng/mL ±

1.51 SE; Tukey – Kramer HSD). The effect of age on CBG capacity was also dependent on season (F1, 205 = 5.50, P = 0.02), with CBG capacity lower in juveniles during the breeding season when most juveniles captured are fledglings

(adult CBG = 98.14 nM ± 3.85 SE; juvenile CBG = 67.0 nM ± 8.15 SE), but similar during non-breeding seasons when most juveniles are independent (Tukey –

Kramer HSD). Thus juvenile stress responses were higher than adults during non- breeding and lower during breeding. Differences in residual CORT characteristics were observed between the sexes. Female Gouldian finches had consistently higher mean residual CORT compared to males (female residual CORT = 4.99 ng/mL ± 1.44 SE; male residual CORT = -0.16 ng/mL ± 1.37 SE; Tukey – Kramer

HSD). Females had similar CBG capacity compared to males in all seasons (F1, 188

= 0.36, P = 0.55).

61

10

Gouldian Finch 5 Long-tailed Finch Masked Finch 0

-5 Residual CORT Residual (ng/mL) -10

Breeding Breeding

Non-breeding Non-breeding Season

Figure 2.2: Mean residual CORT levels (ng/mL) for Gouldian, Long-tailed, and

Masked finches during two non-breeding seasons (2007 and 2008) and two breeding seasons (2008 and 2009). Symbols represent means with standard error bars. Error values are small, and bars are not often visible.

62

d>b>y y>b=d d=y>b b=d=y 30

25 Bradshaw 20 Delamere 15 Yinberri 10

5

0

-5

-10 Mean(ng/mL) ResidualCORT -15

Breeding Breeding

Non-breeding Non-breeding

Season

Figure 2.3: Mean residual CORT levels (ng/mL) for Gouldian finches sampled at three sites during two non-breeding seasons (2007 and 2008) and two breeding seasons (2008 and 2009). Symbols represent means with standard error bars.

The direction of significant differences between sites are shown using abbreviations for each site (Tukey – Kramer HSD).

63 190

170 Gouldian Finch 150 Long-tailed Finch 130

110

90 CBG Capacity CBG (nM)Capacity 70

50

Breeding Breeding

Non-breeding Non-breeding Season

Figure 2.4: Mean CBG Capacity (nM) for adult Gouldian and Long-tailed finches during two non-breeding seasons (2007 and 2008) and two breeding seasons

(2008 and 2009). Symbols are means with standard error bars. Error values are small, and bars are not often visible.

64 Long-tailed finch

The effect of season on residual CORT of Long-tailed finches was dependent on year and site (interaction F2, 266 = 6.88 P = 0.001), while seasonal variation in CBG capacity was weak, but consistent between years (season F1, 133 = 3.09, P = 0.08).

Unlike Gouldian finches, Long-tailed finches had significantly higher residual CORT and CBG capacity during both breeding seasons and lower residual CORT and

CBG capacity during the non-breeding season 2008 (Tukey – Kramer HSD,

Figures 2.2, 2.4). During the non-breeding season 2007, birds had high residual

CORT levels on Yinberri only and similar CBG capacity compared to non-breeding

2008 (Tukey – Kramer HSD, Figures 2.2, 2.4, 2.5). Over all seasons, juvenile

Long-tailed finches had lower residual CORT (Tukey – Kramer HSD) and similar

CBG capacity compared to adults (F1, 134 = 0.72, P = 0.40). Differences in residual

CORT between the sexes were not significant (F1, 141 = 0.00001, P = 0.998).

Female finches had lower CBG capacity compared to males during breeding 2008

(female CBG = 131.2 nM ± 28.3 SE; male CBG = 213.5 nM ± 27.2 SE, Tukey –

Kramer HSD). Females had similar CBG capacity compared to males during other seasons (F1, 39 = 2.02, P = 0.16).

65

30 y>d=b b>y=d b=d=y b=d=y 25 Bradshaw 20 Delamere

15 Yinberri

10

5

0

-5 Mean(ng/mL) ResidualCORT

-10

-15

Breeding Breeding

Non-breeding Non-breeding

Season Figure 2.5: Mean residual CORT levels (ng/mL) for Long-tailed finches sampled at three sites during two non-breeding seasons (2007 and 2008) and two breeding seasons (2008 and 2009). Symbols represent means with standard error bars.

The direction of significant differences between sites are shown using abbreviations for each site (Tukey – Kramer HSD).

66 Masked finch

In Masked finches, the effect of season on stress response was only weakly dependent on year (interaction F1, 213 = 3.07, P = 0.08). This can be seen in stable residual CORT levels across sampling periods except in the non-breeding season

2007, when levels were significantly elevated compared to all other seasons

(Tukey – Kramer HSD, Figure 2.2). There were no site differences in Masked finch

CORT levels (F2, 213 = 0.04, P = 0.96). Interestingly, mean CORT levels were often intermediate to Long-tailed and Gouldian finch levels (Figure 2.2). Juvenile

Masked finches had similar CORT compared to adults during the two seasons measured (non-breeding 2008, breeding 2009; F1, 222 = 0.27, P = 0.60), however, there is little data on juvenile CORT characteristics (n < 10 both seasons). Masked finches had similar CORT levels between the sexes (F1, 94 = 2.51, P = 0.12).

Body condition (muscle score, fat score and body mass)

Gouldian finch

Though muscle scores, fat and body mass of Gouldian finches often differed between seasons, variation in body condition sometimes depended on year, age and sex. The effect of season on body mass was dependent on age (F1, 442 = 8.18,

P = 0.004), with adults having higher body mass during breeding, and juveniles having higher body mass during non-breeding seasons (Tukey – Kramer HSD,

Figure 2.6). Muscle scores were higher during breeding seasons for all Gouldian 67 finches (F1, 439 = 7.26, P = 0.007, Tukey – Kramer HSD, Figure 2.6). The effect of season on fat was dependent on year (F1, 439 = 26.68, P < 0.0001). All finches had highly elevated fat scores during the non-breeding season compared to breeding in the first year of the study, and higher fat scores during breeding during the second year (Tukey – Kramer HSD). No seasonal changes in body condition depended on site (body mass F2, 435 = 1.92, P = 0.14, muscle F2, 439 = 1.54, P = 0.21, fat F2, 439 =

1.38, P = 0.25). Overall, juvenile Gouldian finches tended to have lower muscle scores and body mass compared to adults (Tukey – Kramer HSD, Figure 2.6).

The effect of age on fat scores was nearly dependent on season (F1, 446 = 3.17, P =

0.08) because during the breeding seasons, when juveniles are particularly young, juveniles had significantly higher fat scores than adults (Tukey – Kramer HSD,

Figure 2.6). When juveniles are older in the non-breeding seasons they had similar fat scores compared to adults. The effect of sex on body mass and muscle was dependent on season (body mass F1, 305 = 6.50, P = 0.02; muscle; F1, 309 =

4.47, P = 0.04). Female Gouldian finches tended to have higher muscle scores than males during the breeding seasons, but lower scores during the non-breeding seasons (Tukey – Kramer HSD, Figure 2.6). Females also had higher body mass than males during the breeding seasons (but similar mass during non-breeding seasons Tukey – Kramer HSD), despite their significantly smaller mean body size

(head to bill measure F1, 312 = 31.76, P < 0.0001; tarsus F1, 194 = 9.05, P = 0.003), and the fact that body mass was corrected for size. This is likely an effect of egg production. Fat scores showed no predictable seasonal differences between the sexes (Figure 2.6).

68 a) Body mass b) Muscle score 1.5 2.5

1.0 2.0

0.5 1.5

0.0 1.0

-0.5 0.5

Muscle Score (0Muscle - 3) Residual Body Mass (g) Mass Body Residual -1.0 0.0

Breeding Breeding Breeding Breeding

Non-breeding Non-breeding Non-breeding Non-breeding Season Season

c) Fat score 2.5 Female 2.0 Male Adult 1.5 Juvenile

1.0

Fat Fat Score (0 - 5) 0.5

0.0

Breeding Breeding

Non-breeding Non-breeding Season

Figure 2.6: Mean Gouldian finch residual body mass (a), muscle scores (b) and fat scores (c). Plots depict means for two non-breeding seasons (2007 and 2008) and two breeding seasons (2008 and 2009) and include values for differing ages and sexes. Standard error bars are shown when variation is large enough to be visible.

69 Long-tailed finch

Body masses of Long-tailed finches were higher during breeding seasons compared to non-breeding seasons, irregardless of year or site (season F1, 433 =

25.21, P < 0.0001, year and site P > 0.05; Figure 2.7). Muscle scores were significantly higher during breeding during the second year of the study at all sites

(interaction between year and season F1, 433 = 8.71, P = 0.003; Figure 2.7). The effect of season on fat scores was dependent on year and site (F2, 433 = 5.86, P =

0.003). Fat scores were very high during the non-breeding season compared to breeding during the first year of the study at Bradshaw and Delamere, but similarly low between seasons at Yinberri (Tukey – Kramer HSD). Also, fat scores were only slightly higher during breeding compared to non-breeding at all sites during the second year of the study (Tukey – Kramer HSD, Figure 2.7). Juvenile Long- tailed finches had consistently lower muscle scores and body mass compared to adults (Tukey – Kramer HSD, Figure 2.7). Fat scores were statistically similar between ages (F1, 443 = 0.71, P = 0.40; Figure 2.7). Similar to breeding Gouldian finches, Long-tailed finch females had significantly smaller body size (head to bill measure F1, 148 = 45.60, P < 0.0001; tarsus F1, 173 = 3.38, P = 0.07) but similar real body mass (F1, 147 = 1.20, P = 0.28) compared to males, possibly due to egg production. Correcting mass for bird size did not cause sex differences to be significant (residual body mass F1, 146 = 1.94, P = 0.16). Muscle was similar between sexes (F1, 146 = 1.10, P = 0.30). The effect of sex on fat scores was dependent on season (F1, 144 = 5.66, P = 0.02), with fat scores significantly higher in breeding females (Tukey – Kramer HSD, Figure 2.7). 70 (a) Body mass (b) Muscle score

1.0 2.5

0.5 2.0

0.0 1.5

-0.5 1.0

0.5

-1.0 Muscle Score (0Muscle - 3)

Residual Body Mass (g) Mass Body Residual 0.0 -1.5

Breeding Breeding Breeding Breeding

Non-breeding Non-breeding Non-breeding Non-breeding Season Season

(c) Fat score 2.5 Female 2.0 Male Adult 1.5 Juvenile

1.0

Fat Score Fat (0Score - 5) 0.5

0.0

Breeding Breeding

Non-breeding Non-breeding Season

Figure 2.7: Mean Long-tailed finch residual body mass (a), muscle scores (b), and fat scores (c). Plots depict means for two non-breeding seasons (2007 and 2008) and two breeding seasons (2008 and 2009) and include values for differing ages and sexes. Standard error bars are shown when values are large enough to be visible.

71

Masked finch

Masked finch muscle scores were higher during breeding seasons compared to non-breeding seasons (muscle F1, 349 = 31.52, P < 0.0001; Tukey – Kramer HSD,

Figure 2.8). This seasonal difference in muscle was not affected by year or site

(interaction with year F1, 349 = 0.17, P = 0.68, site F2, 349 = 1.09, P = 0.34). There was no seasonal changes in residual body mass (F1, 346 = 1.24, P = 0.26). The effect of season on fat score was dependent on year (F1, 357 = 16.12, P < 0.0001), with fat scores higher during non-breeding compared to breeding during the first year, and similarly low during both seasons in the second year (Tukey – Kramer

HSD, Figure 2.8). Seasonal differences in fat were not dependant on site (F2, 349 =

2.65, P = 0.07). Juvenile Masked finches had lower muscle scores and body mass compared to adults (Tukey – Kramer HSD, Figure 2.8). Fat scores were similar between ages (F1, 357 = 3.37, P = 0.07). Neither sex had consistently different muscle or fat scores (muscle F1, 100 = 0.21, P = 0.65, fat F1, 100 = 0.73, P = 0.39).

Similar to other finch species, uncorrected body mass showed that bird mass was similar between Masked finch sexes despite the fact that this species is strongly sexually dimorphic in body size measures (body mass F1, 100 = 0.44, P = 0.51; head to bill measure F1, 100 = 33.24, P < 0.0001; tarsus F1, 48 = 13.99, P = 0.0005). When mass is corrected for size, females are nearly significantly heavier (residual body mass F1, 100 = 3.65, P = 0.06).

72 a) Body mass b) Muscle score 14 2.5

2.0 13 1.5

1.0 Mass (g)Mass 12

0.5 Muscle Score (0 - 3) Score Muscle 11 0.0

Breeding Breeding Breeding Breeding

Non-breeding Non-breeding Non-breeding Non-breeding Season Season

c) Fat score 2.5 Female 2.0 Male Adult 1.5 Juvenile

1.0

Fat Score Fat (0 Score - 5) 0.5

0.0

Breeding Breeding

Non-breeding Non-breeding

Season

Figure 2.8: Mean Masked finch residual body mass (a), muscle scores (b) and fat scores (c). Plots depict means for two non-breeding seasons (2007 and 2008) and two breeding seasons (2008 and 2009) and include values for differing ages and sexes. Standard error bars are shown where values are large enough to be visible.

73 Haematocrit and haemoglobin

Gouldian finch

The effect of season on Gouldian finch haemoglobin levels depended on year (F1,

382 = 44.14, P < 0.0001) with higher haemoglobin levels during the non-breeding season in the first year, and higher levels during breeding in the second year of the study (Tukey – Kramer HSD, Figure 2.9). Seasonal changes in haemoglobin levels also depended on site (F2, 382 = 7.95, P = 0.0004). Delamere birds had significantly higher haemoglobin levels in the first non-breeding season and both breeding seasons compared to the second non-breeding season, while Yinberri birds showed higher levels in the first non-breeding season compared to other seasons, and Bradshaw birds only had higher levels during the second breeding season compared to other seasons (Tukey – Kramer HSD). Seasonal differences in haematocrit depended on site (F2, 403 = 11.73, P < 0.0001), with Bradshaw birds showing significantly higher levels during breeding seasons, Delamere birds showing higher levels during non-breeding seasons, and Yinberri birds showing no seasonal trends (Tukey – Kramer HSD). Gouldian finch adults had higher haematocrits but similar haemoglobin levels compared to juvenile birds

(haematocrit Tukey – Kramer HSD, haemoglobin F1, 392 = 0.03, P = 0.87; Figure

2.9). Females had similar haematocrit and haemoglobin levels compared to males

(haematocrit F1, 292 = 0.004, P = 0.95; haemoglobin F1, 274 = 0.01, P = 0.91; Figure

2.9).

74 215 210 205 Female 200 Male 195 Adult 190 Juvenile

Haemoglobin (ng/mL) Haemoglobin 65.0

62.5

60.0

57.5

Haematocrit 55.0

Breeding Breeding Non-breeding Non-breeding Season

Figure 2.9: Mean Gouldian finch haemoglobin (top graph) and haematocrit for two non-breeding seasons (2007 and 2008) and two breeding seasons (2008 and

2009). Plots depict means for different ages and sexes. Standard error bars are shown where values are large enough to be visible.

75 Long-tailed finch

Long-tailed finch haematocrit differed between breeding and non-breeding seasons depending on site and year (both interacted with season F2, 413 = 4.63, P = 0.01;

Figure 2.10). Mean haematocrit levels for Delamere birds were higher during the first year non-breeding season compared to other seasons, haematocrit levels were higher during the second year breeding season at Yinberri, and all levels were similar between seasons on Bradshaw. Seasonal differences in Long-tailed finch haemoglobin depended on year (F1, 354 = 26.88, P < 0.0001). Haemoglobin was similar between seasons during the first year of the study and higher during breeding compared to the non-breeding season during the second year (Tukey –

Kramer HSD, Figure 2.10). Seasonal differences in haemoglobin did not depend on site (F2, 354 = 1.71, P = 0.18). Haematocrit and haemoglobin of different aged birds were not significantly different (haematocrit F1, 423 = 0.64, P = 0.42; haemoglobin F1, 364 = 0.05, P = 0.81; Figure 2.10). There were no significant gender differences in either haematocrit or haemoglobin values (haematocrit F1, 146

= 0.58, P = 0.45; haemoglobin F1, 131 = 1.06, P = 0.30; Figure 2.10).

76 215 210 205 Female 200 Male 195 Adult 190 Juvenile

65 Haemoglobin (ng/mL)Haemoglobin

60

55

Haematocrit Haematocrit 50

Breeding Breeding

Non-breeding Non-breeding Season

Figure 2.10: Mean Long-tailed finch haemoglobin (top graph) and haematocrit for two non-breeding seasons (2007 and 2008) and two breeding seasons (2008 and

2009). Plots depict means for different ages and sexes. Standard error bars are shown where values are large enough to be visible.

77 Masked finch

Similar to the other two finch species, Masked finch haematocrit and haemoglobin sometimes differed between sampling periods, but there were no consistent significant differences between these measures during breeding and non-breeding seasons (Figure 2.11). Haematocrit varied by year only (F1, 340 = 7.54, P = 0.006), with higher haematocrits measured during the second year of the study (Tukey –

Kramer HSD). The effect of season on haemoglobin depended on year and site

(interaction with year F1, 291 = 11.25, P = 0.0009 and site F2, 291 = 3.89, P = 0.02), with Masked finch haemoglobin similar between seasons in the first year at

Bradshaw and Delamere, but higher during the non-breeding seasons at Yinberri, and higher during the breeding season in the second year of the study on all properties (Tukey – Kramer HSD). There was a nearly significant effect of age on haematocrit depending on season (F1, 347 = 2.92, P = 0.08). This is because

Masked finch adults had similar haematocrit and haemoglobin compared to juveniles, except during the non-breeding season when juvenile haematocrit was low compared to adults (Tukey – Kramer HSD, Figure 2.11). There were no significant differences between haematocrit and haemoglobin readings of the two sexes (haematocrit F1, 100 = 0.52, P = 0.47; haemoglobin F1, 90 = 0.04, P = 0.84;

Figure 2.11).

78 215 210 205 200 Female 195 Male 190 Adult

Haemoglobin (ng/mL) 65 Juvenile

60

55

50 Haematocrit

Breeding Breeding

Non-breeding Non-breeding Season

Figure 2.11: Mean Masked finch haemoglobin (top graph) and haematocrit for two non-breeding seasons (2007 and 2008) and two breeding seasons (2008 and

2009). Plots depict means for different ages and sexes. Standard error bars are shown where values are large enough to be visible.

79 Interactions between health indices

Gouldian finch

No Gouldian finch health measures were strongly correlated, with most correlations only explaining 1 - 5% of health measure variation. Body mass was positively

2 correlated to fat and muscle scores (r = 0.06, F2, 443 = 15.31, P < 0.0001); however, this regression model only explained 6% of the variation in these measures, suggesting the three measures are not strongly correlated. One-factor regression model analyzing fat and muscle suggests that these two measures are

2 slightly more correlated (r = 0.12, F1, 451 = 61.41, P < 0.0001) than a two-factor regression model incorporating body mass. There was a significant negative correlation between residual CORT and muscle score for adult Gouldian finches (r2

= 0.03, F1, 158 = 5.20, P = 0.02). Thus birds with lower muscle scores sometimes had higher CORT levels; however, residual CORT showed no correlation with fat score or body mass (P > 0.05). Residual CORT had a weak positive correlation

2 with haematocrit (r = 0.01, F1, 287 = 3.74, P = 0.05), and no correlation with

2 haemoglobin (r = 0.0001, F1, 267 = 0.01, P = 0.90). Body mass, fat and muscle scores were all weakly negatively correlated with haematocrit, thus individuals with lower haematocrit sometimes had higher measures of body condition (mass r2 =

2 2 0.007, F1, 407 = 3.08, P = 0.08; fat r = 0.02 F1, 415 = 7.01, P = 0.008; muscle r =

0.004 F1, 415 = 1.75, P = 0.18). However, only body mass had a weakly negative

2 correlation with haemoglobin (r = 0.007, F1, 387 = 3.06, P = 0.08).

80 Long-tailed finch

Muscle and fat scores and body mass were all positively correlated to each other

2 (r = 0.10, F2, 442 = 25.60, P < 0.0001). Residual CORT had a significant positive

2 correlation with haemoglobin (r = 0.04, F1, 231 = 8.56, P = 0.004), and a weak

2 positive correlation with haematocrit (r = 0.01, F1, 275 = 3.08, P = 0.08). Residual

2 CORT had a significant negative correlation with fat (r = 0.02, F1, 276 = 7.18, P =

0.008) and non-significant results of one-factor regression with other body condition measures (P > 0.05). Both haematocrit and haemoglobin were not correlated with any condition measures (P > 0.05). Thus there was no correspondence between haematological state and body condition in this species.

Masked finch

Muscle and fat scores and body mass were all positively correlated to each other

2 (r = 0.02, F1, 355 = 5.44, P = 0.005). Residual CORT had no correlation to body mass, muscle, fat, haematocrit or haemoglobin (P > 0.05). Body mass was

2 negatively correlated with haemoglobin (r = 0.02, F1, 299 = 5.72, P = 0.02), but not haematocrit (P > 0.05). Thus individuals with lower haemoglobin tended to have higher mass, but other measures of body condition seemed unrelated.

81 2.4 Discussion

Seasonal stress response variation

Despite all three finch species experiencing similar climatic and habitat variation, the results of this study suggest that Gouldian finches have unique seasonal changes in their stress response and CBG capacity. While most small passerines have their highest stress levels during breeding seasons (Astheimer et al. 1994;

Breuner & Orchinik 2002b; Romero 2002; Wada et al. 2006b), Gouldian finches did not follow this pattern. Breeding Gouldian finches had consistently lower stress responsiveness (as measured by residual CORT) and higher CBG capacity compared to non-breeding and moulting individuals sampled during the late dry/early wet season. This seasonal pattern of CORT levels and CBG capacity should translate into lower levels of free CORT during breeding (less CORT would be available, being bound to CBGs) and comparatively higher free CORT during non-breeding or moulting periods in Gouldian finches.3 Interestingly, overall mean

Gouldian finch CBG capacity is unusually low in comparison to the overall mean

CBG capacity of the common Long-tailed finch, suggesting that Gouldian finches retain a smaller reservoir of CORT bound to CBGs in general with greater variation

3 The actual calculation of free CORT would involve the determination of CBG affinity for CORT, which can differ by species, season, year and site (Malisch & Breuner 2010). Because of the extra months of expensive laboratory analysis needed to collect this data, and the lack of consensus among researchers as to the biological action of free CORT, we determined that this calculation was beyond the scope of this study and possibly unnecessary for the general examination of the dynamics between CORT and CBGs presented in this preliminary work.

82 in free CORT levels compared to Long-tailed finches. Long-tailed finches also had seasonally variable CORT levels but with a very different pattern to Gouldian finches. Long-tailed finches tended to have both higher stress responses and CBG capacity during breeding seasons and lower levels during the non-breeding seasons, similar to the pattern reported for most other passerines (Astheimer et al.

1994; Romero & Wingfield 1998; Romero & Remage-Healey 2000; Breuner &

Orchinik 2001, 2002a). This suggests that Long-tailed finches maintained low free

CORT levels all year but a larger CORT reservoir during the breeding season. The pattern of matching high levels of CORT and CBG levels during the breeding seen in Long-tailed finches is only one of five possible patterns that birds balance CORT and CBG capacity while breeding (Table 2.1).

Table 2.1: Basic patterns of CORT and CBG levels found in birds during breeding compared to other times of the year. The five patterns listed are based on those found in the literature (cited in text) and in this study.

CORT/CBG Pattern CORT Levels CBG Levels

1. Matching high levels Higher than mean Higher than mean

2. CBG highly elevated Higher than mean Extremely elevated

3. CORT suppression Lower than mean Similar to mean

4. Chronic stress Lower than mean Lower than mean

5. Gouldian finch Lower than mean Higher than mean

83

Differences in life history traits of the three finch species studied may underlie their different physiological adaptations and resulting patterns of CORT and CBG capacity. As Gouldian finches are short lived and extremely productive breeders it is very likely that their sustained breeding effort lowers individual long-term survival probability despite the abundance of food resources during the breeding season.

Short-lived birds and those with short breeding seasons are thought to minimize stress reactivity during nesting because stress induced interruptions to breeding would severely reduce their lifetime breeding potential. For example, the arctic breeding White-crowned sparrow (Zonotrichia pugetensis) has only a brief summer in which to raise one brood and this species has lower free CORT levels due to their higher CBG capacity in comparison to temperate breeding sparrows which have long summers in which they can raise multiple broods (Breuner et al. 2003).

Increased CBG capacity mitigating high CORT levels is the second pattern of balancing stress levels during breeding seen in the literature. Open desert breeding birds in North America were found to suppress the release of CORT without differing in CBG capacity compared to nearby birds breeding in the shady riparian habitat (Wingfield et al. 1992). This is a third pattern seen in the literature.

Results of this study show that Gouldian finches have lower stress levels because of their low CORT release as well as high CBG levels during breeding. This appears to be either an adaptation to promote the maintenance of breeding behaviour throughout the season, a product of lowered stress due to abundant food, or alternatively, symptoms of chronic stress during breeding. However, based on current knowledge, chronically stressed birds tend to reduce their CBG 84 capacity and down-regulate CORT response (Fleshner et al. 1995; Rich & Romero

2005; Cyr et al. 2007). Thus breeding Gouldian finches do not fit the profile of chronically stressed birds (fourth known pattern of CORT/CBG balance), and their stress response pattern represents a fifth and alternative pattern balancing CORT and CBG capacity in breeding passerines (Table 2.1).

In contrast, Gouldian finches sampled during the non-breeding season had much higher residual CORT levels in response to handling stress and lower CBG capacity, suggesting that free CORT is predictably high during this time of year.

Increasing plasma CORT levels have been shown to increase foraging behaviour in a variety of bird species (Wingfield et al. 1983; Wingfield & Silverin 1986;

Astheimer et al. 1995; Lynn et al. 2003a). It is probable that increased CORT drives Gouldian finches to forage more actively and therefore increases their odds of survival in this season of relatively poor food resource availability. Because nearly all individuals are undergoing a complete body and flight feather moult over a short time period (4 - 8 weeks) during the non-breeding season, it is critical that nutritional needs are met. Lowered body condition or any stressor with the potential of interrupting feeding would therefore be expected to elicit an extreme stress response. The increased reactivity of Gouldian finches to stress during the non-breeding season could be a response to lowered resource availability during the metabolically challenging moult period. An increase in energy demands and metabolic rate has been documented in many bird species during moult, and would be expected to occur in the Gouldian Finch (Klaassen 1995; Swaddle & Witter

1997). King penguins (Aptenodytes patagonica) have an increase in stress 85 responsiveness at the end of their moult, most likely due to nutritional deficiencies during fasting (Cherel et al. 1988a). The increased CORT seen in Gouldian finches would promote increased feeding or time spent searching for productive foraging sites, and may be a response to low caloric intake during moult, similar to fasting in King penguins. Interestingly, several studies of birds undergoing moult in resource rich environments show that these species have lower stress levels and responses during moult, opposite to the Gouldian finch (Astheimer et al. 1995;

Romero & Wingfield 1999; Romero & Remage-Healey 2000; Rich & Romero

2001). It is possible that the CORT increase seen in non-breeding Gouldian finches may not be necessary during the resource rich breeding season and could even be counter productive during breeding through the interruption of breeding behaviour. Alternatively, high stress responses coupled with lowered CBG capacity during non-breeding could be a symptom of chronic stress due to a persistent poor diet, as similar patterns are often seen in naturally and experimentally fasted birds (Cherel et al. 1988b; Astheimer et al. 1995; Vleck et al.

2000; Lynn et al. 2003a). Captive Gouldian finches have been shown to have higher stress response levels when fed a nutrient poor diet compared to captive finches fed a diet more complete in nutrients (Prike 2008). Predictable seasonal chronic stress in non-breeding Gouldian finches would suggest that they may be poorly adapted to current non-breeding (late dry/early wet) season conditions in northern Australia. The idea that Gouldian finches may be naturally sensitive to environmental changes is predicted from lifestyle traits. The nomadic lifestyle of the Gouldian finch may help these birds to access naturally dispersed resources; however, this highly energetic lifestyle and their diet specialization could now be 86 limiting this species survival probability if resources are now less available or more widely dispersed due to habitat degradation.

Unlike the rare Gouldian finch, the common Long-tailed finch showed a balanced stress response during most seasons. Long-tailed finches had lower CORT levels and lower CBG capacity in response to a stressor during the non-breeding season compared to the higher CORT levels and higher CBG capacity measured during the breeding season. This ‘matching’ of CORT and CBG levels should translate into a more constant background of low free CORT during both seasons. Studies on the seasonal modulation of stress responses show that many other passerines also ‘match’ high total CORT and CBG during breeding and ‘match’ low levels during non-breeding or moult periods (Breuner & Orchinik 2001, 2002b; Wada et al. 2006b). In two studies, Wingfield has theorized that higher potential stress levels during breeding are helpful in promoting adult survival behaviour over breeding behaviour during stressful climatic or starvation events (Wingfield et al.

1983; Wingfield & Silverin 1986). However, Long-tailed finches monitored during non-breeding season 2007 had low CBG capacity but very high CORT levels at the

Yinberri Hills site, while the other two sites had low CORT levels. This should equate to higher free CORT in late 2007 compared to 2008 for Long-tailed finches at Yinberri Hills, but similarly low free CORT during non-breeding seasons for finches on other sites. This pattern of stress responsiveness and CBG capacity the Yinberri Hills during 2007 is similar to the pattern seen in Gouldian finches during all non-breeding seasons. The unusually high free CORT during this season may be an indication of unusually poor environmental conditions in the 87 Yinberri Hills or a symptom of birds recovering from a period of chronic stress.

These patterns of stress in Long-tailed finches at one site suggest that this species’ stress response changes not only in relation to biological state (breeding vs. non- breeding) but CORT response and CBG capacity also varied in response to the changing environmental conditions in different years and at different sites. Long- tailed finches are relatively long-lived and opportunistic in their timing of breeding and moult compared to Gouldian finches (Tidemann & Woinarski 1994; Franklin et al. 1998; Higgins et al. 2006). It is likely that their ability to time breeding and moult during optimal environmental conditions makes them less likely to show evidence of either suppression of the stress response during breeding or chronic stress during the non-breeding season.

Gouldian finch patterns of stress response suggested this species may also slightly vary the pattern of higher stress responses during non-breeding compared to breeding seasons in different years and at different sites. Gouldian finches monitored in the Yinberri Hills had moderate stress response levels during both seasons in the first year, but a high/low pattern similar to other sites in the second year. Thus Gouldian finches do not always show symtoms of cronic stress during the non-breeding (late dry/early wet) season. If Gouldian finches are chronically stressed some years and this translates into lowered survival of finches during these years, this could explain why the abundance of this endangered species is estimated to fluctuate widely between years in the Yinberri Hills (O'Malley 2006a;

Price et al. In preparation).

88 Though CBG capacity was not measured in Masked finches, the seasonal patterns in stress induced CORT levels highlight a possible intermediate response compared to Gouldian and Long-tailed finches. Similar to both other species,

Masked finches had high levels of total CORT during non-breeding season 2007.

However, Masked finches do not have the fluctuating high and low levels seen in other finch species between 2008 and 2009. Interestingly, Masked finch CORT levels were relatively stable between all seasons, except non-breeding 2007. Due to the similarity in life history traits between Masked and Long-tailed finches, it is likely that Masked finches would ‘match’ total CORT and CBG capacity similarly to

Long-tailed finches (Franklin et al. 1998; Dostine & Franklin 2002; Higgins et al.

2006). Further analysis of Masked finch stress responses and CBG capacity is needed to confirm this hypothesis.

Seasonal body condition variation

All finches showed seasonal variation in body condition. In particular, birds had higher muscle scores and body mass during the breeding season, possibly as a result of improved food resources during the breeding season. Though finches rarely carried much fat, Gouldian, Long-tailed and Masked finches all had significantly higher fat scores during non-breeding seasons, particularly in 2007.

Knowing that seed resources dwindle by the end of the non-breeding season

(Woinarski & Tidemann 1992; Dostine & Franklin 2002), these results suggest that finches are storing fat to compensate for brief periods of lower food availability.

Theoretically, birds with year-round home-ranges and reliable food resources will 89 carry less fat, as larger reserves increase mass and decrease flight agility

(McNamara & Houston 1990; Houston & McNamara 1993; Witter & Cuthill 1993;

Brodin 2001). It is interesting to note that Gouldian, Long-tailed and Masked finch maximum stress response values were recorded the same season as maximum fat scores (non-breeding 2007). Increased stress levels are sometimes positively correlated with increased fat storage in passerines (Wingfield & Silverin 1986; Gray et al. 1990), however, Long-tailed finch residual CORT levels showed a negative correlation with fat scores, suggesting that Long-tailed finches with lower fat stores have increased stress responses, similar to migrating or arctic breeding northern hemisphere passerines (Schwabl et al. 1991; Wingfield et al. 1994; Jenni et al.

2000; Muller et al. 2007). It is possible that poor resource availability during non- breeding (late dry/early wet) seasons encouraged birds to trade flight agility for increased fat stores as a safeguard against starvation, similar to many other bird species exposed to unpredictability in food access (Ekman & Hake 1990; Hurly

1992; Bednekoff & Krebs 1995; Witter et al. 1995; Smith & Metcalfe 1997; Cuthill et al. 2000). Captive studies have shown that increases in body mass of less than

7% can lower the speed and agility of Zebra finches by as much as 30% (Metcalfe

& Ure 1995). This decrease in flight performance is very likely to translate into a higher probability of predation and increased energy demands. Compared to

2008, Gouldian finches displayed 7% higher mean body mass (based on mass not corrected for bird size) during the non-breeding season 2007 in which elevated fat scores and stress responses were also recorded. The greater variety in the diets of Long-tailed and Masked finches, including insects along with a diverse range of seeds, may provide them greater access to protein and reduce their need for fat 90 storage. Telemetry and observational studies of the nomadic behavior of Gouldian finches suggests that a majority of their movements may be a response to the seasonal patchiness of their preferred grass-seeds (Dostine et al. 2001; Dostine &

Franklin 2002; Legge 2009). If Gouldian finches must travel further and forage longer in order to find and feed on the dwindling sources of their preferred grass seeds during the non-breeding season, fat storage may be more advantageous to this specialist species.

Correlations between health indices

Gouldian finch stress response, haematocrit and condition measures were often weakly correlated. For example, Gouldian finches with lower body condition measures had higher haematocrits as well as higher residual CORT levels but no difference in CBG capacity. This equates to higher free CORT in birds with lower condition measures. Increased free CORT is likely to drive these individuals to forage more and expend more energy, and could therefore explain their higher haematocrit levels. Under high temperatures and no free water, elevated haematocrit levels could also be a response to dehydration. However, significant increases in haematocrit of pigeons (up to 12%) can take more than 17 hours to be detected, even at high air temperatures (Arad et al. 1989; Carmi et al. 1994). It is likely that part of this lag in plasma response to dehydration is due to fluid balance changes: animals moving water out of tissues and into the blood stream in response to dehydration (Khvoinitskaya 1959; Bernstein 1971; Dawson et al. 1983;

Carmi et al. 1993). In cases of dehydration in captivity, the osmolality of bird 91 plasma increases linearly with haematocrit (Carmi et al. 1994). The finding that the finch plasma used in this study showed no correlation between osmolality and haematocrit suggests the study animals were not dehydrated. This is not surprising, as all finches were captured near drinking water sources in the first three hours after sunrise, when the air temperatures are cool relative to later in the day. Though we cannot be sure that birds were able to drink before early morning capture, it is likely that they would have drank before roosting the night before, and would not yet be under high risk of dehydration due to cool morning temperatures.

Higher haematocrit and haemoglobin levels in birds are also often attributed to increases in energy demands, such as cold stress and increased activity during breeding or migration (Carpenter 1975; Breuer et al. 1995; Horak et al. 1998a; Box et al. 2002; Fair et al. 2007). However, many studies have also found a significant negative correlation between haematocrit, diet quality and body condition, e.g. fasting birds (or those with poor diets) had poor condition and lower than normal haematocrits (Merila & Svensson 1997; Saino et al. 1997; Piersma et al. 2000a;

Owen & Moore 2006; Fair et al. 2007). Higher than normal Gouldian finch haematocrits were correlated with higher stress responses and lowered body condition, which suggests that increased energy demands, not poor diet, was likely to be driving these increases in haematocrit levels.

Long-tailed finch health indices were not often correlated. Unlike Gouldian finches,

Long-tailed finch residual CORT levels were not related to mass or muscle measures, most likely because of the low variation in mass or muscle in this species. Based on the analysis of individual bird data, Long-tailed and Masked 92 finch haematocrit and haemoglobin were not consistently related to any body condition measures, however higher Long-tailed finch haematocrit and haemoglobin were correlated with higher CORT levels during all sampling periods.

Together, these results suggest that increased Long-tailed and Masked finch

CORT levels do not indicate poor body condition as seen in Gouldian finches but the positive interaction between haematocrit and CORT levels is similar in

Gouldian and Long-tailed finches. This interaction between CORT and haematocrit implies that stress responses increase with increasing energy demands; however, the correlation is weak, accounting for very only 1 - 2% of the variation while factors such as site, season, age and sex account for 10 – 40% of stress response variation in these species.

Variation in health indices due to age and sex

All finch species displayed age and sex differences in residual CORT levels, body condition and haematocrit, which must be taken into consideration if researchers are to compare health indices between individuals. However, the differences between ages and sexes are not consistently observed across species, seasons and years. This suggests that both environmental conditions and individual traits are interacting to determine individual condition.

Juvenile Gouldian and Long-tailed finches occasionally had significantly lower residual CORT levels, body condition and haematocrit levels compared to adults.

Though comparisons were not always significant, mean values suggest that 93 juvenile residual CORT is lower than adults when finches are young (during the breeding season) and can increase to levels similar to adults later in their first year

(during the non-breeding/late dry season). Other species also show an increase in stress responsiveness with age (Sims & Holberton 2000; Ouillfeldt et al. 2004;

Wada et al. 2006a). Juvenile Gouldian, Long-tailed and Masked finches also had lower muscle scores and body mass, but higher fat scores compared to adults, especially when they are recently fledged. Juvenile finches sometimes displayed lower haematocrit and haemoglobin concentrations, but the influence of age was inconsistent. It is possible that this is a reflection of the mixed age of juveniles sampled during the study. Previous research has shown that juveniles do tend to have lower haematocrit readings than adults for a certain period after hatching but they soon produce enough mature red-blood cells to match adult levels (Sturkie

1986; Fair et al. 2007). As a result, the difference between adult and juvenile haematocrits will be more pronounced when the majority of juveniles sampled are still quite young. The differences between health measures of juvenile and adult finches were not as apparent in the non-breeding season when first year birds are older (no longer fledglings, but independent) and moulting, and environmental conditions are likely to be more physiologically challenging.

In all species, female finches often showed a tendency to have higher mean residual CORT levels compared to males, especially during the breeding season

(though these results were not often significant, never in Masked finches).

Females also tended to have higher body mass than males during the breeding season despite their smaller body size. This is most likely an effect of the 94 abundance of females sampled during the egg laying stage of breeding. Adult female Gouldian finches often had lower muscle scores and body mass compared to males during the non-breeding season. Because muscle and mass measures are typically controlled by diet (particularly food availability or foraging success) lower muscle and mass measures suggest that female Gouldian finches are more nutritionally challenged compared to males during the non-breeding season

(Birkhead et al. 1999; Leary et al. 1999; Strong & Sherry 2000). Long-tailed and

Masked finches tend to have shorter breeding seasons or raise fewer broods per year compared to Gouldian finches (Tidemann & Woinarski 1994). This may explain why Long-tailed and Masked finches have few significant and consistent differences between the body conditions of different sexes, while female Gouldian finches seem to have lowered condition compared to males during non-breeding seasons. For all three species of finches, no significant differences in haematocrit or haemoglobin concentration were observed between the sexes. These hematological measures may therefore be more affected by individual health or environmental conditions (e.g. season, year and site) than by sex.

Utility of health indices

All finch species showed within-species variation in residual CORT levels, CBG capacity, body condition, haematocrit or haemoglobin that could not be explained by seasonality, age or sex alone. It is possible that environmental variation between seasons, study areas and years may drive some of this unexplained variation in individual health indices. As shown by the differing pattern of stress- 95 responsiveness in Gouldian and Long-tailed finches monitored in the Yinberri Hills during the non-breeding seasons in 2007 and 2008, some bird populations show dramatic inter-annual variation in stress measures. Comparing the annual patterns of stress in breeding and non-breeding bird populations provides insight into how well birds are coping with a particular set of environmental conditions. This information could be especially helpful in determining the risk of population declines for groups or species found in different environments. Stress measures have been shown to predict the survival probability of individual Galapagos marine iguanas (Amblyrhynchus cristatus) (Romero & Wikelski 2001), and stress is likely to be a good survival predictor in birds. Stress levels also predict breeding success, as birds found to show evidence of chronic nutritional or psychological stress had lowered reproductive output compared to birds without high stress

(Silverin 1986; Kitaysky et al. 1999; Cyr & Romero 2006; Bonier et al. 2007; Buck et al. 2007). Assuming measures of stress can predict survival and breeding potential in individuals or populations, it is probable that stress measures can predict the probability of increase or decrease in animal populations. This assumption is a contentious topic in eco-physiology (Romero 2004; Breuner et al.

2008), but if valid, the finding that Gouldian finches have a consistent pattern of elevated CORT responses during the non-breeding season, while common grass- finch species do not, predicts that Gouldian finches are more at risk of lower survival during the non-breeding season. This also translates into a higher risk of population decline for this species.

96 Similar to stress response measures, seasonal variation in finch body condition changes were not always repeated between years or sites, suggesting changes in habitat quality between years could be a cause of body condition variation. These indices were found to be good predictors of survival and breeding potential in a number of bird species; however, factors such as season, social status or sex often had to be taken into account before condition measures were found to be significant predictors of survival (Piper & Wiley 1990; Magrath 1991; Gosler 1996;

Suorsa et al. 2004). If information on sex, social status and season can be collected on individuals concurrently with body condition measures, these indices have a high potential for use in studies comparing the health status of individuals or populations. However, further study of the predictive value of these indices in determining reproductive success and survival probability would be necessary to more thoroughly validate health measures as a surrogate to direct measures of breeding success and survival.

Interestingly, mean haematocrit values for each finch species did not vary consistently over the course of the study, suggesting there is more individual variation or consistent intra-population variation in this measure than seasonal, site or yearly variation. There were few significant age and sex differences in haematocrit in these finch species, signifying further that these measures could be useful in studies describing the effects of differences in individual condition.

However, current reviews of this index suggest that links between haematocrit, breeding potential and survival are often circumstantial, and that this measure is not easily interpreted without detailed information on factors likely to affect 97 haematocrit directly such as age, sex, nutrition, parasitism, elevation and energy demands (Fair et al. 2007).

The results of this study highlight the potential value of measures of stress responsiveness, body condition and haematocrit in quantifying bird population and individual responses to environmental quality: a much needed tool in conservation ecology research. Mean measures of CORT response and CBG capacity had high variation between seasons, while mean measures of body condition, haemoglobin and haematocrit had seasonal changes of a smaller magnitude. This suggests that stress measures are more likely to change relative to season, and could be useful in other studies of seasonality of physiology in birds. Body condition and haematological state did not change much relative to season in these bird species.

For example, mean muscle scores for all species ranged from 1.8 to 2.4, all of which are relatively high values. However individual values for birds within seasons, sites and years ranged more widely, suggesting these measures could be useful when comparing condition between populations and individuals. Health measures such as CORT and CBG may have limited use in other studies because of the large blood samples needed for analysis; however, almost all health indices are limited by relatively large sample sizes necessary to determine normal ranges of values for each species before inferences can be made about which individuals or populations are more or less ‘healthy.’ While body condition measures (e.g. muscle score) are easily interpreted as above or below a mean value for a species, measures such as stress responsiveness, immune status and responsiveness, plasma metabolite and oxidative stress indices are unstudied in most birds and 98 therefore a larger body of data must be collected so that a context for these indices can be created for each new study species.

Several health measures commonly used in avian studies were not measured in this study. Measures that monitor immune status, such as heterophil to lymphocyte (H/L) ratios, have two key advantages over CORT and CBG measures in that they use small amounts of blood (5 - 10 uL) and ratios do not increase quickly with capture stress (Davis et al. 2008). Their close relationship with CORT supports H/L ratios as a valid surrogate to measuring baseline plasma CORT as a measure of stress and measures of H/L ratios have also been linked with bird survival and reproductive potential (Gross & Siegel 1983; Friedl & Edler 2005;

Lobato et al. 2005; Kilgas et al. 2006b; Nadolski et al. 2006). However, H/L ratios can be difficult to interpret, as levels can be a reflection of both individual immunocompetence and/or the presence of infections (Norris & Evans 2000). This key disadvantage directed the decision to study finch stress responses directly in this study. The main advantage of using more dynamic challenges (stress response, immune response) as opposed to monitoring indices (body condition, haematocrit, H/L ratio and metabolite levels) is that response measures can not only provide insight into an individual’s current status but also provide information on that individual’s ability to cope with stressors (Saino et al. 1997; Birkhead et al.

1998; Romero & Wikelski 2001; Navarro et al. 2003; Romero 2004; Cichon &

Dubiec 2005; Moreno et al. 2005; Moller & Haussy 2007; Breuner et al. 2008).

Measures of immune responsiveness, such as injection of novel antigens or phytohaemagglutinin and subsequent measurement of humoral or cell-mediated 99 responses, are commonly used to quantify the immune response of individual birds

(Ots et al. 2001; Navarro et al. 2003). Experimentally induced high immune responses have been linked to decreased reproductive effort and survival potential

(Ilmonen et al. 2000; Ots et al. 2001), while naturally higher immunocompetence has been linked to better diet, body condition and sometimes higher reproductive effort and survival (Saino et al. 1997; Birkhead et al. 1998; Navarro et al. 2003;

Cichon & Dubiec 2005; Moreno et al. 2005; Moller & Haussy 2007). These measures were not used due to the invasiveness of the procedure (injection) and long handling times needed to monitor immune challenge (up to 12 hours).

Measures of monitoring innate immunity using blood samples to quantify bacteria killing ability of the plasma and measures of oxidative stress levels have recently been trialed in field studies, but are logistically and financially difficult to implement in most field situations (Matson et al. 2005; Matson et al. 2006; Cohen et al. 2007).

The possible insight these tools may provide in determining the health status of individuals suggest further validation of the links between indices and bird survival and breeding potential through captive and field based studies would be highly worthwhile.

Conclusion

The dramatic differences in the seasonal stress response patterns of the declining

Gouldian finch and common Long-tailed and Masked finches shown in this study highlight the usefulness of dynamic stress measures in determining species interaction with seasonal environments. Combined with parallel seasonal declines 100 in body condition, this information also helps determine a critical time of year for survival in a rare species. The results of this study suggest that the early non- breeding season is more physiologically stressful than peak breeding season in

Gouldian finches. This suggests that conservation-based monitoring studies for this endangered species should focus research efforts during the non-breeding season in order to better discern the main causes of finch stress. Further research on finch health during other seasons could assist in determining whether other times of the year are as stressful as the late dry/early wet non-breeding season, as well as help evaluate the utility of different indices in determining the health status of birds.

101 3.0 Chapter Three: The effects of land management on grass cover in tropical savanna, northern Australia

3.1 Introduction

Unlike other savannas around the world, Australia’s tropical savannas are relatively unfragmented and unmodified (Dyer et al. 2001; Bradstock et al.; Woinarski et al.

2007). Despite this landscape integrity, many native plants and animals are declining across Australian savannas (Bowman et al. 2001; Woinarski et al. 2001;

Franklin et al. 2005; Firth et al. 2010). Changes to fire regimes since the introduction of pastoralism and loss of traditional aboriginal burning practices are heavily implicated in these biodiversity declines (Crowley & Garnett 2000; Russell-

Smith et al. 2003b; Yates & Russell-Smith 2003). Central and northwestern savanna is now subject to more extensive, frequent and intense fires (Gill et al.

2000; Vigilante 2001; Russell-Smith et al. 2003b). The resulting changes to the structure and floristic composition of vegetation are likely to affect the shelter and food resources available to animals. For example, patches of unburnt old growth heath and Triodia spp. (spinifex) are the preferred habitat for disappearing species such as the Leichhardt's Grasshopper (Petasita ephippigera) and White-throated

Grasswren (Amytornis woodwardt) (Woinarski 1992a; Lowe 1995). Similarly, granivorous birds, small mammals and reptiles have declined across northern

Australia (Woinarski & Ash 2002; James 2003; Andersen et al. 2005; Franklin et al.

2005), because changed fire patterns and grazing have changed the distribution and productivity of grasses that are important food or habitat for these animals.

102 Overgrazing and fire suppression have resulted in the loss of key grasses important for the Golden-shouldered Parrot (Psephotus chrysopterygius) on Cape

York Peninsula, Queensland (Garnett 1992; Crowley & Garnett 1998).

Overgrazing during drought years during the past century has removed large areas of tall grass and allowed woody plants to take their place (Crowley & Garnett 1998;

Fensham & Fairfax 2003; Sharp & Whittaker 2003). Lower grass biomass has reduced fire frequency and intensity, further promoting woody plants by allowing them to grow unencumbered and further shade out the understorey, making the re- establishment of grasses unlikely (Crowley & Garnett 2000). In this way, the degradation of grasslands may have also caused the extirpation of the grass seed- eating Gouldian finch (Erythrura gouldiae) in Queensland. This species still occurs in scattered populations across the Northern Territory and in the Kimberley region of Western Australia where pastoralism has yet to reach the intensity of use found in parts of Queensland and the grass layer may be less degraded (O'Malley

2006b).

Though grazing and fire are similar in the way they periodically defoliate grasses, grasses tend respond to fire in species-specific ways. For example, perennial grasses do not generally benefit from burning during the wet season (Ash & McIvor

1998). However, the perennial grass Heteropogon contortus responds to fire with increased productivity in the first year after fire (Orr et al. 1997). Heteropogon contortus continues to grow particularly well if burnt every third year, but may decline in productivity if burnt too frequently. Other perennial grasses such as

103 Themeda triandra, respond poorly to any amount of grazing or fire (Walker et al.

1997a; Ash & Corfield 1998). Annual grasses are less productive when burnt during flowering or seed ripening, both of which often peaks during the wet season

(Andrew & Mott 1983; Lonsdale et al. 1998; Crowley & Garnett 1999). Thus the highest diversity and abundance of grasses are likely to be retained within a landscape that has a variable fire history, with the majority of burns being small, patchy and occurring during the dry season.

Similar to fire, grazing can decrease survivorship and productivity of many important savanna grasses (Ash et al. 1997; Letnic 2004). However, cattle are much more selective than fire, and have the potential to reduce the abundance and productivity of preferred grass species quickly, especially in heavy stocking situations (Sharp & Whittaker 2003). Cattle grazing can shift the grass layer species to a suite of grasses unpalatable to cattle within two wet seasons if grazing levels are not properly managed (Ash et al. 1997; Bastin et al. 2003). These effects are well studied for perennial grasses, but grazing effects on less palatable annual grasses and forbs are poorly understood in comparison (Letnic 2004).

This study reports the changes in productivity and abundance of annual grasses, perennial grasses and forbs in response to grazing and fire on three properties with different land tenures (pastoral, aboriginal, Defence) in the Northern Territory.

Although each land tenure was represented by a single property in this study, the sampling within each property captures variation that is usually associated with site effects because the properties are very large (several thousand square kilometres). 104 The general aim of this study is to provide insight into the effects of the Northern

Territory’s three common types of land management on the grass layer.

3.2 Methods

Study site descriptions

This study was conducted on three properties characterized by a predominantly native grass layer in the Victoria River District and Top End of the Northern

Territory, Australia (Figure 1.1a). Bradshaw Field Training Area (Bradshaw FTA, latitude -15.5S, longitude 130.4E; area of 4050 km2) is a Defence property with little to no cattle or other introduced herbivores present and therefore no grazing by exotic herbivores. Kangaroo grazing levels are low. The property is well protected from unplanned fires from the south and west by the Victoria River and Timor Sea.

Some prescribed burning is conducted annually in order to create fire breaks for protection of infrastructure, and also to safeguard biodiversity from the impacts of unplanned fires originating outside of the north-eastern property boundary.

Delamere Station (latitude -15.7S, longitude 131.6E, area of 4000 km2) is a pastoral lease owned by the Australian Agricultural Company (AACo) that supports a breeding population of Brahman cattle. The property has a low fire frequency as a result of a fire management plan that uses little prescribed burning and the property experiences few unplanned fires. Delamere is shielded from ignitions to the east and south by a Defence property and similarly managed cattle station. 105 Gregory National Park lies to the west and has an active prescribed burning program. The section of the park that borders Delamere is the remote southern side of the Victoria River, which acts as an effective fire break. Occasionally, parts of Delamere have high stocking rates (3.5 head/km2) and stock are moved between paddocks some years or lured to sections of paddocks (effectively spelling other sections within these paddocks) using diet supplement licks in other years. Paddocks are large (over 100 km2), therefore stocking rates were not a reliable measure of cattle density at a local scale. Many first-year cattle are moved to other AACo properties or sold by the end of the dry season. Kangaroos (Euros and smaller wallaby species) are numerous, possibly due to dingo control and an increased availability of water-points, thus rocky escarpment areas usually free of cattle are often under moderate to high grazing pressure from kangaroos.

Yinberri Hills (latitude -14.1S, longitude 132.0E; area of 200 km2) is aboriginal land jointly managed by the Jawoyn Association, National Parks and the Vista Gold mining company. No cattle are stocked in the Yinberri Hills, and only low numbers of introduced herbivores, mostly horses and pigs, are found on the property.

Kangaroos are moderately common and are likely to exert only a small amount of grazing pressure on the property. Frequent patchy fires burn extensive areas each year due to a mix of prescribed burning (for protection of mining infrastructure and general land management) by National Parks and Jawoyn Association land managers, and massive unplanned fires that originate on neighboring properties

(Nitmiluk National Park rangers, personal communication). Yinberri Hills is only 60 km north of Katherine and is bordered on the west by the Stuart Highway (common 106 ignition source). Pine Creek is 50 km to the north and to the west is Nitmiluk

National Park. Unplanned fires enter Yinberri Hills across all borders at almost any time of year.

Study plots and monitoring

This study was part of a larger study investigating the impact of land management on granivorous birds; thus, survey locations were selected based on bird feeding and nesting activity in order to gauge the effects of fire and grazing on areas important to grass-dependent birds. On each property, twenty to twenty-eight circular plots of 100 m diameter were selected in representative breeding and foraging habitats used by small granivorous birds. On each property, half of the plots were situated in rocky hill or escarpment habitat dominated by Triodia spp. grasses and a sparse canopy of smooth barked gums (Eucalyptus brevifolia or E. tintinnans), and half were placed in low-land or floodplain habitat dominated by a variety of perennial and annual grasses and a canopy of box (E. tectifica) or bloodwoods. Shrubs and small trees make up a very small proportion of the vegetation, except in drainage areas and in patchy Calytrix sp. thickets. Soils in high rocky country tended to be well-drained sandy soils, while floodplains were dominated by sandy-clay or red-clay soils. Bradshaw and Yinberri Hills plots were

20% clay soils, while 60% of soils on Delamere were red-clay. Plots were located within 1 km of unsealed vehicle tracks but 5 to 10 km from any sealed roads. Each plot was visited twice a year, once in the early dry season (May or June 2007,

2008 and 2009), and again in the late dry/early wet season (between October and 107 November, 2008 and 2009). All properties were visited within a period of three weeks during these two sampling periods of the dry, non-growing season. In the savanna region of northern Australia, there is typically no rainfall between May and

December, thus plants are dormant and do not typically grow or recover from defoliation during the two sampling periods each year (Ridpath 1985; Ash & McIvor

1998; Woinarski et al. 2007). Each plot was surveyed a total of five times on

Yinberri Hills and Delamere. Plots were only visited four times on Bradshaw FTA due to access restrictions (no sample for early 2007).

Estimation of grazing and burning levels

The amount of each plot showing evidence of ground layer defoliation due to grazing or fire was estimated visually on a scale of 0 to 4. The score for each plot was estimated by walking a transect through the centre of the plot and assigning a score of 0 for no to 5% defoliation, a score of 1 for 5 - 25% defoliation, 2 for 25 -

50% defoliation, 3 for 50 - 75% defoliation, and 4 for 75 – 100% defoliation. This rough method of grazing impact estimation was chosen because the large size of paddocks caused stocking rates to be an inaccurate measure of cattle density estimates on a local scale. Grazing defoliation was recorded when grass tussocks showed trimming or complete removal of grass blades by herbivores. Defoliation due to burning is easily identified in northern Australian savanna, as black or gray ash is retained on the ground and on grass tussock bases after burning. This burning evidence is not washed away until the following wet season (late

December or January). Separating the defoliation effects of fire and herbivory was 108 straightforward in this study because fires were actively suppressed on Delamere, the only property with significant numbers of large exotic herbivores (cattle).

Burning and grazing occurring on plots before sampling during the early dry season was recorded again each non-breeding season, as grass does not grow or recover from defoliation between these two sampling periods. Thus the area burnt and grazing impact of each plot recorded during the non-breeding season is essentially a total for the entire year. Therefore, this study uses the late dry season measure for each year when describing differences between properties, as opposed to plotting each early dry and late dry season separately. Differences in cover estimates between early and late seasons are only used to calculate the effect of grazing on the grass cover at Delamere within a year and the effect of fire on grass cover on Yinberri and Bradshaw grass cover within a year.

Rainfall data

Rainfall data from Australian Bureau of Meteorology records was used to describe wet season rainfall totals for each property. Weather stations representative of each site were within 20 km of study plots and rainfall totals were collected daily by remote sensors or station staff.

109 Estimation of percent ground cover

The percentage of each 100 m plot covered by annual grass, perennial grass, bare soil, leaf litter, forbs and logs was estimated using a non-random stepping transect.

At each plot, an observer took 100 haphazard steps starting at the centre of the plot and contained within the plot area, recording the ground cover type under each footfall; the total footfalls for each ground cover type were calculated as a percentage out of 100 footfalls (Evans & Love 1957). Ground cover types included annual grass, perennial grass, forb, bare soil, litter, rock, and logs.

Annual grasses were defined as any grass species that completes its life cycle within one year e.g. Sarga intrans and Schizachyrium fragile. Perennial grasses were defined as any grass species that takes more than one year to complete their life cycle. Common perennials on study plots included Triodia bitextura and

Heteropogon contortus.

Measures of grass productivity

The seed production of individual plants has been shown to increase with both plant height and width in Alloteropsis semialata, and is likely to also influence the seeding of unstudied perennial grasses (Crowley & Garnett 2001). In order to give a rough estimate of overall grass productivity at each plot, the mean height of all grasses within each plot was estimated visually to the nearest 0.5 m. Though subjective, this measue was estimated by the same observer each season and 110 was intended only to document large differences in overall grass height between properties and years. A more detailed estimate of the productivity of dominant perennial grass species (identified by similarity analysis; Table 4) was estimated separately each season. The maximum crown width of 15 randomly selected individual clumps of each dominant perennial grass species within the plot was measured to the nearest cm. The maximum crown width was estimated by using a tape measure to determine the maximum above-ground distance between tillers on each individual plant. The mean crown width for each species of grass was calculated as an average of the 15 individuals measured in each plot. These mean widths were used to represent the productivity of each grass species for each plot.

Statistical analysis

Mixed design ANOVA was used to analyze differences in grazing, burning, ground cover and grass productivity among years (repeated measures variable) and among tenures (between group variable; PASW Statistics SPSS). When repeated measures data violated the assumption of sphericity, degrees of freedom were corrected using Greenhouse-Geisser estimates of sphericity. Pairwise comparisons were used to determine the direction of significant differences between tenures and between years within tenures. All comparisons were corrected using Bonferroni adjustment. Simple regression analysis was used to examine correlations between grass and forb measures within properties and the grazing and burning measures recorded on the properties. Regression analysis

111 was also used to examine changes in grass cover between the early and late dry seasons or between years within properties.

Similarity analysis (SIMPER) of grass species cover on each property was used to determine the relative contribution of each grass species to the average Bray-

Curtis dissimilarity between properties (Primer 6.1.10, Primer-E Hd. 2007). This analysis determines which species contribute the most to the dissimilarities amongst pairs of plots within each tenure and between years based on Bray-Curtis

Indices. It also calculates which species contribute most to Bray-Curtis Indices of

Dissimilarity between tenures and years. These dominant grass species (those identified to constitute 90% of dissimilarities) were then assumed to have the most influence on differences in total perennial grass cover between properties, and productivity estimates (crown measures) of these species were then analyzed using mixed design ANOVA (PASW Statistics SPSS).

3.3 Results

Site specific grazing, burning and rainfall levels

Grazing impact measures differed between tenures, but the extent of differentiation was dependent on year (interaction F3.6, 201 = 14.42, P < 0.0001). Grazing impacts on Delamere were significantly higher than grazing impacts on other properties during all years (effect of tenure F2, 67 = 44.59, P < 0.0001). On Delamere, grazing levels were higher during 2007 compared to other years (F2, 66 = 37.99, P < 0.0001; 112 Figure 3.1). The effect of tenure on the area burnt was also dependent on year

(F3.4, 201 = 3.80, P = 0.009). Burning impacts were higher in the Yinberri Hills compared to Delamere in all years and similarly high on Bradshaw in 2007 and

2009 (Figure 3.1). Rainfall totals were highest in the Yinberri Hills all years, and usually lowest on Delamere (Table 3.1). The wet season rainfall for 2006 - 07 was below average for all properties, and longer in duration compared to other wet seasons. In subsequent years, properties tended to have higher rainfall, but a shorter wet season.

113

d>y>b d>y>b d>y>b 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0

2007 2008 2009 Grazing Impact on Scale Grazing of Impact Scale on 0 to 4 Year

3.5 y>b>d Bradshaw FTA 3.0 Yinberri Hills

Delamere Station 2.5

2.0

1.5

1.0 0.5

Area Burnt on scale of 0 to 4 to 0 Areascaleof onBurnt 0.0 2007 2008 2009 Year

Figure 3.1: Grazing and burning measure levels on each property during 2007,

2008 and 2009. Bars are means +/- SE, and significant differences between properties with values above 0 are signified with abbreviations for sites (P < 0.05).

114 Table 3.1: Rainfall totals and length of wet season based on data for properties from the Australian Bureau of Meteorology.

Rainfall Totals (mm)

Length of Wet Bradshaw Delamere Yinberri

Season FTA Station Hills

Oct 06 - June 07 680 600 1100

Oct 07 - March 08 840 980 1370

Oct 08 - March 09 1290 615 1320

Average Annual (1981-2011) (1997-2011) (2003-2011)

Rainfall 1000 900 1200

Cover patterns by year and land tenure

Overall grass cover and height differed between properties inconsistently among years (cover F4, 201 = 7.53, P > 0.0001; height F4, 201 = 13.13, P > 0.0001). Detailed analysis shows significantly higher grass cover on Bradshaw compared to other properties in 2007, no differences between cover on properties in 2008, and significantly lower cover on Bradshaw compared to Delamere (Yinberri Hills intermediate between the two) in 2009 (Figure 3.2). Grass height was significantly higher on Bradshaw and the Yinberri Hills compared to Delamere in 2007, no differences were seen between properties in 2008, and Delamere was significantly

115 higher than the Yinberri Hills (Bradshaw intermediate) in 2009 (Figure 3.3). Within property analysis across years showed that cover and overall productivity (ie. height) of grasses was lowest on Delamere in 2007 (Figures 3.2, 3.3). Yinberri

Hills grass height was higher than average in 2007, while overall grass cover did not differ by year on this property (Figures 3.2, 3.3). Bradshaw grass cover estimates were slightly lower than average in 2009 (Figure 3.2), but grass height did not significantly differ between years on this property (Figure 3.3). Forb cover differences between properties was weakly dependent on year (F3.6, 201 = 14.42, P

< 0.0001). Forb cover was consistently higher on Delamere compared to other properties (F2, 201 = 44.59, P < 0.0001), with peak levels in 2009 and minimum levels in 2007 on this property (Figure 3.4).

116 70 b>d=y b

10 Grass Cover Grass % of plots 0 2007 2008 2009 Year

Figure 3.2: Mean percent cover of grasses on plots by property and year. Bars are means +1 standard error, significant differences between properties are designated with abbreviations for sites (P < 0.05).

1.0 y>b>d d>y Bradshaw FTA Delamere Station Yinberri Hills

0.5

Mean Height Mean Estimate of 0.0

Grass Grass Cover (to nearest 0.5 m) 2007 2008 2009 Year

Figure 3.3: Mean height of grasses on plots by property and year. Means were calculated using estimation of grass height to the nearest 0.5m for each plot, which were then pooled and averaged for the property each year. Bars are means +1 standard error, significant differences between properties are designated with abbreviations for sites (P < 0.05). 117 d>y d>y>b 12 d>y Bradshaw FTA 10 Delamere Station 8 Yinberri Hills

6

4

2 Forb Cover % of plots 0 2007 2008 2009 Year

Figure 3.4: Mean percent forb cover of plots by property and year. Bars are means

+1 standard error, significant differences between properties are designated with abbreviations for sites (P < 0.05).

Cover patterns within tenures: effects of grazing and burning

Within property regression analysis identified a significant negative effect of grazing on perennial grass cover for all years at Delamere (F1, 70 = 50.40, P <

0.0001) and negative effects between all grasses and forbs and the area burnt at

Yinberri Hills (perennial grass F1, 76 = 25.65, P < 0.0001, annual grass F1, 76 =

39.44, P < 0.0001, and forbs F1, 76 = 12.51, P = 0.0007). At Bradshaw, analysis showed significant negative effects of burning on grass cover, (perennial grass F1,

118 58 = 56.91, P < 0.0001, annual grass F1, 58 = 22.92, P < 0.0001) but not forb cover

(F1, 58 = 2.58, P = 0.11). At Delamere only, there was a significant positive relationship between grazing impact and percent cover of forbs (F1, 70 = 8.99, P =

0.004).

Annual grass cover

Annual grass cover differences between properties were dependent on year (F2.9,

200 = 4.10, P = 0.004). Annual cover was significantly lower on Delamere compared to Bradshaw and the Yinberri Hills in 2007 and lower than Bradshaw only in 2008 (Figure 3.5). Regression analysis comparing annual cover decreases between the early dry and late dry sampling periods each year showed highly significant decreases in annual grass cover between the two seasons in 2007 on

Delamere (F1, 42 = 15.34, P = 0.0003) and the Yinberri Hills (F1, 50 = 19.95, P <

0.0001). Early dry season data was not collected for Bradshaw in 2007 due to access restrictions. No property showed significant decreases in annual cover between early and late sampling in 2008 (Bradshaw F1, 38 = 0.62, P = 0.44,

Delamere F1, 46 = 1.98, P = 0.16, Yinberri Hills F1, 50 = 0.30, P = 0.72).

119

30 b=y>d b=y>d b=d

20 Yinberri Hills

10

0 2007 2008 2009 annualgrass Mean% cover Year

Figure 3.5: Mean percent annual grass cover on plots by property and year. Bars are means +1 standard error, significant differences between properties are designated with abbreviations for sites (P < 0.05).

120 Perennial grass cover and productivity

Differences in perennial grass cover between properties depended on year

(interaction F4, 201 = 8.66, P < 0.0001). Perennial grass cover was significantly lower in the Yinberri Hills in 2008 compared to Delamere, and both Yinberri Hills and Bradshaw had lower perennial cover than Delamere in 2009 (Figure 3.6).

Though Yinberri Hills perennial grass cover tended to be lower than other properties all years, mean perennial cover estimates on the property were not significantly different between years (Figure 3.6). At Delamere, perennial grass cover did change between years, and was significantly lower during 2007 compared to other years (Figure 3.6). Regression analysis comparing perennial cover decreases between the early dry and late dry sampling periods each year showed statistically significant decreases in the perennial cover only at Delamere, and only in 2007 (F1, 42 = 7.07, P = 0.01).

Similarity analysis (SIMPER) of cover estimates for all grass species showed that the characteristic species which constituted 90% of the dissimilarity among grass cover on properties included 14 genera of grasses (listed in Table 3.2). The species composition of dominant grasses on Delamere was slightly different from the grass compesition on Bradshaw and the Yinberri Hills. These differences were largely due to increased cover of the annual Sarga spp. and a high cover of the perennial Triodia spp. on Bradshaw and Yinberri (Mean cover estimates for each grass species on properties are presented in Table 3.2). At Delamere, the main characteristic grasses (those that constituted 90% of the cover estimate 121 dissimilarity) were perennial Triodia spp., Heteropogon spp. and Sehima nervosum. On all properties, the main annual grass that was deemed a characteristic species using similarity analysis was Sarga spp. (Table 3.2). The effect of tenure on the productivity of all dominant characteristic perennial species depended on year (interaction Triodia spp., F2, 57 = 3.00, P = 0.03, C. fallax F2, 50 =

6.36, P < 0.0001; Eriachne spp. F2, 57 = 6.36, P < 0.0001, Heteropogon spp. F2, 50 =

4.04, P = 0.03, T. triandra , 57 = 3.00, P = 0.03). Of the dominant perennial grasses, all genera except Triodia spp. and T. triandra had lowest productivity on Delamere compared to other properties during 2007 (productivity estimates are shown in

Figure 3.7). In 2008, all genera except Heteropogon spp. and T. triandra had low productivity on the Yinberri Hills compared to Bradshaw (Figure 3.7). All three dominant grass species on Bradshaw had low productivity measures in 2009 compared to other years on this property (Figure 3.7).

122

50 b=d>y b=y

30

20

10

0 2007 2008 2009

Avg % PerennialGrass % Avg Cover plots on Year

Figure 3.6: Mean percent perennial grass cover of plots by property and year.

Bars are means +1 standard error, significant differences between properties are designated with abbreviations for sites (P < 0.05).

123

Table 3.2: Characteristic annual and perennial grass species as identified by similarity analysis. Average percent

cover is mean cover of each species out of the total grass cover for each property and all years (Total grass cover

Figure 3.2). Cattle and wildlife use of grasses is based on published studies (Garnett & Crowley 1994; Ash et al.

1997; Crowley & Garnett 2001; Dostine & Franklin 2002).

Average Percent Cover Growth Selectively Seeds eaten Grass Genus species Type Bradshaw Delamere Yinberri Hills Grazed? by Wildlife? Sarga intrans, timorense Annual 19 3 31 No Often Schizachyrium fragile, pachyarthron Annual 6 2 5 No Often Triodia bitextura, pungens Perennial 44 23 20 No Often Heteropogon contortus, triticeus Perennial 0 17 8 Yes Rarely Chrysopogon fallax Perennial 9 4 6 Yes Often Eriachne obtusa, glauca Perennial 19 4 12 No Often Themeda triandra Perennial 0 4 1 Yes Rarely Sehima nervosum Perennial 0 19 0 Yes No Sarga plumosum Perennial 0 2 7 Yes Often Aristita spp. Perennial 1 4 2 No Rarely Cymbopogon bombycinus Perennial 0 3 0 Yes No Alloteropsis semialata Perennial 1 0 3 Yes Often Eulalia aurea Perennial 0 2 0 Yes No Rottoboellia cochinchinensis Perennial 0 3 0 No Rarely

124 Triodia spp. Chrysopogon fallax 30 b>y bd b>d=y

20 20

10 10 Mean crown diameterMean (cm) 0 crown diameterMean (cm) 0 2007 2008 2009 2007 2008 2009 Year Year Eriachne spp. 30 b=y>d b>d=y Bradshaw FTA Delamere Station 20 Yinberri Hills

10

Mean crown diameterMean (cm) 0 2007 2008 2009 Year

Heteropogon spp. Themeda triandra dy 30 40

30 20 20 10 10

Mean crown diameterMean (cm) 0 0 2007 2008 2009 crown diameter Mean (cm) 2007 2008 2009 Year Year

Figure 3.7: Mean perennial grass productivity (mean crown diameter in cm) within plots by property and year. Graphs represent the productivity of a) Triodia spp., b)

Chrysopogon fallax, c) Eriachne spp., d) Heteropogon spp., and e) Themeda triandra. Bars are means +1 standard error, significant differences between properties are designated with abbreviations for sites (P < 0.05).

125 3.4 Discussion

General ground cover differences between properties

Both overall grass cover and grass height were higher on Bradshaw compared to

Delamere in 2007. It is very likely that Delamere station had particularly low grass cover and productivity compared to other properties that year because of the high grazing levels recorded. Interestingly, these trends seem to be reversed in 2009, when Bradshaw had unusually high burning levels and Delamere had lower grazing levels. These general differences in grass cover and productivity between two study properties may simply demonstrate the loss of grass in response to grazing and fire in these areas. Results of detailed analysis show that grazing and fire affect different components of the grass layer on different tenures.

General ground cover characteristics within properties

At Delamere, there seems to be retention of annual grasses and forbs, and loss of perennial grasses in areas of higher grazing pressure. This is shown by both the positive relationship found between grazing levels and annual grass and forb cover and the negative relationship found between grazing and perennial grass cover on this property. It is likely that perennials were being selectively grazed and thus there was more open ground available for use by annual grasses and forbs. The loss of perennial grass cover in preference to annual grasses and forbs is a well described effect of grazing (Landsberg et al. 1997; Ash & Corfield 1998; Bastin et 126 al. 2003), and is a sign of pasture degradation (Ash et al. 1997; Fensham & Skull

1999; James et al. 1999; Sharp & Whittaker 2003).

At Yinberri Hills, where grazing levels were low, differences between plots were opposite to those found on Delamere: plots with higher grazing levels were characterized by lower annual grass and forb cover and higher perennial grass cover and productivity. Healthy or high-quality pasture is often described as having high perennial grass cover and seasonal growth (Ash et al. 1997; Landsberg et al.

1997; Ash & Corfield 1998). Thus, the association of grazing with these characteristics at Yinberri Hills is most likely a result of a grazing animal preference to feed at high-quality sites for grazers. The grazing animals present simply grazed where perennial grasses were most abundant and productive, without large-scale effects on grass abundance and productivity at these sites.

Fires were very rare on Delamere. Burning at Yinberri Hills and Bradshaw was associated with decreased annual grass, perennial grass and forb cover at survey plots within the same year. These results illustrate the non-selective pressure of burning. Fire removed large areas of cover, similar to heavy grazing, but fire was not associated with increased forb abundance.

127 Differences in specific ground cover types between properties

Annual Grasses

On all properties, annual grasses (particularly annual Sarga spp.) had lower coverage during seasons of higher grazing impact scores or burning scores on study plots in comparison to seasons with lower grazing and less burning on plots.

However, cattle do not graze on annual Sarga spp. (Andrew & Mott 1983; Lonsdale et al. 1998). Cattle did occur in moderate to high densities in areas where this grass was common on Delamere (personal observation). The lower coverage of

Sarga spp. on Delamere compared to other properties could have been partially due to the effect of trampling by cattle, which has also been shown to cause soil compaction and sealing (Mott et al. 1979). Mott suggested that both compaction and sealing limit seed penetration and germination during subsequent years, and could theoretically limit the productivity of these annual grasses (1979).

Alternatively, high incidence of nutrient rich red clay soil on Delamere may preferentially benefit the productivity of perennial grasses and forbs, which could compete with and limit the productivity of annual grasses (Ash et al. 1997).

However, this study did not measure soil properties; therefore further research is required to identify factors other than grazing that could be causing lowered annual grass cover on Delamere.

In contrast, fire does not contribute to compaction and can allow annual grasses to increase coverage by defoliating competing perennial grasses (Russell-Smith et al. 128 2003a). This may partially explain why Sarga spp. cover was consistently higher at

Yinberri Hills, which tended to have lower perennial grass cover compared to other properties. A relatively large proportion of this property was burnt each year before the early dry season, which removed large areas of both annual and perennial grass cover. However, because the soil at this site was not compacted by cattle

(this property has almost no feral cattle), a higher proportion of annual grass seeds may be able to germinate during the wet season and set seed before burning in the early dry season in comparison to areas that do have compacted soils (Mott et al.

1979). Annual grass seed stores may be reduced by burns, however, they usually occur in such high abundance that only burning during the growing season has been shown to reduce the viability of annual grass populations (Andrew & Mott

1983; McIvor & Gardner 1991; Crowley & Garnett 2000). In contrast, most perennial grasses benefit from years without burning by increasingly shading out annual grasses and forbs, and accumulating more energy stores for use in flowering and seed production (Walker et al. 1997a; Crowley & Garnett 2001;

Bennett et al. 2002). Thus, annual grasses proliferate with frequent burning of ground cover. This scenario is often used to explain why annual Sarga spp. are now dominating parts of the northern Australian savanna which are prone to frequent late dry season fires (Yibarbuk et al. 2001; Russell-Smith et al. 2003a).

The observation that annual grasses maintain a high abundance in the frequently burnt Yinberri Hills fits this explanation.

129 Perennial Grasses

Perennial grass cover tended to be lower at the frequently burnt Yinberri Hills and similarly low on Bradshaw when a large propotion of plot areas were burnt in 2009.

These results show how fire effectively defoliates perennial grass. Despite the higher levels of burning in the Yinberri Hills, this site maintained a relatively stable percent of perennial grass cover on survey plots over three years. The Yinberri

Hills also had high amounts of rainfall all three years. The high rainfall in the

Yinberri Hills is one possible explanation for the persistence of grass cover on this property. An alternate explanation is that grass species surveyed in the Yinberri

Hills were not impacted equally each year due to most fires occurring on sites dominated by annual grasses. An exception to this trend occurred in 2008 in the

Yinberri Hills, when all perennial grasses except Heteropogon spp. and T. triandra had their lowest productivity. During this year, the largest areas of this property were burnt. These two genera had relatively high productivity in 2008 because T. triandra sites were not burnt until 2009, and Heteropogon spp. sites were mostly burnt in 2007, after which this species seemed to have increased productivity in

2008. Rangeland studies have shown that Heteropogon spp. sometimes has increased growth the first year after burning (Orr et al. 1997; Williams et al. 2003).

Interestingly, C. fallax productivity increased dramatically with a slight decrease in burning in 2009. This result could suggest that C. fallax may benefit from occasional burning with increased productivity the year after fire, similar to

Heteropogon spp. Thus, because different sections of the Yinberri Hills were burnt each year, total perennial grass cover seemed stable. These results provide 130 further evidence of the species-specific responses of grasses to fire, and the possible benefits to maintaining perennial grass cover by patchy burning regimes.

Though perennial grass cover was often high on Delamere compared to other properties, these results do not necessarily mean that grazing has less of an effect on these grasses than fire. For example, perennials were found to have their lowest coverage on Delamere when grazing levels were highest in the late dry

2007, and nearly all dominant perennial grasses had lower crown size estimates on Delamere compared to the Yinberri Hills. This suggests that there are more individual perennial grass clumps on Delamere; however, each individual plant is more likely to be grazed and have a resulting smaller size than those surveyed on other properties. Delamere has lower average rainfall in comparison to other properties. It is possible that Delamere’s low rainfall partially explains the lower crown size of perennial grasses on this property. However, the major drop in

Delamere’s perennial grass cover in 2007 happened between the early and late dry season when grazing, not rainfall, is the limiting factor on grass clump crown size.

Despite small crown size (which implies low productivity) of perennial grasses on

Delamere Station, perennial grass cover was retained during the three year study.

While nearly all northern Australian soil types are low in nitrogen and phosphorus, red clay and cracking clay soils retain more nutrients (Ash et al. 1997). Much of the Yinberri Hills and Bradshaw FTA was characterized by sandy soils likely to have poor nutrient content and retention compared to areas on Delamere which 131 were characterized by red clay soils. High nutrient clay soils may facilitate high survivorship and resilience of plants on sections of Delamere despite sometimes high grazing pressure.

Implications for savanna conservation management

Lowered annual and perennial grass cover and productivity in response to heavy grazing and frequent fire has negative implications for a wide range of native flora and fauna (Burbidge & McKenzie 1989; Garnett 1992; Landsberg et al. 1997;

Woinarski & Ash 2002; Franklin et al. 2005). Threatened populations of parrots and finches are known to feed on Sarga, Triodia, Chrysopogon, Eriachne, and

Alloteropsis spp. grass seeds in preference to other genus of grass seeds (Garnett

& Crowley 1994; Lewis 2007). These prefered grasses have been shown to have lower annual growth and potentially produce less seed in response to grazing and frequent fire (Mott & Andrew 1985; Ash & Corfield 1998; Crowley & Garnett 2001;

Lewis 2007). Lowered grass productivity and cover could mean fewer seed resources for granivores such as the endangered Gouldian finch and Golden- shouldered parrot, and unsuitable habitat for many declining small mammals and reptiles (Burbidge & McKenzie 1989; Garnett 1992; Landsberg et al. 1997;

Woinarski & Ash 2002; Franklin et al. 2005). If biodiversity in Australia’s tropical savannas is to be conserved, the impacts of grazing and fire on ground cover need to be minimized where sensitive native species are still found.

132 The results of this study suggest some variability in grass cover could be due to grazing and fire. Due to the differences in rainfall, grass species assemblages and soil types between properties, this study cannot confidently attribute differences in grass productivity to grazing or fire alone. Also, several of the measurements used in this study are simple estimates of grazing and burning pressures on grasses and more detailed research is needed to accurately define the responses of important grass species to grazing and fire. Ideally, this would be done using an experimental design incorporating specific treatments of fire and grazing on vegetation survey plots of similar soil, rainfall and grass species composition, which could then be compared to non-treated control plots. The possibility of some grass variability being controlled by grazing or fire implies there is potential for land managers to limit declines in grass cover and productivity by manipulating the amount of grazing and fire on their property. Research has suggested managers can increase grass productivity by using techniques such as spelling pastures during wet seasons and leaving patches of habitat unburnt for longer than 2 years

(Ash et al. 1997; Crowley & Garnett 2001; Bastin et al. 2003). Modeling trade-offs between grazing and spelling have suggested that the benefits of long-term increases in grass cover to cattle production and savanna biodiversity are likely to outweigh short-term profits made by over-grazing or burning of much of Australia’s tropical savanna (Ash et al. 2004). The challenge for today’s conservationists is communicating these values to pastoralists and providing effective incentives for managers to use sustainable practices.

133 4.0 Chapter Four: Variation in the health of declining and non-declining

Australian savanna finch populations in relation to land tenure

4.1 Introduction

Current research suggests that the impacts of introduced herbivores and altered fire patterns are causing detrimental change to savanna biota in tropical Australia, as seen in dramatic declines in a number of native species (Russell-Smith et al.

1998; Franklin 1999; Woinarski et al. 2001; Andersen et al. 2005; Firth et al. 2010).

Granivorous birds are an example of a guild affected by changes in savanna land management. Surveys have detected substantial declines in the abundance and distribution of 11 species of granivores, half of which are grass-finches (Franklin

1999). Declines in grass-finches occurred first in the eastern savannas and then tracked the advance of pastoralism westwards across the continent (Franklin 1999;

Franklin et al. 2005). Grazing and frequent fire can lower the abundance and seed production of annual and perennial grasses important in finch diets (Ash et al.

1997; Ash & Corfield 1998; Crowley & Garnett 2001; Sharp & Whittaker 2003).

This has led to speculation that the impacts of introduced herbivores and changed fire regimes have caused finch declines (Tidemann 1989; Dostine & Franklin 2002;

Franklin et al. 2005).

The nationally endangered Gouldian finch (Erythrura gouldiae) mainly feeds on annual grass seeds in the dry season and perennial grass seeds in the wet season

(Garnett & Crowley 2000; O'Malley 2006a). This species has a highly seasonal 134 lifestyle, with birds breeding through much of the mid-wet and early dry season and moulting over a short period during the late dry season. Gouldian finch populations are currently restricted to areas characterized by high annual rainfall, high topographic variation, and areas with a shorter history of grazing (Franklin et al.

2005; O'Malley 2006b).

However, several species that share the same habitat with Gouldian finches, such as the Long-tailed finch (Poephila acuticauda) and Masked finch (P. personata) are not reported as having suffered range contractions or population declines. These species may be less sensitive to changes in grass because they can supplement their diets with insects and a wider range of seeds, and have more flexible breeding and moulting strategies than Gouldian finches (Tidemann & Woinarski

1994; Franklin et al. 1998; Dostine & Franklin 2002).

There is little empirical evidence for the mechanism of declines in Gouldian finches; thus, we are left with few insights into how land management could reverse these trends. At a phenomenological level, attempts have been made to describe Gouldian finch population responses to fire by monitoring numbers of birds visiting limited water sources at the end of the dry season each year.

However, waterhole counts do not typically generate useful population trend data, due to the unpredictable ranging behaviour of these finches (O'Malley 2006b, a;

Price et al. In preparation). Individual Gouldian finches often have home ranges of over 200 km2, and entire populations occasionally move large distances,

135 presumably in response to regional variation in climate and food resources (Legge

2009; Lewis 2007).

To circumvent the difficulties involved in attempting to monitor the abundance of nomadic species, measures of health can be used as alternate indicators of population stability or change. The relative number of unhealthy or stressed animals can indicate a population’s probability of decline, as sick or chronically stressed individuals are less likely to survive and breed (Romero & Wikelski 2001;

Friedl & Edler 2005; Cyr & Romero 2006; Kilgas et al. 2006b). Body condition indices such as mass, muscle and fat scores are relatively easy to obtain and are generally found to be good predictors of survival (Lundberg 1985; Magrath 1991;

Houston & McNamara 1993; Brown 1996; Merino et al. 2001; Stevenson & Woods

2006). The physiological correlates of haematocrit (the percent of red-blood cells to plasma) makes this index ideal for the study of health variation in birds.

Clinically ill birds have below average haematocrit levels, a symptom of anaemia and nutritional stress (Sturkie 1986; Averbeck 1992). Haematocrit levels have been linked to individual survival (Dubiec & Cichon 2001), breeding success

(Andersson & Gustafsson 1995), and higher levels have commonly been correlated with superior body condition, elevated testosterone, immunological health, and lack of ectoparasites (Dawson & Bortolotti 1997b; Potti et al. 1999;

O'Brian et al. 2001; Sergent et al. 2004; Dudaniec et al. 2006; Cuervo et al. 2007).

Physiological ecologists have also found that differences between basal and stress-induced plasma levels of corticosterone (CORT; the main avian stress hormone) can be reliable predictors of survival or lifetime reproductive potential for 136 individuals and populations (Romero & Wikelski 2001; Blount et al. 2003).

Increases in stress levels or response often track decreases in body condition

(Kitaysky et al. 1999; Pravosudov et al. 2001; Lynn et al. 2003a; Ouillfeldt et al.

2004; Kern et al. 2005). These measures of individual condition, haematological state and stress have been used successfully to predict the likelihood of bird survival and breeding success, as well as explaining the physiological consequences of habitat disturbance (Suorsa et al. 2004; Kilgas et al. 2006b;

Nadolski et al. 2006; Stevenson & Woods 2006; Bonier et al. 2007). Monitoring when and where unhealthy birds are found potentially offers insight into the process of population decline.

This study tests the assumption that the Gouldian finch is sensitive to the environmental effects of grazing and fire by monitoring health indices of populations in three land tenures with differing impacts of grazing and fire. Based on current knowledge, we predict that populations of declining Gouldian finches living in areas prone to grazing or frequent intense fires are more likely to show symptoms of poor health compared to populations living in areas with lower levels of these disturbances. The health of two sympatric, non-declining finch species was also monitored to determine whether the patterns of population health are unique to Gouldian finches, or characterize a more general response shared by all finch species.

137 4.2 Methods

Study species

The Gouldian finch, Long-tailed finch and Masked finch are all endemic to northern

Australia’s tropical savanna, which has distinct wet (late December to April) and dry (April to December) seasons. These finch species are found in a variety of grassy woodland habitats dominated by a ground cover of perennial and annual grasses; the seeds of which they rely on for food. Gouldian finches are prolific breeders, often nesting continuously from January through September (Tidemann

& Woinarski 1994; Tidemann et al. 1999). Long-tailed and Masked finch breeding attempts during this period are more sporadic, and most birds finish nesting by about June, much earlier in the dry season. Gouldian finches have a short moult period where both adults and juveniles moult all body and flight feathers within two months during the late dry season, around November (Franklin et al. 1998). Long- tailed and Masked finches have been recorded moulting in all months of the year except December, and commonly interrupt early dry season body or flight feather moult to resume nesting (Tidemann & Woinarski 1994). All species are locally nomadic, with individuals sometimes flying long distances each day (up to 15 kilometres) in search of food, water and nesting sites (O'Malley 2006b; Legge

2009). Gouldian finches may have larger home ranges (up to 200 km2 based on band resight data) and may be prone to periodic long-distance movements outside of their typical home range (Legge 2009). Regional movements (greater than 25 km) of Gouldian finch populations are not recorded most years and may not occur 138 in all areas of their range (Lewis 2007; Price et al. In preparation). The typical movements of Long-tailed and Masked finches are less well studied, but it is likely that their daily flight and home ranges are considerably smaller than the Gouldian finch (Higgins et al. 2006; Legge 2009).

Study site and survey seasons

Study sites were selected where all three study species occurred regularly in the vicinity and where the properties encompassed particular land management types.

Three properties in the Northern Territory Australia were sampled twice a year;

Bradshaw Field Training Area (latitude -15.5S, longitude 130.4E), a Defence property with very low grazing and burning levels, Delamere Station (latitude -

15.7S, longitude 131.6E), a pastoral property with moderate to high grazing levels, and the Yinberri Hills (latitude -14.1S, longitude 132.0E), aboriginal land with a history of extensive and intense bushfires (Figure 1.1a). Bradshaw Field Training

Area (Bradshaw FTA), and Delamere Station are part of the northern Victoria River

District, while the Yinberri Hills is within the southern section of the Top End.

The habitat on all properties was characterized by grassy understorey of perennial grasses such as Triodia spp., Heteropogon contortus, H. triticeus, Themeda triandra, Sehima nervosum, Chrysopogon fallax, and annual grasses such as

Sarga spp. and Schizachyrium spp. The Victoria River District’s sparse tree layer was dominated by Northern White Gums (Eucalyptus brevifolia) and Darwin Box

(E. tectifica), while the Yinberri Hills was dominated by Salmon Gum (E. tintinnans) 139 and Corymbia confertiflora. Shrubs and small trees make up a very small proportion of the vegetation, except in drainage areas and in patchy thickets of

Calytrix spp. Soils in high rocky country tended to be well-drained sandy soils, while floodplains were dominated by red-clay soils. Cracking clay or black soil plains were present in most areas, however, finches were not found to utilize this habitat during monitoring sessions.

In order to better understand the effect of land tenure on the health of finches, we looked at heath measure differences between tenures during the most stressful time of year for finches. Results of analysis of seasonal trends in health measures for these finch species suggested that the peak breeding season is not as stressful as the moulting period of the non-breeding season (Chapter 2). Body condition measures were consistently higher during breeding, and poorer during non- breeding and moulting. Therefore we compared health measures of birds captured on different land tenures during the non-breeding season only. In total 193

Gouldian, 234 Long-tailed and 202 Masked finches are included in this analysis.

Each property was sampled once in both 2007 and 2008 during the late dry/early wet season (late September to mid December), when annual grass seed food resources are limited due to sprouting, finches begin to switch to other food sources and all Gouldian finches are moulting. Populations of each finch species were located within an area of 5 - 10 km2 on each property each sampling period.

The search areas were relatively large due to the locally nomadic behaviour of these species. During each sampling visit to each study site, the goal was to capture and measure 30 - 70 individuals of each species, thereby sampling a 140 range of ages and sexes. No recaptured individuals from this or other studies were included in analysis and therefore all birds included in analysis were naïve to the capture process.

Estimation of grazing and burning levels

On each property, 20 - 28 circular plots of 100 m diameter were selected near areas where finches were sighted feeding or nesting. Because this study is part of a larger project monitoring birds during both breeding and non-breeding seasons, plots were placed in representative breeding and foraging habitats used by finches.

Finches were observed roosting, socializing or feeding at many of the nesting sites during non-breeding seasons. Thus nesting site vegetation was included in the analysis for this study. On each property, half of the plots were situated in rocky hill or escarpment habitat dominated by Triodia spp. grasses and a sparse canopy of smooth barked gums (E. brevifolia or E. tintinnans), and half were placed in low- land or floodplain habitat dominated by a variety of perennial and annual grasses and a canopy of Box or Bloodwoods.

To estimate grazing or burning levels, the percentage of each plot showing evidence of ground layer defoliation was visually estimated on a scale of 0 to 4 during each season. The score for each plot was estimated by walking a transect through the centre of the plot and assigning a score of 0 for no to 5% defoliation, a score of 1 for 5 - 25% defoliation, 2 for 25 - 50% defoliation, 3 for 50 - 75% defoliation, and 4 for 75 – 100% defoliation. This rough method of grazing impact 141 estimation was chosen because the large size of paddocks caused stocking rates to be an inaccurate measure of cattle density estimates on a local scale. Grazing defoliation was recorded when grass tussocks showed trimming or complete removal of grass blades by herbivores. Defoliation due to burning is easily identified in northern Australian savanna, as black or gray ash is retained on the ground and on grass tussock bases after burning. This burning evidence is not washed away until the following wet season (late December or January).

Separating the defoliation effects of fire and herbivory was straightforward in this study because fires were actively suppressed on Delamere, the only property with significant numbers of large exotic herbivores (cattle).

Finch capture and handling

Birds were captured during fine weather and within 4 hours of sunrise using mist nets at sites used by finches for drinking or foraging. Sites were between 5 to 20 km from any sealed road or human settlement, thus the possibility of human caused disturbance (other than our trapping effort) was minimal. For each bird, the length of time between capture (when the bird hit the mist net) and blood sampling was measured to the nearest minute to account for the effect of capture stress on total plasma corticosterone concentration (CORT). Birds were held in clean cotton cloth bags between extraction from the mist nets and blood sampling. Blood samples were taken by venipuncture of the left brachial vein with a 26 - gauge needle. A blood volume of 50 - 140 μL was collected into heparinised haematocrit tubes. The blood samples were kept on ice until they could be spun down in a 142 centrifuge, the haematocrit measured, and the plasma removed and frozen at -20C

(within 6 hours of sampling).

Each bird was banded with a uniquely numbered aluminium leg band and assessed for the level of body and flight feather moult, and the presence of any external reproductive characters (e.g. brood patch, egg in oviduct). This assessment information was used to interpret any changes in health measures that were due to moulting or breeding status and to confirm that all finches were moulting during sampling in the non-breeding seasons. Almost all finches were found to be moulting flight and body feathers during sampling. A small number of non-moulting birds (Gouldian n = 8, Long-tailed n = 12, Masked n = 16) were excluded from analysis. Adult Gouldian finches were sexed by plumage, while other species are not reliably sexed by physical characteristics. A proportion of adult Long-tailed (33 %) and Masked finches (28 %) were sexed using genetic material in blood samples (Griffiths et al. 1998).

Measurement of finch condition in the field

The text for this section of the methods is identical to chapter 2, and is therefore not repeated here. It is my intention to include that text when this chapter is submitted for review by an academic journal.

143 Outline of methods: Finch condition measures taken in the field included mass

(weight in grams corrected for finch sized using head to bill measures), fat scores and muscle scores.

Measurement of health indices from plasma

The text for this section of the methods is identical to chapter 2, and is therefore not repeated here.

Outline of methods: Total CORT and CBG capacity were determined from field- collected plasma samples. Plasma CORT was measured for 5μL aliquots of whole plasma from each bird in an EIA Kit ACETM Competitive Enzyme Immunoassay for corticosterone (Cayman Chemical Co. Ann Arbor, Michigan, USA). Plasma CBG capacity was measured using a radioligand-binding assay with tritiated corticosterone (described in Lynn, 2003). Due to the difficulty in safely bleeding small finches, not all birds had blood samples with sufficient amounts of plasma to test for CORT and CBG. Thus only 124 Gouldian, 139 Long-tailed and 110

Masked finches were tested for CORT, and 101 Gouldian and 60 Long-tailed finches were tested for CBG capacity.

Statistical analysis

Mixed design ANOVA analyses tested whether grazing and burning levels differed with year (repeated measure variable) and land tenure (between group variable, 144 PASW Statistics SPSS). When repeated measures data violated the assumption of sphericity, degrees of freedom were corrected using Greenhouse-Geisser estimates of sphericity. Contrast analysis using pairwise comparisons were corrected using Bonferroni adjustment. Though land tenure was the dominant factor, year did have a significant effect within the model. Thus differences in grazing and burning levels of study plots within properties were plotted separately for each year to show the change in grazing or burning over time. Two-factor

ANOVA analyses were used to test whether each finch health index varied with year (2007 and 2008) and land tenure (conservation, pastoral or aboriginal, JMP

5.2, SAS Institute). Where interaction effects were significant, Tukey-Kramer HSD tests were used to identify where the differences lay (SAS Institute 1995).

4.3 Results

Grazing and burning impacts

Land tenure differences in grazing and burning levels depended on year (grazing

F3.6, 201 = 14.42, P < 0.0001, burning F3.4, 201 = 3.80, P = 0.009), because the levels of grazing and burning changed within properties between years, not among properties (Figure 4.1). Thus each property was generally classed as grazed (no fire on pastoral land), fire prone (no grazing on aboriginal land), or conservation managed (little to no fire or grazing on Defence land).

145 Because the effect of land tenure on grazing and burning was weakly dependent on year, changes between years in grazing and burning levels at each site were looked at to determine if these effects coincided with differences in finch health.

There was a weakly significant tendency for higher levels of grazing during 2007 on pastoral land (F2, 69 = 4.88, P = 0.01; Figure 4.1). Other properties had minimal levels of grazing throughout the study. The pastoral property had no burning and the Defence site had low levels of burning, while the aboriginal property had a tendency for higher levels of burning during the second year of the study (Figure

4.1).

146

3.5 d

a d

a Defence 3.0 Pastoral 2.5 Aboriginal 2.0 1.5 1.0 0.5 0.0

2007 2008 Grazing Grazing Impact ofScale on 0 to 4 Year

3.5 d=a>p d=a>p 3.0

2.5

2.0

1.5

1.0

0.5

Area Burnt on scale of 0 to 4 Area to 0 Burntscaleon of 0.0 2007 2008

Year

Figure 4.1: Grazing and burning levels on each property during 2007 and 2008.

Bars are means +/- SE, and significant differences between land tenures are signified with abbreviations for tenure (with Bonferroni adjustment, P < 0.05).

147 Gouldian finch health indices

Year and land tenure explained 9 - 25% of the variation in Gouldian finch health indices. Body mass and muscle scores differed between land tenures, and the effect was weakly dependent on year (interaction term; body mass F2, 182 = 3.11, P

= 0.05, muscle F2, 187 = 2.7, P = 0.07; Figure 4.2), however, each independent variable was not significant. In contrast, the main effects of year and tenure on fat scores were highly significant (year F2, 187 = 25.85, P < 0.0001, tenure F2, 187 =

5.85, P = 0.003). Body condition measures for Gouldian finches sampled on

Defence land during 2007 were consistently higher than those sampled on other properties. Body mass and muscle scores were significantly higher on Defence land compared to pastoral land during 2007 (aboriginal land had similarly low mass to pastoral land, but moderate muscle scores between the two); while fat scores were higher on Defence land compared to aboriginal land during both years

(pastoral land was moderate between the two; Tukey – Kramer HSD, Figure 4.2).

Fat was also higher on all properties during 2007 compared to 2008.

Similar to body mass and muscle, the differences between mean haematocrit and stress responses between properties were dependent on year (interaction term; haematocrit F2, 172 = 13.49, P < 0.0001, and stress F2, 122 = 5.19, P = 0.007).

Haematocrit was significantly elevated in birds sampled on pastoral and aboriginal land compared to Defence managed land during 2007, while there were no significant tenure differences in 2008 (Tukey – Kramer HSD, Figure 4.3). Gouldian finches sampled at the pastoral site had higher stress responses compared to 148 other properties; however this trend was only significant for 2007 (Tukey – Kramer

HSD, Figure 4.4). CBG capacity varied with year (F1, 99 = 6.30, P = 0.01), but not tenure (F2, 99 = 3.04, P = 0.06), with CBG levels higher during 2008 compared to

2007. In summary, measures of body condition, haematocrit and stress response all differed between Defence land and the two more disturbed habitats (pastoral or aboriginal) during 2007 (Table 4.1).

149

a) Body Mass b) Muscle

1.0 d>p=a 2.5 d>p

0.5 2.0 1.5 0.0 1.0 -0.5

0.5 Residual mass (g)mass Residual

-1.0 (0score muscle to 3) 0.0 2007 2008 2007 2008 Year Year

c) Fat

2.5 d>a d=p>a Defence 2.0 Pastoral Aborignal 1.5

1.0

0.5 fat (0score to 5) 0.0 2007 2008 Year

Figure 4.2: Body condition measures for Gouldian finches sampled during non- breeding seasons 2007 and 2008. Graphs refer to measures of finch a) body mass, b) pectoral muscle score, and c) furcular fat score. Measures are means +/-

SE. Significant differences between land tenures are signified with abbreviations for land tenures (Tukey – Kramer HSD, P < 0.05).

150

70 d

Pastoral 65 Aboriginal

60

55

50

Heamatocrit (% red cells/plasma) blood 2007 2008 Year

Figure 4.3: Haematocrit levels for Gouldian finches sampled during non-breeding seasons 2007 and 2008. Measures are means +/- SE. Significant differences between tenures are signified with abbreviations for land tenures (Tukey – Kramer

HSD, P < 0.05).

151

d

a d

20 Defence Pastoral 15 Aboriginal 10

5

0 Residual Cort (ng/mL) Cort Residual -5

-10 2007 2008 Year

Figure 4.4: Residual CORT levels (ng/mL) for Gouldian finches sampled during non-breeding seasons 2007 and 2008. Measures are means +/- SE. Significant differences between land tenures are signified with abbreviations for land tenures

(Tukey – Kramer HSD, P < 0.05).

152

Table 4.1: Direction of significant differences in Gouldian finch health indices between tenures during the two years of the study (determined using Tukey –

Kramer HSD).

Gouldian Finch

Non-breeding Season 2007 Non-breeding Season 2008 Heath Index

Body mass Defence > Pastoral = Aboriginal No Tenure Differences

Defence > Pastoral Muscle No Tenure Differences Aboriginal not different Defence > Aboriginal Fat Defence = Pastoral > Aboriginal Pastoral not different

Haematocrit Pastoral = Aboriginal > Defence No Tenure Differences

CORT Pastoral > Defence > Aboriginal No Tenure Differences

CBG No Tenure Differences No Tenure Differences

153 Long-tailed finch health indices

Year and land tenure together explained a significant amount of variation (5 - 26%) in Long-tailed finch health indices. Body mass was higher on pastoral land both years (effect of tenure F2, 228 = 9.43, P < 0.0001, Tukey – Kramer HSD, Figure 4.5).

Muscle score was not significantly different between land tenures during any year

(F2, 228 = 0.88, P = 0.42). The effect of tenure on fat depended on year (F2, 228 =

5.98, P = 0.003), with birds on aboriginal land being less fat than other tenures in

2007, and birds being less fat on the Defence property compared to other tenures in 2008 (Tukey – Kramer HSD, Figure 4.5). Thus body condition measures were inconsistently different between tenures (Table 4.2).

The effect of tenure on Long-tailed finch haematocrit and stress response was dependent on year (haematocrit F2, 212 = 3.55, P = 0.03, stress response F2, 137 =

6.98, P = 0.001). Haematocrits were elevated in birds sampled on pastoral and aboriginal land compared to Defence land during non-breeding in 2007, but not different between tenures in 2008 (Tukey – Kramer HSD, Figure 4.6). Long-tailed finch stress responses were significantly higher on aboriginal land compared to

Defence in 2007, but not significantly different between tenures in 2008 (Tukey –

Kramer HSD, Figure 4.7). Year and tenure had no effect on CBG capacity (year

F1, 60 = 0.37, P = 0.55, tenure F2, 60 = 0.52, P = 0.60). Thus Long-tailed finch health indices did often differ between land tenures, but unlike Gouldian finches, health indices did not differ consistently between land tenures during any one year (Table

4.2). 154 a) Body Mass 1.0 d=a

0.5

0.0 mass (g) mass -0.5

-1.0 2007 2008 Year

b) Fat

2.5 d=p>a d=a

1.0

0.5 fat (0 score to 5) 0.0 2007 2008 Year

Figure 4.5: Long-tailed finch a) body mass and b) furcular fat score sampled during non-breeding seasons 2007 and 2008. Measures are means +/- SE. Significant differences between land tenures are signified with abbreviations for land tenures

(Tukey – Kramer HSD, P < 0.05).

155 70 d

60

55

50 2007 2008

Year Haematocrit (% red blood cells/plasma)

Figure 4.6: Haematocrit levels for Long-tailed finches during non-breeding seasons

2007 and 2008. Measures are means +/- SE. Significant differences between land tenures are signified with abbreviations for land tenures (Tukey – Kramer

HSD, P < 0.05).

156 25

20 d

5

0

-5 Residual CORT Residual (ng/mL) -10

-15 2007 2008 Year

Figure 4.7: Residual CORT levels (ng/mL) for Long-tailed finches sampled during non-breeding seasons 2007 and 2008. Measures are means +/- SE. Significant differences between land tenures are signified with abbreviations for land tenures

(Tukey – Kramer HSD, P < 0.05).

157 Table 4.2: Direction of significant differences in Long-tailed finch health indices between tenures during the two years of the study (determined using Tukey –

Kramer HSD).

Long-tailed Finch Non-breeding Season 2007 Non-breeding Season 2008 Health Index

Body mass Pastoral > Defence = Aboriginal Pastoral > Defence = Aboriginal

Muscle No Tenure Differences No Tenure Differences

Pastoral > Defence Fat Defence = Pastoral > Aboriginal Aboriginal not different

Haematocrit Pastoral > Aboriginal > Defence No Tenure Differences

CORT Aboriginal > Pastoral = Defence No Tenure Differences

CBG No Tenure Differences No Tenure Differences

158 Masked finch health indices

Year and land tenure explain a smaller amount of variation (2 - 15%) in Masked finch health indices than found in other finch species. The effect of land tenure on body mass was dependent on year (F2, 196 = 8.73, P = 0.0002), with pastoral and

Defence site birds heavier than aboriginal land birds during 2008 only (Tukey –

Kramer HSD, Figure 4.8). The effect of tenure on muscle and fat scores was not significant (P > 0.05). However, year had an effect on overall fat scores (F1, 196 =

24.52, P < 0.0001), with fat scores on all properties higher during 2007.

Haematocrit did not significantly differ between birds sampled on different land tenures during any year (F2, 189 = 1.01, P = 0.37). Haematocrit did differ by year

(F1, 189 = 6.1, P = 0.01), with overall haematocrits higher during 2008. Though no differences between tenures or years were significant, Masked finch stress responses were slightly elevated during the non-breeding season 2007 compared to 2008 (Figure 4.9). Together these results show that Masked finch health indices varied little in response to land tenure (Table 4.3).

159 a) Body Mass

1.0 d=p>a Defence 0.5 Pastoral Aboriginal

0.0 mass (g) mass -0.5

-1.0 2007 2008 Year

Figure 4.8: Body mass for Masked finches sampled during non-breeding seasons

2007 and 2008. Measures are means +/- SE. Significant differences between land tenures are signified with abbreviations for land tenures (Tukey – Kramer

HSD, P < 0.05).

160

25

20 Defence

15 Pastoral

10 Aboriginal

5

0

-5 Residual Cort (ng/mL)Residual

-10

-15 2007 2008

Year

Figure 4.9: Residual CORT levels (ng/mL) for Masked finches sampled during non- breeding seasons 2007 and 2008. Measures are means +/- SE. There were no significant differences between land tenures (Tukey – Kramer HSD, P < 0.05).

161 Table 4.3: Direction of significant differences in Masked finch health indices between tenures during the two years of the study (determined using Tukey –

Kramer HSD).

Masked Finch Non-breeding Season 2007 Non-breeding Season 2008 Health Index

Body mass No Tenure Differences Pastoral = Defence > Aboriginal

Muscle No Tenure Differences No Tenure Differences

Fat No Tenure Differences No Tenure Differences

Haematocrit No Tenure Differences No Tenure Differences

CORT No Tenure Differences No Tenure Differences

162 4.4 Discussion

Health variation between tenures in a declining species

The results confirm that health measures of endangered Gouldian finch populations do differ between land tenures, and differ from the health measures of non-declining species in a manner which suggests habitat disturbance does influence the health of this species. Health indices (muscle, body mass, fat, haematocrit and stress response) suggested poorer health for birds sampled on pastoral land during the non-breeding season 2007, which coincides with the highest level of grazing across years and study sites. The short-term increase in grazing on this property in 2007 coincided with decreased body condition, higher stress responses and unusually high haematocrit for Gouldian finches. During both non-breeding seasons, Gouldian finches sampled on pastoral land had higher residual CORT but similar CBG capacity to other land tenures, which should translate into higher free CORT levels for this population. When grazing levels on this property decreased in 2008, Gouldian finch health measures were similar between pastoral and ungrazed properties. Gouldian finches sampled on aboriginal land prone to frequent fire also showed higher than average stress levels during 2008, which coincides with the time when the largest amount of land was burnt on this property. Research has shown that Gouldian finches are more likely to be stressed or show poor body condition during the non-breeding season compared to the breeding season (Chapter 2). As expected, changes in finch health measures between years on pastoral and aboriginal land did suggest that 163 increases in grazing and fire can lower the health indices of Gouldian finches during the stressful non-breeding (late dry/early wet) season. Mean measures of fat and muscle suggest that Gouldian finches monitored on pastoral land in 2007 had some fat and muscle however, 10% were observed to be carrying no fat and had muscle contour suggested they were highly ematiated and possibly starving.

In general, birds are not typically observed to show signs of ematiation, except during extreme climatic events or during migration, when birds are forced to fast

(Sturkie 1986). The fact that Gouldian finches are completing an energetically demanding full body moult during a likely time of lowered food availability could explain why these birds had such low condition in habitats with grazing and fires - disturbances which are likely to further deplete food (grass seed) resources important to moulting Gouldian finches.

Health variation between tenures in a non-declining species

In contrast, Long-tailed finch and Masked finch health measures did not consistently differ between land tenures during this study. While measures of body condition did vary between Long-tailed and Masked finch populations at some point during the study, a difference in one measure did not often follow the same pattern of differences in other measures for a given sampling period. For example, body mass for Long-tailed finches sampled on conservation managed Defence land was significantly lower than birds sampled on grazed pastoral land in late

2007. However, their muscle scores and fat were similar between the two land tenures. Likewise, Masked finch body mass was significantly lower on aboriginal 164 land compared to other land tenures during 2008, but mean muscle and fat scores were not different between land tenures. Variation in body condition was also unlikely to coincide with similar changes in haematocrit or stress. One example of a possible exception is seen during 2007, where lowered body mass and fat scores in Long-tailed finches on aboriginal land coincided with moderately high haematocrit and significantly higher stress response measures compared to birds sampled on other land tenures. However, Long-tailed finch haematocrit and stress did not differ between land tenures in 2008, though mass and fat scores were still depressed on aboriginal land. Masked finch stress response differences also rarely coincided with body condition differences. In most birds, individuals in good condition tend to have lowered stress responses (Schwabl et al. 1991; Marra &

Holberton 1998; Kitaysky et al. 1999; Muller et al. 2007). Thus the lack of coinciding lowered body condition and heightened stress responses in these species suggests that Long-tailed and Masked finches stress response levels may be too low in general to be an indicator of condition, or simply vary more in response to factors other than body condition, such as moult status.

These differences in the Long-tailed and Masked finch health measures that appear to vary erratically with land tenure are difficult to interpret. All birds were observed to be moulting; however it was not feasible to accurately measure the stage of moult for each individual. This is because Long-tailed and Masked finches regularly interrupt or suspend moult, possibly in response to lowered food resources, or in order to resume breeding condition (Tidemann & Woinarski 1994;

Franklin et al. 1998). Study of possible changes in individual bird health profiles 165 during active compared to suspended moult may help determine if differences in moult stage are likely to interact with differences in finch health measures between populations. Though body condition, haematocrit and stress response were often statistically different between populations sampled on different land tenures, all measures may be within ‘healthy’ limits for these species. This is supported by three observations. Firstly, Long-tailed and Masked finch mean muscle scores were between scores of ‘2’ and ‘3,’ indicating that all populations had above average contour for passerine species (Lindstrom et al. 2000). Secondly, mean haematocrit varied little in general and was typically below 60%, while Gouldian finch levels varied more widely and were often 62% and higher. These lower haematocrit levels suggest that common finches may have lower energy demands in general compared to declining Gouldian finches, as haematocrit levels higher than 60 are indicative of birds involved in energetically demanding activities, such as migration (Carpenter 1975; Fair et al. 2007). Lastly, mean Long-tailed and

Masked finch stress responses were only moderately variable compared to the 20- fold differences seen between Gouldian finch populations. Thus statistically different measures may not always indicate better or poorer health for finch populations unless values consistently fall outside a normal range for wild individuals of that species.

However, the actual haematocrit values of both declining and non-declining species in this study are considerably higher than normal values for caged birds and more sedentary wild finches such as the Crimson finch (Neochmia phaeton), which has a mean haematocrit value of 55% (Fair et al. 2007; Milenkaya 2010). 166 Mean haematocrit values of 35 – 55% have been recorded for caged birds, including an average value of 55% for captive Zebra finches (Birkhead et al. 1999;

Fair et al. 2007). The factors commonly known to increase the haematocrit levels of bird individuals or species include increases in altitude, energy demands such as increased flying and dehydration (Fair et al. 2007). Because all three of the study species reside at the same altitude and plasma osmolality tests suggested birds were not dehydrated (Chapter 2), it is most likely that the high haematocrit levels observed in these species are due to high energy demands. Gouldian,

Long-tailed and Masked finches with high haematocrits often occupy the same areas as Crimson finches which have lower haematocrits, and have been recorded flying further each day compared to Crimson finches (Legge 2009; Milenkaya

2010). These observations support the hypothesis that high haematocrits in

Gouldian, Long-tailed and Masked finches were dure to higher energy demands in comparison to other bird species. In turn, the possibility that all species in this study have relatively high energy demands suggests that this aspect of their lifestyles is similar, and that other lifestyle traits, such as diet and timing of breeding or moult may be more likely to be driving differences in health measures between these species.

Interestingly, common finch health measures vary more consistently between seasons or years than between land tenures, while Gouldian finch health measures did not. Though all finch species body mass and muscle scores tended to be higher during the breeding season and fat scores tended to be higher during the non-breeding season, these trends were more likely to be significant during 167 both years of the study for Long-tailed and Masked finches, while only significant during one year for Gouldian finches (Chapter 2). Long-tailed finch stress responses and CBG capacity tended to be higher during breeding (Chapter 2) similar to other passerines (Breuner & Orchinik 2002b), while Masked finch stress responses were only higher than average during the non-breeding season 2007.

In contrast, Gouldian finch stress responses were elevated during both non- breeding seasons compared to breeding, and were particularly high on disturbed land tenures. This suggests that regional environmental variation such as rainfall and overall seasonal growth of food plants could be having more influence on the health of common finches than localized grazing or fire impacts that seem to influence the health of the rare Gouldian finch. The pronounced seasonal pattern of variation in health indices of common finches and their consistently high body condition scores agrees with the prediction that these species are not likely to show variation in health in response to localized land management differences.

Their ability to adapt to habitat variability is most likely due to their more opportunistic breeding and moulting strategies and diverse diet compared to the rare Gouldian finch.

Health variation between individuals

Certain heath indices measured in both declining and non-declining finches varied considerably between individuals within seasons and populations. For example, residual CORT levels in finches ranged from 15 ng/mL below mean to more than

35 ng/mL above mean, and total plasma CORT concentrations varied from 4 – 80 168 ng/mL. This range is similar to most passerines studied, which have total CORT levels ranging between 5 and 90 ng/mL total CORT (Astheimer et al. 1995;

Breuner et al. 1999; Pravosudov et al. 2001; Wada et al. 2006b). The range in

CORT levels between individual finches in this study is not fully explained by differences between properties or years. It is also not attributable to differences in breeding status or moult, as all birds were not in breeding condition and all birds were moulting. Factors not measured, such as an individual’s social status, interactions with predators or a history of nutritional stress not apparent though body condition measures may be affecting individual stress responses (Romero

2004; Malisch & Breuner 2010).

Condition measures did not vary greatly between individual finches within populations: fat typically varied between scores 1 - 2, muscle 1.5 - 2.5, and mass varied only 0.5 - 1g (based on mass not corrected for bird size). One exception includes a number of Gouldian finches measured to have fat scores of 3 or higher during 2007 (n = 18), which indicates that these birds were carrying 10 – 50% more fat than birds with lower scores. Also during 2007, a number of Gouldian finches sampled on pastoral land were observed to be carrying no fat and had muscle scores suggesting these individuals were ematiated. Part of this extreme variation in Gouldian finch body condidtion can be explained by land tenure: 61% of birds with fat scores of 3 and over were sampled on defense land, while 100% of birds with no fat and low muscle scores were sampled on pastoral land. Still, 22 % of birds with fat scores of 3 and higher were sampled on pastoral land, highlighting the considerable range in body condition between individuals within this population 169 during 2007. This large variation in condition may be related to variation in individual finch ability to adapt to environmental conditions particular to pastoral land, such as an ability to find and exploit new food resources when preferred seeds are less available (Donnelly & Sullivan 1998; Cuthill et al. 2000). No

Gouldian finches had fat scores of three or higher during 2008. This suggests that some of the variation in fat storage was related to annual variation in the benefit of carrying fat, which could be caused by factors such as perceived levels of food availability and predation pressure (McNamara & Houston 1990; Hurly 1992;

Cuthill et al. 2000).

Though mean haematocrit varied little between populations and years, haematocrit levels varied considerably between individuals. It is typical for wild birds to exibit between 15 – 25% variation in haematocrit within a population (Fair et al. 2007).

Gouldian, Long-tailed and Masked finches exhibited a range of haematocrit values between 55 and 70, which is a range of about 20% and within the normal range for avian species. Similar to other health indices, factors not measured in this study such as individual differences in parasite loads, activity levels or slight differences in nutritional status may have caused some of the individual variation in haematocrit (Fair et al. 2007). Differences in the health measures of individuals was not the focus of this population based study, however, the questions raised by the observation of high levels of individual variation in some health indices suggests a need for experimental studies which can determine the factors most likely to affect the health of individual wild finches.

170 Conservation implications

The results of this study suggest that measures of body condition, haematocrit, and stress can be useful indicators of the response of avian populations to environmental conditions. These measures were helpful in determining the response of a declining species, the Gouldian finch, to the environmental disturbances of grazing and fire. The results of this study show significantly lowered health in Gouldian finches during a period of increased grazing on a pastoral property. Pastoral practices have been assumed to have an impact on the abundance and distribution of this declining species, based on studies documenting the decline of granivores across Australia using large scale presence- absence data (Franklin et al. 2005). However, it is unclear as to how much of an impact grazing could have on a local scale and what the actual limiting processes might be. The finding that Gouldian finch condition is lowered and stress levels are higher in a population living in regularly grazed pastoral habitat during the non- breeding season suggests that grazing affects finch food resources that are important during the transition between the late dry and wet season. This food resource, grass seed, is also known to be limited by fire (Mott & Andrew 1985;

Crowley & Garnett 2001; Russell-Smith et al. 2003a). Several of the health indices used (haematocrit and stress indices in particular) did suggest that birds living in fire-prone landscapes on aboriginal land were also prone to lowered condition compared to birds living in conservation areas with no grazing and fire. However, the differences in body condition were not as extreme compared to those found in birds sampled in grazed pastoral habitat. Higher rainfall, higher annual grass cover 171 and patchy burning patterns on the fire prone property may result in more resources being available in comparison to a pastoral property with lower rainfall averages and a more evenly distributed grazing pattern (Chapter 3). Incidentally,

Gouldian finches on pastoral land were observed feeding on Aristida spp. grass seeds which are not a typical component of their diet (Dostine & Franklin 2002).

These observations were made in 2007, when finch health measures indicated higher stress and poor body condition. Gouldian finches monitored on fire-prone aboriginal land were observed to be feeding on their preferred grass seed, Sarga spp. These observations support the hypothesis that the availability of preferred seed was lower on the pastoral property in comparison to the fire-prone site.

Because finches are highly mobile, they would be relatively well equipped to find areas of high seed availability within an unevenly burnt landscape compared to an evenly grazed region with fewer patches of preferred seed. However, the patchiness of grazing or burning or actual seed abundance on properties was not measured. A more detailed study of possible factors affecting seed abundance of grasses on different properties is warranted, as are experimental studies examining what nutritional factors are likely to cause poor health in finches.

This study provides evidence that even one season of high grazing pressure can significantly lower the health of Gouldian finches within that year. Multiple seasons of heavy grazing and extensive burning can dramatically or irreversibly change the savanna grass layer, which could in turn cause these areas to be unsuitable for the long-term persistence of Gouldian finches and less profitable for grazing (Friedel

1997; Crowley & Garnett 1998; Bennett et al. 2002; Sharp & Whittaker 2003). By 172 managing savannas with moderate levels of grazing and fire, both wildlife and pastoral enterprises could benefit from higher grass productivity and abundance.

Through the retention of savanna biodiversity, local communities could also benefit though increased tourism in high quality habitats and the creation of land-care employment opportunities. Thus the conservation of highly productive savanna habitat has potential long-term benefits for endangered wildlife such as the

Gouldian finch, as well as local human communities.

173 5.0 Chapter Five: Does variation in Gouldian finch health indices predict variation in Star finch and Black-throated finch health indices?

5.1 Introduction

Many of Australia’s granivorous birds are in decline, with large range contractions documented in half of the native grass-finches in the state of Queensland (Franklin

1999). At least one subspecies of grass-finch is likely to be extinct: the Southern

Star finch (Neochmia ruficauda ruficauda). This subspecies once ranged throughout savanna habitat in south-eastern Queensland but has not been sighted since the early 1990s (Holmes 1998). The southern subspecies of the Black- throated finch (Poephila cincta cincta) has also suffered substantial range contractions in the past 50 years and it is now represented by only a few remnant populations in south-eastern Queensland (Barrett et al. 2003). In contrast, the northern subspecies of the Star finch (N. r. clarescens) persists on the Cape York

Peninsula in small isolated populations of uncertain status (Holmes 1996; Dorricott

& Garnett 2007). The Cape York Peninsula region of north-eastern Australia is still largely undeveloped compared to other regions on the east coast, particularly the highly populated area of south-eastern Queensland. This may explain why, the northern subspecies of the Black-throated finch (P. c. atropygialis) persists in the

Cape York Peninsula and seems to have a stable population and distribution

(Garnett & Crowley 2000). There is a general consensus among ecologists that a combination of factors, mainly over-grazing, changed fire regimes and land clearing has led to the declines of savanna grass-finches in northern Australia, with 174 declines following the advance of pastoralism from the south-east towards the west

(Garnett 1992; Franklin et al. 2005; Dorricott & Garnett 2007).

Gouldian finches have nearly disappeared from Queensland, with only one confirmed breeding population sighted on the western border of the state in the past decade (Holmes 1995; O'Malley 2006a; Maute & Venables 2009). Unlike most grass-finches, the Gouldian finch is nationally listed as endangered and seems particularly sensitive to environmental change due to its specialized grass- seed diet and strict breeding and moulting schedules (Tidemann & Woinarski 1994;

Dostine & Franklin 2002). This sensitivity is reflected in the unique seasonal pattern of variation in body condition, haematocrit and stress hormone (CORT) levels of Gouldian finch populations compared to co-occurring finch species that are not declining. In general, Gouldian finches have lowered measures of body condition (body mass and flight muscle contour) and higher fat scores, haematocrit and CORT values during the non-breeding season compared to the breeding season (Chapters 2, 4). The likely cause of poor condition and higher stress during the non-breeding (late dry/early wet) season is possibly decreased availability of preferred grass seed (Chapter 3) combined with the increased energy demands of moulting. Gouldian finches also had lowered body condition and higher CORT levels in areas showing habitat degradation from higher levels of grazing or burning. Non-declining Long-tailed and Masked finches found in the same habitat did not display these changes in condition or hormonal indicators of stress in response to seasonal, yearly or site-specific environmental change.

Instead, these species tended to be in good condition year round and have higher 175 stress during the breeding season compared to non-breeding, as reported for many passerine species in the northern hemisphere (Breuner & Orchinik 2002b;

Romero 2002). It is unclear if the pattern of health index variation observed in

Gouldian finches is unique to the species or is representative of other species that are also sensitive to environmental change.

Star finches have several characteristics in common with Gouldian finches, including living in similar habitats, having a moderately specialized diet (small range of seeds but supplement their diet with insects) and strict timing of moult

(most individuals moult during one or two months of the year (Todd et al. 2003;

Higgins et al. 2006). In contrast, Black-throated finches have a relatively diverse diet (larger range of seeds and other food sources) and a more opportunistic moulting strategy compared to Gouldian or Star finches (Higgins et al. 2006). Both

Star and Black-throated finches have more opportunistic breeding strategies than

Gouldian finches, involving lengthening or shortening of the breeding season, possibly depending on food resource availability (Higgins et al. 2006).

All finch species in this study are found in grassy woodland habitats dominated by a ground cover of perennial and annual grasses; the seeds of which they rely on for food. These seeds are likely to be in short supply by the end of the dry season, and if finches are not adapted to or able to find alternate food resources, they are likely to show signs of poor condition and stress during this time of year. If Star and Black-throated finches are as sensitive to environmental conditions as

Gouldian finches, their populations are predicted to show evidence of higher stress 176 and reduced body condition during the non-breeding season compared to the breeding season. If there is yearly variation in the quality of habitat among the finch populations being sampled, it is predicted that they will show correspondingly poor health indices in heavily grazed or frequently burnt habitats, as seen in

Gouldian finches. This study investigates whether moderately specialized Star and

Black-throated finches show similar patterns in health indices in response to seasonal and environmental fluctuation to those shown in Gouldian finches. The future status of remaining populations of threatened Star and Black-throated finches in northern Queensland is uncertain and this study aims to predict the vulnerability of these threatened sub-species to further declines.

5.2 Methods

Study species populations

Three subspecies of Australian endemic finches were studied. The northern subspecies of the Star finch was sampled at two sites on the Cape York Peninsula;

Lakefield National Park and near the community of Pormpuraaw (Figure 1.1b).

The southern sub-species of the Black-throated finch (P. c. cincta) was sampled at

Laudham Park Station and the northern sub-species of the Black-throated finch was sampled at Lakefield National Park. The Laudham population is one of the few remaining groups of the threatened subspecies P. c. cincta, while birds sampled at Lakefield are the relatively common subspecies P. c. atropygialis.

177 Study site characteristics and survey seasons

Study sites containing the study species were chosen over a range of habitat types at three locations in northern Queensland, Australia. Lakefield National Park

(latitude -14.38S, longitude 143.58E) is a conservation reserve with moderate grazing and burning levels (few introduced herbivores and burning of some areas repeatedly every two to six years). Laudham Park Station (latitude -19.29S, longitude 146.41E) is a pastoral property with high levels of grazing (> 5 head/km2) and little fire. Pormpuraaw (latitude -14.55S, longitude 141.37E) is an area of aboriginal land with a history of annual extensive hot bushfires, weed infestations and moderate to high grazing levels (Figure 1.1b). The habitat on all properties was characterized by grassy understorey of perennial grasses such as

Heteropogon contortus, Sarga plumosum, Themeda arguens, Oryza australiensis,

Panicum maximum, Urochloa mosambicensis and annual grasses such as

Schizachyrium spp, Brachiaria spp., and Themeda quadrivalvis. The very sparse tree layer at sub-coastal sites within Lakefield was dominated by tea trees

(Melaleuca viridiflora), while Laudham was more heavily wooded by dominant

Eucalyptus platyphylla, E. erythrophloia and Melaleuca viridiflora. Pormpuraaw is a coastal area with a patchy but often dense tree layer dominated by beach casuarina (Casuarina equisetifolia) and a variety of mangrove and colonizing rainforest trees. Soils at Pormpuraaw and Laudham tended to be well-drained sandy soils, while floodplains in Lakefield were often cracking clay or black soils.

These general observations at sites allowed the characterization of Lakefield habitat as infrequently grazed and burnt intact coastal grassland, Laudham habitat 178 as highly grazed woodland savanna, and Pormpuraaw habitat as frequently burnt open casuarina woodland with small patches of dense coastal rainforest.

Populations of each finch species were located and sampled twice a year within an area of 5 - 10 km2 on each of their respective study properties (two per species).

The search areas were relatively large due to the semi-nomadic behaviour of these species. Individuals from each population were captured, measured and released during both the early non-breeding season (November – early December 2007 and

2008) and the end of the breeding season (June 2008 and 2009). During each visit to a study site, 30 - 40 individuals of each species were captured and measured. In total, 300 Star finches and 228 Black-throated finches were sampled for this study. All individuals were naïve birds that had not been previously captured.

Finch capture and handling

The text for this section of the methods is identical to chapter 2, and is therefore not repeated here. It is my intention to include that text when this chapter is submitted for review by an academic journal.

Notes for the species studied in this chapter:

Adult Star finches were sexed by plumage, while all other finches are not reliably sexed by physical characteristics. A proportion of adult Black-throated finches were sexed using genetic material in blood samples (Griffiths et al. 1998). 179

Measurement of finch condition in the field

The text for this section of the methods is identical to chapter 2, and is therefore not repeated here. It is my intention to include that text when this chapter is submitted for review by an academic journal.

Outline of methods:

Finch condition measures taken in the field included mass (weight in grams), fat scores and muscle scores. Mass was corrected for bird structural size using a regression against head to bill measurements, based on the high correlation

2 between head to bill measures and bird mass (Star finch r = 0.08 F1, 298 = 25.21, P

2 < 0.0001; Black-throated finch r = 0.13 F1, 226 = 34.59, P < 0.0001).

Measurement of health indices from plasma

The text for this section of the methods is identical to chapter 2, and is therefore not repeated here. It is my intention to include that text when this chapter is submitted for review to an academic journal.

Outline of methods:

Total CORT and CBG capacity were determined from field-collected plasma samples. Due to the difficulty in safely bleeding small finches, not all birds had blood samples with sufficient amounts of plasma to test for CORT and CBG. Thus 180 only 169 Star and 145 Black-throated finches were tested for CORT, and 103 Star finches were tested for CBG capacity. 4 Plasma CORT was measured for 5μL aliquots of whole plasma from each bird in an EIA Kit ACETM Competitive Enzyme

Immunoassay for corticosterone (Cayman Chemical Co. Ann Arbor, Michigan,

USA). Regression of Star and Black-throated finch CORT to handle time suggested a significant correlation between these measures (Star finch r2 = 0.27

2 F1, 167 = 15.44, P < 0.0001; Black-throated finch r = 0.23 F1, 143 = 19.60, P <

0.0001), thus residuals taken from these regressions were used in analysis of differences in stress response between seasons and sites (Figure 5.1).

Star finch plasma CBG capacity was measured using a radioligand-binding assay with tritiated corticosterone (described in Lynn, 2003).

4 Black-throated finch CBG capacity was not measured due to the time and budget constraints of this preliminary study. 181 a) Star finch

120 150 100 80 100 60 40 20

CORT(ng/ 50 Residual ml) CORT(ng/mL) 0 -20 -40 0 10 20 30 40 50 60 70 80 90 20 30 40 50 60 handle time (minutes) CORT (ng/ml) Predicted b) Black-throated finch

200 150

150 100

100 50

CORT(ng/

ml) Residual ml) CORT(ng/mL) 50 0

0 -50 10 20 30 40 50 60 70 80 90 30 40 50 60 70 80 handle time (minutes) CORT (ng/ml) Predicted

Figure 5.1: Graphs on the left are regression plots of a) Star finch and b) Black- throated finch CORT levels (ng/mL) to handling time (minutes). Resulting residuals for each species are shown on graphs on the right.

182 Statistical analysis

Health measurements were taken from naïve individuals each season and found to be normally distributed. Therefore ANOVA models of health index response to year (first year; non-breeding 2007 and breeding 2008, second year; non-breeding

2008 and breeding 2009), season (breeding vs. non-breeding) and site (two sites per species) were used to examine how well these factors explained variation in finch health indices (SAS institute JMP5.2). Because interaction effects were often significant Tukey-Kramer HSD tests were used to identify where differences lay between sites within each year (SAS Institute 1995).

5.3 Results

Star finches

Seasonal patterns of Star finch health index variation were often dependent on year. The effect of season on body mass, fat score and stress response all varied by year (body mass F1, 292 = 13.28, P = 0.0003; muscle F1, 292 = 5.85, P = 0.02; fat

F1, 292 = 103.86, P < 0.0001; residual CORT F1, 167 = 6.820, P = 0.01). Measures of body mass and fat were higher during the non-breeding season compared to breeding while muscle scores were higher during breeding in the first year of the study; however, no body condition measures differed seasonally in the second year (Tukey – Kramer HSD, Figure 5.2). The effect of season on haematocrit was dependent on site (F1, 285 = 15.89, P < 0.0001), with higher haematocrits measured 183 during non-breeding compared to breeding seasons on Lakefield only (Tukey –

Kramer HSD, Figure 5.3). Stress responses tended to be higher during the first year of the study compared to the second, and consistently higher during non- breeding seasons compared to breeding seasons (Tukey – Kramer HSD, Figure

5.4). This last trend is mainly controlled by the extremely elevated stress response measures recorded during the non-breeding season 2007. No seasonal difference was seen in CBG levels (F1, 101 = 0.25, P = 0.62; Figure 5.5). These results suggest that Star finches tended to have higher stress, haematocrit, and sometimes lower muscle condition in the non-breeding season compared to breeding, particularly during non-breeding in 2007.

184 a) body mass b) muscle

1.0 3 L

0.5 2

0.0 1 Residual Residual mass (g)

-0.5 (0score muscle to 3) 0

Breeding Breeding Breeding Breeding Non-breeding Non-breeding Non-breeding Non-breeding Season Season

c) fat

4 L

P L

2

1 fat score (0 to 5) 0

Breeding Breeding

Non-breeding Non-breeding Season

Figure 5.2: Body condition measures for Star finches sampled during the non- breeding seasons 2007 and 2008, and the breeding seasons 2008 and 2009.

Graphs refer to measures of finch a) body mass, b) pectoral muscle score, and c) furcular fat score. Bars are means +/- SE. Significant differences between properties for each season determined by Tukey – Kramer HSD are designated with abbreviations for sites (P < 0.05). 185

65 L>P L>P Lakefield Pormpuraaw 60

55

Haematocrit 50

45

Breeding Breeding

Non-breeding Non-breeding Season

Figure 5.3: Haematocrit levels for Star finches sampled during the non-breeding seasons 2007 and 2008, and the breeding seasons 2008 and 2009. Bars are means +/- SE. Significant differences between properties determined by Tukey –

Kramer HSD are designated with abbreviations for sites (P < 0.05).

186 75 65 55 Lakefield 45 Pormpuraaw 35 25 15 5

-5 Residual CORT (ng/mL)Residual -15 -25 -35 Non-breeding Breeding Non-breeding Breeding

Season

Figure 5.4: Residual CORT levels (ng/mL) for Star finches sampled during the non- breeding seasons 2007 and 2008, and the breeding seasons 2008 and 2009. Bars are means +/- SE. No significant differences between properties were found.

187 300 Lakefield Pormpuraaw 200

100 CBG Capacity CBG (nM)Capacity 0

Breeding Breeding

Non-breeding Non-breeding Season

Figure 5.5: CBG capacity (nM) for Star finches sampled during the non-breeding seasons 2007 and 2008, and the breeding seasons 2008 and 2009. Bars are means +/- SE. No significant differences between properties were found.

188 Differences in Star finch health measures between sites were sometimes dependent on season. Body mass did not vary between sites (F1, 292 = 1.36, P =

0.24), while both muscle and fat scores varied between sites in some seasons

(interaction between site and season for muscle F1, 292 = 4.81, P = 0.03; and fat F1,

292 = 17.48, P < 0.0001). Fat scores were higher on Pormpuraaw compared to

Lakefield during non-breeding seasons and lower on Pormpuraaw during breeding season 2008 only (Tukey – Kramer HSD, Figure 5.2). Muscle scores were higher on Pormpuraaw during the 2009 breeding season only (Tukey – Kramer HSD,

Figure 5.2). Haematocrit was elevated for birds sampled at Lakefield compared to

Pormpuraaw during non-breeding seasons (Tukey – Kramer HSD, Figure 5.3).

Stress response levels did not differ by site (F1, 167 = 0.39, P = 0.53, Figure 5.4).

The effect of site on CBG capacity depended on season (F1, 95 = 5.62, P = 0.02), however the Tukey-Kramer tests could not detect where differences lay (Figure

5.5). Higher haematocrit and lower fat scores at Lakefield during non-breeding were the only consistent health index differences between the two Star finch populations (Table 5.1).

189 Table 5.1: Direction of significant differences in Star finch health indices between sites (Lakefield National Park = L; Pormpuraaw = P) during the two years of the study (determined using Tukey – Kramer HSD). The abbreviation ‘ns’ refers to health indices found to have no significant differences between sites.

Star Finch

Health Index Non-breeding Breeding Non-breeding Breeding

2007 2008 2008 2009

Body mass ns ns ns ns

Muscle ns ns ns L < P

Fat L < P L > P L < P ns

Haematocrit L > P ns L > P ns

CORT ns ns ns ns

CBG ns ns ns ns

190 Black-throated finches

Seasonal patterns in Black-throated finch health measures were sometimes dependent on year. The effect of season on body mass and muscle scores was independent of site and year (residual body mass F1, 220 = 4.56, P = 0.03, muscle

F1, 220 = 26.83, P < 0.0001), with mass and muscle scores consistently higher during breeding seasons compared to non-breeding seasons (Tukey – Kramer

HSD). Black-throated finch seasonal fat score and haematocrit variation were dependent on year (fat F1, 220 = 12.66, P = 0.0005, haematocrit F1, 217 = 8.19, P =

0.005). Both fat scores and haematocrit were greater during breeding compared to the non-breeding season during the second year of the study, while fat was higher during the non-breeding season during the first year of the study at Laudham only

(Tukey – Kramer HSD). The effect of season on stress response was dependent on both site and year (residual CORT F1, 143 = 7.61, P = 0.007), with non-breeding season stress levels consistently higher than breeding seasons both years, but this trend was significantly more pronounced in the first year of the study and on

Laudham (Tukey – Kramer HSD, Figure 5.8).

Most Black-throated finch measures of body condition, haematocrit and stress did not consistently differ between sites. The effect of site on body mass was dependent on year (F1, 220 = 9.03, P = 0.003), with Laudham mass lower than

Lakefield during breeding 2009 only, and similar between sites during other seasons (Tukey – Kramer HSD, Figure 5.6). This interaction is also influenced by the differing annual variation in body mass between the two properties; mass was 191 similar across all years for birds sampled at Lakefield, while birds at Laudham had lower body masses during the second year of the study (Tukey – Kramer HSD,

Figure 5.6). The effect of site on fat was significant (F1, 220 = 14.90, P = 0.0001), with levels often higher on Laudham, especially during the non-breeding season

2007 (Figure 5.5, Table 5.2). Stress levels were significantly higher for birds sampled at Laudham during non-breeding 2007 and again during breeding 2009

(Tukey – Kramer HSD, Figure 5.8). Thus the only consistent difference between the health indices of the two Black-throated finch populations was lower fat scores for birds monitored at Lakefield during non-breeding seasons (Table 5.2).

192 a) body mass b) muscle 3 1.0 L>Lp

0.5 2

0.0 1

-0.5 Residual mass (g)mass Residual muscle score (0 score to muscle 3) 0 -1.0

Breeding Breeding Breeding Breeding

Non-breeding Non-breeding Non-breeding Non-breeding Season Season

c) fat 3 L

1 fat (0score to 5) 0

Breeding Breeding

Non-breeding Non-breeding Season

Figure 5.6: Body condition measures for Black-throated finches sampled during the non-breeding seasons 2007 and 2008, and the breeding seasons 2008 and 2009.

Graphs refer to measures of finch a) body mass, b) pectoral muscle score, and c) furcular fat score. Bars are means +/- SE. Significant differences between properties for each season determined by Tukey – Kramer HSD are designated with abbreviations for sites (P < 0.05).

193 65 L>Lp Lakefield Laudham Park 60

55

Haematocrit 50

45

Breeding Breeding

Non-breeding Non-breeding Season

Figure 5.7: Haematocrit levels for Black-throated finches sampled during the non- breeding seasons 2007 and 2008, and the breeding seasons 2008 and 2009. Bars are means +/- SE. One significant difference between properties determined by

Tukey – Kramer HSD is designated with abbreviations for sites (P < 0.05).

194 L

-10 Residual CORT (ng/mL)Residual -20 -30 -40 Non-breeding Breeding Non-breeding Breeding

Season

Figure 5.8: Residual CORT levels (ng/mL) for Black-throated finches sampled during the non-breeding seasons 2007 and 2008, and the breeding seasons 2008 and 2009. Bars are means +/- SE. Significant differences between properties determined by Tukey – Kramer HSD are designated with abbreviations for sites (P

< 0.05).

195

Table 5.2: Direction of significant differences in Black-throated finch health indices between sites (Lakefield National Park = Lake; Laudham = Laud) during the two years of the study (determined using Tukey – Kramer HSD). The abbreviation ‘ns’ refers to health indices found to have no significant differences between sites.

Black-throated

Finch Non-breeding Breeding Non-breeding Breeding

Health Index 2007 2008 2008 2009

Body mass ns ns ns Lake > Laud

Muscle ns ns ns ns

Fat Lake < Laud ns Lake < Laud Lake < Laud

Haematocrit ns ns Lake > Laud ns

CORT Lake < Laud ns ns Lake < Laud

196 5.4 Discussion

Seasonal variation in stress

The pattern of health index variation for both Star and Black-throated finches sometimes mimicked that noted in the endangered Gouldian finch. Seasonal differences in stress levels were similar, in that all species had significantly higher

CORT levels during non-breeding seasons in comparison to breeding seasons.

This differs from the usual pattern of reduced CORT levels found in non-breeding periods in non-declining finch populations studied in other parts of Australia

(Chapter 2) and in the Northern Hemisphere (Romero 2002). Notably, high levels of CORT have also been recorded outside the breeding season in Mountain chickadees (Poecile gambeli) experiencing unpredictable food supply during times of food shortage (Pravosudov et al. 2001). This suggests that the Gouldian, Black- throated and Star finch populations are all experiencing resource limitation during non-breeding (late dry/early wet) seasons. It is very likely that the grass seeds that finches rely on for food are in shorter supply than historically during the non- breeding (late dry/early wet) season due to changes in the grass layer caused by grazing and changed fire regimes. Recent studies in Queensland concluded that a combination of overgrazing and fire suppression is allowing the proliferation of woody trees and shrubs at the expense of native grasses (Neldner et al. 1997;

Crowley & Garnett 1998, 2001; Garnett et al. 2005). Other studies have directly measured a lower abundance of grass seeds produced by individual plants after grazing or burning (Crowley & Garnett 2001; Letnic 2004). The overall lowered 197 abundance and productivity of native grasses is more likely to limit seed availability towards the end of the dry season, whereas early in the dry season grass seed is probably sufficient in most habitats to maintain health in breeding finch populations

(Andrew & Mott 1983; Crowley & Garnett 1999, 2001). Thus the results of this study could suggest that declining Gouldian, Star and Black-throated finch populations are persisting in degraded habitats, because as food at these sites becomes less available during non-breeding seasons, finches show consistently poorer muscle measures and increased CORT levels. Non-declining Long-tailed and Masked finches did not show higher stress levels during non-breeding seasons, suggesting that these species were better able to adapt to possible changes in food availability compared to declining species.

Seasonal variation in finch health indices

Interestingly, Star finch seasonal fat and haematocrit patterns completely matched the Gouldian finch profile during the first year of the study. Star finch fat, haematocrit and stress were extremely elevated during the non-breeding season

2007, especially for birds sampled at Lakefield. Many non-migratory passerines, including grass-finches, rarely carry much fat (Blem 1976; Breuer et al. 1995;

Gosler 1996; Ewenson et al. 2001; Rozman et al. 2003). It is theorized that this is due to the disadvantage extra weight poses to flight agility, energy demands and predator avoidance (Witter & Cuthill 1993; Brodin 2001). The tendency of Star finches that carry large fat reserves to have high haematocrit suggests that increased weight may increase energy demands, an interpretation that agrees with 198 optimal weight theories (Witter & Cuthill 1993; Brodin 2001). Studies of other passerines show that birds experiencing unpredictable access to food or more limited foraging opportunities show higher levels of fat storage than when food is more accessible (Ekman & Hake 1990; Cuthill et al. 2000; Cresswell 2003). Thus the pattern of higher Star finch fat scores and haematocrit recorded during the stressful non-breeding season fits the general profile of birds experiencing food limitations and is very similar to the non-breeding season pattern of health indices found in Gouldian finches (Chapter 4). However, unlike stress measures patterns seen in Gouldian finches, Star finch CBG capacity was elevated during the high stress non-breeding season 2007. This ‘matching’ of high CORT and CBG capacity is a pattern found in common finches and other passerines (Breuner &

Orchinik 2002b; Wada et al. 2006b), but unlike the opposing CORT and CBG levels seen in Gouldian finches during non-breeding seasons (Chapter 2). Overall, the Star finch health profile was very similar to the generalized health profile of

Gouldian finches during non-breeding 2007, except that Star finches have much higher CBG levels that may help lower the biological activity of their high CORT levels.

In contrast to Star and Gouldian finches, Black-throated finches did not display highly elevated fat and haematocrit levels during the stressful non-breeding season in 2007. Though fat levels often fluctuated between seasons, no Black-throated finch population carried fat levels as high as those of Star finches (Black-throated finch mean scores ≤ 2, while Star finch means ≤ 3.5), and fat levels did not vary consistently between non-breeding and breeding seasons. Also dissimilar to Star 199 finches, no Black-throated finch population showed higher haematocrit during non- breeding seasons. This suggests Star finch populations were more likely to be responding to variation in environmental quality like Gouldian finches and may be more susceptible to or more likely to experience environmental changes than

Black-throated finches.

Site differences finch health indices

Similar to Gouldian finches, Star finches showed differences in haematocrit and fat levels between populations. The elevated haematocrit values of Star finches at

Lakefield during non-breeding may indicate that Lakefield birds have a more energetically demanding lifestyle than birds at Pormpuraaw. Though the habitat at

Lakefield is relatively intact grassland, there is very little cover from predators and there are larger distances between waterholes compared to the densely vegetated and wetter habitat at Pormpuraaw. Star finches at Lakefield may therefore fly further to find food, water and cover each day, which could explain their higher haematocrit (e.g. energy demand) levels. Increased activity could also cause these birds to be less able to accumulate fat. It is unclear from the results of this study as to whether a more energetically demanding lifestyle would cause a population to be more sensitive to environmental change than populations with a more sedentary lifestyle. However, both populations did display variation in condition, haematocrit and stress between seasons and years, suggesting that finches are responding to frequently changing environmental conditions.

200 Black-throated finch health indices did differ between the two study populations, although less consistently than seen in Gouldian and Star finches. It is possible that the much higher CORT levels and fat scores of Black-throated finches sampled at Laudham compared to Lakefield during the non-breeding season 2007 are indicative of a higher sensitivity to seasonal or environmental change for this particular population. Unlike Gouldian finches however, higher Black-throated finch stress levels as indicated by residual CORT did not correspond with lower body mass, muscle scores or differing haematocrit levels between the two Black- throated finch populations. Though health indices rarely differed between the two populations, the threatened subspecies sampled at Laudham exhibited greater variation in health measures than the common northern subspecies at Lakefield.

Measures of Laudham finch body mass, fat, haematocrit and CORT levels varied significantly between seasons or years, while birds sampled at Lakefield only showed significant variation in CORT levels throughout the study. If Laudham is a more unstable environment compared to Lakefield, this would explain why Black- throated finches at Laudham show more variation in health indices. This could be attributed to possible differences in the quality of finch habitat due to grazing, as

Laudham is a working cattle station with moderate to high stocking rates, while

Lakefield is a National Park with few introduced herbivores, including cattle.

Inter-annual patterns in finch health indices

Surprisingly, all Queensland finch populations showed higher than mean stress levels in 2007 suggesting a regional environmental factor was responsible. Early 201 in 2007, wet season rainfall totals were average, or above average, for all sites.

However, an unusually high amount of rain (> 20 mm at Pormpuraaw and

Lakefield, > 100 mm at Laudham) was recorded across the dry savanna country of northern Queensland in late June. This habitat does not typically receive any significant rain between May and October (Commonweath of Australia, Bureau of

Meterology 2011). It is possible that unseasonal rain may have disrupted the abundance of grass seeds later in the year by promoting premature and extensive germination, rotting or burial of seeds that would have otherwise made up part of the surface seed bank that is available to finches (Andrew & Mott 1983; McIvor &

Gardner 1991; Crowley & Garnett 1999). Because the amount and timing of rainfall was not sufficient to cause seeding of grasses, it is not likely that the seeds lost to germination, decomposition or burial were replaced within that dry season

(McIvor & Gardner 1991). Thus it is possible that a widespread rain event led to a shortfall of preferred seed in the late dry season 2007, which in turn would lead to higher stress and poorer body condition among all four populations of Queensland grass-finches in this study.

Conclusion

Both populations of Star finches were prone to large fluctuations in body condition, haematocrit and stress between seasons and years similar to the health measure fluctuations experienced by Gouldian finches. However, only one population of

Black-throated finches showed significant variation in these health indices. Thus, not all Queensland finches are likely to be as sensitive to environmental changes 202 as Gouldian finches. However, there are sufficient similarities in the pattern of annual variation in health measures among all species to identify food limitation as a likely cause of their lowered health indices during some seasons. The drop in body mass and muscle scores and increase in CORT levels and haematocrit seen in all Queensland populations in late 2007 suggests that all finches responded poorly to a regional change in environmental conditions. In order to buffer remaining populations from similar detrimental circumstances, such as unseasonal rain, overgrazing or wildfire, care could be taken to minimize finch stress during the non-breeding season by increasing grass seed production in the remaining habitats of threatened subspecies. Monitoring the subsequent levels of variation in finch health measures could then give managers an indication of the effectiveness of improved conservation land management.

203 6.0 Chapter Six: Conclusion and General Discussion

6.1 Savanna finch health variation offers insight into conservation priorities

Background and summary of findings

Over a quarter of Australia’s granivorous birds have declined in abundance or experienced range contractions over the last 50 years (Franklin et al. 2005). It is widely assumed that cattle grazing and changed fire regimes have lowered the quality of savanna habitats for some grass-seed eating species but the relative impacts of changed land use on declining and common finches remain poorly defined because of difficulties in monitoring the abundance of these semi-nomadic birds. In this thesis, I have instead compared health measures of declining and non-declining finch species to determine if the timing and severity of changes in health indicators coincide with differing land tenures. Monitoring of the health of finches in the northern Australian savanna suggests that all species experience more physiological stress during non-breeding seasons compared to breeding seasons. Evidence of poorer health was found in declining species studied on grazed and frequently burnt land. Correlations between poor finch health, grazing, burning and lowered grass productivity on pastoral and aboriginal land tenures has important implications for conservation management of finches.

204 Seasonal heath changes in finches and conservation implications

All finches displayed signs of poorer health during non-breeding compared to breeding seasons. This is seen in the lowered muscle scores of all species and lowered mass of most species during the non-breeding or late dry/early wet season

(Figure 6.1). Significantly higher stress responses during non-breeding were found for declining Gouldian, Star and Black-throated finches but not non-declining Long- tailed and Masked finches (Figure 6.2). Thus declining finch species showed more evidence of poor health during non-breeding seasons than non-declining finch species.

205

a) c) 0.4 0.4

0.2 0.2

0.0 0.0

-0.2 -0.2 Residual mass (g) mass Residual (g)mass Residual

-0.4 -0.4 Breeding Non-breeding Breeding Non-breeding

b) 2.5 d) 2.5 Gouldian Star Finch Finch Long-tailed Black-throated Finch Finch Masked 2.0 2.0 Finch

muscle scoremuscle (0-3) scoremuscle (0-3) 1.5 1.5 Breeding Non-breeding Breeding Non-breeding

Season Season

Figure 6.1: Seasonal changes in mean body mass and muscle scores of finches

(Taken from data used in Chapters 2 and 5). Graphs represent, a) Gouldian,

Long-tailed, and Masked finch body mass, b) Gouldian, Long-tailed, and Masked finch mucle scores, c) Star and Black-throated finch body mass, d) Star and Black- throated finch muscle scores. Measures are means +/- SE.

206 a) b) 10 10

5 5

0 0

-5 -5

Residual CORT (ng/mL)Residual Residual CORT (ng/mL)Residual -10 -10 Breeding Non-Breeding Breeding Non-Breeding d) 10 c) 30 20 5 10

0 0 -10

-5 -20 -30

CORT (ng/mL)Residual -10 Breeding Non-Breeding CORT (ng/mL)Residual -40 Breeding Non-Breeding

e) 30 20 10 0

-10 -20

-30 Residual CORT (ng/mL) Residual -40 Breeding Non-Breeding

Figure 6.2: Seasonal changes in mean stress response levels of finches (Taken from data used in Chapters 2 and 5). Graphs represent stress response levels for a) Gouldian finches, b) Long-tailed finches, c) Masked finches d) Star finches, and e) Black-throated finches. Measures are means +/- SE. Note the much larger range of stress response scores in d) and e) which necessitate the larger range of values on the x-axis of these graphs. 207

There are two main implications of these findings: first, all species’ survival probability may be lower during the non-breeding and moulting season than peak breeding season, and second, declining species are more likely to have significantly lowered survival than non-declining species. Researchers wishing to learn more about what environmental factors are limiting finch survival in the

Northern Territory and Queensland should therefore design studies that focus on the early non-breeding season, as mortality is likely to be higher during this time of year compared to peak breeding. The Gouldian finch Recovery Plan has called for the design and implementation of more effective population monitoring at key sites

(O'Malley 2006a). The present research suggests annual health monitoring would be more likely to show which populations are likely to recover or decline by focusing survey efforts during the early non-breeding season, when finches are switching from sprouting annual grass seed food to ripening perenial grasses or other food sources. It is possible that other times of the year could also be limiting, particularly the wet season and early breeding. During this time of year, finch diets are composed of mostly perennial grass seed, which is likely to be reduced by the effects of grazing and fire. However, monitoring finches during the wet season is logistically difficult in remote areas due to wet weather limitations, lack of site access and possible dispersal of finch flocks. It may therefore be less problematic to monitor finches early in the wet or late dry season in many remote areas.

Conservation managers wishing to protect finch habitat and promote healthier finch populations should focus land care around improving preferred finch grass-seed 208 food resources during the late dry and wet seasons. The diets of declining finch species are relatively well studied (Garnett & Crowley 1994; Dostine & Franklin

2002; Todd et al. 2003; Garnett et al. 2005) and can be used to determine the priority of which grasses to protect and promote. For example, the Gouldian finch recovery plan has already proposed improving fire and grazing management at key sites, and detailed guidelines would most likely include protecting larger areas of the annual Sarga spp., and Trioda spp., and Chrysopogon fallax, Alloteropsis spp. and other important perennial grasses from the defoliating effects of grazing and fire (O'Malley 2006a). However, this will be logistically difficult, even on conservation managed properties due to the large scale necessary and lack of funds and infrastructure currently available in northern Australia (Woinarski &

Fisher 2003; Woinarski et al. 2007).

Heath differences between populations on different land tenures

Declining populations of finches showed more variation between land tenures compared to non-declining populations, some of which suggested birds on land not managed for conservation were in poorer health than those on land managed for conservation. All Gouldian finch body measures were lower, haematocrit was abnormally elevated and stress responses were sometimes higher on pastoral and aboriginal land during the non-breeding (late dry/early wet) season 2007 (Figures

6.3, 6.4 and 6.5). Long-tailed and Masked finches health measures had no consistent differences between tenures, though both species had higher fat scores,

Long-tailed finches had higher haematocrit and Masked finches had higher stress 209 on all land tenures during the non-breeding season 2007 (Chapter 4). These results suggest that 2007 was a difficult season for all Northern Territory finches and a particularly stressful season for Gouldian finches persisting on pastoral and aboriginal land. However, few significant differences were found between any

Northern Territory finch populations on different land tenures during 2008 (Chapter

4). Star and Black-throated finch populations in Queensland also exhibited high yearly variation in body condition and stress measures (Chapter 5), suggesting high variation in the probability of decline between years. If this is true, several years of poor environmental conditions could be extremely detrimental to any declining finch population, as all species are relatively short lived (Higgins et al.

2006). This annual variation in finch population condition highlights the need for recovery efforts that are well funded and consistent between years and the fact that short-term monitoring may not be sufficient in determining finch population responses to land management changes.

210

a) Body Mass b) Muscle

1.0 d>p=a 2.5 d>p

0.5 2.0 1.5 0.0 1.0 -0.5

0.5 Residual mass (g)mass Residual

-1.0 (0score muscle to 3) 0.0 2007 2008 2007 2008 Year Year

c) Fat

2.5 d>a d=p>a Defence 2.0 Pastoral Aborignal 1.5

1.0

0.5 fat (0score to 5) 0.0 2007 2008 Year

Figure 6.3: Body condition measures for Gouldian finches sampled during the non- breeding seasons 2007 and 2008 (Reproduced from Figure 4.2). Graphs refer to measures of finch a) body mass, b) pectoral muscle score, and c) furcular fat score. Measures are means +/- SE. Significant differences between land management types are signified with abbreviations for sites (Tukey – Kramer HSD,

P < 0.05).

211

70 d

60

55

50

Heamatocrit (% red cells/plasma) blood 2007 2008

Year

Figure 6.4: Haematocrit levels for Gouldian finches sampled during the non- breeding seasons 2007 and 2008 (Reproduced from Figure 4.3). Measures are means +/- SE. Significant differences between land management types are signified with abbreviations for sites (Tukey – Kramer HSD, P < 0.05).

212

d

a d

Aboriginal 10

5

0

Residual Cort (ng/mL) Cort Residual -5

-10 2007 2008 Year

Figure 6.5: Residual CORT levels (ng/mL) for Gouldian finches sampled during the non-breeding seasons 2007 and 2008 (reproduction of Figure 4.4). Measures are means +/- SE. Significant differences between land tenures are signified with abbreviations for sites (Tukey – Kramer HSD, P < 0.05).

213 In Queensland, Star and Black-throated finches show the same symptoms of high stress during non-breeding seasons as Gouldian finches, which could be an indicator that all three species are reacting strongly to seasonal changes in habitat quality. Unlike Gouldian finch indices, Star and Black-throated finch body condition, haematocrit and stress levels did not often differ between properties

(Chapter 5). This similarity of health measures between land tenures suggests that all Queensland properties had limited resources during non-breeding seasons or these species are better adapted to degraded environments compared to Gouldian finches. However, some properties showed high seasonal variation in health measures while others did not. Star finches sampled at Lakefield National Park

(conservation managed land) had higher variation in haematocrit, while those sampled at Pormpuraaw (aboriginal land) had higher variation in fat. Black- throated finches monitored at Laudham (pastoral land) had seasonal variation in all health measures while those monitored at Lakefield National Park only had only seasonally variable stress measures. The higher health variation between seasons on some properties suggests the environment is more variable on those properties and may be slightly lower in quality to finches.

There are several important implications of the finding that declining finch species show evidence of poorer health on land not managed for conservation. Firstly, declining species populations currently existing in disturbed areas are more likely to be prone to decline than those in conservation managed areas. Secondly, declining species may be in danger of further population declines if land currently managed for conservation is converted to pastoral or fire-prone tenures. 214 Fortunately, both the federally funded Gouldian finch Recovery Plan as well as plans for declining finches in the state of Queenland aim at encouraging land management practices that are benefictial to finches (O'Malley 2006a; Dorricott &

Garnett 2007). For example, the Gouldian Finch Recovery Plan calls for land managers on key properties to implement small, patchy, early dry season burns as a method of creating fire breaks which can protecting important finch habitat from larger, more intense wildfires during the late dry season (O'Malley 2006a). Both

Gouldian and Black-throated Finch plans encourage pastoral managers to spell finch habitats from grazing during the wet season in order to allow perennial grasses to seed and recover from grazing that occurs during the dry season

(O'Malley 2006a; Black-thorated Finch Recovery Team et al. 2007). The findings of the present study suggest that recovery efforts should continue to give high priority to decreasing grazing and fire impacts on sites important to declining finch populations.

Poor finch health coincides with lower grass cover and productivity

Coinciding differences were found between finch health and grass layer properties on differing land tenures in the Northern Territory. Land not managed for conservation had lower grass layer cover and productivity than land managed for conservation (Figures 6.6 and 6.7). Grass cover and productivity were especially low on land not managed for conservation during 2007, the year when Gouldian finch health measures were the poorest on these land tenures (Figures 6.3, 6.4 and 6.5). 215 70 c>p=a c>a Conservation land 60 Pastoral land 50 Aboriginal land 40 30 20

10 Grass Cover %Cover Grass of plots 0 2007 2008 Year

Figure 6.6: Mean percent cover of grasses on plots by land management type and year (taken from Figure 3.2). Bars are means +1 standard error, significant differences between properties are designated with abbreviations for sites (P <

0.05).

1.0 a>c>p Conservation land 0.8 Pastoral land Aboriginal land 0.6

0.4

0.2

0.0 2007 2008

(m) Cover Grass of Height Mean Year

Figure 6.7: Mean height of grasses on plots by land management type and year

(taken from Figure 3.3). Height is a general measure of grass productivity. Bars are means +1 standard error, significant differences between land types are designated with abbreviations for sites (P < 0.05).

216

Gouldian finches are known to feed almost exclusively on the grass seeds of the annual Sarga spp. during the dry season (Dostine & Franklin 2002). The finding that Sarga spp. cover was significantly lower during 2007 on pastoral land compared to other tenures and lowest during this year compared to other years

(Chapter 3), coincides with the poor health measures seen in Gouldian finches on this property (Figures 6.3, 6.4 and 6.5). While birds on conservation and aboriginal managed land were feeding on Sarga spp., Gouldian Finches on pastoral land were observed feeding mainly on grass seeds other than Sarga spp. (Aristida spp. and Schizachyrium spp.) during the late dry 2007 (personal observation). Because

Sarga spp. is not used as forage by cattle (Andrew & Mott 1983), and the areas that loss cover of Sarga spp. were not burnt, I speculate that the removal of this grass cover may have been assisted by cattle trampling the grass and soil, which could have also assisted in the burial of much of the Sarga spp. seeds beneath the soil and making them unavailable to finches. Unseasonal rainfall can also assist in the reduction of grass seed availability (Andrew & Mott 1983), and coincidentally, there was over 25 mm of rainfall across northern Australia in late June 2007, during the middle of the dry season. Both unseasonal rain and disturbance by cattle may have reduced grass seed availability below average and are therefore identified as possible causes of the poor body condition, increased haematocrit

(energy demands) and extremely elevated stress responses of Gouldian finches on this tenure (Figures 6.3, 6.4 and 6.5). More study on the factors that affect Sarga spp. seed availability during the dry season is needed in order to indentify the main causes of reduced seed availability on pastoral land. 217

Notably, all finch species showed some evidence of increased stress responses during the non-breeding season of 2007 (Chapters 2 and 5). It is possible that unseasonal rainfall during this dry season reduced the availability of grass seeds across most of the savannas of northern Australia. If true, the lower abundance of food is a likely cause of increased stress in finches. The two finch populations with the highest stress responses during late 2007 were Gouldian and Black-throated finch populations on highly grazed pastoral land (Chapters 4 and 5). The finding that Gouldian finch populations on regularly burnt land had similarly poor body condition and high energy demands (Figures 6.3 and 6.4) but lower stress responses than finches on grazed pastoral land during this season (Figure 6.5) suggests that the effect of grazing on one property had more of an effect on finch stress responses than did fire on a different property. Research in South Africa has suggested that grazing can limit the seed production and seed banks of certain grasses in favor of small annual grasses such as Aristida spp. and Urochloa spp.

(O'Connor & Pickett 1992). These findings lead me to speculate that although both grazing and fire remove important grass cover, grazing is sometimes more detrimental to grass seed availability because of the added effect of seed trampling, or some other unknown factor. Again, more research on the actual effects of fire or grazing on grass seed availability are necessary to fully understand the probability of either effect on the late dry season survival probability of finches. Historically, recovery plans have focused predominantly on the reduction of inappropriate fire regimes at key finch conservation sites (O'Malley

2006a; Dorricott & Garnett 2007; Team et al. 2007); however, the high stress 218 measures and poor body condition of declining finches on pastoral land found in this study is good evidence that the effects of grazing are just as detrimental to finches and similarly high priority should be given to protecting finch habitat from these effects.

Though sometimes circumstantial, the results of the present study suggest that lower abundance and productivity of grasses may lower the health of the finches that rely on these grasses for food. Protection of key grass species from disturbance is therefore likely to be imperative in preserving and encouraging the recovery of declining finch species. Interestingly, several of the grass species likely to be important to finches are not important cattle forage in the areas where these finches were found, particularly the annual Sarga spp. and Triodia spp.

(Craig 1994; Ash et al. 1997; Rice & Westoby 1999). Both Sarga spp. and Triodia spp. often occur in areas where soils cannot support grass species favored by cattle, or in high rocky country not easily accessed by cattle or humans (Fensham

& Skull 1999; Dyer et al. 2001). This lack of conflicting interests between pastoral and native fauna use suggests that these grasses could be better protected on this type of land tenure to the advantage of biodiversity. Similarly, protecting these grasses from frequent burning may not necessarily conflict with the management interests of traditional owners. The challenge to conservationists is to design incentives for landowners to protect biodiversity on agricultural and traditional land.

219 6.2 Improving the use of physiological measures for gauging health status of free-living birds

I aimed to monitor the health of Australian grass-finches across the northern savanna to better understand the processes contributing to granivore population declines. By conducting this research I added to the knowledge of how physiological measures of bird health co-vary and change in relation to environmental conditions. The study has further validated the use of physiological indices in assessing bird health status but more research is needed to identify variation in the health of other free-living bird populations in different habitats and during different seasons.

The factors and circumstances affecting a bird’s physiological status are extremely varied, making interpretation of individual health indices difficult. Convincing interpretation of each health index involves cross correlation with as many bird and habitat variables that can reasonably be collected in the field. Examples of variables known to affect interpretation of health indices include individual bird age, sex, diet, breeding status, social status, parasite load and stress history, and environmental factors such as climate, weather, food availability and distribution, and intensity of predation (Maxwell & Robertson 1998; Romero 2004; Stevenson &

Woods 2006; Fair et al. 2007). This interpretation is also reliant on understanding how these variables are likely to affect health measures and the temporal correspondence between these measures and habitat or physiological disturbance

(Maxwell & Robertson 1998; Romero 2004; Stevenson & Woods 2006; Fair et al. 220 2007). By combining physical and physiological information on the individuals being sampled with information on habitat characteristics at the time, we may be able to identify which heath indices discriminate habitat variability best. For example, the results of this study suggest that Gouldian and Black-throated finch stress response measures taken during the non-breeding season vary between land tenures more often than body condition measures (Chapters 4 and 5). Thus stress reponse may be a more sensitive measure to use when studying differences in the health of non-breeding finches living in different habitats. Ideally, the extent to which these indices accurately reflect the overall health of individual animals as well as whole populations must also be determined. More study is needed in order to validate and better describe the connection between many new health indices

(stress, immune function) and actual individual survival probability and breeding success (Tella et al. 2002; Hanssen et al. 2004; Romero 2004). Information on the predictive value of these new measures will greatly aid and substantiate future and existing interpretations of health indices.

Many physiological studies of free-living birds have focused on migratory species and those living in temperate or highly seasonal environments. In such circumstances, physiological changes are often dramatic, occurring synchronously in populations in response to seasonally predictive environmental cues, notably photoperiod (Lindstrom & Piersma 1993; Holberton 1999; Romero & Rich 2007).

The relatively small amount of information on the physiology of tropical and arid- adapted birds suggests that lower environmental predictability of food resources leads to greater flexibility in the population timing of life history events such as 221 breeding and moult (Foster 1975; Hau 2001). These birds tend to adjust body condition, stress levels and immunity in different patterns and in response to different stressors than their higher latitude conspecifics (Rozman et al. 2003;

Martin et al. 2005; Wada et al. 2006b; Martin et al. 2007). This study has shown that not all tropical grass-finches adjust stress levels differently from temperate birds; the non-declining Long-tailed finch has a similar seasonal pattern of CORT response and CBG capacity balance to northern latitude passerines (Figure 6.2).

However, declining Gouldian finches have a seasonal pattern opposite to most common temperate bird species (Figure 6.2). It is likely that many more specific differences exist between temperate and tropical, or declining and non-declining birds, but more study is needed to identify them.

Though there is information on how birds react to changes in food availability, much of the research involves experimental removal of food to induce fasting or comparison of migratory birds using habitats of varying quality (Stuebe & Ketterson

1982; Rees et al. 1985; Ekman & Hake 1990; Totzke et al. 1999; Lindstrom et al.

2000; Brown & Sherry 2006). Free-living birds that do not migrate long distances and captive birds are likely to be responding to different constraints due to the fact that they do not actually choose between habitats but adapt to local environments as conditions change. This could result in a wide range of differences in physiological patterns between captive, free-living, migratory and resident birds.

Some of these differences have already been examined experimentally. Juncos

(Junco hyemalis) translocated from high to low elevation, and wild Zebra finches

(Taenopygia guttata) brought into captivity both decreased their body weight and 222 fat storage to match the lower fat and mass of birds already adjusted to the new environments (Rogers et al. 1993; Ewenson et al. 2001). These examples demonstrate the common response of birds to reduced activity in the wild and captivity. Similarly, it is thought that stress, haematological condition and immune responses will vary according to life-style characteristics. Lifestyle differences are implicated in several of the health pattern differences seen between grass-finch species. For example, differences in health variation between the declining

Gouldian finch and non-declining Long-tailed finch can be explained using differences in their lifestyle traits. In general, Gouldian finches have a more strictly timed style of breeding and moulting, as well as a more specialized diet compared to Long-tailed finches (Tidemann & Woinarski 1994; Franklin et al. 1998; Dostine &

Franklin 2002). This more strictly timed and specialized lifestyle is likely to put more physiological demands on finches and may explain why Gouldian finches show more evidence of seasonally poor health on land tenures with disturbed habitat compared to the generalist Long-tailed finch (Chapters 2 and 4).

The relative changes in bird physiology due to increased energy demands for free- living birds during different breeding stages and during migration are relatively well known (Morton 1994; JenniEiermann & Jenni 1996; Horak et al. 1998b; Lindstrom et al. 2000; Landys-Ciannelli et al. 2002; Kern et al. 2007). This is especially true of the physiological changes necessary for migratory flights. We know that birds must increase their blood oxygen carrying capacity, closely regulate stress levels to ensure fat and muscle accumulation before flights, and optimize metabolite levels and body condition (Schwabl et al. 1991; Lindstrom et al. 2000; Piersma et 223 al. 2000b; Landys-Ciannelli et al. 2002; Landys et al. 2004; Guglielmo et al. 2005;

Jenni et al. 2006). It is possible that non-migrants may also show similar patterns of haematological, stress, body condition, and metabolite optimization during periods of elevated energy demands. For example, a Star finch population living in open grassland habitat, which would require them to fly long distances for food and cover each day, tended to have higher haematocrit and lower fat scores compared to relatively sedentary population living in dense cover near abundant food

(Chapter 5). Based on the theory of optimized mass, I hypothesized that a more energetic lifestyle could be the cause of the grassland population’s high haematocrit and could also make carrying higher fat reserves more maladaptive and/or difficult to obtain (Lima 1986). Studies of captive Zebra finches have found lowered body condition to be associated with increased activity when food availability was decreased (Birkhead et al. 1998). However, the decreased food availability alone could account for the changes in bird health, while increased activity might simply reflect the attempts of these caged birds to locate food. Thus, distinguishing the effect of food shortage versus higher activity on condition indices requires careful experimental design. Studies on free-living birds and captive individuals presented with rigorous increases in exercise without deficient diets may be needed to tease apart the effects of increased energy expenditure and decreased food availability.

The acute stress response in birds is also well studied under a wide range of conditions in captive and free-living birds. However, questions still remain regarding the identification of chronic stress in free-living birds and what conditions 224 lead to biologically costly stress levels. For example, chronic psychological disturbance of captive wild-breeding birds was associated with suppression of their baseline and stress-induced CORT secretion, as well as reduced breeding success

(Cyr & Romero 2006). It is likely that these treatments mimic conditions experienced by free-living birds in areas with high predation pressure. It remains to be seen whether birds presented with stressors which are less acute would display the same symptoms of stress-induced down-regulation of the glucocorticoid axis. The results of this study show that three declining finch species (Gouldian,

Star and Black-throated finches) had high CORT levels in response to capture and sometimes lowered CBG levels during a possible period of low food availability and while birds were not breeding (Figure 6.2, Chapter 4 and 5). This mimics the pattern seen in several studies of the effects of lowered food availability of both wild and captive birds (Astheimer et al. 1995; Pravosudov et al. 2001; Lynn et al.

2003a). Increased stress response combined with lowered CBG levels may represent the typical pattern of response to chronically lowered food availability during non-breeding. However, food-restricted populations of breeding Kittiwakes

(Rissa tridactyla) show evidence of stress response suppression compared to well- fed populations (Kitaysky et al. 1999). More detailed knowledge of the effects of changes in food availability during breeding and non-breeding periods would help answer the question of whether down-regulation of stress is a symptom of exposure to all chronic stressors during all life stages or whether avian responses to chronic stressors differ depending on the type of stressor and seasonal timing.

225 Studies have attempted to describe haematological change in response to climate, dehydration, parasite infection, diet and energy demands (Fair et al. 2007).

However, we still know little about the time scale involved between disturbance and response of haematological measures. Dehydration causes significant changes in plasma volume and haemoglobin levels in captive pigeons (Columba livia) and chickens (Gallus domesticus) only after many hours of food and water deprivation at high ambient temperatures (Carmi et al. 1994; Zhou et al. 1999). However, few of the haematological effects of dehydration in free-living birds have been described, leaving researchers unsure of the significance of variation in blood volume or haemoglobin levels under conditions where birds may be prone to dehydration. Though the lack of correlation between finch plasma osmolality and haematocrit readings suggest finches in this study were not suffering from dehydration (Chapter 2), other studies may benefit from more detailed information on what conditions are likely to cause significant variation in haematological measures due to dehydration in other passerines.

Blood parasites also seem to have a variety of effects on blood cell numbers and volume. While some parasites cause increases in haematocrit, others cause decreases; again leaving researchers unsure about interpretation of measures in infected birds (Fair et al. 2007). For example, almost nothing is known about the physiological effects of blood parasites found in Australian finches because of both a lack of research and the extreme rarity of infections found in wild finch populations (Tidemann et al. 1992). In contrast, there is a wealth of information on how diet and energy demands affect relative blood measures of temperate-zone 226 and migratory species with low parasite infection rates. This is due to the large number of studies on haematological measures in relation to fasting, breeding effort, and migration biology based in the Northern Hemisphere. Further studies are needed to validate these measures as useful tools in the understanding of tropical and arid climate species likely to experience different environmental extremes than Northern Hemisphere birds.

Ecologists are often able to find differences in the physiological state of individuals and populations. Without clearer information on what each physiological state or pattern of changes in physiological state means in terms of relative bird productivity and survival, the usefulness of these discoveries is diminished. Further validation of avian condition measures in relation to their breeding success, survival and environmental conditions will greatly benefit the fields of conservation ecology and evolutionary biology. Changes in body condition, metabolic status, haematological condition and stress level are some of the most obvious biomarkers for establishing health indices. What is needed, however, is to validate these indices in a controlled environment under captive conditions to ensure they accurately identify an animal’s health, breeding potential and survival probability under free- living conditions.

6.3 Concluding summary

This study attempts to better understand the processes contributing to granivore population declines in order to inform the management of Australia’s tropical 227 savannas. The results of this study suggest grazing and fire can lower the cover and productivity of grasses important in finch diets (Chapter 3). Low cover and productivity of grass layers on frequently grazed and burnt land coincided with

Gouldian finch health measures that predicted lower finch productivity and survival on disturbed land compared to conservation managed land (Chapter 4). This finding is consistent with my first hypothesis; that high levels of grazing and burning are likely to be detrimental to the health of finches. The finding that non-declining

Long-tailed and Masked finches did not show as much evidence of poor health during stressful seasons and on grazing or fire prone land compared to declining

Gouldian, Star or Black-throated finches is consistent with my second hypothesis; that declining finches are more sensitive to seasonal changes in habitat quality and the environmental effects of grazing and fire (Chapters 2, 4 and 5). By corroborating these hypotheses, we are closer to determining when and where these finch species are experiencing suboptimal conditions for breeding and survival and are better able to aid in their conservation. More generally, this research adds to the knowledge of how physiological measures of bird health co- vary and change in relation to environmental conditions. Such knowledge has been especially lacking in the southern hemisphere, and in tropical environments.

Further investigation of how physiological indices can be used in concert to identify health differences among populations living in different habitats and during different seasons would be helpful in validating many of the assumptions of this and other ecophysiology studies.

228 7.0 References

Alexander, S. L., and C. H. G. Irvine. 1998. The effect of social stress on adrenal

activity in horses: the importance of monitoring corticosteroid-binding

globulin capacity. Journal of Endocrinology. 157, 425-432.

Allen, L. R. L., and I. D. Hume. 1997. The importance of green seed in the nitrogen

nutrition of the zebra finch Taeniopygia guttata. Austral Ecology. 22, 412-

418.

Andersen, A. N. 1991. Responses of ground-foraging ant communities to three

experimental fire regimes in a savanna forest of tropical Australia.

Biotropica. 23, 575-585.

Andersen, A. N., G. D. Cook, L. K. Corbett, M. M. Douglas, R. W. Eager, J.

Russell-Smith, S. A. Setterfield, R. J. Williams, and J. C. Z. Woinarski. 2005.

Fire frequency and biodiversity conservation in Australian tropical savannas:

implications from the Kapalga fire experiment. Austral Ecology. 30, 155-167.

Andersson, M. S., and L. Gustafsson. 1995. Glycosylated haemoglobin: A new

measure of condition in birds. Proceedings: Biological Sciences. 260, 299-

303.

Andrew, M., and J. Mott. 1983. Annuals with transient seed banks: the population

biology of indigenous Sorghum species of tropical north-west Australia.

Australian Journal of Ecology. 8, 265-276.

Arad, Z., M. Horowitz, U. Eylath, and J. Marder. 1989. Osmoregulation and body

fluid compartmentalization in dehydrated heat exposed pigeons. American

229 Journal of Physiology, Regulatory Integrative and Comparative Physiology.

251, R377-R382.

Ash, A., J. Mcivor, J. Mott, and M. Andrew. 1997. Building grass castles:

Integrating ecology and management of Australia's tropical tall-grass

rangelands. The Rangeland Journal. 19, 123-144.

Ash, A. J., J. Gross, and S. M. Smith. 2004. Scale, heterogeneity and secondary

production in tropical rangelands. African Journal of Range and Forage

Science. 21, 137-145.

Ash, J. A., and J. P. Corfield. 1998. Influence of pasture condition on plant

selection patterns by cattle: its implications for vegetation change in a

monsoon tall-grass rangeland. Tropical Grasslands. 32, 178-187.

Ash, J. A., and J. G. McIvor. 1998. How season of grazing and herbivore selectivity

influence monsoon tall-grass communities of northern Australia. Journal of

Vegetation Science. 9, 123-132.

Astheimer, L. B., W. A. Buttemer, and J. C. Wingfield. 1992. Interactions of

corticosterone with feeding, activity and metabolism in passerine birds.

Ornis Scandinavica. 23, 355-365.

Astheimer, L. B., W. A. Buttemer, and J. C. Wingfield. 1994. Gender and seasonal

differences in adrenal response to ACTH challenge in an arctic passerine,

Zonotrichia leucophrys gambelii. General Comparative Endocrinology. 94,

33-43.

Astheimer, L. B., W. A. Buttemer, and J. C. Wingfield. 1995. Seasonal and acute

changes in adrenocorticol responsiveness in an arctic-breeding bird.

Hormones and Behaviour. 29, 442-457. 230 Astheimer, L. B., and W. A. Buttemer. 2002. Changes in latitude, changes in

attitude: a perspective on ecophysiological studies of Australian birds. Emu.

102, 19-27.

Averbeck, C. 1992. Hematology and blood-chemistry of healthy and clinically

abnormal great black-backed gulls (Larus marinus) and herring gulls (Larus

argentatus). Avian Pathology. 21, 215-223.

Barrett, G., A. Silcocks, S. Barry, R. Cunningham, and R. Poulter 2003. The New

Atlas of Australian Birds. Royal Australian Ornithologists' Union, Victoria,

Australia.

Bastin, G. N., J. A. Ludwig, R. W. Eager, A. C. Liedloff, R. T. Andison, and M. D.

Cobiac. 2003. Vegetation changes in a semiarid tropical savanna, northern

Australia: 1973-2002. Rangeland Journal. 25, 3-19.

Bayliss, P., and K. M. Yeomans. 1989. The distribution and abundance of feral

livestock in the Top End of the northern territory (1985-1986), and their

relation to population control. Australian Wildlife Research. 16, 651-676.

Bednekoff, P. A., and J. R. Krebs. 1995. Great tit fat reserves: effects of changing

and unpredictable feeding day length. Functional Ecology. 9, 457-462.

Bennett, L. T., T. S. Judd, and M. A. Adams. 2002. Growth and nutrient content of

perennial grasslands following burning in semi-arid, sub-tropical Australia.

Plant Ecology. 164, 185-199.

Bernstein, M. 1971. Cutaneous water loss in small birds. The Condor. 73, 468-469.

Birkhead, T. R., F. Fletcher, and E. J. Pellatt. 1998. Sexual selection in the zebra

finch Taeniopygia guttata: condition, sex traits and immune capacity.

Behavioural Ecology and Sociobiology. 44, 179-191. 231 Birkhead, T. R., F. Fletcher, and E. J. Pellatt. 1999. Nestling diet, secondary sexual

traits and fitness in the zebra finch. Proceedings of the Royal Society B:

Biological Sciences. 266, 385.

Bleeker, M., S. A. Kingma, I. Szentirmai, T. Székely, and J. Komdeur. 2005. Body

condition and clutch desertion in penduline tit Remiz pendulinus. Behaviour.

142, 1465-1468.

Blem, C. R. 1976. Patterns of lipid storage and utilization in birds. American

Zoologist. 16, 671-684.

Blount, J. D., N. B. Metcalfe, K. E. Arnold, P. F. Surai, G. Devevey, and P.

Monaghan. 2003. Neonatal nutrition, adult antioxidant defences and sexual

attractiveness in the zebra finch Proceedings of the Royal Society B:

Biological Sciences. 270, 1691-1696.

Bonier, F., P. R. Martin, K. S. Sheldon, J. P. Jensen, S. L. Foltz, and J. C.

Wingfield. 2007. Sex-specific consequences of life in the city. Behavioural

Ecology. 18, 121-129.

Booth, C. E., and P. F. Elliott. 2002. Hematological responses to hematozoa in

North American and neotropical songbirds. Comparative Biochemistry and

Physiology - Part A: Molecular & Integrative Physiology 133, 451-467.

Bortolussi, G., J. Mcivor, J. J. Hodgkinson, S. G. Coffey, and C. A. Holmes. 2005.

The northern Australian beef industry, a snapshot. 4. Condition and

management of natural resources. Australian Journal of Experimental

Agriculture. 45, 1109-1120.

232 Bowman, D. M. J. S., O. Price, P. J. Whitehead, and A. Walsh. 2001. The

'wilderness effect' and the decline of Callitris intratropica on the Arnhem

Land Plateau, northern Australia. Australian Journal of Botany. 49, 665-672.

Box, J., A. Lill, and J. Baldwin. 2002. Seasonal variation in body mass and blood

oxygen carrying capacity of the superb fairy-wren (Malurus cyaneus)

Australian Journal of Zoology. 50, 313-323.

Bradstock, R. A., J. E. Williams, and M. A. Gill 2002. Flammable Australia: the fire

regimes and biodiversity of a continent. Cambridge University Press,

Cambridge, New York.

Breuer, K., A. Lill, and J. Baldwin. 1995. Hematological and body-mass changes of

small passerines overwintering in south-eastern Australia. Australian

Journal of Zoology. 43, 31-38.

Breuner, C. W., J. C. Wingfield, and L. M. Romero. 1999. Diel rhythms of basal and

stress-induced corticosterone in a wild, seasonal vertebrate, Gambel's

white-crowned sparrow. Journal of Experimental Zoology. 284, 334-342.

Breuner, C. W., and M. Orchinik. 2001. Seasonal regulation of membrane and

intracellular corticosteroid receptor in the house sparrow brain. Journal of

Neuroendocrinology. 13, 412-420.

Breuner, C. W., and M. Orchinik. 2002a. Plasma binding proteins as mediators of

corticosteroid action in vertebrates. Journal of Endocrinology. 175, 99-112.

Breuner, C. W., and M. Orchinik. 2002b. Downstream from corticosterone:

seasonality of binding globulins, receptors and behavior in the avian stress

response. Pages 385-399 in A. Dawson, editor. Avian Endocrinology.

Narosa Publishing, New Deli and London. 233 Breuner, C. W., M. Orchinik, T. P. Hahn, S. L. Meddle, I. T. Moore, N. T. Owen-

Ashley, T. S. Sperry, and J. C. Wingfield. 2002. Differential mechanisms for

regulation of the stress response across lattitudinal gradients. Am J Physiol

Regul Integr Comp Physiol. 285, R594-R600.

Breuner, C. W., M. Orchinik, T. P. Hahn, S. L. Meddle, I. T. Moore, N. T. Owen-

Ashley, T. S. Sperry, and J. C. Wingfield. 2003. Differential mechanisms for

regulation of the stress response across latitudinal gradients. American

Journal of Physiology, Regulatory and Integrated Comparative Physiology.

285, R594-R600.

Breuner, C. W., S. E. Lynn, G. E. Julian, J. M. Cornelius, B. J. Heidinger, O. P.

Love, R. S. Sprague, H. Wada, and B. A. Whitman. 2006. Plasma-binding

globulins and acute stress response. Hormone and Metabolic Research. 38,

260-268.

Breuner, C. W., S. H. Patterson, and T. P. Hahn. 2008. In search of relationships

between the acute adrenocortical response and fitness. 157, 288-295.

Brodin, A. 2001. Mass-dependent predation and metabolic expenditure in wintering

birds: is there a trade-off between different forms of predation? Animal

Behaviour. 62, 993-999.

Brooks, T. M., R. A. Mittermeier, C. G. Mittermeier, G. A. B. da Fonseca, A. B.

Rylands, W. R. Konstant, P. Flick, J. Pilgrim, S. Oldfield, G. Magin, and C.

Hilton-Taylor. 2002. Habitat Loss and Extinction in the Hotspots of

Biodiversity. Conservation Biology. 16, 909-923.

Brown, D. R., and T. W. Sherry. 2006. Food supply controls the body condition of a

migrant bird wintering in the tropics. Oecologia. 149, 22-32. 234 Brown, M. E. 1996. Assessing body condition in birds. Pages 67-135 in V. J.

Nolan, and E. D. Ketterson, editors. Current ornithology. Plenum Press,

New York.

Buchanan, K. L. 2000. Stress and the evolution of condition-dependent signals.

Trends in Ecology and Evolution. 15, 156-160.

Buck, L. C., K. M. O'Reilly, and D. S. Kildaw. 2007. Interannual variability of black-

legged kittiwake productivity is reflected in baseline corticosterone. General

and Comparative Endocrinology. 150, 430-436.

Burbidge, A. A., and N. L. McKenzie. 1989. Patterns in the modern decline of

Western Australia's vertebrate fauna: causes and conservation implications.

Biological Conservation. 50, 143-198.

Buttemer, W. A. 1992. Differential overnight survival by Bumpus' House Sparrows:

An alternate interpretation. Condor. 94, 944-954.

Carmi, N., B. Pinshow, M. Horowitz, and M. H. Bernstein. 1993. Birds conserve

plasma-volume during thermal and flight-incurred dehydration. Physiological

Zoology. 66, 829-846.

Carmi, N., B. Pinshow, and M. Horowitz. 1994. Plasma volume conservation in

pigeons - Effects of air temperature during dehydration. American Journal of

Physiology, Regulatory Integrative and Comparative Physiology. 36, R1449-

R1453.

Carpenter, F. L. 1975. Bird hematocrits: effects of high altitude and strength of

flight. Comparative Biochemistry and Physiology. 50A, 415-417.

235 Carrascal, L. M., J. C. Senar, I. Mozetich, F. Uribe, and J. Domenech. 1998.

Interactions among environmental stress, body condition, nutritional status,

and dominance in Great Tits The Auk. 115, 727-738.

Cherel, Y., J. Leloup, and Y. LeMaho. 1988a. Fasting in the King Penguin. II.

Hormonal and metabolic changes during molt. American Journal of

Physiology. 23, R178-R184.

Cherel, Y., J. P. Robin, O. Walch, H. Karmann, P. Netchitailo, and Y. LeMaho.

1988b. Fasting in the King Penguin. I. Hormonal and metabolic changes

during breeding. American Journal of Physiology 23, R170-R177.

Christidis, L., and W. E. Boles 2008. Systematics and taxonomy of Australian birds.

CSIRO Publishing, Collingwood, VIC.

Churchhill, T. B., and J. A. Ludwig. 2004. Changes in spider assemblages along

grassland and savanna grazing gradients in northern Australia. The

Rangeland Journal. 26, 3-16.

Cichon, M., and A. Dubiec. 2005. Cell-mediated immunity predicts the probability of

local recruitment in nestling Blue Tits. Journal of Evolutionary Biology. 18,

962-966.

Clinchy, M., L. Zanette, R. Boonstra, J. C. Wingfield, and J. N. M. Smith. 2004.

Balancing food and predator pressure induces chronic stress in songbirds.

Proceedings of the Royal Society B: Biological Sciences. 271, 2437-2479.

Cohen, A., K. Klasing, and R. E. Ricklefs. 2007. Measuring circlulating antioxidants

in wild birds. Comparative Biochemistry and Physiology, Part B. 147, 110-

121.

236 Craig, A. B. 1994. Aspects of post-fire regeneration in soft spinifex (Triodia

pungens) communities near Newman, Western Australia. Pages 47–50.

Australian Rangeland Society. 8th Biennial Conference, Katherine, NT.

Working Papers, Australian Rangeland Society, Perth.

Creel, S. 2001. Social dominance and stress hormones. Trends in Ecology and

Evolution. 16, 491-497.

Cresswell, W. 2003. Testing the mass-dependent predation hypothesis: in

European blackbirds poor foragers have higher overwinter body reserves

Animal Behaviour. 65, 1035 - 1044.

Crowley, G., and S. Garnett. 1998. Vegetation change in the grasslands and

grassy woodlands of east-central Cape York Peninsula, Australia. Pacific

Conservation Biology. 4, 132-148.

Crowley, G., and S. Garnett. 1999. Seeds of the annual grasses Schizachyrium

spp. as a food resource for tropical granivorous birds. Australian Journal of

Ecology. 24, 208-220.

Crowley, G., and S. Garnett. 2000. Changing fire management in the pastoral

lands of Cape York Peninsula of northeast Australia, 1623-1996. Australian

Geographical Studies. 38, 10-26.

Crowley, G., and S. Garnett. 2001. Growth, seed production and effect of

defoliation in an early flowering perennial grass, Alloteropsis semialata

(Poaceae), on Cape York Peninsula, Australia. Australian Journal of Botany.

49, 735-743.

Cuervo, J. J., A. Moller, and F. De Lope. 2007. Haematocrit is weakly related to

condition in nestling Barn Swallows Hirundo rustica. Ibis. 149, 128-134. 237 Cuthill, I. C., S. A. Maddocks, C. V. Weall, and E. K. M. Jones. 2000. Body mass

regulation in response to changes in feeding predictability and overnight

energy expenditure. Behavioural Ecology. 11, 189-195.

Cyr, N. E., and L. M. Romero. 2006. Chronic stress in free-living European

Starlings reduces corticosterone concentrations and reproductive success.

General and Comparative Endocrinology. 151, 82-89.

Cyr, N. E., K. Earle, C. Tam, and L. M. Romero. 2007. The effect of chronic

psychological stress on corticosterone, plasma metabolites, and immune

responsiveness in European Starlings General and Comparative

Endocrinology. 154, 59-66.

Davis, A. K., D. L. Maney, and J. C. Maerz. 2008. The use of leukocyte profiles to

measure stress in vertebrates: a review for ecologists. Functional Ecology.

22, 760-772.

Dawson, A., and P. D. Howe. 1983. Plasma corticosterone in wild starlings

(Sturnus vilgaris) immediately following capture and in relation to body

weight during the annual cycle. General Comparative Endocrinology. 51,

303-308.

Dawson, R. D., and G. R. Bortolotti. 1997a. Are avian hematocrits indicative of

condition? American kestrels as a model Journal of Wildlife Management.

61, 1297-1306.

Dawson, R. D., and G. R. Bortolotti. 1997b. Variation in hematocrit and total

plasma proteins of nestling American kestrels (Falco sparverius) in the wild

Comparative Biochemistry and Physiology A: Physiology. 117, 383-390.

238 Dawson, T. J., R. M. Herd, and E. Skadhauge. 1983. Water turnover and body

water distribution during dehydration in a large arid-zone bird, the Emu,

Dromaius novaehollandiae Journal of Comparative Physiology B:

Biochemical, Systemic, and Environmental Physiology. 153, 235-240.

Deviche, P., J. C. Wingfield, and B. R. Sharp. 2000. Year-class differences in the

reproductive system, plasma prolactin and corticosterone concentrations,

and onset of prebasic molt in male dark-eyed juncos (Junco hyemalis)

during the breeding period. General and Comparative Endocrinology. 118,

425-435.

Deviche, P., C. W. Breuner, and M. Orchinik. 2001. Testosterone, corticosterone,

and photoperiod interact to regulate plasma levels of binding globulin and

free steroid hormone in Dark-Eyed Juncos, Junco hyemalis. General and

Comparative Endocrinology. 122, 67-77.

Devictor, V., R. Julliard, and F. Jiguit. 2008. Distribution of specialist and generalist

species along spatial gradients of habitat disturbance and fragmentation

Oikos. 117, 0-0.

Donnelly, R. E., and K. A. Sullivan. 1998. Foraging proficiency and body condition

of juvenile American dippers Condor. 100, 385-388.

Dorricott, K. I., and S. T. Garnett. 2007. National Recovery Plan for the White-

bellied Subspecies of the Crimson Finch (Neochmia phaeton evangelinae)

and the Northern Subspecies of the Star Finch (Neochmia ruficauda

clarescens). Report to the Australian Government Department of the

Environment and Water Resources, Canberra. QPWS, Brisbane.

239 Dostine, P. L., G. C. Johnson, D. C. Franklin, Y. Zhang, and C. Hempel. 2001.

Seasonal use of savanna landscapes by the Gouldian Finch, Erythrura

gouldiae, in the Yinberrie Hills area, Northern Territory. Wildlife Research.

28, 445-458.

Dostine, P. L., and D. C. Franklin. 2002. A comparison of the diet of three finch

species in the Yinberrie Hills area, Northern Territory. Emu. 102, 159-164.

Dubiec, A., and M. Cichon. 2001. Seasonal decline in health status of Great Tit

(Parus major) nestlings. Canadian Journal of Zoology. 79, 1829-1833.

Dudaniec, R. Y., S. Kleindorfer, and B. Fessl. 2006. Effects of the introduced

ectoparasite Philornis downsi on haemoglobin level and nestling survival in

Darwin's Small Ground Finch (Geospiza fuliginosa) Austral Ecology. 31, 88-

94.

Dyer, R., R. Jaklyn, P. Partridge, J. Russell-Smith, and D. Williams, editors. 2001.

Savanna burning: Understanding and using fire in northern Australia.

Tropical Savannas CRC, Darwin.

Edwards, A., R. Kennett, O. Price, J. Russell-Smith, G. Spiers, and J. C. Z.

Woinarski. 2003. Monitoring the impacts of fire regimes on vegetation in

northern Australia: an example from Kakadu National Park. International

Journal of Wildland Fire. 12, 427-440.

Eeva, T., E. Lehikoinen, and M. Nikinmaa. 2003. Pollution-induced nutritional

stress in birds: An experimental study of direct and indirect effects

Ecological Applications. 13, 1242-1249.

240 Ekman, J. B., and M. K. Hake. 1990. Monitoring starvation risk: adjustments of

body reserves in greenfinches (Carduelis chloris L.) during periods of

unpredictable foraging success. Behavioural Ecology. 1, 62-67.

Ekman, J. B., and K. Lilliendahl. 1993. Using priority of food access: fattening

strategies in dominance-structured willow tit (Parus montanus) flocks.

Behavioural Ecology. 4, 232-238.

Evans, R., and R. Love. 1957. The step-point method of sampling - a practical tool

in range research. Journal of Range Management. 10, 208 - 212.

Evans, S. M., and M. E. Fidler 1986. The Gouldian Finch. Blandford Press, Poole,

UK.

Ewenson, E. L., R. A. Zann, and G. R. Flannery. 2001. Body condition and immune

response in wild zebra finches: effects of capture, confinement and captive-

rearing Naturwissenschaften. 88, 391-394.

Fair, J., S. Whitaker, and B. Pearson. 2007. Sources of variation in haematocrit.

Ibis. 149, 535-552.

Fensham, R. J., and S. D. Skull. 1999. Before cattle: a comparative floristic study

of eucalyptus savanna grazed by macropods and cattle in north

Queensland, Australia. Biotropica. 31, 37-47.

Fensham, R. J., and R. J. Fairfax. 2003. Assessing woody vegetation cover

change in north-west Australian savanna using aerial photography.

International Journal of Wildland Fire. 12, 359-367.

Ferdinands, K., K. Beggs, and P. J. Whitehead. 2005. Biodiversity and invasive

grass species: multiple-use or monoculture? Wildlife Research. 32, 447-457.

241 Firth, R. S. C., B. W. Brook, J. C. Z. Woinarski, and D. A. Fordham. 2010. Decline

and likely extinction of a northern Australian native rodent, the Brush-tailed

Rabbit-rat Conilurus penicillatus. Biological Conservation. in press.

Fisher, R., T. Vigilante, C. Yates, and J. Russell-Smith. 2003. Patterns of

landscape fire and predicted vegetation response in the North Kimberley

region of Western Australia. International Journal of Wildland Fire. 12, 369-

379.

Fleshner, M., T. Deak, R. L. Spencer, M. L. Laudenslager, L. R. Watkins, and S. F.

Maier. 1995. A long term increase in basal levels of corticosterone and a

decrease in corticosteroid-binding globulin after acute stressor exposure.

Endocrinology. 136, 5336-5342.

Ford, H. A., G. W. Barrett, D. A. Saunders, and H. F. Recher. 2001. Why have

birds in the woodlands of southern Australia declined? Biological

Conservation. 97, 71-88.

Foster, M. S. 1975. Overlap of molting and breeding in some tropical birds.

Condor. 77, 304-314.

Franklin, D. C., P. L. Dostine, and S. C. Tidemann. 1998. Post-juvenile moult

strategies of co-existing Gouldian, Long-tailed and Masked Finches. Corella.

22, 73-79.

Franklin, D. C. 1999. Evidence of disarray amongst granivorous bird assemblages

in the savannas of northern Australia, a region of sparse human settlement.

Biological Conservation. 90, 53-68.

242 Franklin, D. C., J. C. Z. Woinarski, and R. A. Noske. 2000. Geographic patterning

of species richness among granivorous birds in Australia. Journal of

Biogeography. 27, 829-842.

Franklin, D. C., P. J. Whitehead, G. Pardon, J. Matthews, P. McMahon, and D.

McIntyre. 2005. Geographic patterns and correlates of the decline of

granivorous birds in northern Australia. Wildlife Research. 32, 399-408.

Fraser, B. F., V. Lawson, S. Morrison, P. Christophersen, S. McGreggor, and M.

Rawlinson. 2003. Fire management experiment for the declining Partridge

Pigeon, Kakadu National Park. Ecological Management & Restoration. 4,

94-102.

Friedel, M. H. 1997. Discontinuous change in arid woodland and grassland

vegetation along gradients of cattle grazing in central Australia. J Arid Env.

37, 135-164.

Friedl, T. P., and R. Edler. 2005. Stress-dependent trade-off between

immunological condition and reproductive performance in the polygynous

red bishop (Euplectes orix) Evolutionary Ecology. 19, 221-239.

Garnett, S., and G. Crowley. 1994. Wet season feeding by four species of

granivorous birds in the Northern Territory. Australian Bird Watcher. 15,

306-309.

Garnett, S. T. 1992. Threatened and extinct birds of Australia. Royal Australasian

Ornithologists Union Report 82, Melbourne and ANPWS, Canberra.

Garnett, S. T., and G. M. Crowley. 2000. The Action Plan for Australian Birds 2000.

Environment Australia, Canberra.

243 Garnett, S. T., J. R. Clarkson, A. Felton, G. N. Harrington, and A. N. D. Freeman.

2005. Habitat and diet of the Star Finch (Neochmia ruficauda clarescens) in

the early wet season at Princess Charlotte Bay, Cape York Peninsula,

Australia. Emu. 105, 81-85.

Gavett, A. P., and J. S. Wakeley. 1986. Blood constituients and their relation to diet

in urban and rural house sparrows. The Condor. 88, 279-284.

Gayathri, K. L., and S. N. Hegde. 2006. Alteration in haematocrit values and

plasma protein fractions during the breeding cycle of female pigeons,

Columba livia Animal Reproduction Science. 91, 133-141.

Gibbs, J. P., and E. J. Stanton. 2001. Habitat fragmentation and arthropod

community change: carrion bettles, phoretic mites and flies. Ecological

Applications. 11, 79-85.

Gill, A. M., and R. A. Bradstock. 1995. Extinction of biota by fires. Pages 309-322

in R. A. Bradstock, T. D. Auld, D. A. Keith, R. T. Kingsford, D. Lunney, and

D. P. Siverston, editors. Conserving Biodiversity: Threats and Solutions.

Surrey Beatty, Sydney.

Gill, A. M., P. G. Ryan, P. H. R. Moore, and M. Gibson. 2000. Fire regimes of world

heritage Kakadu National Park, Australia. Austral Ecology. 25, 616-625.

Goldstein, D. L., and A. Zahedi. 1990. Variation in osmoregulatory parameters of

captive and wild house sparrows (Passer domesticus). The Auk. 107, 533-

538.

Gosler, A. G. 1996. Environmental and social determinants of winter fat storage in

the Great Tit Parus major. The Journal of Animal Ecology. 65, 1-17.

244 Gray, J. M., D. Yarian, and M. Ramenofsky. 1990. Corticosterone, foraging

behaviour, and metabolism in dark-eyed juncos, Junco hyemalis. General

and Comparative Endocrinology. 79, 375-384.

Griffiths, R., M. C. Double, K. Orr, and R. J. G. Dawson. 1998. A simple DNA test

to sex most birds. Molecular Ecology. 7, 1071-1075.

Gross, J., and H. S. Siegel. 1983. Evaluation of the heterophil/lymphocyte ratio as

a measure of stress in chickens. Avian Diseases. 27, 972-979.

Guglielmo, C. G., D. J. Cerasale, and C. Eldermire. 2005. A field validation of

plasma metabolite profiling to assess refueling performance of migratory

birds. Physiological and Biochemical Zoology. 78, 116-125.

Hanssen, S. A., D. Hasselquist, I. Folstad, and K. E. Erikstad. 2004. Costs of

immunity: immune responsiveness reduces survival in a vertebrate

Proceedings of the Royal Society B: Biological Sciences. 271, 925-930.

Harvey, S., J. G. Phillips, A. Rees, and T. R. Hall. 1984. Stress and adrenal

function. Journal of Experimental Zoology. 232, 633-646.

Hau, M. 2001. Timing of breeding in variable environments: Tropical birds as

model systems Hormones and Behaviour. 40, 281-290.

Helms, C. W., and W. H. J. Drury. 1960. Winter and migratory weight and fat field

studies on some North American buntings. Bird-Banding. 31, 1-40.

Higgins, P. J., J. M. Peter, and S. J. Cowling, editors. 2006. Handbook of

Australian, New Zealand & Antarctic Birds (HANZAB) Volume 7. Boatbills to

Starlings. Oxford University Press, Melbourne.

245 Hoffmann. 2003. Responses of ant communities to experimental fire regimes on

rangelands in the Victoria River District of the Northern Territory. Austral

Ecology. 28, 182-195.

Hoi-Leitner, M., M. Romero-Pujante, H. Hoi, and A. Pavlova. 2001. Food

availability and immune capacity in serin (Serinus serinus) nestlings

Behavioural Ecology and Sociobiology. 49, 333-339.

Holberton, R. L. 1999. Changes in patterns of corticosterone secretion concurrent

with migratory fattening in a neotropical migratory bird. General and

Comparative Endocrinology. 116, 49-58.

Holmes, G. 1995. Survey of Gouldian Finch in Queensland during dry season of

1995 - with review of distribution and status. Unpublished report to

Conservation Commission of the Northern Territory, Darwin.

Holmes, G. 1996. Distribution and status of the Southern Star Finch. Sunbird. 26,

49-59.

Holmes, G. 1998. A review of the distribution, status and ecology of the Star Finch

Neochmia ruficauda in Queensland. . Australian Bird Watcher. 17, 278-289.

Horak, P., S. Jenni-Eiermann, I. Ots, and L. Tegelmann. 1998a. Health and

reproduction: the sex-specific clinical profile of great tits (Parus major) in

relation to breeding. Canadian Journal of Zoology. 76, 2235-2244.

Horak, P., I. Ots, and A. Murumagi. 1998b. Haematological health state indices of

reproducing Great Tits: a response to brood size manipulation Functional

Ecology. 12, 750-756.

246 Houston, A. I., and J. M. McNamara. 1993. A theoretical investigation of the fat

reserves and mortality levels of small birds in winter. Ornis Scandinavica.

24, 205-219.

Houston, D. C., D. Donnan, P. Jones, I. Hamilton, and D. Osborne. 1995. Changes

in the muscle condition of female zebra finches, Poephila guttata, during

egg-laying and the role of protein storage in bird skeletal muscle. Ibis. 137,

322-328.

Hurly, T. A. 1992. Energetic reserves of marsh tits (Parus palustris): food and fat

storage in response to variable food supply. Behavioural Ecology. 3, 181-

188.

Ilmonen, P., T. Taarna, and D. Hasselquist. 2000. Experimentally activated

immune defence in female pied flycatchers results in reduced breeding

success. Proceedings of the Royal Society B: Biological Sciences 267,

1471-2954.

Ilmonen, P., D. Hasselquist, A. Langefors, and J. Wiehn. 2003. Stress,

immunocompetence and leukocyte profiles of pied flycatchers in relation to

brood size manipulation Oecologia. 136, 148-154.

Immelman, K. 1982. Australian Finches. Angus and Robertson, Sydney.

James, C., J. J. Landsberg, and S. R. Morton. 1999. Provision of watering points in

the Australian arid zone: a review of effects on biota. Journal of Arid

Environments. 41, 87-121.

James, C. D. 2003. Response of vertebrates to fenceline contrasts in grazing

intensity in semi-arid woodlands of eastern Australia. Austral Ecology. 28,

137-151. 247 Jansen, A., and A. I. Robertson. 2001. Relationships between livestock

management and the ecological condition of riparian habitats along an

Australian floodplain river. Journal of Applied Ecology. 38, 63-75.

Jansen, A., and M. Healey. 2003. Frog communities and wetland condition:

relationships with grazing by domestic livestock along and Australian

floodplain river. Biological Conservation. 109, 207-219.

Jenni, L., S. Jenni-Eiermann, F. Spina, and H. Schwabl. 2000. Regulation of

protein breakdown and adrenocortical response to stress in birds during

migratory flight American Journal of Physiology, Regulatory Integrative and

Comparative Physiology. 278, R1182-R1189.

Jenni, L., S. Muller, F. Spina, A. Kvist, and A. Lindstrom. 2006. Effect of endurance

flight on haematocrit in migrating birds. Journal of Ornithology. 147, 531-

542.

JenniEiermann, S., and L. Jenni. 1996. Metabolic differences between the

postbreeding, moulting and migratory periods in feeding and fasting

passerine birds. Functional Ecology. 10, 62-72.

Jennings, D. H., M. C. Moore, R. Knapp, L. Matthews, and M. Orchinik. 2000.

Plasma steroid-binding globulin mediation of differences in stress reactivity

in alternative male phenotypes in tree lizards, Urosaurus ornatus. General

and Comparative Endocrinology. 120, 289-299.

Kauffman, J. B., and W. C. Krueger. 1984. Livestock impacts on riparian

ecosystems and streamside management implication, A review. Journal of

Range Management. 37, 430-438.

248 Kern, M., W. Bacon, D. Long, and R. J. Cowie. 2005. Blood metabolite and

corticosterone levels in breeding adult Pied Flycatchers Condor. 107, 665-

677.

Kern, M., W. Bacon, D. Long, and R. J. Cowie. 2007. Blood metabolite levels in

normal and handicapped pied flycatchers rearing broods of different sizes

Comparative Biochemistry and Physiology A-Molecular & Integrative

Physiology. 147, 70-76.

Khvoinitskaya, M. A. 1959. The effect of high environmental temperature on the

water distribution in the body. Journal Bulletin of Experimental Biology and

Medicine. 47, 578-581.

Kilgas, P., R. Mand, M. Magi, and V. Tilgar. 2006a. Hematological parameters in

brood-rearing great tits in relation to habitat, multiple breeding and sex

Comparative Biochemistry and Physiology A-Molecular & Integrative

Physiology. 144, 224-231.

Kilgas, P., V. Tilgar, and R. Mand. 2006b. Hematological health state indices

predict local survival in a small passerine bird, the Great Tit (Parus major)

Physiological and Biochemical Zoology. 79, 565-572.

King, J. R., and D. S. Farner. 1966. The adaptive role of winter fattening in the

white-crowned sparrow with comments on its regulation. American

Zoologist. 100, 403-418.

Kitaysky, A. S., J. C. Wingfield, and J. F. Piatt. 1999. Dynamics of food availability,

body condition, and physiological stress response in breeding Black-legged

Kittiwakes. Functional Ecology. 13, 577-584.

249 Klaassen, M. 1995. Moult and basal metabolic costs in males of two subspecies of

stonechats: the European Saxicola torquata rubicula and the East African S.

t. axillaris. Oecologia. 104, 424-432.

Klukowski, L. A., J. M. Cawthorn, E. Ketterson, and J. Nolan. 1997. Effects of

experimentally elevated testosterone on plasma corticosterone and

corticosteroid-binding globulin in dark-eyed junco (Junco hyemalis). General

and Comparative Endocrinology. 108, 141-151.

Knapp, R., and M. C. Moore. 1997. Male morphs in tree lizards have different

testosterone responses to elevated levels of corticosterone. General and

Comparative Endocrinology. 107, 273-279.

Koivula, K., M. Orell, and S. Rytkonen. 1996. Winter survival and breeding success

of dominant and subordinate Willow Tits, Parus montanus. Ibis. 138, 624-

629.

Landsberg, J. J., C. James, S. R. Morton, T. J. Hobbs, J. Stol, A. Drew, and H.

Tongway. 1997. The effects of artificial sources of water on rangeland

biodiversity. CSIRO Division of Wildlife and Ecology, Canberra.

Landsberg, J. J., S. Lavorel, and J. Stol. 1999. Grazing response groups among

understory plants in arid rangelands. J Veg Sci. 10, 683-696.

Landys-Ciannelli, M. M., J. Jukema, and T. Piersma. 2002. Blood parameter

changes during stopover in a long-distance migratory shorebird, the bar-

tailed godwit, Limosa lapponica taymyrensis Journal of Avian Biology. 33,

451-455.

Landys, M. M., J. C. Wingfield, and M. Ramenofsky. 2004. Plasma corticosterone

increases during migratory restlessness in the captive White-crowned 250 Sparrow, Zonotrichia leucophrys gambelli Hormones and Behaviour. 46,

574-581.

Layme, V. M. G., A. P. Lima, and E. Magnusson. 2004. Effects of fire, food

availability and vegetation on the distribution of the rodent Bolomys lasiurus

in an Amazonian savanna. Journal of Tropical Ecology. 20, 183-187.

Leary, J., K. A. Sullivan, and N. Hillgarth. 1999. Relationships among dominance,

foraging proficiency, and condition in juvenile Dark-eyed Juncos. The Auk.

116, 1136-1141.

Legge, S. M. 2009. Unpublished data on Gouldian and Long-tailed finch

movements using telemetry and color-band resight data. Personal

communication. Australian Wildlife Conservancy.

Letnic, M. 2004. Cattle grazing in a hummock grassland regenerating after fire: the

short-term effects of cattle exclusion on vegetation in south-western

Queensland. Rangeland Journal. 26, 34-48.

Lewis, M. 2007. Foraging responses of the endangered Gouldian finch to temporal

differences in seed availability in northern Australian savanna grasslands in

J. A. Bissonette, and I. Storch, editors. Temporal Dimensions of Landscape

Ecology: Wildlife Responses to Variable Resources. Springer, New York,

NY.

Liedloff, A. C., M. B. Coughenour, J. A. Ludwig, and R. Dyer. 2001. Modelling the

trade-off between fire and grazing in a tropical savanna landscape, northern

Australia. Environment International. 27, 173-180.

Lima, S. L. 1986. Predation risk and unpredictable feeding conditions:

determinants of body mass in birds. Ecology. 67, 377-385. 251 Lindquist, S., and D. A. Parsell. 1993. The function of heat-shock proteins in stress

tolerance: degradation and reactivation of damaged proteins. Annu. Rev.

Genet. 27, 437-496.

Lindstrom, A., and T. Piersma. 1993. Mass changes in migrating birds - the

evidence for fat and protein storage re-examined. Ibis. 135, 70-78.

Lindstrom, A., A. Kvist, T. Piersma, A. Dekinga, and M. W. Dietz. 2000. Avian

pectoral muscle size rapidly tracks body mass changes during flight, fasting

and fuelling. Journal of Experimental Biology. 203, 913-919.

Lobato, E., J. Moreno, S. Moreno, J. J. Sanz, and E. Arriero. 2005. Haematological

variables are good predictors of recruitment in nestling Pied Flycatchers

(Ficeduld hypoleuca) Ecosience. 12, 27-34.

Lonsdale, W. M., R. W. Braithwaite, A. M. Lane, and J. Farmer. 1998. Modelling

the recovery of an annual savanna grass following a fire-induced crash.

Australian Journal of Ecology. 27, 509-513.

Love, O. P., E. H. Chin, K. E. Wynne-Edwards, and T. D. Williams. 2005. Stress

hormones: A link between maternal condition and sex-biased reproductive

investment American Naturalist. 166, 751-766.

Lowe, L. 1995. Preliminary investigations of the biology and management of

Leichhardt's grasshopper, Petasida ephippigera White. Journal of

Orthoptera Research. 4, 219-221.

Ludwig, J. A., M. B. Coughenour, A. C. Liedloff, and R. Dyer. 2001. Modelling the

resilience of Australian savanna systems to grazing impacts. Environment

International. 27, 167-172.

252 Lundberg, P. 1985. Dominance behaviour, body weight, and fat variations, and

partial migration in European blackbirds, Turdus merula. Behavioural

Ecology and Sociobiology. 17, 185-189.

Lynn, S. E., C. W. Breuner, and J. C. Wingfield. 2003a. Short-term fasting affects

locomotor activity, corticosterone, and corticosterone binding globulin in a

migratory songbird. Hormones and Behaviour. 43, 150-157.

Lynn, S. E., K. E. Hunt, and J. C. Wingfield. 2003b. Ecological factors affecting the

adrenocortical response to stress in Chestnut-collared and McCown's

longspurs (Calcarius ornatus, Calcarius mccownii). Physiological and

Biochemical Zoology. 76, 566-576.

Maclouf, J., J. Grassi, and P. Pradelles. 1987. Development of enzyme-

immunoassay techniques for the measurement of eicosanoids. Pages 355-

364 in L. Thomas, J. Walden, and H. N. Hughes, editors. In International

Conference on Prostaglandin and Lipid Metabolism in Radiation Injury.

Penum, New York.

MacMillen, R. E. 1990. Water economy of granivorous birds: A predictive model.

Condor. 92, 379-392.

MacMillen, R. E., and R. V. Baudinette. 1993. Water economy of granivorous birds:

Australian parrots. Functional Ecology. 7, 704-712.

Magrath, R. D. 1991. Nestling weight and juvenile survival in the Blackbird, Turdus

merula. Journal of Animal Ecology. 60, 335-351.

Malisch, J. L., and C. W. Breuner. 2010. Steroid-binding proteins and free steroids

in birds. Molecular and Cellular Endocrinology. 316, 42-52.

253 Marra, P. P., and R. L. Holberton. 1998. Corticosterone levels as indicators of

habitat quality: effects of habitat segregation in a migratory bird during the

non-breeding season. Oecologia. 116, 284-292.

Martin, L. B., J. Gilliam, P. Han, K. Lee, and M. Wikelski. 2005. Corticosterone

suppresses cutaneous immune function in temperate but not tropical House

Sparrows, Passer domesticus General and Comparative Endocrinology.

140, 126-135.

Martin, L. B., M. Wikelski, and P. Wikelski. 2007. Greater seasonal variation in

blood and ecto parasite infections in a temperate than a tropical population

of House Sparrows Passer domesticus in North America. Ibis. 149, 419-423.

Martin, T. G., and H. P. Possingham. 2005. Predicting the impact of livestock

grazing on birds using foraging height data. Journal of Applied Ecology. 42,

400-408.

Martin, T. G., S. McIntyre, C. P. Catterall, and H. P. Possingham. 2006. Is

landscape context important for riparian conservation? Birds in grassy

woodland. Biological Conservation. 127, 201-214.

Matson, K. D., R. E. Ricklefs, and K. C. Klasing. 2005. A hemolysis-

hemagglutination assay for characterizing constitutive innate humoral

immunity in wild and domestic birds. Developmental and Comparative

Immmunology. 29, 275-286.

Matson, K. D., B. I. Tieleman, and K. Klasing. 2006. Capture stress and the

bactericidal competence of blood and plasma in five species of tropical

birds. Physiological and Biochemical Zoology. 79, 556-564.

254 Maute, K., and B. Venables. 2009. A significant sighting. Pages 1-2. Southern Gulf

Links.

Maxwell, M. H., and G. W. Robertson. 1998. The avian heterophil leucocyte: a

review. Worlds Poultry Science Journal. 54, 155-178.

Mazerolle, D. F., and K. A. Hobson. 2002. Physiological ramifications of habitat

selection in territorial male ovenbirds: consequences of landscape

fragmentation. Oecologia. 130, 356-363.

McDonald, I. R., A. K. Lee, K. A. Than, and R. W. Martin. 1986. Failure of

glucocorticoid feedback in males of a population of small marsupials

(Antechinus swainsonii) during the period of mating. Journal of

Endocrinology. 108, 63-68.

McIntyre, S., K. M. Heard, and T. G. Martin. 2003. The relative importance of cattle

grazing in subtropical grasslands: does it reduce or enhance plant

biodiversity. Journal of Applied Ecology. 40, 445-457.

McIvor, J., and C. J. Gardner. 1991. Soil seed densities and emergence patterns in

pastures in the seasonally dry tropics of north-eastern Australia. Australian

Journal of Ecology. 16, 159-169.

McNamara, J. M., and A. I. Houston. 1990. The value of fat reserves and the

tradeoff between starvation and predation. Acta Biotheretica. 38, 37-61.

Meeson, N., A. I. Robertson, and A. Jansen. 2002. The effects of flooding and

livestock on post-dispersal seed predation in river red gum habitats. Journal

of Applied Ecology. 39, 247-258.

Merila, J., and E. Svensson. 1997. Are fat reserves in migratory birds affected by

condition in early life? Journal of Avian Biology. 28, 279-286. 255 Merino, S., J. Martinez, A. Moller, A. Barbosa, F. de Lope, and F. Rodriguez-

Caabeiro. 2001. Physiological and haematological consequences of a novel

parasite on the red-rumped swallow, Hirundo daurica International Journal

for Parasitology. 31, 1187-1193.

Metcalfe, N. B., and S. E. Ure. 1995. Diurnal variation in flight performance and

hence potential predation risk in small birds. Proceedings: Biological

Sciences. 261, 395-400.

Milenkaya, O. 2010. Information on the blood parameters and home range of

Crimson Finches (Neochmia phaeton) studied in the Kimberley Region,

Western Australia. Personal Communication.

Moller, A. P., and C. Haussy. 2007. Fitness consequences of variation in natural

antibodies and complement in the Barn Swallow Hirundo rustica. Functional

Ecology. 21, 363-371.

Moore, M. C., C. W. Thompson, and C. A. Marler. 1991. Reciprocal changes in

corticosterone and testosterone levels following acute and chronic handling

stress in the tree lizard, Urosaurus ornatus. General and Comparative

Endocrinology. 81, 217-226.

Moreno, J., S. Merino, J. J. Sanz, E. Arriero, J. Morales, and G. Tomas. 2005.

Nestling cell-mediated immune response, body mass and hatching date as

predictors of local recruitment in the Pied Flycatcher Ficedula hypoleuca.

Journal of Avian Biology. 36, 251-260.

Morton, M. L. 1994. Hematocrits in montane sparrows in relation to reproductive

schedule. Condor. 96, 119-126.

256 Mott, J., B. J. Bridge, and W. Arndt. 1979. Soil seals in tropical tall grass pastures

of northern Australia. Australian Journal of Soil Restoration. 30, 483-494.

Mott, J., and M. Andrew. 1985. The effect of fire on the population dynamics of

native grasses in tropical savannas of north-west Australia. Proceedings of

the Ecological Society of Australia.

Muller, C., S. Jenni-Eiermann, J. Blondel, P. Perret, S. P. Caro, M. Lambrechts,

and L. Jenni. 2007. Circulating corticosterone levels in breeding Blue Tits,

Parus caeruleus, differ between island and mainland populations and

between habitats General and Comparative Endocrinology. 154, 128-136.

Munday, P. L. 2004. Habitat loss, resource specialization, and extinction on coral

reefs Global Change Biology. 10, 1642-1647.

Nadolski, J., J. Skwarska, A. Kalinski, M. Banbura, R. Sniegula, and J. Banbura.

2006. Blood parameters as consistent predictors of nestling performance in

Great Tits (Parus major) in the wild Comparative Biochemistry and

Physiology A-Molecular & Integrative Physiology. 143, 50-54.

Navarro, C., A. Marzal, F. De Lope, and A. Moller. 2003. Dynamics of an immune

response in house sparrows Passer domesticus in relation to time of day,

body condition and blood parasite infection Oikos. 101, 291-298.

Neldner, V. J., R. J. Fensham, J. R. Clarkson, and J. P. Stanton. 1997. The natural

grasslands of Cape York Peninsula, Australia: Description, distribution and

conservation status. Biological Conservation. 81, 121-136.

Newton, I. 2004. The recent declines of farmland bird populations in Britain: an

appraisal of causal factors and conservation actions. Ibis. 146, 579-600.

257 Norris, K., and R. Evans. 2000. Ecological immunity: life history trade-offs and

immune defence in birds. Behavioural Ecology. 11, 19-26.

O'Brian, E. L., B. L. Morrison, and L. S. Johnson. 2001. Assessing the effects of

haematophagous ectoparasites on the health of nestling birds: haematocrit

vs haemoglobin levels in House Wrens parasitized by blow fly larvae

Journal of Avian Biology. 32, 73-76.

O'Brien, E. L., B. L. Morrison, and L. S. Johnson. 2001. Assessing the effects of

haematophagous ectoparasites on the health of nestling birds: haematocrit

vs haemoglobin levels in House Wrens parasitized by blow fly larvae

Journal of Avian Biology. 32, 73-76.

O'Connor, T. G., and G. A. Pickett. 1992. The influence of grazing on seed

production and seed banks of some African savanna grasslands. Journal of

Applied Ecology. 29, 247-260.

O'Dwyer, T. W., W. A. Buttemer, and D. M. Priddel. 2007. Differential rates of

offspring provisioning in Gould's petrels: are better feeders better breeders?

Australian Journal of Zoology. 55, 155-160.

O'Malley, C. 2006a. Recovery plan for the Gouldian Finch (Erythrura gouldiae).

WWF-Australia, Sydney and Parks and Wildlife, NT Department of Natural

Resources, Environment and the Arts, NT Goverment, Palmerston.

O'Malley, C. 2006b. Appendix to the National Recovery Plan for the Gouldian Finch

(Erythrura gouldiae) Background Information.

http://www.environment.gov.au/biodiversity/threatened/publications/pubs/e-

gouldiae-background.pdf.

258 Orr, D. M., C. J. Paton, and A. T. Lisle. 1997. Using fire to manage species

composition in Heteropogon contortus (black speargrass) pastures 1.

Burning regimes. Australian Journal of Agricultural Research. 48, 795-802.

Ots, I., A. B. Kerimov, E. V. Ivankina, T. A. Ilyina, and P. Horak. 2001. Immune

challenge affects basal metabolic activity in wintering great tits. Proceedings

of the Royal Society B: Biological Sciences. 268, 1175-1181.

Ouillfeldt, P., J. F. Masello, and E. Mostl. 2004. Blood chemistry in relation to

nutrition and ectoparasite load in Wilson's storm-petrels Oceanites

oceanicus Polar Biology. 27, 168-176.

Owen, J. C., M. K. Sogge, and M. D. Kern. 2005. Habitat and sex differences in

physiological condition of breeding Southwestern Willow Flycatchers

(Empidonax trailliextimus). The Auk. 122, 1261-1270.

Owen, J. C., and F. R. Moore. 2006. Seasonal differences in immunological

condition of three species of thrushes. The Condor. 108, 389-398.

Partecke, J., I. Schwabl, and E. Gwinner. 2006. Stress and the city: Urbanization

and its effects on the stress physiology in European Blackbirds. Ecology. 87,

1945-1952.

Piersma, T., and J. M. Everaarts. 1996. Build-up of red blood cells in refuelling bar-

tailed godwits in relation to individual migratory quality. Condor. 98, 363-

370.

Piersma, T., A. Koolhaas, A. Dekinga, and E. Gwinner. 2000a. Red blood cell and

white blood cell counts in sandpipers (Philomachus pugnax, Calidris

canutus): effects of captivity, season, nutritional status, and frequent

bleedings Canadian Journal of Zoology. 78, 1349-1355. 259 Piersma, T., J. Reneerkens, and M. Ramenofsky. 2000b. Baseline corticosterone

peaks in shorebirds with maximal energy stores for migration: A general

preparatory mechanism for rapid behavioural and metabolic transitions? .

General and Comparative Endocrinology. 120, 118-126.

Piper, W. H., and R. H. Wiley. 1990. The relationship between social dominance,

subcutaneous fat, and annual survival in wintering White-throated Sparrows

(Zonotrichia albiocollis). Behavioural Ecology and Sociobiology. 26, 201-

208.

Potti, J., J. Moreno, S. Merino, O. Frias, and R. Rodriquez. 1999. Environmental

and genetic variation in the haematocrit of fledgling Pied Flycatchers

Ficedula hypoleuca Oecologia. 120, 1-8.

Pradelles, P., J. Grassi, and J. Maclouf. 1985. Enzyme immunoassays of

eicosanoids using acetylcholinesterase as label: An alternative to

radioimmunoassay. Analytical Chemistry. 57, 1170-1173.

Pravosudov, V. V., A. S. Kitaysky, J. C. Wingfield, and N. S. Clayton. 2001. Long-

term unpredictable foraging contidtions and physiological stress response in

mountain chickadees (Poecile gambeli). General and Comparative

Endocrinology. 123, 324-331.

Price, O., A. Edwards, G. Connors, J. C. Z. Woinarski, G. Ryan, A. Turner, and J.

Russell-Smith. 2005a. Fire heterogeneity in Kakadu National Park, 1980-

2000. Wildlife Research. 32, 425-433.

Price, O., B. Rankmore, D. Milne, C. Brock, C. Tynan, L. Kean, and L. Roeger.

2005b. Regional patterns of mammal abundance and their relationship to

260 landscape variables in eucalypt woodlands near Darwin, northern Australia.

Wildlife Research. 32, 435-446.

Price, O., C. Palmer, D. Hooper, and M. Lewis. In preparation. A population

monitoring program for the endangered Gouldian Finch in the Yinberrie

Hills, northern Australia.

Prike, S. 2008. Diet study in captive grass-finches suggests high nutrient diets

increase breeding success and survival probability of offspring. Personal

communication. Manuscript in preparation.

Recher, H. F., and L. Lim. 1990. A review of current ideas of the extinction,

conservation and management of Australia's terrestrial vertebrate fauna.

Proceedings of the Ecological Society of Australia. 16, 287-301.

Rees, A., S. Harvey, and J. G. Phillips. 1985. Adrenocortical responses to novel

stressors in acutely or repeatedly starved chickens General and

Comparative Endocrinology. 59, 105-109.

Rice, B., and M. Westoby. 1999. Regeneration after fire in Triodia R. Br. Australian

Journal of Ecology. 24, 563-572.

Rich, E. L., and L. M. Romero. 2001. Daily and photoperiod variations of basal and

stress-induced corticosterone concentrations in house sparrows (Passer

domesticus). Journal of Comparative Physiology B. 171, 543-547.

Rich, E. L., and M. L. Romero. 2005. Exposure to chronic stress downregulates

corticosterone responses to acute stressors. American Journal of

Physiology, Regulatory and Integrated Comparative Physiology. 288,

R1628-R1636.

261 Ridpath, M. G. 1985. Ecology of the wet-dry tropics: how different? Proceedings of

the Ecological Society of Australia. 13, 191-202.

Rogers, C. M. 1987. Predation risk and fasting capacity: do wintering birds

maintain optimal body mass? Ecology. 68, 1051-1061.

Rogers, C. M., E. D. Ketterson, and V. J. Nolan. 1993. Geographic variation in

winter fat of dark-eyed juncos: displacement to a common environment.

Ecology. 74, 1183-1190.

Romero, L. M., K. K. Soma, and J. C. Wingfield. 1998. Changes in pituitary and

adrenal sensitivities allow the snow bunting (Plectrophenax nivalis), an

arctic-breeding songbird, to modulate corticosterone release seasonally.

Journal of Comparative Physiology B. 168, 353-358.

Romero, L. M., and J. C. Wingfield. 1998. Seasonal changes in adrenal sensitivity

alter corticosterone levels in Gambel's White-crowned Sparrows

(Zonotrichia leucophyrs gambelii). Comparative Biochemistry and

Physiology C: Pharmacology, Toxicology and Endocrinology. 119, 31-36.

Romero, L. M., and J. C. Wingfield. 1999. Alterations in hypothalamic-pituitary-

adrenal function associated with captivity in Gambel's white-crowned

sparrows (Zonotrichia leucophrys gambelii). Comparative Biochemistry and

Physiology B: Biochemistry and Molecular Biology. 122, 13-20.

Romero, L. M., and L. Remage-Healey. 2000. Daily and seasonal variation in

response to stress in captive starlings (Sturnus vularis): corticosterone.

General and Comparative Endocrinology. 119, 52-59.

Romero, L. M., and M. Wikelski. 2001. Corticosterone levels predict survival

probabilities of Galapagos marine iguanas during El Nino events. 262 Proceedings of the National Academy of Sciences of the United States. 98,

7366-7370.

Romero, L. M. 2002. Seasonal changes in plasma glucocorticoid concentrations in

free-living vertebrates. General and Comparative Endocrinology. 128, 1-24.

Romero, L. M., and M. Wikelski. 2002. Exposure to tourism reduces stress-induced

corticosterone levels in Galápagos marine iguanas. Biological Conservation.

108, 371-374.

Romero, L. M. 2004. Physiological stress in ecology: lessons from biomedical

research. Trends in Ecology and Evolution. 19, 249-255.

Romero, L. M., and E. L. Rich. 2007. Photoperiodically-induced changes in

hypothalamic–pituitary–adrenal axis sensitivity in captive house sparrows

(Passer domesticus) Comparative Biochemistry and Physiology - Part A:

Molecular & Integrative Physiology 147, 562-568.

Rossiter, N. A., S. A. Setterfield, M. M. Douglas, and L. B. Hutley. 2003. Testing

the grass-fire cycle: alien grass invasion in the tropical savannas fo northern

Australia. Diversity and Distributions. 9, 169-176.

Rozman, J., D. Runciman, and R. A. Zann. 2003. Seasonal variation in body mass

and fat of zebra finches in south-eastern Australia. Emu. 103, 11-19.

Rubenstein, D. R. 2007. Stress hormones and sociality: integrating social and

environmental stressors Proceedings of the Royal Society B: Biological

Sciences. 274, 967-975.

Russell-Smith, J., P. G. Ryan, D. Klessa, G. Waight, and R. Hardwood. 1998. Fire

regimes, fire-sensitive vegetation and fire management of the sandstone

263 Arnhem Plateau, monsoonal northern Australia. Journal of Applied Ecology.

35, 829-846.

Russell-Smith, J., P. J. Whitehead, G. D. Cook, and J. L. Hoare. 2003a. Response

of Eucalyptus-dominated savanna to frequent fires: lessons from

Munmarlary, 1973-1996. Ecological Monographs. 73, 349-375.

Russell-Smith, J., C. Yates, A. Edwards, G. E. Allan, G. D. Cook, P. Cooke, R.

Craig, B. Heath, and R. Smith. 2003b. Contemporary fire regimes of

northern Australia, 1997-2001: change since aboriginal occupancy,

challenges for sustainable management. International Journal of Wildland

Fire. 12, 283-297.

Saino, N., S. Calza, and A. Moller. 1997. Immunocompetence of nestling Barn

Swallows in relation to brood size and parental effort Journal of Animal

Ecology. 66, 827-836.

Sanchez-Guzman, J. M., A. Villegas, C. Corbacho, R. Moran, A. Marzal, and R.

Real. 2004. Response of haematocrit to body condition changes in northern

bald ibis Geronticus eremita. Comparative Biochemistry and Physiology A:

Physiology. 139, 41-47.

Sapolsky, R. M., L. M. Romero, and A. U. Munck. 2000. How do glucocorticoids

influence stress responses? Integrating permissive, suppressive,

stimulatory, and preparative actions. Endocrine Reviews. 21, 55-89.

Schoech, S. J., and R. Bowman. 2003. Does differential access to protein influence

differences in timing of breeding of florida scrub-jays (Aphelocoma

coerulescens) in suburban and wildland habitats? . The Auk. 120, 1114-

1127. 264 Schwabl, I., F. Bairlein, and E. Gwinner. 1991. Basal and stress-induced

corticosterone levels of Garden Warbler, Sylvia borin, during migration.

Journal of Comparative Physiology B. 161, 576-580.

Seaman, D. A. A., C. G. Guglielmo, R. W. Elner, and T. D. Williams. 2006.

Landscape-scale physiology: Site differences in refueling rates indicated by

plasma metabolite analysis in free-living migratory sandpipers. The Auk.

123, 563-574.

Senar, J. C., J. Domenech, and F. Uribe. 2002. Great tits (Parus major) reduce

body mass in response to wing area reduction: a field experiment.

Behavioural Ecology. 13, 725-727.

Sergent, N., T. Rogers, and M. Cunningham. 2004. Influence of biological and

ecological factors on hematological values in wild Little Penguins, Eudyptula

minor. Comparative Biochemistry and Physiology A: Physiology. 138, 333-

339.

Sharp, B. R., and R. J. Whittaker. 2003. The irreversible cattle-driven

transformation of a seasonally-flooded Australian savanna. Journal of

Biogeography. 30, 783-802.

Sharp, B. R., and D. M. J. S. Bowman. 2004. Net woody vegetation increase

confined to seasonally inundated lowlands in an Australian tropical savanna,

Victoria River District, Northern Territory. Austral Ecology. 29, 667-683.

Siegel, H. S. 1980. Physiological stress in birds. BioScience. 30, 529-534.

Silverin, B. 1986. Corticosterone-binding proteins and behavioural effects of high

plasma levels of corticosterone during the breeding period in the pied

flycatcher. General and Comparative Endocrinology. 64, 67-74. 265 Sims, C. G., and R. L. Holberton. 2000. Development of the corticosterone stress

response in young Northern mockingbirds (Mimus polyglottos). General and

Comparative Endocrinology. 119, 193-201.

Smith, R. D., and N. B. Metcalfe. 1997. Diurnal, seasonal and altitudinal variation in

energy reserves of wintering snow buntings. Journal of Avian Biology. 28,

216-222.

Soler, M., M. Martín-Vivaldi, J. M. Marín, and A. P. Møller. 1999. Weight lifting and

health status in the black wheatear Behavioural Ecology. 10, 281-286.

Stevenson, R. D. 2005. Ecophysiology and conservation: the contribution of

endocrinology and immunology -- introduction tho the symposium.

Integrative and Comparative Biology. 45, 1-3.

Stevenson, R. D., and W. A. J. Woods. 2006. Condition indices for conservation:

new uses for evolving tools. Integrated Comparative Biology. 46, 1169-

1190.

Strong, A. M., and T. W. Sherry. 2000. Habitat-specific effects of food abundance

on the condition of ovenbirds wintering in Jamaica. Journal of Animal

Ecology. 69, 883-895.

Stuebe, M. M., and E. D. Ketterson. 1982. A study of fasting in tree sparrows

(Spizella arborea) and dark-eyed juncos (Junco hyemalis): ecological

implications. The Auk. 99, 299-308.

Sturkie, P. D. 1986. Avian Physiology. Springer Verlag, New York, NY.

Suorsa, P., E. Huhta, A. Nikula, M. Nikinmaa, A. Jantti, H. Helle, and H.

Hakkarainen. 2003. Forest management is associated with physiological

266 stress in an old-growth forest passerine. Proceedings of the Royal Society

B: Biological Sciences. 270, 963-969.

Suorsa, P., H. Helle, V. Koivunen, E. Huhta, A. Nikula, and H. Hakkarainen. 2004.

Effects of forest patch size on physiological stress and immunocompetence

in an area-sensitive passerine, the Eurasian treecreeper (Certhia familiaris):

an experiment Proceedings of the Royal Society B: Biological Sciences.

271, 435-440.

Swaddle, J. P., and M. S. Witter. 1997. The effects of molt on the flight

performance, body mass, and behavior of European starlings (Sturnus

vulgaris): an experimental approach. Canadian Journal of Zoology. 75,

1135-1146.

Tassicker, A. L., A. S. Kutt, E. Vanderduys, and S. Mangru. 2006. The effects of

vegetation structure on the birds in a tropical savanna woodland in north-

eastern Australia. The Rangeland Journal. 28, 139-152.

Team, B.-t. F. R., D. o. E. a. C. C. (NSW), and Q. P. a. W. Service. 2007. National

recovery plan for the Black-throated finch southern subspecies (Poephila

cincta cincta). Report to the Department of the Environment and Water

Resources, Canberra. Department of Environment and Climate Change

(NSW), Hurstville and Queensland Parks and Wildlife Service, Brisbane.

Tella, J. L., A. Scheuerlein, and R. E. Ricklefs. 2002. Is cell-mediated immunity

related to the evolution of life-history strategies in birds? . Proceedings of

the Royal Society B: Biological Sciences. 269, 1059-1066.

Tidemann, S. C. 1989. Relationships between finches and pastoral practices in

northern Australia. in J. Pinowski, and J. D. Summers-Smith, editors. 267 Granivorous Birds in the Agricultural Landscape. PWN - Polish Scientific

Publishers, Warsaw.

Tidemann, S. C., M. Calley, and C. Burgoyne. 1992. An investigation of blood

smears of northern Australian finches. Emu. 92, 114-117.

Tidemann, S. C., and J. C. Z. Woinarski. 1994. Moult characteristics and breeding

seasons of gouldian, Erythrura gouldiae, masked, Poepila personata and

long-tailed finches, P. acuticauda in savannah woodland in the northern

territory. Emu. 94, 46-52.

Tidemann, S. C. 1996. Causes of the decline of the Gouldian Finch, Erythrura

gouldiae. Bird Conservation International. 6, 49-61.

Tidemann, S. C., C. Lawson, R. Elvish, J. Boyden, and J. Elvish. 1999. Breeding

biology of the Gouldian Finch Erythura gouldiae, and endangered finch of

northern Australia. Emu. 99, 191-199.

Todd, M. K., A. Felton, and S. T. Garnett. 2003. Morphological and dietary

differences between common and uncommon subspecies of Crimson Finch,

(Neochmia phaeton), and Star Finch, (Neochmia ruficauda), in northern

Australia. Emu. 103, 141-148.

Totzke, U., M. Fenske, O. Huppop, H. Raabe, and N. Schach. 1999. The influence

of fasting on blood and plasma composition of herring gulls (Larus

argentatus) Physiological and Biochemical Zoology. 72, 426-437.

Tracy, C. R., K. E. Nussear, T. C. Esque, K. Dean-Bradley, C. R. Tracy, L. A.

DeFalco, K. T. Castle, L. C. Zimmerman, R. E. Espinoza, and A. M. Barber.

2006. The importance of physiological ecology in conservation biology.

Integrated Comparative Biology. 46, 1191-1205. 268 Travis, J. M. J. 2003. Climate change and habitat destruction: a deadly

anthropegenic cocktail. BioScience. 270, 467-473.

Vigilante, T. 2001. Analysis of explorers' records of Aboriginal landscape burning in

the Kimberley region of Western Australia. Australian Geographical Studies.

39, 135-155.

Vleck, C. M., and J. Priedkalns. 1985. Reproduction in zebra finches - hormone

levels and effect of dehydration. The Condor. 87, 37-46.

Vleck, C. M., N. Vertalino, D. Vleck, and T. L. Bucher. 2000. Stress, corticosterone,

and heterophil to lymphocyte ratios in free-living Adlie Penguins. The

Condor. 102, 392-400.

Wada, H., T. P. Hahn, and C. W. Breuner. 2006a. Development of stress reactivity

in white-crowned sparrow nestlings: Total corticosterone response

increases with age, while free corticosterone response remains low. General

and Comparative Endocrinology. 150, 405-413.

Wada, H., I. T. Moore, C. W. Breuner, and J. C. Wingfield. 2006b. Stress

responses in tropical sparrows: Comparing tropical and temperate

Zonotrichia Physiological and Biochemical Zoology. 79, 784-792.

Walker, B. G., P. D. Boersma, and J. C. Wingfield. 2005. Field endocrinology and

conservation biology. Integrated Comparative Biology. 45, 12-18.

Walker, B. H., J. L. Langridge, and F. McFarlane. 1997a. Resilience of an

Australian savanna grassland to selective and non-selective perturbations.

australian Journal of Ecology. 22, 125-135.

269 Walker, B. H., J. L. Langridge, and F. McFarlane. 1997b. Resilience of an

Australian savanna grassland to selective and non-selective perturbations.

Australian Journal of Ecology. 22, 125-135.

Werner, P. A., I. D. Cowie, and J. S. Cusack. 2006. Juvenile tree growth and

demography in response to feral water buffalo in savannas of northern

Australia: and experimental field study in Kakadu National Park. Australian

Journal of Botany. 54, 283-296.

Whitehead, P. J., J. Russell-Smith, and J. C. Z. Woinarski. 2005. Fire, landscape

heterogeneity and wildlife management in Australia's tropical savannas:

introduction and overview. Wildlife Research. 32, 369-375.

Williams, J. B. 1996. A phylogenetic perspective of evaporative water loss in birds.

The Auk. 113, 457-472.

Williams, P. R., R. A. Congdon, A. C. Grice, and P. J. Clarke. 2003. Effect of fire

regime on plant abundance in a tropical eucalypt savanna of north-eastern

Australia. Austral Ecology. 2003, 327-338.

Williams, R. J., A. D. Griffiths, and G. E. Allan. 2002. Fire regimes and biodiversity

in the savannas of northern Australia. Pages 281-304 in R. A. Bradstock, J.

E. Williams, and A. M. Gill, editors. Flammable Australia: The Fire Regimes

and Biodiversity of a Continent. Cambridge University Press, Cambridge.

Williams, T. D., W. O. Challenger, J. K. Christians, M. Evanson, O. P. Love, and F.

Vezina. 2004. What causes the decrease in haematocrit during egg

production? Functional Ecology. 18, 330-336.

270 Wingfield, J. C., M. C. Moore, and D. S. Farner. 1983. Endocrine responses to

inclement weather in naturally breeding populations of White-crowned

Sparrows (Zonotrichia leucophrys pugetensis). The Auk. 100, 56-62.

Wingfield, J. C., and B. Silverin. 1986. Effects of corticosterone on territorial

behaviour of free-living male song sparrows, Melospiza melodia. Hormones

and Behaviour. 20, 405-417.

Wingfield, J. C., C. M. Vleck, and M. C. Moore. 1992. Seasonal changes of the

adrenocortical response to stress in birds of the Sonoran desert. Journal of

Experimental Zoology. 264, 419-428.

Wingfield, J. C., P. Deviche, S. Sharbough, L. B. Astheimer, R. L. Holberton, R.

Suydam, and K. Hunt. 1994. Seasonal changes in the adrenocortical

responses to stress in Redpols, Acanthis flammea, in Alaska. Journal of

Experimental Zoology. 270, 372-380.

Wingfield, J. C., K. M. OReilly, and L. B. Astheimer. 1995. Modulation of the

adrenocortical responses to acute stress in arctic birds: a possible

ecological basis. American Zoologist. 35, 285-294.

Witter, M. S., and I. C. Cuthill. 1993. The ecological costs of avian fat storage.

Philosophical transactions of the royal society of London, Series B. 340, 73-

92.

Witter, M. S., J. P. Swaddle, and I. C. Cuthill. 1995. Periodic food availability and

strategic regulation of body mass in the European Starling, Sturnus vulgaris.

Functional Ecology. 9, 568-574.

Woinarski, J. C. Z., and R. W. Braithwaite. 1990. Conservation foci for Australian

birds and mammals. Search. 21, 65-69. 271 Woinarski, J. C. Z. 1992a. The conservation status of the white-throated grasswren

Amytornis woodwardi, an example of problems in status designation. The

Northern Territory Naturalist. 13, 1 - 5.

Woinarski, J. C. Z. 1992b. Biogeography and conservation of reptiles, mammals

and birds across north-western Australia: an inventory and base for planning

an ecological reserve system. Wildlife Research. 19, 665-705.

Woinarski, J. C. Z., and S. C. Tidemann. 1992. Survivorship and some population

parameters for the endangered Gouldian Finch Erythrura gouldiae and two

other finch species at two sites in tropical northern Australia. Emu. 92, 33-

38.

Woinarski, J. C. Z., and A. Fisher. 1995. Wildlife of Lancewood (Acacia shirleyi)

thickets and woodlands in northern Australia. 1. Variation in vertebrate

species composition across the environmental range occupied by

lancewood vegetation in the Northern Territory. Wildlife Research. 22, 379-

411.

Woinarski, J. C. Z. 1999. Prognosis and framework for the conservation of

biodiversity in rangelands: building on the north Australian experience.

Pages 639-645 in D. Elderidge, and Freudenberger, editors. Proceedings of

People and Rangelands: Building the Future. Proceedings of the VIth

International Rangelands Congress. VI International Rangelands Congress,

Inc., Aitkenvale, QLD.

Woinarski, J. C. Z., C. Brock, A. Fisher, D. Milne, and B. Oliver. 1999. Response of

birds and reptiles to fire regimes on pastoral land in the Victoria River

District, NT. Rangeland Journal. 21, 24-38. 272 Woinarski, J. C. Z., D. Milne, and G. Wanganeen. 2001. Changes in mammal

populations in relatively intact landscapes of Kakadu National Park,

Northern Territory, Australia. Austral Ecology. 26, 360-370.

Woinarski, J. C. Z., A. N. Andersen, T. B. Churchhill, and A. J. Ash. 2002.

Response of ant and terrestrial spider assemblages to pastoral and military

land use, and to landscape position, in a tropical savanna woodland in

northern Australia. Austral Ecology. 27, 324-333.

Woinarski, J. C. Z., and A. J. Ash. 2002. Responses of vertebrates to pastoralism,

military land use and landscape position in an Australian tropical savanna.

Austral Ecology. 27, 311-323.

Woinarski, J. C. Z., and A. Fisher. 2003. Conservation and the maintenance of

biodiversity in the rangelands. The Rangeland Journal. 25, 157-171.

Woinarski, J. C. Z., M. Armstrong, O. Price, J. McCartney, A. D. Griffiths, and A.

Fisher. 2004a. The terrestrial vertebrate fauna of Litchfield National Park,

Northern Territory: monitoring over a 6-year period and response to fire

history. Wildlife Research. 31, 587-596.

Woinarski, J. C. Z., and C. P. Catterall. 2004. Historical changes in the bird fauna

at Coomooboolaroo, northern Australia, from the early years of pastoral

settlement (1873 to 1999). Biological Conservation. 116, 379-401.

Woinarski, J. C. Z., J. Risler, and L. Kean. 2004b. Response of vegetation and

vertebrate fauna to 23 years of fire exclusion in a tropical Eucalyptus open

forest, Northern Territory, Australia. Austral Ecology. 29, 156-176.

273 Woinarski, J. C. Z., R. J. Williams, O. Price, and B. Rankmore. 2005. Landscapes

without boundaries : Wildlife and their environments in northern Australia.

Wildlife Research. 32, 377-388.

Woinarski, J. C. Z., J. C. McCosker, G. Gordon, B. Lawrie, C. James, J.

Augusteyn, L. Slater, and T. Danvers. 2006. Monitoring change in the

vertebrate fauna of central Queensland, Australia, over a period of broad-

scale vegetation clearance, 1973-2002. Wildlife Research. 33, 263-274.

Woinarski, J. C. Z., B. Mackey, H. Nix, and B. Traill 2007. The Nature of Northern

Australia. Australian National University, Canberra, ACT.

Yates, C., and J. Russell-Smith. 2003. Fire regimes and vegetation sensitivity

analysis: an example from Bradshaw Station, monsoonal northern Australia.

International Journal of Wildland Fire. 12, 349-358.

Yibarbuk, D., P. J. Whitehead, J. Russell-Smith, D. Jackson, C. Godjuwa, A.

Fisher, P. Cooke, D. Choquenot, and D. M. J. S. Bowman. 2001. Fire

ecology and Aboriginal land management in central Arnhem Land, northern

Australia: a tradition of ecosystem management. Journal of Biogeography.

28, 325-355.

Zanette, L., M. Clinchy, and J. N. M. Smith. 2006. Food and predators affect egg

production in song sparrows. Ecology. 87, 2459-2467.

Zhou, W. T., M. Fujita, and S. Yamamoto. 1999. Thermoregulatory responses and

blood viscosity in dehydrated heat-exposed broilers (Gallus domesticus)

Journal of Thermal Biology. 24, 185-192.

Zysling, D. A., T. J. Greives, C. W. Breuner, J. M. Casto, G. E. Demas, and E. D.

Ketterson. 2006. Behavioural and physiological responses to experimentally 274 elevated testosterone in female dark-eyed juncos (Junco hyemalis carolinensis). Hormones and Behaviour. 50, 200-207.

275