Managing Diversity in the Riverina Rice Fields—

Home , African bullfrog, Limnodynastes, Myrmecophagy, Rio Grande leopard frog, Spotted grass frog, Willandra National Park

Reconciling Farming with Wildlife —Managing diversity in the Riverina rice fields—

RIRDC Publication No. 10/0007

RIRDCInnovation for rural Australia

Reconciling Farming with Wildlife: Managing Biodiversity in the Riverina Rice Fields

by J. Sean Doody, Christina M. Castellano, Will Osborne, Ben Corey and Sarah Ross

April 2010

RIRDC Publication No 10/007 RIRDC Project No. PRJ-000687

ISBN 1 74151 983 7 ISSN 1440-6845

Reconciling Farming with Wildlife: Managing Biodiversity in the Riverina Rice Fields Publication No. 10/007 Project No. PRJ-000687

The information contained in this publication is intended for general use to assist public knowledge and discussion and to help improve the development of sustainable regions. You must not rely on any information contained in this publication without taking specialist advice relevant to your particular circumstances.

While reasonable care has been taken in preparing this publication to ensure that information is true and correct, the Commonwealth of Australia gives no assurance as to the accuracy of any information in this publication.

The Commonwealth of Australia, the Rural Industries Research and Development Corporation (RIRDC), the authors or contributors expressly disclaim, to the maximum extent permitted by law, all responsibility and liability to any person, arising directly or indirectly from any act or omission, or for any consequences of any such act or omission, made in reliance on the contents of this publication, whether or not caused by any negligence on the part of the Commonwealth of Australia, RIRDC, the authors or contributors.

The Commonwealth of Australia does not necessarily endorse the views in this publication.

This publication is copyright. Apart from any use as permitted under the Copyright Act 1968, all other rights are reserved. However, wide dissemination is encouraged. Requests and inquiries concerning reproduction and rights should be addressed to the RIRDC Publications Manager on phone 02 6271 4165.

Researcher Contact Details

J. Sean Doody Department of Botany and Zoology, Australian National University, ACT 2601

Phone: 0418 599 719 Fax: 02 6125 5573 Email: [email protected]

In submitting this report, the researcher has agreed to RIRDC publishing this material in its edited form.

RIRDC Contact Details

Rural Industries Research and Development Corporation Level 2, 15 National Circuit BARTON ACT 2600

PO Box 4776 KINGSTON ACT 2604

Phone: 02 6271 4100 Fax: 02 6271 4199 Email: [email protected]. Web: http://www.rirdc.gov.au

Electronically published by RIRDC in April 2010 Print-on-demand by Union Offset Printing, Canberra at www.rirdc.gov.au or phone 1300 634 313

Foreword

Although the establishment of the Australian rice industry has led to an altered pastoral landscape, it has also resulted in the seasonal availability of extensive aquatic habitats in a previously drier landscape. Consequently, rice-farming areas are of considerable importance to a range of vertebrate fauna.

The study aimed to begin farm management of biodiversity using the knowledge gained in the previous study. An experimental approach for incorporation into the Environmental Champions Program was used in tandem with monitoring and further research on selected significant species. However, the re-gripping of Australia’s worst drought on record had a significant impact on the project, through retraction of funding, and through the environmental impact on revegetation efforts on farm.

Nevertheless, significant outcomes were achieved, and this publication details findings from (1) baseline monitoring of vertebrates on rice farms; (2) a study of frogs as natural pest control for rice crops; and (3) a study of how an iconic snake species utilises farms. The project concludes by integrating knowledge gained in these studies with that of other studies to formulate management strategies for rice farmers and other stakeholders in the region.

This report is an addition to RIRDC’s diverse range of over 2000 research publications and it forms part of our Rice R&D program, which aims to improve the profitability and sustainability of the Australian rice industry through the organisation, funding and management of a research, development and extension program that is both market and stakeholder driven.

Most of RIRDC’s publications are available for viewing, free downloading or purchasing online at www.rirdc.gov.au. Purchases can also be made by phoning 1300 634 313.

Peter O’Brien Managing Director Rural Industries Research and Development Corporation

iii

Acknowledgments

We sincerely thank the following farmers for allowing continuous access to their properties, for advice and general assistance during the study: Barry Kirkup, David Pike, Gary and Marg Knagge, Jim and Caroline Maskus, June and Russell Mason, Michael McAliece, Mick Anderson and Rob Houghton. Rob, Jayne, Alice and Jacob Houghton and Barry, Gillian and Renee Kirkup also played instrumental roles in data collection on their homesteads. We also thank Marc Serafin for providing accommodation and Max O’Sullivan for conducting the bird surveys.

Russell and Robin Ford were invaluable to us in many ways, including general needs, accommodation, local advice and exceptional cooperation. Without their help the study would not have been possible. We sincerely appreciate the time, energy and resources put forth by Russell Ford and we thank him for the construction of the experimental rice bays.

For assistance in the field, we thank Rohan Rewhinkel, Imi Moore, Dave Rhind, Mitchell Sidwell, Geoff Sidwell, Tim Ross, Emily Ford, James Knight, Paul Read, Kirsten Wheatley, Glenn Murray, Dave Steer, Christy Davies, Brett Stewart and Brian Green. We are also grateful to Lesley Ishiyama and Dave Rhind for their assistance in the laboratory. Geoff Heard and Peter Robertson shared their knowledge of Morelia spilota with us and provided supporting literature. We are indebted to the Gardiner family for their support and hospitality and to David Pederson for assisting with the data analysis.

The study would not have been possible without the assistance and support of Janelle McGufficke. We thank those showing an interest in the study and other members of the RIRDC Rice R&D Committee, members of the Biodiversity Steering Committee for the rice industry and Rice Research Australia Pty Ltd.

The Rural Industries Research and Development Corporation, Ricegrowers’ Association of Australia and Rice Research Australia funded this study.

Contents

Foreword ...... iii Acknowledgments...... iv Tables and Figures ...... vii Executive Summary ...... x Introduction ...... 1 Reconciliation Ecology ...... 1 Importance of Biodiversity ...... 2 Revegetation ...... 2 Significant Species ...... 3 Frog Species Under Study ...... 4 Dietary Habits of Frogs ...... 7 Feeding Mechanisms ...... 7 Ontogenetic Dietary Shifts ...... 8 Foraging Strategies ...... 8 Prey Selection ...... 9 Diet Overlap in Sympatric Species ...... 10 Prey Availability ...... 10 Top Predators ...... 11 Space Use by Wildlife ...... 11 The Carpet Python ...... 12 Objectives ...... 13 Vertebrate Biodiversity ...... 13 Selected Significant Species Research ...... 13 Methodology ...... 14 Study Location ...... 14 Vegetation ...... 15 Climate ...... 15 Water ...... 16 Rice Farming ...... 16 Study Design and Period ...... 16 Sampling Techniques ...... 17 Visual Encounter Surveys ...... 17 Bird Counts ...... 18 Artificial Cover Surveys ...... 19 Frog Surveys ...... 19 Frog Dietary Study ...... 20 Abundance ...... 20 Prey Item Identification ...... 20 Quantifying Prey Items...... 22 Food Consumption Rates ...... 23 Body Size vs. Prey Size ...... 23 Ontogenetic Dietary Shift...... 23 Seasonal Availability of Prey ...... 23 Statistical Analyses ...... 24 Data Collection for Carpet Pythons ...... 24 Radio-telemetry ...... 24 Habitat Use ...... 25 Habitat Characterisation and Analyses ...... 25 Tree Measurement ...... 27 Diet and Relative Prey Density ...... 27

Activity and Movements Analyses ...... 28 Home Range Analysis ...... 28 Statistical Analysis ...... 28 Results ...... 29 Vertebrate Surveys ...... 29 VESs and Artificial Cover Surveys ...... 29 Bird Counts ...... 29 Frog Study ...... 29 Abundance ...... 29 Dietary Items ...... 29 Food Consumption Rates ...... 39 Body Size and Food Intake...... 39 Prey Consumption ...... 39 Ontogenetic Dietary Shift...... 49 Sex Differences in Diet ...... 52 Seasonal Availability of Prey ...... 52 Carpet Python Study ...... 65 Activity and Movements ...... 65 Home Range Estimates ...... 65 Patterns of Habitat Use ...... 69 Patterns in Macro- and Microhabitat use ...... 69 Microhabitat Selection ...... 70 Patterns in Tree Use ...... 72 Characteristics of Trees ...... 74 Diet and Relative Prey Density ...... 78 Discussion ...... 81 Vertebrate Surveys ...... 81 Low Detectability for Vertebrate Surveys ...... 81 Differences in Bird Species Between Sites ...... 81 Frog Study ...... 81 Species Abundance ...... 81 Dietary Analysis ...... 82 Body Size Versus Prey Size ...... 83 Ontogenetic Dietary Shifts ...... 83 Sex Related Differences in Diet ...... 83 Dietary Differences Between Species ...... 83 Seasonal Availability of Prey ...... 84 Rice Pest Consumption ...... 84 Carpet Python Study ...... 85 Activity Patterns ...... 85 Movement Patterns ...... 86 Home Range ...... 87 Overall Patterns in Habitat Use ...... 88 Patterns in Macro- and Microhabitat Use ...... 89 Microhabitat Selection ...... 90 Patterns in Tree Use ...... 90 Characteristics of Trees Used ...... 90 Diet and Relative Prey Density ...... 91 Implications for Conservation ...... 91 Recommendations ...... 93 Monitoring On-farm Vertebrate Wildlife ...... 93 The Importance of Frogs in Controlling Rice Pests ...... 93 Ecology of Carpet Pythons in an Agricultural Landscape ...... 93 References ...... 95

Tables

Table 1 Structural variables used in the analysis of microhabitat selection by M. spilota with associated abbreviations and sampling radii ...... 26 Table 2 The species detected by the three sampling methods used in this study ...... 30 Table 3 Species of birds observed in revegetated and controls sites according to season ...... 32 Table 4 Monthly summary of the numbers of frogs collected, number of stomachs flushed and the samples extracted for the study period...... 37 Table 5 Composition of the diet of L. tasmaniensis collected between January and May 2007 ...... 40 Table 6 The composition of the diet of L. fletcheri collected between January and May 2007 ...... 42 Table 7 The frequency of occurrence in the diet of L. tasmaniensis and L. fletcheri of specific invertebrates known to be pests in rice crops ...... 43 Table 8 Comparison of diet composition (numeric proportion in %) between two age groups (size classes) for both study species...... 49 Table 9 Spearman rank correlation coefficients between adult and juvenile L. tasmaniensis and L. fletcheri when examining the prey types consumed ...... 53 Table 10 Spearman's rank correlation coefficients between the frequency the prey orders are consumed by males and females of L. tasmaniensis and L. fletcheri ...... 54 Table 11 Correlation between the availability of prey caught in pitfall traps and sweep samples and the prey found in the stomach contents of L. tasmaniensis and L. fletcheri ...... 56 Table 1 Comparison of frequency of occurrence (%) of prey in L. tasmaniensis stomachs with relative abundance of prey collected in pitfall traps and sweep samples at Old Coree...... 57 Table 13 Comparison of frequency of occurrence (%) of prey in L. flectheri stomachs with relative abundance of prey collected in pitfall traps and sweep samples at Leeton...... 59 Table 14 Comparison of frequency of occurrence (%) of prey in L. fletcheri stomachs with relative abundance or prey collected in pitfall traps and sweep samples at Old Coree ...... 61 Table 15 Comparison of frequency of occurrence (%) of prey in L. fletcheri stomachs with relative abundance ...... 63 Table 16 Movement and space use patterns in Morelia spilota. Note: values are means and standard errors are given in parentheses ...... 66 Table 17 Means (± SE) of variables used in the analysis of microhabitat selection for the random locations and locations of male, female and juvenile M. spilota at Willandra ...... 71 Table 18 Distances between the four group centroids in the discriminant space and their statistical significance for the analysis of habitat selection by M. spilota ...... 71 Table 19 Summary statistics for the three discriminant functions and their pooled within- groups correlations (r) with the discriminating variables used in the analysis of habitat selection by M. spilota ...... 72 Table 20 Summary of tree use by radio-tracked M. spilota...... 74 Table 21 Abundance and availability of trees compared with usage of trees by M. spilota ...... 76 Table 22 Prey items of M. spilota from Willandra Homestead and Willandra National Park, as determined from faecal examination and direct observation...... 79

vii

Figures

Figure 1 Limnodynastes tasmaniensis adult (L. Fuscko) ...... 6 Figure 2 Limnodynastes fletcheri juvenile (www.wikipedia.com)...... 6 Figure 3 Foam nest created by L. tasmaniensis (L. Fucsko)...... 7 Figure 4 Adult Carpet Python (Morelia spilota) ...... 12 Figure 5 Map of the rice-growing region within the NSW Riverina showing the three main irrigation districts: Murrumbidgee (top), Coleambally (middle) and Murray Valley (bottom) ...... 14 Figure 6 Total monthly rainfall received at the two study locations during the study period...... 15 Figure 7 An example of how secretive species use natural cover objects ...... 18 Figure 8 Example of the technique used in artificial cover surveys...... 19 Figure 9 Overview of the experimental rice bays at Old Coree. Three bays with fences and pitfall traps (white with grey borders) and three bays without fences (solid grey)...... 21 Figure 10 Fenced rice bays at Old Coree ...... 21 Figure 11 Equipment used in stomach flushing procedure ...... 22 Figure 12 Stomach contents collected in sieve for later analysis ...... 22 Figure 13 Mean number of individuals detected in revegetated (n = 7) and control sites (n = 7) in each season...... 31 Figure 14 Total number of L. tasmaniensis captured at Old Coree from the three experimental rice bays in 2007 ...... 35 Figure 15 The total number of L. fletcheri captured at Old Coree from the three experimental rice bays in 2007 ...... 35 Figure 16 The total number of L. tasmaniensis captured from the four 40 m drift fences at three study farms in Leeton and Old Coree in 2007 ...... 36 Figure 17 The total number of L. fletcheri captured from the four 40 m drift fences at three study farms in Leeton and Old Coree in 2007 ...... 36 Figure 18 Monthly variations in the frequency of occurrence of the major larval prey taxa in the diet of L. tasmaniensis...... 44 Figure 19 Monthly variations in the frequency of occurrence of the major larval prey taxa in the diet of L. fletcheri ...... 44 Figure 20 Monthly variation in the frequency of occurrence of the major terrestrial prey taxa in the diet of L. tasmaniensis ...... 45 Figure 21 Monthly variation in the frequency of occurrence of the major terrestrial prey taxa in the diet of L. fletcheri ...... 45 Figure 22 Monthly variations in the mean total volume of prey items found in the stomachs of L. tasmaniensis...... 46 Figure 23 Monthly variations in the mean total volume of prey items found in the stomachs of L. fletcheri ...... 46 Figure 24 Relationship between frog body size and smallest prey item in stomachs of L. tasmaniensis ...... 47 Figure 25 Relationship between frog body size and smallest prey item in stomachs of L. fletcheri ...... 47 Figure 26 Relationship between frog body size and largest prey item in stomachs of L. tasmaniensis ...... 48 Figure 27 Relationship between frog body size and largest prey item for L. fletcheri ...... 48 Figure 28 Frequency of occurrence of 10 major prey orders in adult and juvenile L. tasmaniensis ...... 51 Figure 29 Frequency of occurrence of 10 major prey orders in the stomachs of adult and juvenile L. fletcheri ...... 51

viii

Figure 30 The frequency of occurrence (%) of aquatic prey orders from sweep samples at Old Coree and Leeton and stomach samples from January to March 2007 ...... 55 Figure 31 Seasonal movement frequencies for M. spilota from the woodland of Willandra National Park and Willandra Homestead ...... 67 Figure 32 Frequency distribution of movement distance intervals for M. spilota from the woodland of Willandra National Park and Willandra Homestead ...... 67 Figure 33 Seasonal distances moved by M. spilota from the woodland of Willandra National Park and Willandra Homestead ...... 68 Figure 34 Positions (± SE) of the group centroids of random locations and locations of male, non-gravid female and juvenile M. spilota on the two discriminant axes in the analysis of habitat selection...... 73 Figure 35 Characteristics of trees selected by radio-tracked M. spilota compared with available trees in the same area ...... 77 Figure 36 Relative abundance (number of individuals per trap night) of small mammals in Willandra National Park and around Willandra Homestead...... 80

Executive Summary

What the report is about

This document reports on the major findings of three projects investigating biodiversity issues in the rice-growing industry. Specifically, it reports on (1) the initiation of a monitoring program for on-farm vertebrates associated with revegetation efforts; (2) the role of frogs in controlling rice pests; and (3) how a focal species of conservation concern, the carpet python, utilizes a agricultural landscapes.

Who the report is targeting

The report targets all stakeholders the in rice-growing areas of the Riverina, including RIRDC, RGA, Rice Research Pty Ltd., irrigation companies, NSW Parks and Wildlife, on-ground land management organisations (e.g., Landcare, Greening Australia), the scientific community, and the general community.

Background

‘Reconciliation ecology’ is the science of conserving biodiversity in areas where we live, work and play. Although the establishment of the Australian rice industry has led to an altered pastoral landscape, it has also resulted in the seasonal availability of extensive modified aquatic habitats in a previously drier, altered landscape. Consequently, rice-farming areas are of considerable importance to a range of vertebrate fauna groups, as demonstrated by our previous RIRDC-sponsored research (UCA-6A; Sustainable Management of On-farm Biodiversity in the Rice-growing Industry). In those studies we demonstrated that vertebrate diversity was underpinned by two major factors: flooded rice bays and remnant vegetation patches. The former were important drivers of abundance in frogs, snakes, turtles and waterbirds, while the latter were critical for richness of reptiles, frogs, mammals, and woodland birds.

In the present project, we further investigated these links by conducting three studies. The first study was the initiation of a monitoring program for on-farm vertebrates associated with revegetation efforts; the second study determined the role of frogs in controlling rice pests; while the third study determined how a focal species of conservation concern, the carpet python, utilized an agricultural landscape.

Our previous studies and those of others identified remnant vegetation as critical to biodiversity, especially in woodland birds and small reptiles. Thus, revegation efforts would be beneficial to on- farm biodiversity. Therefore, in conjunction with the industry’s Environmental Champions Program, we initiated a long-term vertebrate monitoring program associated with revegetation efforts.

Previous research estimated that, on average, hundreds of millions of spotted grass frogs (Limnodynastes tasmaniensis) were produced annually from rice bays in the Riverina Bioregion of New South Wales (NSW), and these frogs probably represent the highest biomass of any vertebrate on rice farms. This highlighted the possibility that these frogs, which consume insects, may be controlling economically important pests of rice crops.

The inland race of the carpet python (Morelia spilota metcalfei) is of conservation concern due to its apparent decline west of the dividing range. This decline is believed to be caused by land clearing, and remaining populations in the Riverina region are generally restricted to areas with significant vegetation, of which most are along rivers and creeks. The present study used radio-telemetry to determine how these snakes used agricultural landscapes.

Methods

For the first study we used a variety of survey techniques to establish a baseline relative abundance of vertebrates on seven rice farms, including standardised bird counts, artificial shelter surveys for reptiles and frogs, and visual encounter surveys.

In the second study we created experimental rice bays to determine frog abundance, and we used stomach flushing to determine what pests frogs were consuming. The main study species were the spotted grass frog (Limnodynastes tasmaniensis) and the barking marsh frog (L. fletcheri).

In the third study we used radiotelemetry to determine how carpet pythons (Morelia spilota) utilise agricultural landscapes. Specifically we studied habitat selection, movements, diet, and thermoregulation in 17 pythons in two landscape types: homestead and woodland.

Results/key findings

The re-gripping of Australia’s worst drought on record had a significant impact on the project through (1) the environmental impact on revegetation efforts on farms; (2) by influencing the detectability of animals during surveys; and (3) through retraction of funding by RIRDC. The majority of plantings failed within a few months of the onset of the monitoring due to water restrictions and lack of rainfall. Moreover, arid conditions resulted in extremely low counts of small vertebrates during visual encounter surveys and artificial cover surveys. Our third method, standardized bird counts, met with better success, although bird diversity was very likely also influenced by drought conditions. Nevertheless, we have included some comparisons of bird counts between revegetated areas and control sites, because a few of the revegetated sites were located in remnant vegetation.

We tested the idea that frogs control rice pests by stomach flushing > 1000 frogs captured in rice bays. Both the spotted grass frog and the barking marsh frog (L. fletcheri) were generalist and opportunistic predators, consuming mainly beetles (Coleoptera), true bugs (Hemiptera), and ants (Hymenoptera). Sex or age-class did not influence dietary composition, but larger frogs consumed larger prey items. It was estimated that these frogs consumed about eight insects per day. In addition to economically harmless invertebrates, these frogs consumed a wide-range of pest species and may play a significant role as biological control agents. Rice plant pests consumed included the stink bug (Eysarcoris trimaculatus), water snails (Glytophysa sp. and Isidorella newcombi), and paddy bug (Leptocorisa acuta).

We are unable to locate sufficient numbers of snakes for study on rice farms (e.g., Old Coree Homestead), and so we conducted the study on 17 pythons at Willandra National Park, NSW. Although rice is not grown at Wilandra, its landscape offered the opportunity to achieve our objectives. There were clear differences in habitat and resource use between snakes inhabiting the ‘park’ (riparian woodland) and those inhabiting human-modified environments (around the homestead). Homestead snakes moved less often during spring and summer than park snakes, possibly due to a more plentiful food supply there. Small mammal trapping revealed that the House Mouse, Mus musculus, was more plentiful around the homestead than in the surrounding park. However, home ranges did not differ in size between homestead and park. Habitat use differed significantly between landscape types: homestead snakes spent most of their time within buildings, whereas ‘park’ or woodland snakes inhabited mainly trees, logs, or thickets. Dietary composition also differed between homestead and park snakes, with the former consuming mainly exotic species (house mouse and European rabbit), while park snakes consumed a variety of mainly native fauna (birds, eggs, and a lizard). Compared to pythons from coastal areas, pythons in the present study had considerably smaller home ranges, although daily movements were up to four times higher. Snakes at Willandra were much more arboreal and used tree hollows more extensively than those in northeastern and southeastern NSW.

Implications for relevant stakeholders

Extreme drought conditions rendered our monitoring results inadequate for longer-term comparisons. Overall, this program was unsuccessful due to drought, and should not be re-initiated prior to the cessation of drought conditions.

Assuming that hundreds of millions of frogs are produced in a given year, billions of insects would be consumed annually in rice-growing areas of the Riverina. This suggests a role for frogs as natural biological control agents of rice pests. Thus, frogs are likely economically beneficial to rice farmers, under current water practices. A reduction in frog abundance would likely cause a boom in certain pests that might only be controlled with the application of insecticides. Current frog diversity may thus facilitate the current low level of pesticide use in Australian rice crops. Decreases in the availability of water in farms will result in a reduction of frog abundance and an increase in pests, although we do not know by how much. Dry cropping will result in a significant reduction in frog abundance compared to rice cropping.

Our findings of how carpet pythons utilise agricultural landscapes highlight the need for protecting riparian zones, especially overstorey, in those landscapes. Also, they indicate a role for buildings as snake habitat, provided that these buildings are in close proximity to woodland. Finally, these findings suggest that exotic species are a critical resource for some populations of snakes, complicating issues underlying invasive species control.

Recommendations

Monitoring study:

(1) Do not re-initiate monitoring prior to the cessation of drought conditions;

Frogs as pest control:

(1) maintain current levels of water use and practices to maintain frog diversity;

(2) monitor changes in frog abundance associated with major changes in water use/availability;

(3) experimental exclusion of frogs from rice bays would confirm the economical importance of frogs in controlling rice pests (and pests of other crops);

Carpet pythons in agricultural landscapes:

(1) protect remnant vegetation on farms, including that close to homesteads and other buildings

(2) revegetate farms with overstorey and understorey, especially in riparian areas such as rivers, creeks and in black box depressions.

(3) Connect remnants and revegetated areas, especially between riparian areas;

(4) Consider the importance of rabbits, mice, and willow trees for carpet pythons when making decisions regarding controlling those exotic species;

xii

Introduction

Reconciliation Ecology An important recent advance in conservation biology is the realisation that we cannot hope to conserve the earth’s species through protected reserves alone (Soule and Sanjayan, 1998; Rosenweiz, 2003). At least 95% of the earth’s land surfaces have been “refashioned” by humans (Srivasta et al., 1996; Rosenwiez, 2003), and these areas are likely to be subject to further modification. Increasingly isolated habitats patches, in reserves and otherwise, are earmarked for increases in extinction rates through natural (e.g., drought and fire) and unnatural (e.g., greenhouse events) processes (Rosenwiez, 2003).

Given the unlikely prospect of restoring ecosystems or habitats and their biodiversity to their original conditions, we are faced with the problem of reconciling biodiversity conservation with human land use. The philosophy, science, and practice of conservation must be framed against the reality of human-dominated ecosystems, rather than the separation of humanity and nature underlying the modern (traditional) conservation movement (Western and Pearl, 1989; Srivasta et al., 1996; Knight, 1999; Western, 2001). For example, traditionally biodiversity conservation has been thought to be best practiced through the establishment of reserves (Margules and Pressey, 2000; Rodrigues and Gaston, 2001), but this approach may have limitations in relatively heterogeneous production landscapes outside of protected areas (Fischer et al., 2004). Yet, production areas dominate the earth’s surface and are increasing (Srivasta et al., 1996).

The strategic inclusion of a new branch of ecology termed ‘reconciliation ecology’, or the ‘maintenance of species friendly habitats in the places where humans live, work, and play’, may be our only chance of conserving biodiversity (Rosenweiz, 2003). Reconciliation ecology seeks environmentally friendly ways for us to continue to use the land for our benefit. Reconciliation ecology has already enjoyed some success as a conservation strategy (see examples in Rosenweiz, 2003). A major driver in reconciliation ecology is likely to be restoration ecology (see Saunders et al., 1993; Dobson et al., 1997). Restoring or improving habitats for biodiversity, however, requires detailed knowledge of biodiversity and what factors underpin it.

An increasing body of evidence indicates that considerable biodiversity can be maintained in ecosystems that have been modified for human use (Knight, 1999; Rosenweiz, 2003; but see Lemly et al., 2000). One such system is the irrigated landscape. For example, although wildlife use of farms has received very limited study in general (Lindenmeyer et al., 2003), irrigated rice agroecosystems have been shown to harbour considerable waterbird diversity (Fasola, 1986; Heitmeyer et al., 1989; Miller et al., 1989; Remsen et al., 1991; Pain, 1994; Brouder and Hill, 1995; Fasola et al., 1996; Elphick and Oring, 1998; Lane and Fujioka, 1998; Elphick and Oring, 2003; Maeda, 2001; Elphick, 2004), and frog diversity (Maeda and Matsui, 1989; Fujioka and Lane, 1997), and in some cases rice fields have been suggested to be ‘functionally equivalent’ to natural or semi-natural wetlands for some fauna (Fasola and Ruiz, 1996; Fujioka and Lane, 1997; Elphick, 2000). On the other hand, irrigated systems have resulted in the degradation of wetlands and their biodiversity around the world (Gerakis and Kalburtji, 1998; reviewed in Lemly et al., 2000) and the general trend across agricultural landscapes is that, as farming intensity increases, biodiversity decreases (Wood et al., 2000; Donald et al., 2001). In either case, understanding and maintaining biodiversity in rice agroecosystems is of considerable importance, because rice occupies a larger area than any other crop in the world (Fores and Comin, 1992), and about 40 % of the world’s population depends on rice for food (Odum, 1993). This situation will likely be even more critical in the future because the world’s human population is expected to at least double before it stablises, and the need for primary products is expected to triple in the next 50 years (Avery, 1996).

Importance of Biodiversity Biodiversity can be broadly defined as every species of wildlife (plants and animals), the ecosystems in which they live and their various ecological functions (Redford and Richter, 1999). It is usually quantified as the number of species present in a specific area, or species richness, from data collected while conducting inventory and monitoring programs (Hengeveld, 1996). Biodiversity is valuable in many ways. For instance, it has economic value in the form of material goods such as timber and plant and animal based pharmaceuticals, as well as, aesthetic and recreational value (Pimentel et al., 1992). Moreover, plants, animals and other organisms are vital in maintaining healthy ecosystems and play key roles in pollination, seed dispersal, nutrient cycling and pest control (Lovejoy, 1994). In agricultural systems biodiversity can help restore degraded land and water resources by reducing soil erosion and salinity, and through improved nutrient input and cycling lead to increased productivity (Kimber et al., 1999).

Despite the overall reduction in species richness that is associated with land clearing, a great deal of biodiversity can still be found in some agricultural landscapes. For instance, Doody et al. (2004) recorded more than 200 species of birds, mammals and reptiles using rice farms in the Riverina in New South Wales. They reported that many common species inhabited the farms, with some conspicuous groups such as the water birds and frogs being dependent on the seasonal flooding associated with rice production. Moreover, several threatened species, including the Southern Bell Frog (Litoria raniformis) and Superb Parrot (Polytelis swainsonii) were observed and nests of both species were located, suggesting that they were reproducing on the study farms. Although these findings are encouraging, the need to maintain and promote biodiversity in the rice ecosystem is still present since the existing species richness is probably a small fraction of what once occurred in the region.

Patches of remnant vegetation (i.e., the natural vegetation that remains after some portion of an area has been cleared) support a wide-range of wildlife in many different types of landscapes (Kimber et al., 1999). The number and types of animals that use these habitats often depend on the size, shape and location of the remnant (Saunders et al., 1991). For instance, Doody et al. (2004) recorded a greater variety of species in large, connected patches than in small, isolated ones on rice farms in the Riverina. Remnants need to be actively maintained however, in order to prevent insect damage and weed invasion (Hajek and Johnson, 2002).

Revegetation Revegetation is the planting of trees, shrubs and other plants in areas where the natural vegetation has been cleared. It also includes other activities that allow the regrowth of native vegetation through natural processes (Bennett et al., 2000). Most revegetated areas function as windbreaks, shelterbeds, hedgerows and shade trees and are often times fenced-off in order to exclude grazing livestock. There are many benefits that flow on from these activities including the provision of additional habitats for wildlife and movement corridors and linkages for migrating species (Bennett et al., 2000).

Under the stewardship of the Australian rice industry’s Environmental Champions Program (ECP), many rice growers are working to enhance biodiversity and ecosystem health through the planting and maintenance of natural areas (Adcock, 2003). While there is recognition that revegetation activities can help restore various habitats, there have been few studies that have monitored and quantified changes in wildlife populations following revegetation.

Significant Species Australian rice growers produce more than one million tones of rice annually, a large portion of which gets exported to more than 70 countries worldwide (Donovan, 2004). In comparison with other rice producing nations, Australia has relatively few invertebrate pests (Stevens, 2003). This has been mainly due to Australia’s geographic isolation and national quarantine system (Plant Health Australia, 2005). Increases in tourism, imports and exports, and the potential for pests to disperse from overseas however, means that exotic pests could pose a high risk to Australia’s rice industry in the near future (Plant Health Australia, 2005). Ensuring that the rice industry has the capacity to minimize the risks posed by pests and the ability to respond to them effectively is crucial for the sustainability and viability of the industry (Plant Health Australia, 2005).

A pest, as defined by the International Plant Protection Convention, is any species, strain or biotype of plant, animal, or pathogenic agent, injurious to plants or plant products (Plant Health Australia, 2005). Although there are more than 40 invertebrate pest species of rice in New South Wales (Plant Health Australia, 2005), the three most damaging are larval bloodworms (Chironomus tepperi Skuse), aquatic snails (Isidorella newcombi (Adams and Angus) primarily but also Glyptophysa sp.), and aquatic earthworms (Eukerria saltensis Beddard). Other less harmful but still significant pests are leafminers (Hydrellia michelae Bock), tadpole shrimp (Triops australiensis australiensis Spencer and Hall), and the common armyworm (Leucania convecta Walker) (Stevens, 2003). All of these pests either feed on rice plants and/or alter the environment in a manner that adversely affects plant growth. For instance, larval bloodworms and aquatic snails feed on the primary roots of young rice seedlings, whereas aquatic earthworms disrupt seedling establishment by tunneling through soil (Stevens, 2003). Leafminers, armyworms and tadpole shrimp feed on the leaves of rice plants (Godfrey, 2004; Lacy and Stevens, 2005). Damage from these pests can be severe; for example, larval bloodworms can reduce plant establishment by over 85% (Stevens, 2003). Rice pests are generally controlled by a variety of pesticides and other chemicals that may have secondary effects on the environment. For instance, copper sulfate, which is used to control snail infestations, can lead to soil poisoning thereby creating a toxic hazard for grazing livestock (Stevens, 2003).

In addition to the pests already present in Australia, there are several exotic species that threaten the rice industry. According to the National Rice Industry Biosecurity Plan, three of the top four threats are invertebrate pests (Plant Health Australia, 2005). Ranked according to their entry, establishment, and spread potential, and their expected economic impact, the highest risk species are the golden apple snail (Pomacea canaliculata), rice water weevil (Lissorhoptrus oryzophilus), and Khapra beetle (Trogoderma granarium) (Plant Health Australia, 2005). The golden apple snail, for instance, can be imported through the aquarium trade, become established due to its fast growth and reproduction, and spread rapidly through drainage channels in rice growing areas (Plant Health Australia, 2005). The invasion of this species throughout Asia has cost the Asian rice industry billions of dollars in snail control, replanting, and lost rice yields (Naylor, 1996). There are no chemical control measures available in Australia for most exotic rice pests (Plant Health Australia, 2005). Biological controls may help minimize the damage caused by exotic invertebrates, although none have been identified (Goldie, 1998).

Biological control is the use of living organisms as pest control agents (Waage and Greathead, 1988). Classical biological control techniques involve the introduction of an exotic species for the regulation of a pest population (Waage and Greathead, 1988). A case in point is the Australian vadalia beetle (Rodolia cardinalis) that was used to control the cottony cushion scale (Icerya purchasi), a pest that devastated the California citrus industry (Hoffman and Frodsham, 1993). While there are many benefits to biological control including reduced pesticide use, there are some drawbacks. For instance, some biological control agents can become competitors or predators of the native wildlife. A typical example is the mosquito fish, Gambusia holbrooki. This species was released in California for mosquito larvae control but has subsequently caused the extinction of newt (Taricha torosa)

populations throughout the state (Gamradt and Kats, 1996). The use of native biological control agents however, may have little impact on non-target native species and may save time and expense on research and quarantine procedures (Sheldon and Creed, 1995). Sheldon and Creed (1995) demonstrated that the North American aquatic weevil (Euhrychiopsis lecontei) successfully controlled the Eurasian watermilfoil (Myriophyllum spicatum), a weed that had invaded aquatic ecosystems in the United States. The use of naturally occurring species to suppress pests is termed natural control (Sheldon and Creed, 1995).

An important natural control of invertebrate populations is insectivorous frogs (Hirai and Matsui, 1999). Feeding trial experiments and stomach content analyses have been used to study prey selection in anurans (Freed, 1980; Donnelly, 1991). In some species, diet composition is influenced by age and sex (Dendrobates pumilio; Donnelly, 1991), body size (Lechriodus fletcheri; Lima et al., 2000), seasonal availability of prey (Rana nigromaculata; Hirai and Matsui, 1999) and prey activity (Hyla cinerea; Freed, 1980).

Research on predator-prey relationships and dietary preferences can be used to evaluate the role of frogs as pest control (Hirai and Matsui, 1999). For instance, Hirai and Matsui (1999) reported finding at least six species of rice pests in the stomach contents of Pond Frogs (Rana nigromaculata) collected from rice fields in Japan. They also observed a high similarity in the diets of Japanese Treefrogs (Hyla japonica) and Pond Frogs when these species foraged simultaneously in rice fields (Hirai and Matsui, 2002). Frogs may also be an important natural control of pests in agroecosystems in Argentina (Attademo et al., 2005). Attademo et al. (2005) reported that frogs in Argentina feed on invertebrate pests of soybean. Since insectivorous frogs consume crop pests they may be economically beneficial (Hirai and Matsui, 1999; Attademo et al., 2005).

In rice farm ecosystems in New South Wales, Spotted Grass Frogs (Limnodynastes tasmaniensis) use rice bays and remnant patches of vegetation for survival and reproduction (Doody et al., 2004). Limnodynastes tasmaniensis is a wide-ranging species that can be observed throughout Tasmania, Victoria, eastern South Australia, New South Wales and eastern Queensland (Cogger, 2000). It uses a variety of permanent and temporary wetlands ranging from swamps and lagoons to flooded grasslands and farm dams. Adult frogs shelter beneath rocks and logs along side watercourses, among emergent vegetation or floating debris, and in cracks in the ground (Hanley, 1999; Cogger, 2000). Their main breeding season is late spring and summer although breeding can be observed throughout the year (Hanley, 1999). Mating pairs construct aquatic foam nests characteristic of the members of this genus (Tyler and Davies, 1979). Nests float at the surface of the water, contain between 80 and 1500 eggs, and measure between 50 and 80 mm in diameter (Hanley, 1999). Tadpoles and juvenile frogs remain in the ponds or farm dams in which they hatched for several months; during this time they metamorphose and grow. Although the food habitats of L. tasmaniensis tadpoles are unknown, they do have upper and lower labile teeth indicating that they may be carnivorous at this life stage (Anstis, 2002). Adult and postmetamorphic juvenile L. tasmaniensis feed primarily on invertebrates (Tyler, 1992).

Frog Species Under Study

Limnodynastes tasmaniensis and Lymnodynastes fletcheri belong to the taxonomic family Myobatrachidae. The Myobatrachidae is found only in Australia and New Guinea and encompasses mainly terrestrial species such as the burrowing (Heleioporus and Neobatrachus), marsh (Limnodynastes and Crinia), and stream dwelling frogs (Taudactylus and Rheobatrachus) (Duellman and Treub, 1985). Unique to the family were two species of gastric-brooding frogs, Rheobatrachus silus and Rheobatrachus vitellinus, both of which are now extinct.

Limnodynastes tasmaniensis is commonly referred to as the Spotted Grass Frog or Spotted Marsh Frog. It is a relatively small species with adult males ranging from 31-42 mm and females between 32- 47 mm in length (Barker et al., 1996). The colouration of L. tasmaniensis is highly variable; it can be

light brown to olive green and has irregular shaped dark spots and blotches on its dorsal surface. Most individuals have a mid-dorsal stripe which can vary in colour from cream to bright red (Barker et al., 1996). The skin on this species is basically smooth, but there are patches with low, tiny warts. The frog’s ventral surface is usually white in colour. Mature males have a bright yellow throat, which changes colour when distended (Barker et al., 1996). Limnodynastes tasmaniensis is widely distributed throughout eastern Australia. Its geographic range includes southern Queensland, western South Australia, New South Wales, Victoria and eastern Tasmania (Cogger, 2000). Limnodynastes fletcheri is commonly referred to as the Barking Marsh Frog, or the Long-thumbed Frog and is slightly bigger than L. tasmaniensis (Figure 5). Adult males range from 37-46 mm in length and females from 38-55 mm in length (Barker et al., 1996). The colouration of L. fletcheri is light gray, or brown with dark blotches on its dorsal surface. These blotches are more irregular in shape than the blotches on L. tasmaniensis. Most individuals have pinkish red patches on the back of each upper eyelid. The skin on the frog’s dorsal surface is smooth with low round warts and the ventral surface is usually white and smooth. This species also has a somewhat rounded snout (Barker et al., 1996). This species is not as widely spread as L. tasmaniensis. The former occurs in parts of Southern Queensland, New South Wales, Victoria and South Australia.

Limnodynastes tasmaniensis and L. fletcheri both occupy a diversity of natural and human created habitats including woodlands, shrublands, grasslands, agricultural fields and urban landscapes (Barker et al., 1996). The two species usually live in vegetation along the edges of creeks and dams within these habitat types and have been observed sheltering beneath logs, rocks and in large cracks in the ground (Barker et al., 1996).

Both species have prolonged breeding seasons, which given appropriate rains can occur all year round. The main breeding season however, extends from August to March (Barker et al., 1996). Mating can take place in a variety of wet habitats including puddles of rain water, farm dams, flooded grasslands, still ponds and swamps (Barker et al., 1996). Male frogs call to attract reproductive females while being partly concealed by vegetation at the edges of shallow water (Barker et al., 1996). The female swims to the male and he mounts her. As the eggs and sperm are released during amplexus, the female constructs a foam nest, which is unique to this genus). The foam nest is created by the female paddling the water with her forearms in an alternating sequence (Tyler and Davies, 1979) (Figure 6). Fleshy flanges located between two or more of her fingers create tiny air bubbles. These air bubbles are forced backwards and eventually disperse around her body. They become trapped in the mucus that accompanies the extrusion of the eggs and thus a foamy mass is created (Martin, 1970; Tyler and Davies, 1979). It has been suggested that foam nests protect the eggs from desiccation, provide oxygen to the developing embryos, and act as a thermal insulator or solar reflector thereby protecting the eggs from extreme temperatures (Villa et al., 1982; Duellman and Trueb, 1986). It has also been suggested that foam nests enable some frogs to lay their eggs before there is water (Roberts and Seymour, 1989). The ability to lay the eggs early may give tadpoles an advantage since their development can occur in the absence of aquatic predators (Roberts and Seymour, 1989). The eggs usually hatch about three days after they are deposited (Barker et al., 1996). A typical L. tasmaniensis egg mass may contain an average of 600 eggs and is approximately 80 mm in diameter (reference). The average number of eggs deposited in an egg mass by L. fletcheri females is approximately 300 (Barker et al., 1996). The tadpoles may take up to five months to metamorphose and disperse away from their aquatic habitat (Barker et al., 1996).

Figure 1 Limnodynastes tasmaniensis adult (L. Fuscko).

Figure 2 Limnodynastes fletcheri juvenile (www.wikipedia.com).

Figure 3 Foam nest created by L. tasmaniensis (L. Fucsko).

Dietary Habits of Frogs The majority of the prey items consumed by frogs are terrestrial invertebrates (Tyler, 1976; Forstner et al., 1998; Hirai and Matsui, 1999; 2001; 2002). In general, frogs consume invertebrates included within the taxonomic orders: Coleoptera (beetle), Diptera (flies), Hymenoptera, mainly Formicidae (ants), Lepidopteran larvae (caterpillars) and Arachnae (spiders) (Clarke, 1974; Darst et al., 2005; Sole et al., 2005). However, studies have shown that some frogs also feed on aquatic invertebrates (Jenssen and Klimstra, 1966; Siqueira et al., 2006). A dietary study of the Green Frog (Rana clamitans) reported that when terrestrial prey items were scarce due to cold temperatures, R. clamitans fed primarily on aquatic organisms such as Water Scavenger Beetles (Hydrophilidae), Dragon Fly Nymphs (Libellulidae), and Snails and Slugs (Pulmonata) (Jenssen and Klimstra, 1966). Studies on the Rock River Frog (Thoropa miliaris) showed that this species consumed both terrestrial and aquatic organisms (Siqueira et al., 2006). Alternatively, Thoropa species that inhabit rocky marine environments consumed marine invertebrates (Sazima, 1971 as cited in Siqueira et al., 2006). Moreover, there are records of large-bodied frog species consuming vertebrates, including other frogs, small mammals and reptiles (Tyler, 1976). For example, it has been reported that the Bullfrog (Rana catesbeiana) may consume small frogs (Stewart and Sandison, 1972), mice, voles (Duellman and Trueb, 1985), snakes (Minton, 1949) and birds (Howard, 1950).

Feeding Mechanisms Frogs use their muscular tongues to facilitate capturing prey items (Porter, 1972). The tongue is attached to the mouth anteriorly with the tip of the tongue facing the pharynx. To capture prey, the tongue is flicked out of the mouth so that the top of the tongue strikes the item (Tyler, 1976). The sticky coating on the tongue is secreted from mucous glands in the throat, nasal cavity and on the tongue itself. The sticky coating aids in securing prey items once they are captured. Frogs have primitive teeth that are not used for chewing, but are simply used to grasp a prey item until it is swallow whole (Porter, 1972).

Due to the fact that frogs swallow prey items whole, they are considered to be gape limited predators (Tyler, 1976). Gape limited predators are restricted to prey items that can fit entirely into their mouths. In general, frogs will ingest their prey in the direction that it is stuck to their tongues. This is often the greatest dimension parallel to a frog’s gape since they are unable to position prey so that the prey’s smallest dimension corresponds to the width of the mouth (Lima et al., 2000). For most species of frogs there is a significant correlation between the size of the frog, measured by snout-vent length (SVL) or mouth width, and the size of the prey item (Toft, 1980). For example, Forstner et al. (1998) studied the Southern Leopard Frog (Rana sphenocephala) and the Bronze Frog (Rana clamitans) and found that there was a positive correlation between prey size and frog size for both species.

Ontogenetic Dietary Shifts As a result of being gape-limited predators, frogs often incorporate larger food items into their diet as they grow. This change is referred to as an ontogenetic dietary shift (Toft, 1980). Hirai and Matsui (1999) studied the Pond Frog (Rana nigromaculata), which inhabits rice paddies in Japan and found that it exhibits an ontogenetic dietary shift as it matures. They observed that juvenile frogs showed a strong preference for Hymenoptera (Formicidae) and Dipteran and a weak preference for larger prey items. In contrast, subadult and adult frogs preferred larger prey items including Coleopterans (beetles) and Orthopterans (grasshoppers and crickets) (Hirai, 2002).

Ontogenetic dietary shifts may occur due to larger frogs requiring more energy and nutrition for growth, maintenance and reproduction than smaller frogs (Schoener, 1971). If a large-bodied frog consumes small prey items it will need to consume a much greater quantity of them to obtain enough nutrients and energy (Heatwole and Heatwole, 1968). Therefore, optimal foraging theory predicts that a large frog will consume large prey items when available (Schoener, 1971). Siqueira et al. (2006) found significant relationships between maximum prey size and jaw width for the River Rock Frog (Thoropa miliaris). In this study, grasshoppers, beetles and caterpillars were volumetrically important prey items. Siqueria et al. (2006) concluded that there may be an adjustment between mouth and prey size in this species as the frog grows and develops. Ontogenetic dietary shifts may also reflect modifications in foraging strategies (Lima, 1998). For example, an adult frog may be more mobile, able to travel longer distances and therefore encounter more potential prey items than immature individuals (Toft, 1981).

Foraging Strategies Frogs exhibit two principal foraging strategies: sit-and-wait and widely foraging (Huey and Pianka, 1981). Foraging strategies have been linked to morphological and physiological adaptations, as well as, the sensory mode and learning ability of a species (Huey and Pianka, 1981). Frogs which use sit- and-wait foraging methods generally have a wider mouth for a given body size (Toft, 1980). They also tend to have very good eyesight and can recognize prey at a distance (Toft, 1985). They have a tendency to consume large, mobile invertebrates that are usually soft bodied (Toft, 1980). The Horned Leaf-frog (Proceratophrys appendiculate) is a sit-and-wait predator that inhabits the Atlantic Rainforest in southeastern Brazil (Boquimpani-Freitas and Rocha, 2002). It consumes a wide variety of prey including grasshoppers and crickets (Orthoptera), cockroaches (Blattodea), beetles (Coleoptera), spiders (Aranae), terrestrial snails and slugs (Pulmonata) and frogs (Anura). The most important prey numerically, volumetrically and also in frequency was Orthopterans (Boquimpani- Freitas and Rocha, 2002).

In order to be able to employ a sit-and-wait strategy, these individuals need to be able to have stored energy so they can move suddenly in order to escape another predator if one is encountered (Huey and Pianka, 1981). These frogs will spot prey and spring at it with their mouths open from a distance. There is generally a lower energy cost associated with a sit-and-wait foraging strategy as the frogs are not actively searching for food; although, their prey may be larger and harder to subdue and handle

(Toft, 1981). Sit-and-wait predators are often cryptic which enables successful foraging by allowing them to go undetected by both predator and prey. Toft (1981) found that frogs in the genus Leptodactylus foraged successfully by hiding under leaf litter and springing upon their prey when it was within close range.

Frogs that use widely foraging methods, often have a narrower mouth width relative to their body size compared to sit-and-wait predators (Huey and Pianka, 1981). Widely foraging frogs actively search for prey items that are slow-moving, hard-bodied and often locally abundant (Toft, 1980). Widely foraging frogs are usually specialist predators and often consume ants and mites in large proportions (Toft, 1980).

To feed effectively using the widely foraging strategy, the frog needs to be able to sustain a steady but low pace of activity during feeding times (Toft, 1985). Widely foraging species often feed while moving over the surfaces of leaves, tree trunks or logs. They capture their prey by leaning forward slightly and then snapping it up with their tongues (Toft, 1981). Widely foraging frogs incur a small cost in terms of the energy they need to expend to capture each prey item. However, because the items are generally slow and small they need to capture significantly more of them at the same time. The prey items are often highly chitinous and therefore difficult to consume in great numbers (Toft, 1981). A widely foraging strategy often places the frog at risk of greater predation than those species that sit- and-wait (Toft, 1985). As a result, some frogs have developed anti-predator defence mechanisms (Toft, 1985). For instance, members of the Dendrobatidae family, which have become increasingly specialised in their dietary intake of ants, have developed toxic skin secretions from the alkaloids in these small arthropods (Toft, 1985). Of the Dendrobatidae family, frogs in the genera Phyllobates and Dendrobates have the most toxic skin (Darst et al., 2005).

Prey Selection Specialisation and selectivity are commonly reported in dietary studies (Toft, 1980). Specialisation and selectivity is reflected in the consumption of a food resource, which occurs in a narrower range than the potential food resources available. The processes of specialisation and selectivity may influence the composition of the diet both simultaneously and separately (Toft, 1980). The assumptions are that the animal has a choice in the prey it consumes and that it is able to curb its natural impulse to capture every moving item even though capturing prey is often an automatic response to movement (Tyler, 1976). The presence or absence of prey, which vary seasonally, influences the frog’s ability to select these items. The presence of certain prey items in the stomach does not necessarily mean the anuran selected to eat it based on its availability (Tyler, 1976). Most frogs will only consume live prey items because the process of capturing prey is almost an automatic response to the prey’s movement (Tyler, 1976). It has been suggested that moving objects elicit the most consistent feeding response in frogs (Freed, 1980). For instance, Freed (1980) observed that the Green Tree Frog (Hyla cinerea) consistently selected the housefly (Musca domestica) in laboratory feeding trial because it was the largest and most active of the prey item offered.

The consumption of small prey items such as ants (myrmecophagy) and mites (acariphagy) is generally correlated with frog body size and microhabitat (Simon and Toft, 1991). Ants and mites are highly abundant organisms in a wide variety of habitats; however, only ground dwelling predators will encounter them in high abundance (Simon and Toft, 1991). Specialists of these small prey items will consume the organism in a higher proportion than it is found in the environment. Toft (1981) suggested ant-specialists consume small, slow-moving, often chitinous or hard-bodied prey items that often have a chemical defence. Many frog species in the families Dendrobatidae and some from the Bufonidae are myrmecophagous (Donnelly, 1991; Hirai and Matsui, 2000); whereas, frogs in the Microhylidae, Myobatrachidae, Ranidae and Sooglossidae have been reported to consume mites in high proportions (Simon and Toft, 1991).

Diet Overlap in Sympatric Species It has been argued that sympatric species of frogs will consume similar prey items as habitat is a defining factor in the abundance and availability of invertebrates (Stewart and Sandison, 1972). However, sympatric species occupying different microhabitats can exhibit prey differences that reflect variability in the invertebrate communities present (Stewart and Sandison, 1972; Van Sluys and Rocha, 1998; Forstner et al., 1998; Franca et al., 2004). Stewart and Sandison (1972) investigated the interactions and dietary differences between the Mink Frog (Rana septentrionalis), Green Frog (R. clamitans) and Bullfrog (R. catesbeiana). These three species show some variation in the microhabitats that they occupy within the same ecosystem. For instance, the Mink Frog inhabits the aquatic zone, the Green Frog lives primarily along the water’s edge, and the Bullfrog can occupy either of these microhabitats. Stewart and Sandison (1972) observed that the differences in prey items consumed by each species reflected the variation in the microhabitat they occupied (Stewart and Sandison, 1972).

Van Sluys and Rocha (1998) studied the feeding habits and microhabitat utilization of the Lesser Tree Frog (Hyla minuta) and a frog in the genus (Pseudopaldicula sp.) in Brazil. Though both species were observed along lakeshores, they differed considerably in the microhabitats they occupied. For example, H. minuta used the vertical and horizontal habitats available such as the ground and standing vegetation, while Pseudopaldicula sp. occupied the ground only (Van Sluys and Rocha, 1998). The two species also differed in body size with H. minuta being significantly larger. The feeding niche overlap for the species was slightly greater than 10% which suggested that despite being sympatric there was a low similarity in the prey items they consumed (Van Sluys and Rocha, 1998).

Prey Availability Prey availability is considered an important factor in defining the prey items consumed by frogs (Hirai and Matsui, 1999). The consumption of prey items will usually reflect the abundance of the prey items in the habitat, which often changes seasonally (Toft, 1980). Jenssen and Klimstra (1966) studied the diet of the Green Frog (Rana clamitans) over 12 months in Southern Illinois in America. The diet of R. clamitans was strongly influenced by both the habitat occupied by this species and the seasonal availability of prey organisms (Jenssen and Klimstra, 1966). The greatest dietary intake occurred during spring and the least during winter. During periods of cold weather individuals that were not hibernating showed an increase in the consumption of aquatic prey items, which was probably due to the cold weather restricting the activity of terrestrial organisms. In contrast, there was a decrease in the consumption of aquatic organisms during spring and summer that was probably due to an increase in the availability of Lepidopteran larvae and moths (Jenssen and Klimstra, 1966).

Parker and Goldstein (2004) studied the diet of the Leopard Frog (Rana berlandieri) during spring and autumn in Texas. Like many other Ranid species, the Leopard Frog is considered to be a generalist, opportunistic predator whose diet is most strongly influenced by prey availability, and thus is likely to reflect seasonal shifts in prey. For instance, there was an apparent increase in the consumption of Lepidopterans in the autumn due to a massive influx of armyworms moving through adjacent agricultural areas.

Seasonal shifts in the dietary composition of some species are also associated with seasonal changes in behaviour. The Green Frog (R. clamitans) migrates to breeding sites in spring, feeding grounds in summer and over-wintering sites in winter (Lamouruex et al., 2002). Prior to the migrations towards over-wintering sites in mid-August, R. clamitans makes repetitive forays away from water for feeding purposes. Increased foraging opportunity in terrestrial environments might explain the forays since food availability was presumed to be limited in aquatic habitat. An increase in mass during the autumn serves as an energy reserve for frogs that need to survive for four or five months without food while over-wintering.

Top Predators Arid and semi arid regions represent the majority of Australia’s land surface area and bear a legacy of degradation and species loss from past policies of land management (Morton et al., 1995; Letnic, 2000). Some of the most distinctive and important vegetation types within Australia’s arid and semi- arid zones are the eucalypt and acacia woodlands (Harrington et al., 1984). Since European settlement, these landscapes have been transformed and degraded by land clearing and overgrazing by livestock. Today, more than half of Australia’s endangered mammal species, about a third of its threatened bird species and about one tenth of its threatened plant species occur in the semi-arid and arid lands (Morton, 1990; Garnett, 1992; Leigh and Briggs, 1992).

Mammals have attracted considerable research and conservation attention in these regions, which is not surprising as this group has suffered range contractions and extinctions to a far greater extent than any other vertebrate group. Reptiles on the other hand, have attracted far less attention in this respect despite about 50 species being identified as requiring conservation attention (Sadlier and Pressey, 1994). The majority of research on reptiles in the arid and semi-arid zone has focused on smaller species of reptiles, particularly those belonging to the Scincidae and Agamidae families (Dickman et al. 1999; Melville 2001; Driscoll, 2004). Large squamate reptiles are significant components of woodland ecosystems, but their ecology is often poorly understood (Bedford, 2003).

Space Use by Wildlife Of the many resources used by wildlife, ‘living space’ is one of the most important. All animals must move to find the resources they need (e.g., food, shelter and mates) while avoiding predators. Understanding the basis for daily and seasonal variation in movements and activity has important implications for understanding the ecology of a species. It is also critical in identifying the appropriate spatial scale and attributes of any area set aside to protect or manage species or populations of conservation concern (Caughley and Sinclair, 1994; Bonnet et al., 1999; Cooke et al., 2004). Knowledge of the factors that shape patterns of movement and space use can help us interpret the consequences of past, present and future land use practices.

Various factors may influence how animals use the landscape (Waser and Wiley, 1979). For instance, using resources that vary unpredictably may require individuals to be vagile and use large areas, while the use of more permanent and predictable resources may allow individuals to be more sedentary and to use smaller areas (Schoener, 1971; Huey and Pianka, 1981; Roe et al., 2004). Territoriality can also influence an animal’s use of the landscape (Maher and Lott, 2000).

The spatial and temporal distribution of resources in the landscape influence patterns of movement and space use in snakes (Gregory et al., 1987). For instance, when resources such as suitable shelter or foraging sites are widely dispersed, snakes must travel long distances between them (King and Duvall, 1990; Madsen and Shine, 1996; Roe et al., 2004). Moreover, body size may influence patterns of movement and space use. For example, larger animals may require more resources and thus may need to travel over larger areas to find them (Whitaker and Shine, 2003; Wilson et al., 2006a). Likewise, reproductive status also plays an important role in the spatial ecology of snakes. In the majority of species, males increase their movements during the mating season while searching for females (e.g. Duvall & Schuett, 1997; Brito, 2003; Whitaker & Shine, 2003). On the other hand, females show decreased activity during gestation or brooding periods (Slip and Shine, 1988a; Brown and Weatherhead, 2000; Pearson et al., 2005). Temperature is perhaps one of the most important factors driving patterns of activity and movement in reptiles (Huey, 1982; Peterson et al., 1993), including snakes that generally exhibit distinct seasonal variations in both (Slip and Shine, 1988b; Heard et al., 2004; Brown et al., 2005).

Both anthropogenic (e.g., land-clearing) and natural disturbances (e.g., hurricanes) influence movements and home range size in snakes (Reinert and Rupert, 1999; Plummer and Mills, 2000;

Wunderle et al., 2004). Despite this, only a handful of studies have examined the spatial ecology of snakes inhabiting human-modified landscapes (Durner and Gates, 1993; Shine and Fitzgerald, 1996; Butler et al., 2005; Wisler et al., 2008). Several authors have made inter-study comparisons (e.g. black rat snakes (Elaphe o. obsoleta), Durner and Gates, 1993 vs. Weatherhead and Hoysak, 1989; carpet pythons (Morelia spilota), Shine and Fitzgerald, 1996 vs. Slip and Shine, 1988b). However, regional differences in climate make it difficult to tease apart the influences of local climatic conditions and thus limit their interpretation. As more and more areas of natural habitat come under human habituation and undergo subsequent modification, a greater understanding of how such processes might influence the ways in which animals use these landscapes is required.

The Carpet Python

The Carpet Pythons (Morelia spilota) is a large, heavy-bodied, non-venomous, ambush foraging species with an average adult length of approximately 2 m and mass of up to 2.5 kg (Figure 30) (Cogger, 2000). These snakes occupy the most diverse habitats of any Australian python, and are found in most of the available habitats within their range, from dry and wet woodlands and forests to rainforests (Cogger, 2000; Wilson and Swan, 2003). They are semi-arboreal and prey upon a variety of small to medium sized mammals, birds and lizards (Slip and Shine 1988c; Shine and Fitzgerald, 1996; Fearn et al., 2001; Pearson et al., 2002).

This species ranges over large areas of the Australian continent. Most studies on this species however were conducted in mesic landscapes (Slip and Shine, 1988b; Shine and Fitzgerald, 1996). Populations in arid and semi-arid zones have attracted little research; despite these areas comprising nearly 70% of Australia’s land surface area (Morton, 1986). Like many mesic areas, the arid and semi-arid areas of Australia are under increasing pressure from anthropogenic disturbances (i.e., land clearing, agriculture, pastoralism, and feral animal introductions); hence, snakes in these areas warrant comparable attention. Moreover, populations of M. spilota have declined across much of inland Australia (Sadlier and Pressey, 1994; Shine, 1994; Shine and Fitzgerald, 1996). Despite this, some populations continue to persist and thrive in areas that have been heavily modified by humans along Australia’s eastern seaboard (Shine and Fitzgerald 1996; Fearn et al., 2001).

Figure 4 Adult Carpet Python (Morelia spilota).

Objectives

The objectives of this study were to 1) demonstrate the potential effectiveness of the Environmental Champions’ (ECP) on-farm management program for vertebrate biodiversity; and, 2) conduct parallel ecological research (e.g., habitat utilisation on rice farms or the farming matrix) on selected significant vertebrate species to facilitate on-farm management decisions. More specifically, the study aimed to:

Vertebrate Biodiversity

1. Monitor changes in species richness and abundance following revegetation activities on rice farms in the Riverina in order to evaluate the effectiveness of revegetation efforts; and,

2. Assist the ECP in the establishment of guidelines for the improvement of on-farm biodiversity.

Selected Significant Species Research

1. Estimate the number of frogs produced in rice bays each year in order to calculate their annual invertebrate consumption;

2. Determine through stomach content analyses the species of rice pests and other invertebrate fauna that frogs consume in the rice agroecosystem;

3. Learn the ecology of Carpet Pythons that occupy agricultural landscapes;

4. Identify patterns of activity, movement and space use for this species;

5. Examine key habitat features required by a semi-arid population; and,

6. Determine the implications for the conservation and management of this top predator.

Methodology

Study Location This study was conducted in the Riverina Bioregion, located in southwest New South Wales (NSW), Australia (Figure 7). Soils of the Riverina are composed of gravel, sand, silt and clay particles that were transported and deposited by ancient rivers or wind (NSW National Parks and Wildlife Service, 2003). The Riverina covers the alluvial fans of the Lachlan, Murrumbidgee and Murray Rivers (NSW National Parks and Wildlife Service, 2003).

Figure 5 Map of the rice-growing region within the NSW Riverina showing the three main irrigation districts: Murrumbidgee (top), Coleambally (middle) and Murray Valley (bottom). Insert shows location of the Riverina Bioregion in Australia. (Source without scale: Doody et al., 2006).

The present landscape includes features such as elevated sandy ridges and their associated stream channels (NSW National Parks and Wildlife Service, 2003). In between the ancient stream formations there are extensive clay plains. The clay plains are a combination of fine sediments carried by the ancient streams and deposits of wind blown material (NSW National Parks and Wildlife Service, 2003).

The land in the Riverina is used to primarily for growing crops, orchard production and grazing livestock. As a result, much of the land has been cleared and extensive irrigation systems have been constructed in the region. Approximately 31% the region’s woodlands have been lost due to human activities (Benson 1999); although, the amount of clearing varies throughout the region, for instance 76% has been cleared in the Murray Province, yet only 4% has been cleared in the Lachlan Province (NSW National Parks and Wildlife Service, 2003). These figures, however, do not include the clearance of native grasses and native shrubs, which were also once widespread.

Vegetation

Prior to clearing, the vegetation was dominated by shrub woodland of Boree (Acacia pedula) and saltbush (Atriplax. Spp.). The shrub woodlands have now been replaced with grasses including: Spear Grass (Stipa spp.), Corkscrew (Stipa setacea), Wallaby Grass (Danthonia spp.), Windmill Grass (Chloris spp.) and Fescue (Vulpia spp.) (NSW National Parks and Wildlife Service, 2003). Historically, the open plains between the ancient stream formations were thinly timbered with Grey Box (Eucalyptus macrocarpa) and occasional Bull Oak (Casuarina luehmannii) (NSW National Parks and Wildlife Service, 2003). Murray Pine (Callitris columellaris), Needlewood (Hakea leucoptera), Rosewood (Heterodendrum oleifolium), Bull Oak, Yellow box (E. mellidora) and Wilga (Geijera parviflora) were found along the ancient stream formations. Sandhills were dominated by Murray Pine and Yellow Box (NSW National Parks and Wildlife Service, 2003).

Climate The Riverina climate is typically described as Mediterranean and semi-arid, it is categorised by hot dry summers and mild wet winters. In a typical year, the average maximum temperature in summer is approximately 32ºC and the average minimum temperature is 17ºC. The average maximum temperature in winter is about 14ºC and the minimum temperature is 2ºC (NSW National Parks and Wildlife Service, 2003). Rainfall is highly variable, the average rainfall is between 30-40 mm per month and 300-450 mm annually (NSW National Parks and Wildlife Service, 2003). For the past five years the region has been subjected to an intensive drought. As a result, the amount of summer crops planted in 2006 and 2007 was dramatically reduced (i.e., 12,000 ha of rice was planted compared to an average 120,000 ha) (Australian Bureau of Agriculture and Resource Economics, 2006). The amount of rainfall during the study period was significantly less than the predicted monthly averages. The average monthly rainfall was 36 mm at Old Coree and 18 mm in Leeton (Figure 8).

80 60

60 40

40 Rainfall (mm) Rainfall (mm) Rainfall 20 20

0 0 JAN FEB MAR APR MAY Month

Old Coree (2007) Leeton (2007) Leeton (average 1971-2000)

Figure 6 Total monthly rainfall received at the two study locations during the study period. The average rainfall for Leeton between 1971–2000 is also indicated as a reference. (Russell Ford provided the data for Old Coree. Data for Leeton was obtained from the Bureau of Meteorology, 2007.)

Water In NSW, the Riverine Plain is drained by the Murray, Murrumbidgee and Lachlan rivers, which flow slowly from east to west (NSW National Parks and Wildlife Service, 2003). Other water resources include groundwater and water stored in aquifers (NSW National Parks and Wildlife Service, 2003). Aquifers deeper than 20-30 m are widespread throughout the region where sandy sediments have been laid down by ancient stream activity (NSW National Parks and Wildlife Service, 2003).

The development of irrigation on the plain in the early 1900’s contributed to the area becoming a successful agricultural region (RIRDC, 2000). Many properties use irrigation to boost the production of crops such as rice, cereals and citrus fruits and also to graze livestock. The region is divided into three main irrigation districts: Murrumbidgee, Coleambally, and Murray Valley (Figure 2.4)(RIRDC, 2000).

Rice Farming In the Riverina, rice is the predominant summer crop with approximately 120, 000 ha grown annually. About 2,500 family operated farms located in the Murrumbidgee, Coleambally and Murray Valley irrigation districts (RIRDC, 2000) (Figure 2.4). Here rice is usually grown in ponded cropping systems. Most rice crops are aerial or directly sown into water 3-5 cm deep. Sowing usually occurs in September and October. Additional water is then released into the rice bays to achieve a depth of about 20 cm by January. The water is then drained off of the crop several weeks before harvest that occurs sometime during March or April.

There are three main ponded cropping systems used on rice farms: natural contour or landformed contour systems, border check systems and furrow systems. Natural contour or landformed contour systems are most commonly used in the Riverina. This system enables the rice bays to be flooded and the water depth to be quickly raised and lowered depending on the stage of the crop (RIRDC, 2000). There are four main rice-based farming systems in southeastern Australia in which rice is rotated with different crops. These farming systems include: rice and winter crop; rice and summer crop; rice and pasture; or continuous rice. Crop rotation is important as different crops may have an unequal effect on soil fertility, moisture, acidity (pH), salinity and structure, as well as, having particular impacts on weed, pest and stubble management (RIRDC, 2000).

Study Design and Period This dietary study on frogs was conducted on three rice farms in the Riverina (NSW): the Old Coree Rice Research Station and two private homesteads in Leeton (Houghton and Kirkup). Old Coree is located in the Murray Valley Irrigation Area approximately 20 km west of Jerilderie (35.20S, 145.41E). The Kirkup and Houghton Homesteads are located in the Murrumbidgee Irrigation Area, south of Leeton (34.33S, 146.23E). Data were collected from spring 2006 to autumn 2007. This period included the majority of the breeding season of both L. tasmaniensis and L. fletcheri, and coincided with the rice-growing season.

This vertebrate biodiversity study was conducted on seven rice farms in the Murrimbidgee Irrigation Area (MIA). These farms were chosen based on three criteria: (1) the presence of on-farm revegetation efforts and (2) the willingness of the farmers to facilitate our surveys. To test the hypothesis that vertebrate biodiversity was positively influenced by the presence of revegetated areas, we chose farms such that they contained both revegetated and non-revegetated (i.e., control) sites. We surveyed for vertebrates (i.e., birds, reptiles and frogs) from 2006 to 2007. To maximise the probability of detection of ectotherms (i.e., reptiles and frogs), surveys for these species were conducted during the warmer months (September-April) each year.

The revegetated sites were fenced off areas that consisted mainly of saltbush, and native tree and shrub plantings with few eucalypt trees. The plantings ranged from 0-5 years old. The non- revegetated, or control sites, consisted of paddocks used for grazing that did not include revegetated areas, but did often contain sparse eucalypts.

This Carpet Python could not be located on any of the rice farms used above; therefore, the study was conducted at Willandra National Park (WNP; 33°12'40. 19’’ S; 144°59'27. 33’’ E) in the Western Division of New South Wales (NSW), Australia. This area is characterised by arid and semi-arid rangelands. Past and present land use practices have extensively modified the natural distribution and condition of vegetation cover (Cambell, 1994). The Willandra region experiences warm to hot summers and cool winters. Mean maximum and minimum monthly temperatures range from 17–34 ºC and 3–17 ºC respectively, while mean monthly rainfall ranges from 22–32 mm with little or no seasonal variation (Bureau of Meteorology, 1901–2006 averages).

Two locations within WNP were used: a highly modified site (homestead of an old sheep farm) and a site with relatively low human impact (woodland). The low impact site (‘Park’) occupies a flat landscape of open grasslands comprised of Stipa sp., Danthonia caespitosa, Chloris truncata, Medicago polymorpha, Maireana aphylla, Eragrostis australasica and Sclerolaena tricuspis. An overstorey of Eucalyptus largiflorens and Acacia stenophylla with a dense shrub understorey of Chenopodium nitrariaceum, E. australasica and Muehlenbeckia florulenta dominates the riparian woodland. The landscape at WNP is typical of much of this part of western NSW (Porteners, 1993). The main watercourse in the park is Willandra Creek, which is a tributary of the Lachlan River, which is part of the Murray-Darling Basin.

The high impact site (‘Homestead’) is situated on the eastern boundary of WNP (Figure 31) and is an old sheep station, which is now managed by the NSW National Parks and Wildlife Service. More than 100 years of land clearing and livestock grazing have modified this area.

Sampling Techniques

Visual Encounter Surveys

Visual Encounter Survey (VES sensu Crump and Scott, 1994) involve field personnel walking through an area for a prescribed time period, systematically searching for animals. This technique is also known as the ‘time-constrained search’ (Campbell and Christman, 1982; Corn and Bury, 1990). An example of a VES is walking through woodlands searching for reptiles and frogs under logs, bark, or other cover objects (Figure 1).

Advantages of this method are: (1) the technique is flexible and can be modified to detect a number of different species; (2) trained surveyors with experience in finding particular species or groups of species can often detect them quickly and efficiently using this method; and (3) the method is inexpensive, requiring little or no materials. Disadvantages include: (1) observer bias is common; an experienced herpetologist for example will find considerably more species and individuals than an inexperienced surveyor (Crump and Scott, 1994); and (2) the technique is not generally effective for species that reside underground or in the canopy (Crump and Scott, 1994).

VES’s during the present study included searching for reptiles and amphibians during the day by searching the ground under cover objects (e.g., bark, logs and hollow trees), inspecting (old) buildings and walking along farm roads and other habitats. We did not attempt to standardise VES’s to compare capture rates among farms because of logistical concerns, observer bias, and variation in habitat types and quality among farms. Thus, we attempted to uncover as many species as possible to satisfy the objective of an inventory, rather than compare absolute numbers among farms.

Figure 7 An example of how secretive species use natural cover objects.

Bird Counts Bird count surveys were conducted for 15 minutes in each revegetated and control plot using a modified version of the area search method (Loyn 1986). Surveys were conducted at a randomly located point within the plot. Using random plot allocation allows a greater portion of the plot to be surveyed over a number of different visits thus increasing the likelihood of obtaining a more complete species count for that plot. In larger plots this method facilitates the inclusion of species that specialise in edge or interior habitats.

The surveys were conducted seasonally during the periods of 1-30 April 2007, 1-31 July 2007, 1–31 October 2007 and 1-31 December 2007. During each period, the 14 sites (two sites on seven farms) were surveyed on two occasions allowing for eight replications at each site during the year. Sites were surveyed in a pre-determined order on the first replication within each trip. On the subsequent replication during the same trip, the sites were re-visited in reverse order to eliminate or reduce any biases associated with the time that the surveys are undertaken. Surveys were not conducted on excessively windy days or days that were affected by precipitation as these conditions deflate bird counts.

Surveys were conducted between a half an hour after sunrise and 1000hrs, to avoid the potentially confounding effect of change in behaviour through the course of the day. Upon arrival at a site, one minute was be spent quietly standing at the beginning of the survey point prior to commencing the survey to allow bird activity to return to normal.

Birds flying over or through the survey area were not included in the count, but were noted separately to indicate that the species was present, but not utilising the habitat within the area. Species presence, abundance and any breeding or unusual behaviour was recorded for each species observed utilising the site.

Artificial Cover Surveys Turning over surface cover objects to look for secretive terrestrial creatures is a common technique (Figure 2). Artificial cover objects are an extension of this search technique with a more formalized sampling design. They are made of many types of materials (e.g., wood, tarpaper shingles, plastic sheets, etc.), but the most common material is nonchemically treated plywood. Artificial cover surveys provide information on: (1) species presence at the time of sampling; (2) life history information, such as data on size-class structure, reproduction, and activity patterns; and (3) habitat information.

Artificial shelter boards were used primarily to detect snakes, frogs and lizards. Boards were 0.6 m x 1.2 m (2’ x 4”) and made of plywood. Boards were deployed in revegetated and control sites on each farm. Ten boards were placed 5 m apart along single 50 m transects per site. These boards were checked twice per season from December 2006-December 2007.

Frog Surveys PVC tubes are used to survey tree frog populations. The tubes are placed on the ground or mounted on trees (Boughton et al., 2000). Tube surveys provide information on: (1) species presence at the time of sampling; (2) life history information, such as when animals arrive at breeding ponds, how long they stay, sex ratios, size-class structure; (3) movement patterns while at the ponds; and (4) information on the direction and distance of dispersal (Boughton et al., 2000).

In the present study, PVC tubes that were 5 cm in diameter were mounted every 5 m for 50 m on metal pickets that comprise the fence lines adjacent to each revegetated and control site on the seven farms. These tubes were checked twice per season from December 2006-December 2007.

Figure 8 Example of the technique used in artificial cover surveys.

Frog Dietary Study

Abundance Six experimental bays were constructed at Old Coree to estimate the number of frogs produced in rice bays each year in the Riverina (Figure 9). The experimental rice bays measured 45 m x 15 m, three of which were enclosed with silt fencing stapled to wooden stakes. The fence was buried approximately 15 cm deep and extended about 45 cm above ground. Fourteen pitfall traps were sunk into the ground at 10 m intervals inside the perimeter of the fenced bays (figure 10). The pitfall traps consisted of 20 L buckets with lids. The traps were closed during the day and opened at dusk to allow animals to be captured during their activity period. The traps were checked twice per night and once at dawn. The species, sex, age (based on size) and trap number were recorded for each individual. Each frog was released following data collection. All frogs produced in the rice bays during the study period that were not of breeding adult size were recorded as juveniles.

Additional sampling of the abundance of frogs occurred at three sites on the rice farms in Leeton (Houghton: 1 site and Kirkup: 2 sites) and also at one site at Old Coree. Each of these additional sites contained a single 30 m long silt fence, with four pitfall traps placed at 10 m intervals. The fence was erected at the edge of a rice bay on each rice farm. Sampling was conducted at the same time as the fenced rice bays at Old Coree.

Prey Item Identification In order to identify the prey items consumed by individuals in the rice bays, frogs were hand-captured when they were foraging along the perimeter of the bays between 1900-2400 hrs (EST). The frogs were transported to the field stations in plastic buckets located at each study location. Their stomachs were flushed using the procedure outlined below. They were then released at their point of capture.

Dietary analyses were conducted on both adult and juvenile frogs. Sample sizes of between 70-100 juvenile/adult frogs of each species were collected for each month. The stomach content of each individual was extracted using a simple stomach flushing technique described by Sole et al. (2005). The tools used included 20 ml syringes attached to flexible tubing (i.e., blood infusion tubing), wooden spatulas, small sieves and sample vials (Figure 11).

The stomach flushing procedure (Figure 11) was conducted within two hours of capture to insure that the items remained undigested as follows:

1. Frogs were gently imobilised by fixing forelimbs with one hand.

2. Mouth was opened using a small wooden spatula and a water filled 20 ml syringe attached to the tubing was introduced into the esophagus and then into the stomach.

3. Water from the syringe forced out the material in the stomach. (This was conducted twice to ensure that all contents were extracted.)

4. Flushed material was collected in a sieve and emptied into a vial of 70% ethanol for later analyses.

5. Frogs were released immediately following the procedure.

The snout-vent length (SVL) and mass of each individual was recorded following flushing. SVL was measured to the nearest 0.1 mm with a caliper rule and mass was measured using an electric balance to the nearest 0.1 g (Hirai and Matsui, 2002).

Figure 9 Overview of the experimental rice bays at Old Coree. Three bays with fences and pitfall traps (white with grey borders) and three bays without fences (solid grey).

Figure 10 Fenced rice bays at Old Coree.

Figure 11 Equipment used in stomach flushing procedure.

Figure 12 Stomach contents collected in sieve for later analysis.

Quantifying Prey Items Prey organisms obtained from the stomach samples were identified to the order and where feasible, to the family taxonomic levels. The items were identified using the following references: ‘The Insects of Australia’ (CSIRO 1991), ‘Australian Beetles ‘(Lawrence and Britton, 1994), ‘The Insects: An outline of Entomology’ (Gullen and Cranston 1999) and The National Rice Industry Biosecurity Plan (Plant Health Australia, 2005). Items were identified to the species level when they were suspected rice pests

(Hirai and Matsui, 1999). Ants (Formicidae) were categorised independently due to the high frequency of occurrence in the stomach contents. The larvae of Coleopterans, Dipterans and Lepidopterans were also counted separately to the adults of these orders, due to the difference in prey size, shape and mobility. The frequency of occurrence (%) and the numeric (%) and volumetric (%) proportion of the prey orders in the stomachs of both study species were also calculated. The volumetric calculations excluded the unidentified prey items.

Food Consumption Rates Feeding trials were conducted in order to determine how many times a frog fills its stomach per day. Fourteen L. tasmaniensis (seven postmetamorphic juveniles and seven adults) were placed in separate 1100 ml containers lined with moist paper towel. The sides of the containers were covered with cardboard to reduce the frog’s field of vision and the potential stress on the animal. The containers were placed in the laboratory, which was maintained at a constant temperature of 22o C. Frogs were not fed for 48 hours prior to the feeding trails to ensure their stomachs were empty before the experiment. During the trials, frogs were fed one cricket. The rate of passage was determined by recording the time the cricket was consumed and the time a fecal pellet was expelled. The frogs were checked every 30 minutes until all had expelled a fecal pellet. The rate of passage was rounded to the nearest hour.

Body Size vs. Prey Size The maximum length and width of each prey item (excluding antennae and cerci) were measured to the nearest 0.1-mm using a calibrated ocular micrometer fitted to a dissecting microscope. Volumes of prey items were calculated using the formula for an ellipsoid:

V = 4/3π(L/2) (W/2)2

V= volume, L= length, W= width

Equation 1 = Volume of prey item calculated using the formula for an ellipsoid.

The volumes of the smallest and largest prey items were compared to the body size of the frog using the snout-vent length measurement.

Ontogenetic Dietary Shift

Types of food items were compared between adult and juvenile L. tasmaniensis and L. fletcheri in order to test for an ontogenetic shift in diet (Hirai and Matsui, 1999). The size classes used for these analyses were juveniles (SVL ≤ 29 mm) and adults (SVL ≥ 30 mm) for L. tasmaniensis, and juveniles (SVL ≤ 36) and adults (SVL ≥ 37 mm) for L. fletcheri.

Seasonal Availability of Prey Seasonal prey availability was estimated by capturing aerial, terrestrial and aquatic invertebrates in monthly samples. Aerial invertebrates were captured by using a light trap and sweeping method on the banks of the rice bays where frogs were also being collected. More specifically, a torch was held above the rice bay for 10 seconds before a net was used to make 20 sweeps through the air to catch the insects this was repeated 10 times. The sampling was conducted after sunset (2100 h) during the frog activity period. Terrestrial invertebrates were passively sampled from the same pitfall traps used to capture the frogs. The pitfall traps remained open between dusk and dawn and invertebrates were removed in the morning. Aquatic invertebrates were sampled by moving a fine meshed net through the water in the tofo (area of water approximately 1 m wide bordering the rice crop) of the rice bays for

approximately 2 min. This sampling procedure was repeated 10 times. All invertebrates collected in the three sampling procedures were stored in ethylene glycol for later analysis. The availability of the invertebrates collected in the pitfall traps and aerial sweep samples were compared against the numeric proportion of invertebrates consumed by each frog species.

Statistical Analyses Statistical analyses were performed using the computer program SPSS (version 14.0, SPSS Inc., 2005). The assumptions of equality of variances and normality were tested for each analyses. When data failed to meet these assumptions, the data were log10 transformed to provide equal variances, or a better distribution of normality. Statistical significance was accepted at α = 0.05 level. Means are presented ± one standard deviation. Statistical analyses performed included regression analyses with Bonferroni multiple comparisons and Spearman’s rank correlations. Spearman’s rank correlations were performed with the ten major prey orders found in the frogs diets only, to avoid bias in the correlations. In the availability vs. consumption correlations the number of Coleopterans captured in the pitfall traps excluded Bombardier Beetles (Pheropsophus verticalis) as a large percentage of the bombardier beetles were captured in the pitfall traps, but were not consumed by either frog species.

Data Collection for Carpet Pythons Study animals were captured by searching through understory vegetation, hollow logs, tree hollows and in the roofs of buildings. Snakes were measured (Snout-vent length) and weighed. Sex was determined by the gentle eversion of the hemipenes and the assessment of the reproductive status of females (i.e., gravid or non-gravid). The latter was determined by gently palping the body for the presence of enlarged ovarian follicles or eggs and later confirmed by nesting activity.

Radio-telemetry Two sizes of implantable radio-transmitters were used depending on the size of the python (Holohil Systems Pty Ltd., Carp, Canada). Individuals > 1.2 m SVL were fitted with SI-2 transmitters (40 × 11 mm, 11 g, with a battery life of 18 months), and those < 1.2 m SVL were fitted with BD-2 transmitters (14 × 9.5 mm, 5 g, with a battery life of 10 months). Transmitters were surgically implanted by a veterinarian (Helen Purdham, BVSc (Hons), Hall Veterinary Surgery, ACT, Australia) following the technique described in Webb and Shine (1997b). In all cases, transmitters weighed < 3% (range = 0.3–2.9) of the pythons’ body mass. Animals were released at their initial capture site.

Radio-tracking commenced one week after release in order to reduce any irregular behaviour resulting from capture and surgeries. Pythons were located once daily (with at least 24 hours between successive locations) for one to three weeks every month from September 2005 to November 2006 with a Regal 2000 telemetry receiver and hand-held three-element Yagi collapsible antenna (Titley Electronics, Ballina, NSW, Australia). The order in which individual animals were tracked among days was changed to avoid temporal autocorrelation in the data (Harris et al., 1990; Kernohan et al., 2001). The location of each individual was recorded using a hand-held Global Positioning System (Garmin ‘etrex’; Garmin International, Inc., Olathe, Kansas, USA) with an error margin of 2–6 m. If the animal was visible, its posture was noted using the methods described in Shine and Fitzgerald (1996).

Seventeen pythons were tracked: eight at the Homestead (three males, three females and two juveniles) and nine in the Park (three males, three females (one of which was brooding eggs) and three juveniles). It was necessary to define a boundary between the homestead complex and the park because of the close proximity between these two areas. An arbitrary boundary of 50 m into the Park from the nearest building or cleared area was set and everything inside of this border was considered “Homestead habitat”. It was necessary to go this distance into the park (instead of using the existing

woodland boundary) in order to include the numerous rubbish piles and old farm machinery associated with the homestead. All of the individuals captured in the vicinity of the Homestead complex were considered “Homestead snakes” even if they spent periods of time in the woodland adjacent to this area because they had access to the homestead buildings. All pythons captured in the woodland were considered “Park snakes”. These Park snakes were captured an average of 6.5 km from the homestead complex (range 1.5–10.6 km).

Habitat Use At each radio-tracked location, we recorded the general cover, macrohabitat and microhabitat types. These classifications were chosen to enable direct comparisons with previous studies of M. spilota along the East Coast of Australia (Slip and Shine, 1988b; Shine and Fitzgerald, 1996), while still accounting for differences associated with landscape type at the study locations. Python locations were first classified according to the cover type in which they were found: (i) open – any situation where the animal was completely visible and in the open (e.g., basking on top of a log, the ground or roof, or moving along any exposed area such as a garden lawn, pasture or other bare surface that offered no form of cover); (ii) filtered cover – any form of cover that provided filtered sunlight where a python was harder to see, but still visible (e.g. lying along a branch in the canopy of a tree or under a sparse shrub), (iii) dense cover – any form of cover where a python was usually not visible, but which still provided small amounts of sunlight (e.g., underneath a dense clump of lignum (Muehlenbeckia florulenta), a pile of fallen branches, or a garden bed); and, (iv) heavy cover – any situation where the python was not directly visible and did not have direct access to sunlight (e.g., inside of a building or tree hollow, down a rabbit warren or under a pile of rubbish).

Macrohabitat types were classified into one of four categories: (i) human-modified – any area associated with human habituation that had undergone considerable modification (e.g., buildings, sheds, rubbish piles, gardens and cleared areas). Woodland areas within 50 m of the homestead complex were considered human-modified, as these areas often contained large piles of rubbish and old machinery and the trees were often pruned and had dead limbs removed; (ii) woodland – areas dominated by an overstory of Eucalyptus largiflorens and Acacia stenophylla and an understory of shrubs (generally Chenopodium nitrariaceum and M. florulenta) and grasses (usually Eragrostis australasica) less than 2 m high; (iv) shrubland – areas dominated by saltbush (Atriplex vesicaria and C. nitrariaceum), grasses (Stipa sp., Danthonia caespitose, Sclerolaena sp., Chloris truncate, Medicago polymorpha, Maireana aphylla, and E. australasica); and, (iv) open – plain or any cleared area (i.e., pasture).

Microhabitat types were classified into one of seven categories: (i) building – this included roof cavities, wall cavities and underneath a building, as well as chicken and garden sheds; (ii) thicket – dense vegetative structures, other than trees. This category included all shrubs, garden beds and reeds (Typha domingensis and T. orientalis); (iii) log – inside, underneath or on top of logs or piles of fallen branches; (iv) tree – any woody perennial plant > 3 m in height (includes dead standing trees); (v) underground – inside of a rabbit burrow or down a crack in a dry watercourse; (vi) rubbish – underneath or inside of any rubbish or old abandoned farm machinery; and, (vii) open – pasture, grass, road or any other bare surface.

Habitat Characterisation and Analyses To examine vegetative and structural cues associated with habitat selection, a number of variables were measured at each python location (Table 15). Habitat characterisation was not conducted at locations where pythons were found actively travelling in order to avoid including instances where pythons may have been disturbed by our approach. Instances where pythons were found in buildings were also excluded because these sites could not be characterised adequately using the prescribed habitat-sampling scheme. When pythons were located more than 3 m up in the trees, habitat variables

were not measured. Habitat was quantified at a given position at the time that the python was located in order to minimise phenological changes between occupancy and sampling time. Locations at which a python was observed more than once were only included one time in the analyses.

Table 1 Structural variables used in the analysis of microhabitat selection by M. spilota with associated abbreviations and sampling radii. Variable Definition Radius

PDBH Mean dbha (cm) of all overstorey trees 5

NDBH Mean dbha (cm) of five nearest overstorey trees (but outside of plot) 30

DTREE Distance (m) to nearest overstorey tree n/a

DH20 Distance (m) to creek n/a

HOVER Height (m) of overstorey vegetation 5

HUNDER Height (m) of understory vegetation 5

%GRND Coverageb (%) of bare ground 5

%GVEG Coverageb (%) of ground vegetation (grasses, herbs, etc.) 5

%UVEG Coverageb(%) of understory vegetation (shrubs, lignum, etc.) 5

%LEAF Coverageb (%) of leaf litter and fallen debris 5

LOG Number of hollow logs > 10 cm in diametre 5

TEMP Microhabitat or shaded air temperature (°C) 5 a = Diameter at breast height and b = Percent coverage estimated using a structural scoring system modified from Specht (1981).

To determine whether pythons were using habitats non-randomly (i.e., selecting specific habitat types over others), habitat use was quantified by repeating the same habitat characterisation at random locations within a 10–200 m radius of the radio-tagged animal. Direction of travel to random sites was determined by rolling an 8-sided dice, each face corresponding to a direction of travel (i.e., 1 = north, 2 = northeast, 3 = east, etc.). Distance of travel was determined by rolling a 20-sided dice, multiplying this by ten and walking that many paces (after Blouin-Demers and Weatherhead, 2001b). The random points were representative of the habitats from which the pythons could choose (Keller and Heske, 2000; Blouin-Demers and Weatherhead, 2001b) since pythons usually moved less than 200 m between relocations (Keller and Heske, 2000; Blouin-Demers and Weatherhead, 2001b). To characterise the habitat at python and random locations, twelve structural variables within circular plots were measured. These plots were centred on the position of the python or random location and sampling radii varied from 5–30 m depending on the variable measured. All distance variables were measured with a tape measure, to either the nearest centimeter or meter.

We used multivariate analysis of variance (MANOVA) to determine if there was a significant difference in habitat centroids of each group and discriminant function analysis (DFA) to examine along which axes the groups differed and which variables contributed most to separation among groups (snake vs. random locations). To facilitate the interpretation of discriminant functions, the

interpretation was limited to the five variables with the highest correlative values with each function (Blouin-Demers and Weatherhead, 2001b). From these correlations it was possible to derive biological interpretations for the discriminant axes. Such interpretations are simplifications based upon only the most highly correlated variables. We removed one variable from pairs of highly correlated (|r| > 0.70) variables from analyses to reduce multicollinearity (see Harvey and Weatherhead, 2006). We conducted Box’s M test prior to all multivariate tests and in all cases the covariance matrices were found to be heterogeneous (p < 0.001). This is often expected with ecological data (James and McCulloch, 1990; Stevens, 1996), but previous authors have defended the analytical value of multivariate tests despite deviation from this assumption (see review in Reinert, 1984b).

Tree Measurement

Trees occupied by radio-tagged pythons were identified to the species level using Cunningham et al. (1981). They were then flagged and given a unique number to assess whether more than one python used them and to determine if individuals exhibited site fidelity. Tree height and diameter at breast height (dbh) were measured with a range finder (Bushnell Elite 1500 Rangefinder, accuracy ± 1 m) and tape measure. The number of hollow branches and stem holes was assessed visually by counting the number of visible hollows from ground level. The same physical characteristics of the first five trees encountered on each of four 2 m wide transects running east, west, north and south from each tree occupied by a radio-tagged python were measured to compare the characteristics of trees that were available but unused (after Webb and Shine, 1997b). The same homestead-woodland boundary described above was used when differentiating between Homestead and Park trees. Used and available trees were also characterised when snakes were observed opportunistically inside of tree hollows but which were not radio-tagged. For all radio-tracked animals we calculated residence times (both minimum and maximum) for the duration of time spent in the same tree. The minimum residence time refers to the number of days that pythons were actually observed in the same tree; while the maximum residence time is the number of days that pythons were potentially in the same tree. The latter includes situations where pythons were located in the same tree on more than one occasion, but it was unable to determine if they actually left and returned to that tree between those trips. All of these cases however, were over-wintering trees, and thus it was unlikely that pythons actually moved in these instances.

Diet and Relative Prey Density All faecal samples produced by freshly captured animals were analysed by a specialist hair identifier (B. Triggs, ‘Dead Finish’, via Genoa, VIC, Australia). Observations of radio-tracked pythons during foraging encounters were also recorded. Small-medium sized mammals are one of the most important components of the diet of M. spilota (Slip and Shine, 1988a; Shine and Fitzgerald, 1996; Pearson et al., 2002a). To determine the relative prey abundance around the Homestead and in the Park, small mammals were captured using Elliott live capture traps (11 × 11 × 30 cm, Elliot Scientific, Upwey, VIC). In each area a grid comprised of five parallel lines spaced 15 m apart and with traps placed at 15 ± 2 m intervals along these lines was used. The best trap location was within 2 m of the sampling point, which was chosen to improve trapping success (Stewart, 1979; Tasker and Dickman, 2002). Traps were baited with a mixture of rolled oats, peanut butter and vanilla essence. A handful of leaf litter was placed inside of traps for insulation. Traps were set in the late afternoon and checked and closed at dawn. Traps containing animals were removed and replaced with clean ones. Traps were opened for four consecutive nights each season, giving a total of 800 trap nights (400 in each location) across the study period.

Activity and Movements Analyses Movement distances were calculated as the minimum straight-line distance between successive locations using the Animal Movement Extension (Hooge & Eichenlaub, 2000) of ArcView GIS (version 3.2; Environmental Systems Research Institute, Inc., Redlands, California, USA). Several movement and activity variables were calculated: the total distance moved by each individual, the mean distance each individual traveled per move (as recommended by Gregory et al., 1987; calculated as: total distance moved / number of moves) and movement frequency (number of moves made / number of relocations). To enable direct comparisons of vagility with results of Shine and Fitzgerald (1996), daily displacements (total distance moved / number of radio tracking days) were also calculated.

Pythons were divided into three groups based on their sex and age: males, females, and juveniles. One female brooded eggs during the course of the study; brooding pythons do not usually leave their eggs except to bask briefly (Slip and Shine, 1988d). We tracked this individual repeatedly (i.e., up to 8 times a day) to determine its behaviour, but it did not leave the eggs for the duration of the brooding period. As such, we did not include this female in any of the movement or space-use analyses during this time, to prevent her lack of movement from biasing any of the female movement and space use estimates.

Home Range Analysis Burt’s (1943) classic definition of home range is the area traversed by an animal in its normal activities of feeding, mating, caring for young, and shelter seeking. Two home range estimators were used in this study (Harris et al., 1990 and Kernohan et al., 2001). Minimum convex polygon (MCP) was calculated to enable comparisons with previous studies and the kernel home range estimator was used to determine location probabilities (Kernohan et al., 2001). For kernel density analysis, we employed the fixed kernel and the least squares cross validation methods (Worton, 1995) to select a bandwidth for the smoothing parameter (h). We used the 95% and 50% kernel density estimates to calculate the total and core activity areas respectively. All home range analysis was undertaken using the Animal Movements Extension (Hooge and Eichenlaub, 2000) in ArcView GIS.

Statistical Analysis Statistical analyses were performed with SAS (version 9.1, SAS Institute Inc., 2004) or SigmaStat (version 3.1, Systat Software Inc., 2004) in accordance with Sokal & Rohlf (1995). In all cases, we examined the assumptions of equality of variances and normality. If the data failed to meet these assumptions, transformations were performed to better approximate normal distributions or equal variances. Tukey-Kramer tests were used for all multiple comparisons. Statistical significance was accepted at α = 0.05 level and means are presented ± one standard error (SE) except where stated otherwise. Because variation among individuals in the number of relocations (telemetry fixes) obtained may potentially contribute to variability in estimates of movement, activity and space use (Kernohan et al., 2001), the number of relocations on all movement and space use estimates was regressed to assess possible relationships.

Results

Vertebrate Surveys

VESs and Artificial Cover Surveys There was considerable variation in the ability of the different sampling methods to detect the targeted species. Visual Encounter Surveys detected all of the species observed, Artificial Cover Surveys detected only small lizards and frogs and PVC tubes were not utilized by any species (Table 1). In total 13 species were observed: six lizards, two snakes, one turtle and four frogs. Overall, few individuals of each species were observed. Although it appears that the mean number of individuals observed in revegetated sites was greater than in control sites, a significant difference was not detectable due to low sample size (Figure 3).

Bird Counts A total of 39 birds species were detected in revegetated and control sites (Table 2). A similar number of species were detected at both: 31 in revegetated and 29 in control. Nine species were only observed in revegetated sites: Blue-faced Honeyeater, Cockatiel, Grey Fantail, Kestrel, Kookaburra, Rufous Whistler, Tree Martin, Yellow Thornbill and the Yellow-rumped Thornbill. Six species were only observed in control sites: Australian Hobby, Black-faced Cuckoo Shrike, Little Friarbird, Littlepied Cormorant, Whistling Kite and Wood Duck.

Frog Study

Abundance A total of 6137 individuals (4557 juveniles and 1580 adults) were captured from the three experimental rice bays at Old Coree with the greatest number being captured in February (Figure 13 and Figure 14). Limnodynastes tasmaniensis were captured more frequently than L. fletcheri (4669 L. tasmaniensis and 1588 L. fletcheri). Juveniles of both species were captured more frequently than adults at Old Coree.

The total number of frogs captured at both study locations in Leeton was 1217 (662 L. tasmaniensis and 555 L. fletcheri). A total of 471 frogs were captured (459 L. tasmaniensis and 12 L. fletcheri) were captured in the single drift fence at Old Coree. The number of frogs captured in these fences was highest in February (Figure 15 and 16).

Dietary Items A total of 1216 frogs were stomach-flushed during the study with 650 and 566 frogs at Old Coree and Leeton, respectively. Stomach contents were extracted from 396 individuals at Old Coree and 397 Leeton (Table 3). The overall proportion of frogs that contained prey items was approximately 60 %.

Table 2 The species detected by the three sampling methods used in this study. X = observed and 0 = not observed. Artificial Cover Visual Encounter Species PVC Tubes Surveys Surveys

Lizards

Bearded Dragon Pogona barbata 0 X 0

Boulenger's Skink Morethia X X 0 boulengeri

Common Dwarf Skink Menetia greyii X X 0

Eastern Blue-tongued Lizard Tiliqua 0 X 0 scincoides

Snake-eyed Skink Cryptoblepharus X X 0 carnabyi

Common Dtella Gehyra variegata X X 0

Snakes 0

Eastern Brownsnake Pseudonaja 0 X 0 textilis

Red-bellied Black Snake Pseudechis 0 X 0 porphyriacus

Turtles 0

Eastern Long-necked Turtle Chelodina 0 X 0 longicollis

Frogs

Plains Froglet Crinia parinsignifera X X 0

Spotted Grass Frog Limnodynastes X X 0 tasmaniensis

Barking Marsh Frog Limnodynastes X X 0 fletcheri

Peron's Treefrog Litoria peroni 0 X 0

7 Reveg 6 Control 5

3 Detected

1 Mean Number of Individuals Individuals of Number Mean

0 Summer Spring Autumn Winter Season

Figure 13 Mean number of individuals detected in revegetated (n = 7) and control sites (n = 7) in each season.

Table 3 Species of birds observed in revegetated and controls sites according to season. Fly-overs and fly-throughs are included.

Species Revegetated Control Autumn Winter Spring Summer

Australian Raven X X X X X

Australian Hobby X X

Black-faced Cuckoo Shrike X X

Black Duck X X X X

Black Shouldered Kite X X X

Blue-faced Honeyeater X X X

Chough X X X 32 Cockatiel X X

Crested Pigeon X X X X X X

Eastern Rosella X X X X X

Galah X X X X X X

Great Egret X X X

Grey Fantail X X

Kestrel X X X

Kookaburra X X X

Little Friarbird X X

Little Pied Cormorant X X X X X

Magpie X X X X X X

Noisy Miner X X X X X X

Pacific Heron X X

Pee Wee X X X X X X

Pied Butcherbird X X X X X X

Red-rumped Parrot X X X X X

Rufous Whistler X X X

Starling X X X X X X

33 Straw-necked Ibis

Species Revegetated Control Autumn Winter Spring Summer

Striated Pardalote X X X X X

Superb Parrot X X X X

Tree Martin X X

Weebill X X X

Whistling Kite X X

White Cockatoo X X X X

White-faced Heron X X X X

White-plumed Honeyeater X X X

Willie Wagtail X X X X X X

Wood Ducks X X X

Yellow Rosella X X X X X X

Yellow Thornbill X X X X X

Yellow-rumped Thornbill X X X

4000

3500

3000

2500 juvenile 2000 adult 1500

Total number of individuals of number Total 1000

500

0 JAN FEB MAR APR Month

Figure 14 Total number of L. tasmaniensis captured at Old Coree from the three experimental rice bays in 2007.

1500

1250

1000 juvenile 750 adult

500 Total number of individuals of number Total

250

0 JAN FEB MAR APR Month

Figure 15 The total number of L. fletcheri captured at Old Coree from the three experimental rice bays in 2007.

450

400

350

300 Farm 1 Leeton 250 Farm 2 Leeton Farm 3 Leeton 200 Old Coree 150

Total number of individuals of number Total 100

0 JAN FEB MAR APR Month

Figure 16 The total number of L. tasmaniensis captured from the four 40 m drift fences at three study farms in Leeton and Old Coree in 2007.

350

300

250

Farm 1 Leeton 200 Farm 2 Leeton Farm 3 Leeton 150 Old Coree

100 Total number of individuals of number Total

0 JAN FEB MAR APR Month

Figure 17 The total number of L. fletcheri captured from the four 40 m drift fences at three study farms in Leeton and Old Coree in 2007.

The diet of L. tasmaniensis was composed of 27 different orders of invertebrate (Table 4). The most frequently consumed prey items for L. tasmaniensis were beetles (Coleoptera), ants (Formicidae) and true bugs (Hemiptera). These three prey types were also consumed in the greatest numeric proportion. The prey types consumed in the greatest volumetric proportion were caterpillars (Lepidoptera larvae), beetles (Coleoptera) and ants (Formicidae). Due to the partial digestion of some items, 9% of the total was unidentifiable. Interestingly, a scorpion was consumed by an individual L. tasmaniensis. This individual had a SVL of 23 mm and weighed 1.6 g; the scorpion consumed was 16 mm long and had a total volume of 33 mm3.

The diet of L. fletcheri was composed of 19 different orders of invertebrate (Table 5). The most frequently consumed prey items for L. fletcheri were ants (Formicidae), beetles (Coleoptera) and sawflies, and wasps and bees (Hymenoptera). These three prey types were also consumed in the greatest numeric proportion. Beetles, millipedes (Diplopoda) and crickets and grasshoppers (Orthoptera) were consumed in the greatest volumetric proportions.

Table 4 Monthly summary of the numbers of frogs collected, number of stomachs flushed and the samples extracted for the study period. Old Coree farm January February March April May

Number of frogs collected 228 131 141 110 40

Number of stomach samples 135 78 70 85 28 extracted

Stomach samples collected (%) 59 59 49 77 60

Total number of L. tasmaniensis 132 72 67 84 22

Total number of L. fletcheri 3 6 3 1 6

Leeton farms

Number of frogs collected 159 115 109 100 83

Number of stomach samples 105 73 79 64 76 extracted

Stomach samples collected (%) 66 63 72 64 91

Total number of L. tasmaniensis 105 70 77 55 49

Total number of L. fletcheri 0 3 2 9 27

Old Coree farm January February March April May

Number of frogs collected 228 131 141 110 40

Number of stomach samples 135 78 70 85 28 extracted

Stomach samples collected (%) 59 59 49 77 60

Total number of L. tasmaniensis 132 72 67 84 22

Total number of L. fletcheri 3 6 3 1 6

Leeton farms

Number of frogs collected 159 115 109 100 83

Number of stomach samples 105 73 79 64 76 extracted

Stomach samples collected (%) 66 63 72 64 91

Total number of L. tasmaniensis 105 70 77 55 49

Total number of L. fletcheri 0 3 2 9 27

The rice pest species consumed by L. tasmaniensis and L. fletcheri included the Rice Stink Bug (Eysarcoris trimaculatus), aquatic snails (Gylptophysa sp. and Isidorella newcombi), Lepidoptera larvae including Armyworms (Leucania convecta, Spodoptera exempta, S. mauritia), Loopers (Mocis frugalis) and Rice Root Aphids (Rhopalosiphum rufiabdominalis) (Table 6). Approximately 5 % of the diet of the L. tasmaniensis and L. fletcheri consisted of rice pest species.

Both species of frogs consumed Lepidopteran, Dipteran and Coleopteran larvae. The proportions they were consumed in were less than 20% in all months. In January the consumption of larvae was highest for L. tasmaniensis and L. fletcher.

Adult invertebrates occurred in the diets of both species of frog in each month sampled (Figure 17 and Figure 18). The consumption of ants (Formicidae) increased after February, where as the consumption of beetles (Coleoptera) was highest in January and February for both species. The proportion of flies (Diptera), spiders (Aranae), grasshoppers (Orthoptera) and true bugs (Hemiptera) in the diet was relatively low in all months.

The mean of the total volume of prey consumed each month by L. tasmaniensis did not differ significantly from January to April. However, the mean of the total volume of prey consumed in May

was significantly higher than in the other four months (multiple comparison Bonferroni; P<0.01) (Figure 19).

There was no significant difference in the volume of prey in the stomachs of L. fletcheri consumed among months (multiple comparison Bonferroni; p = 1.00), noting however, the volume of prey in the stomachs in May had a wider range than in the previous four months (Figure 20).

Food Consumption Rates

The mean number of hours it took for food to pass through adult L. tasmaniensis was 10 hrs (1 SD = 4.2, n = 7) and 12.8 hrs (1 SD = 4.2, n = 6) for juveniles. The minimum time taken to pass the food was 4 hrs and the maximum was 19 hrs. One male did not pass a faecal pellet during the 48 hr feeding trial.

Based on all frogs sampled, the mean daily consumption of invertebrates was 8.2 items (1 SD = 7.4). However, there was a wide range of the number of invertebrates consumed (1 – 69 prey items). The stomach contents of the individual that consumed 69 prey items included 61 ants, 5 beetles and 1 Dipteran larvae.

Body Size and Food Intake There was a significant positive relationship between frog snout-vent length and the smallest volumes of 2 individual prey items consumed for L. tasmaniensis (F 1, 599 = 52.3, P< 0.05, r =0.28 ) (Figure 23).

There was no significant relationship between the smallest volumes of individual prey items consumed 2 for L. fletcheri (F 1, 55 = 0.1 P=0.8, r =0.04) (Figure 24). A significant positive relationship was also detected between frog snout-vent length and the smallest volumes of individual prey items consumed by 2 2 L. tasmaniensis (F 1, 599 = 62.9, P< 0.05, r =0.31) and L. fletcheri (F 1, 55 =26.7, P< 0.05, r = 0.58) (Figure 25 and 26). Large-bodied frogs consumed relatively small prey items, which were also consumed by smaller individuals. The largest volumes of individual prey items were consumed by L. tasmaniensis, which had a snout-vent length between 20-40 mm. The largest frogs (> 40 mm) did not exhibit this relationship, however, they only represented a small proportion of the entire sample of frogs.

Prey Consumption There was a significant correlation between the prey items consumed, and the frequencies at which they were consumed, at Old Coree and Leeton (P< 0.05). This indicates there is no significant difference between the diets of frogs among sites.

Table 5 Composition of the diet of L. tasmaniensis collected between January and May 2007. This table includes 5904 prey items from 733 frogs, with a total volume 61455.5 mm3 (volumetric proportion % excludes unidentified contents). Most abundant prey types found are highlighted. Prey type Frequency of Numeric Numeric Volumetric Volumetric occurrence proportion proportion proportion proportion

% % (mm3) %

Acarina 4.0 74 1.3 2.7 0.0

Hemiptera 20.9 461 7.8 534.4 2.7

Aranae 10.5 105 1.8 908.8 4.6

Blattodea 0.1 1 0.0 0.2 0.0

Coleoptera 57.2 1349 22.8 4439.8 22.5

Cladocera 0.1 2 0.0 1.6 0.0

Coleoptera larvae 3.6 100 1.7 346.5 1.8

Collembola 1.9 28 0.5 3.3 0.0

Dermaptera 1.9 22 0.4 425.7 2.2

Diptera 9.7 129 2.2 362.5 1.8

Diptera larvae 6.0 72 1.2 324.3 1.6

Siphonaptera 0.1 1 0.0 0.1 0.0

Formicidae 45.0 1947 33.0 2319.9 11.8

Hymenoptera 12.6 179 3.0 703.2 3.6

Isoptera 0.7 23 0.4 40.8 0.0

Isopoda 1.0 18 0.3 324.8 1.7

Lepidoptera larvae 11.6 274 4.6 4825.3 24.5

Lepidoptera 0.9 8 0.1 280.1 1.4

Pthiraptera 1.4 31 0.5 3.7 0.0

Prey type Frequency of Numeric Numeric Volumetric Volumetric occurrence proportion proportion proportion proportion

% % (mm3) %

Mantodea 0.2 2 0.0 4.0 0.0

Megaloptera 0.1 1 0.0 3.8 0.0

Diplopoda 0.9 9 0.2 930.4 4.7

Neuroptera 0.3 5 0.1 12.2 0.1

Odonata 3.0 42 0.7 193.7 1.0

Orthoptera 7.2 83 1.4 1089.4 5.5

Plecoptera 3.9 63 1.1 76.2 0.4

Pseudo-scorpiones 0.2 2 0.0 0.9 0.0

Scorpiones 0.1 1 0.0 33.5 0.2

Gastropoda 2.4 33 0.6 424.1 2.2

Unidentifiable 37.7 532 9.0 41725.0 -

Vegetation 19.1 5.1 1114.8 5.7

Thysanoptera 0.2 5 0.1 0.0 0.0

Total - 5904 100.0 61455.5 100.0

Table 6 The composition of the diet of L. fletcheri collected between January and May 2007. This table includes 583 prey items collected from 60 frogs with a total volume 29525.65 mm3 (volumetric proportion % excludes unidentified contents). Most abundant prey types are highlighted. Prey types Frequency of Numeric Numeric Volumetric Volumetric occurrence proportion proportion proportion proportion

% % (mm3) %

Acarina 0.8 1 0.2 0.0 0.0

Hemiptera 12.8 39 6.7 88.3 1.6

Aranae 7.2 20 3.4 72.0 1.3

Coleoptera 29.6 165 28.3 2162.1 39.9

Cladocera 0.8 2 0.3 42.0 0.8

Coleoptera larvae 7.2 13 2.2 82.8 1.5

Collembola 2.4 3 0.5 1.8 0.0

Dermaptera 0.8 1 0.2 70.7 1.3

Diptera 10.4 19 3.3 43.6 0.8

Diptera larvae 3.2 4 0.7 45.3 0.8

Formicidae 42.4 160 27.4 267.6 4.9

Hymenoptera 16.0 43 7.4 100.2 1.9

Isopoda 0.8 1 0.2 8.4 0.2

Lepidoptera 9.6 20 3.4 603.8 11.2 larvae

Megaloptera 0.8 2 0.3 5.2 0.1

Diplopoda 6.4 10 1.7 699.3 12.9

Neuroptera 1.6 2 0.3 5.8 0.1

Odonata 3.2 5 0.9 166.2 3.1

Orthoptera 10.4 14 2.4 663.6 12.3

Prey types Frequency of Numeric Numeric Volumetric Volumetric occurrence proportion proportion proportion proportion

% % (mm3) %

Unidentifiable 22.4 34 5.8 24111.7 -

Gastropoda 2.4 5 0.9 107.8 2.0

Vegetation 8.8 18 3.1 141.7 2.6

Plecoptera 1.6 2 0.3 5.8 0.1

Total - 583 100.0 29525.7 100.0

Table 7 The frequency of occurrence in the diet of L. tasmaniensis and L. fletcheri of specific invertebrates known to be pests in rice crops.

Rice pest species Numeric Numeric proportion proportion

Rice stink bug Eysarcoris trimaculatus 1 0.02

Aquatic snails Glyptophysa sp., Isidorella 37 0.57 newcombi

Lepidoptera Leucania convecta, Mocis 49 0.76 larvae frugalis,

Spodoptera exempta, S. mauritia

Rice root aphid Rhopalosiphum 264 4.07 rufiabdominalis

5 Frequency of occurence (%) Frequency of

0 JAN FEB MAR APR MAY Month Coleoptera larvae Diptera larvae Lepidoptera larvae

Figure 18 Monthly variations in the frequency of occurrence of the major larval prey taxa in the diet of L. tasmaniensis.

10 Frequency of occurence (%) Frequency of

0 JAN FEB MAR APR MAY Month

Coleoptera larvae Diptera larvae Lepidoptera larvae

Figure 19 Monthly variations in the frequency of occurrence of the major larval prey taxa in the diet of L. fletcheri.

Hemiptera 40 Aranae Coleoptera 30 Diptera Formicidae 20 Hymenoptera Orthoptera

Frequency of occurence (%) Frequency of 10

0 JAN FEB MAR APR MAY Month

Figure 20 Monthly variation in the frequency of occurrence of the major terrestrial prey taxa in the diet of L. tasmaniensis.

Hemiptera 40 Aranae Coleoptera 30 Diptera-F Formicidae 20 Hymenoptera Orthoptera

Frequency of occurence (%) Frequency of 10

0 JAN FEB MAR APR MAY Month

Figure 21 Monthly variation in the frequency of occurrence of the major terrestrial prey taxa in the diet of L. fletcheri.

1.2

1.1

1.0

0.9

0.8 Prey volume (mm^3) volume Prey

0.7

0.6

JAN FEB MAR APR MAY

Figure 22 Monthly variations in the mean total volume of prey items found in the stomachs of L. tasmaniensis . Circles indicate the mean and bars indicate one standard deviation.

1.4

1.2

1.0

0.8 Prey volume (mm^3) volume Prey 0.6

0.4

0.2 JAN FEB MAR APR MAY

Figure 23 Monthly variations in the mean total volume of prey items found in the stomachs of L. fletcheri. Circles indicate the mean and bars indicate one standard deviation.

2.0

0.0 Prey volume (mm^3) volume Prey

-2.0

20.0 30.0 40.0 50.0

Figure 24 Relationship between frog body size and smallest prey item in stomachs of L. tasmaniensis.

0.5

0.0

-0.5

-1.0 Prey volume (mm^3) volume Prey

-1.5

-2.0

20.0 25.0 30.0 35.0 40.0 45.0 50.0

Figure 25 Relationship between frog body size and smallest prey item in stomachs of L. fletcheri.

4.0

3.0

2.0

1.0

0.0 Prey volume (mm^3) volume Prey

-1.0

-2.0

20.0 30.0 40.0 50.0

Figure 26 Relationship between frog body size and largest prey item in stomachs of L. tasmaniensis.

3.0

2.0

1.0 Prey volume (mm^3) volume Prey

0.0

20.0 25.0 30.0 35.0 40.0 45.0 50.0

Figure 27 Relationship between frog body size and largest prey item for L. fletcheri.

Ontogenetic Dietary Shift

Adult and juvenile L. tasmaniensis consumed more prey orders than adult and juvenile L. fletcheri (Table 7). The most frequently consumed proportions of prey orders were Coleoptera for L. fletcheri and Hymenoptera (Formicidae) for L. tasmaniensis. Other Hymenoptera (bees, wasps and sawflies) and Hemiptera (true bugs) were consumed in relatively large proportions for both age classes of L. tasmaniensis and L. fletcheri. Adult L. tasmaniensis consumed relatively large proportions of caterpillars (Lepidoptera larvae).

With the exception of Lepidoptera larvae and ants, all the other prey orders were consumed in relatively the same proportions by both species (Figure 27 and Figure 28). Lymnodynastes tasmaniensis adults consumed Lepidoptera larvae in the largest proportion and L. tasmaniensis juveniles consumed ants in the largest proportion.

Table 8 Comparison of diet composition (numeric proportion in %) between two age groups (size classes) for both study species. Leeton L. fletcheri L. tasmaniensis

Juvenile Adult Juvenile Adult

Acarina 0.0 0.6 1.0 2.1

Aranae 4.2 1.3 1.7 2.2

Cladocera 0.5 0.0 0.0 0.0

Coleoptera 27.5 30.6 25.0 15.0

Coleoptera larvae 1.6 3.8 1.1 3.9

Collembola 0.7 0.0 0.6 0.0

Dermaptera 0.2 0.0 0.2 1.2

Diplopoda 1.6 1.9 0.0 0.5

Diptera 2.8 4.5 2.1 2.7

Diptera larvae 0.9 0.0 1.1 1.8

Formicidae 27.5 27.4 35.1 25.5

Gastropoda 0.9 0.6 0.6 0.5

Hemiptera 6.1 8.3 8.7 4.7

Hymenoptera 8.7 3.8 2.6 4.5

Isopoda 0.2 0.0 0.4 0.1

Isoptera 0.0 0.0 0.4 0.2

Leeton L. fletcheri L. tasmaniensis

Lepidoptera 0.0 0.0 0.1 0.3

Lepidoptera larvae 2.8 5.1 1.4 16.5

Mantodea 0.0 0.0 0.0 0.0

Megaloptera 0.5 0.0 0.0 0.0

Neuroptera 0.5 0.0 0.1 0.2

Odonata 0.7 1.3 0.7 0.7

Orthoptera 2.3 2.5 1.2 2.1

Plecoptera 0.5 0.0 0.9 1.6

Pseudoscorpiones 0.0 0.0 0.0 0.1

Pthiraptera 0.0 0.0 0.7 0.0

Siphonaptera 0.0 0.0 0.0 0.1

Thysanoptera 0.0 0.0 0.1 0.0

Vegetation 0.0 1.9 5.2 4.8

Unidentifiable 5.6 6.4 9.1 8.8

40 35 30 25 20 15 10 5

Frequency of occurence (%) Frequency of 0

Aranae Diptera Coleoptera Formicidae Hemiptera Orthoptera Diptera larvae Hymenoptera

Coleoptera larvae Lepidoptera larvae Prey order

L. tasmaniensis juvenile L. tasmaniensis adult

Figure 28 Frequency of occurrence of 10 major prey orders in adult and juvenile L. tasmaniensis.

35 30 25 20 15 10 5

Frequency of occurence (%) Frequency of 0

Aranae Diptera Coleoptera Formicidae Hemiptera Orthoptera Diptera larvae Hymenoptera

Coleoptera larvae Lepidoptera larvae Prey order

L. fletcheri juvenile L. fletcheri adult

Figure 29 Frequency of occurrence of 10 major prey orders in the stomachs of adult and juvenile L. fletcheri.

Correlation coefficients calculated for the 10 major prey orders consumed by L. tasmaniensis and L. fletcheri adults and juveniles indicated no significant relationship (P=0.12). There were significant correlations between L. tasmaniensis juveniles and the other age and species classes and L. fletcheri juveniles and the other age and species classes (P≤ 0.05). These results suggest that the diet is the same between the age classes and species.

Sex Differences in Diet There was a significant correlation between the prey orders consumed by males by and females of both species of frogs (Table 8). The significant correlation suggests there was no difference in the ranked orders of prey items consumed, or the frequency they were consumed by males or females of either species.

Seasonal Availability of Prey

Limnodynastes tasmaniensis and L. fletcheri both consumed a small number of aquatic invertebrates from the rice bays (Figure 29). The aquatic prey orders were all available in a higher proportion in the rice bay environment than consumed. The main families consumed included predaceous diving beetles (Dytisicdae) and water beetles (Hydrophilidae and Hygrobiidae). Aquatic Hemiptera species from the families Corixidae (water boatmen) and Veliidae (water striders) were also consumed. The aquatic coleopteran consumed included adult beetles and larvae.

Spearman’s correlation coefficients were calculated between the ten prey orders most frequently consumed by L. fletcheri and L. tasmaniensis and their abundance in the pitfall trap and sweep net samples (Table 9). The 10 most frequently consumed prey orders were used to avoid bias in the correlation due to nil counts in either the pitfall, sweep or stomach samples. There was a significant correlation between the orders of prey found in the stomach samples of L. fletcheri and L. tasmaniensis (p ≤0.05). A significant correlation occurred between the abundance of terrestrial invertebrates captured in pitfall traps and those found in the stomach contents of L. tasmaniensis (p ≤0.05). The correlation between the abundance of invertebrates captured in the pitfall traps and the stomach contents of L. fletcheri also approached significance (p=0.07). No correlation was found between the abundance of aerial invertebrates and the occurrence of aerial invertebrates and prey items of L. tasmaniensis or L. fletcheri.

The proportion of spiders (Aranae) captured in the pitfall traps at Old Coree was higher than the proportion consumed by L. tasmaniensis and L. fletcheri at both sites (Table 10, 11 and 12). Beetles (Coleoptera) were consumed in higher proportions by L. tasmaniensis at Old Coree than were found in pitfall traps or sweep samples; for L. fletcheri beetles were consumed in lower proportions than occurred in the pitfall traps and sweep samples (Table 13). At Leeton beetles were consumed less frequently by L. tasmaniensis and L. fletcheri than found in the pitfall traps and sweep samples. Bombardier beetles (Pheropsophus verticalis) were caught frequently in pitfall traps in large numbers but, were excluded from the total number of beetles captured as they did not appear to be part of the diet of L. tasmaniensis or L. fletcheri. Ants (Formicidae) were consumed in higher proportions than captured in pitfall traps or sweep samples by L. tasmaniensis at Old Coree and Leeton, and by L. fletcheri at Leeton. Limnodynastes fletcheri consumed ants in a greater proportion than available in the environment during February, March and April. In January and May, ants occurred more frequently in pitfall traps than in stomach contents. Diplopoda occurred more frequently in the environment at both sites than consumed by either species (Table 14).

Table 9 Spearman rank correlation coefficients between adult and juvenile L. tasmaniensis and L. fletcheri when examining the prey types consumed. Significant correlations are indicates as follows: * p<0.05; ** p<0.01; *** p<0.001. L. tasmaniensis L. fletcheri L. tasmaniensis L. fletcheri Adults Adults Juveniles Juveniles

L. tasmaniensis Correlation coefficient 0.52 0.76 0.67

Adults Sig. (2 tailed) 0.12 0.01** 0.03**

N 10 10 10

L. fletcheri Correlation coefficient 0.52 0.84 0.67

Adults Sig. (2 tailed) 0.12 0.00** 0.04** 53

N 10 10 10

L. tasmaniensis Correlation coefficient 0.76 0.84 0.93

Juveniles Sig. (2 tailed) 0.01** 0.00*** 0.00***

N 10 10 10

L. fletcheri Correlation coefficient 0.67 0.67 0.93

Juveniles Sig. (2 tailed) 0.03** 0.04** 0.00***

N 10 10 10

Table 10 Spearman's rank correlation coefficients between the frequency the prey orders are consumed by males and females of L. tasmaniensis and L. fletcheri. Significant correlations are indicates as follows: * p<0.05; ** p<0.01; *** p<0.001. L. tasmaniensis L. tasmaniensis L. fletcheri L. fletcheri Females Males Females Males

L. tasmaniensis Correlation coefficient 0.98 0.84 0.69

Females Sig. (2 tailed) 0.00*** 0.00*** 0.03**

N 10 10 10

L. tasmaniensis Correlation coefficient 0.98 0.84 0.75

Males Sig. (2 tailed) 0.00*** 0.00*** 0.01** 54

N 10 10 10

L. fletcheri Correlation coefficient 0.84 0.84 0.66

Females Sig. (2 tailed) 0.00*** 0.00*** 0.04**

N 10 10 10

L. fletcheri Correlation coefficient 0.69 0.75 0.66

Males Sig. (2 tailed) 0.03** 0.01** 0.04**

N 10 10 10

60.0

50.0

40.0 Old Coree Leeton 30.0 Stomach

20.0 Frequency of occurance (%) 10.0

0.0

Odonata Cladocera CrustaceaDecapoda Hemiptera Tricoptera Amphipoda ColeopteraCollembola Gastropoda Ephemeroptera Aquatic prey orders

Figure 30 The frequency of occurrence (%) of aquatic prey orders from sweep samples at Old Coree and Leeton and stomach samples from January to March 2007.

Table 11 Correlation between the availability of prey caught in pitfall traps and sweep samples and the prey found in the stomach contents of L. tasmaniensis and L. fletcheri. Significant correlations are indicates as follows: * p<0.05; ** p<0.01; *** p<0.001. L. fletcheri L. tasmaniensis Pitfall Sweep

L. fletcheri Correlation Coefficient 0.66 0.26 0.15

Sig. (2-tailed) 0.00*** 0.07 0.31

N 48 48 48

L. tasmaniensis Correlation Coefficient 0.66 0.35 0.24

Sig. (2-tailed) 0.00*** 0.02** 0.10

56 N 48 48 48

Pitfall Correlation Coefficient 0.26 0.35 0.09

Sig. (2-tailed) 0.07 0.02** 0.57

N 48 48 48

Sweep Correlation Coefficient 0.14 0.24 0.09

Sig. (2-tailed) 0.31 0.10 0.57

N 48 48 48

Table 12 Comparison of frequency of occurrence (%) of prey in L. tasmaniensis stomachs with relative abundance of prey collected in pitfall traps and sweep samples at Old Coree. Order January February March April May

Stomach Pitfall Sweep Stom. Pitfall Sweep Stom. Pitfall Sweep Stom. Pitfall Sweep Stom. Pitfall Sweep

Acarina 0.5 0 0 0.8 0 0 0.3 0 0 0 0 0 9.0 0 0

Aranae 3.2 24.5 0.5 1.5 2.6 0 2.1 7.1 0 0.4 7.4 0 3.2 3.9 0

Blattodea 0.1 0 0 0 2.7 0 0 0.7 0 0 0 0 0 0 0

Chilopoda 0 0.3 0 0 0 0 0 0 0 0 1.6 0 0 0 0

Cladocera 0.2 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Coleoptera 14.7 9.9 3.0 23.0 18.8 0 19.2 3.3 27.4 15.5 10.1 19.3 9.0 0 16.1

Coleoptera larvae 1.3 0 0 0.8 0 0 3.2 0 0 0.7 0 0 0 0 0 57 Coleoptera - 0 31.2 0 40.2 0 84.1 0 49.4 0 7.7 (bombardier)

Collembola 0.1 0 0 0.3 0 0 0.8 0 0 0 0 0 1.2 0

Dermaptera 0.1 1.0 0 0 0 0 0 1.3 0 0.4 0.8 0 0 0 0

Diplopoda 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Diptera 5.0 0 92.6 0.8 0 99.7 0.8 0 58.6 1.8 0.8 63.6 4.3 0 81.5

Diptera larvae 3.0 0 0 4.3 0 0 1.6 0 0 0 0 0 0 0 0

Formicidae 17.6 9.6 0 31.9 26.5 0 38.9 2.0 0 39.4 6.6 0 37.0 86.5 0

Gastropoda 0.1 0 0 5.3 0 0 0.5 0 0 0 0 9.2 0 0

Hemiptera 6.6 0.3 1.0 3.0 0.9 0.2 4.0 0.3 9.4 5.0 0.8 12.5 3.2 0 0

Hymenoptera 1.5 0 0 2.3 0 0 0.3 0 0 2.2 0.8 0 0 0 0

Order January February March April May

Stomach Pitfall Sweep Stom. Pitfall Sweep Stom. Pitfall Sweep Stom. Pitfall Sweep Stom. Pitfall Sweep

Isopoda 0 0 0 0.3 0 0 0 0 0 0 0.8 0 5.2 0 0

Isoptera 0.1 0 0 0 0 0 0 0 0 0.1 0 0 0 0 0

Lepidoptera 0.2 0 2.5 0 0 0.2 0 0.3 4.6 0.1 0 4.6 0.3 0 2.5

Lepidoptera larvae 3.4 0 0 3.0 0 0 1.9 0 0 1.0 0 0 1.2 0 0

Megaloptera 0 0 0 0 0 0 0 0 0 0.1 0 0 0 0 0

Neuroptera 0.1 0 0 0 0 0 0.8 0 0 0 0 0 0 0 0

Odonata 0.3 0 0.5 0.3 0.9 0 0.3 0 0 3.9 0.8 0 1.2 0 0

Orthoptera 1.3 21.0 0 2.0 6.8 0 2.4 3.6 0 1.4 5.7 0 0.9 1.9 0

Plecoptera 0.1 0 0 0 0 0 0 0 0 2.9 0 0 4.6 0 0 58

Pthiraptera 0 0 0 0 0 0 0.8 0 0 0 0 0 0 0 0

Pseudoscorpiones 0 0 0 0 0 0 0 0 0 0.1 0 0 0 0 0

Scorpionidae 0.1 2.2 0 0 0.9 0 0 0.7 0 0 7.4 0 0 0 0

Vegetation 1.7 0 0 3.0 0 0 4.3 0 0 14.0 0 0 8.1 0 0

Unidentifiable 11.5 0 0 17.5 0 0 10.7 0 0 10.6 0 0 2.6 0 0

Table 13 Comparison of frequency of occurrence (%) of prey in L. flectheri stomachs with relative abundance of prey collected in pitfall traps and sweep samples at Leeton. Order January February March April May

Stomach Pitfall Sweep Stom. Pitfall Sweep Stom. Pitfall Sweep Stom. Pitfall Sweep Stom. Pitfall Sweep

Acarina 1.4 0 0 2.9 0 0 0.7 0 0 0 0 0 0 0

Aranae 2.0 14.5 0 1.2 13.0 0 0.9 17.4 0 2.4 9.0 0 1.5 2.6 0

Blattodea 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Chilopoda 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Cladocera 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Coleoptera 17.0 50.6 0 35.9 75.4 33.3 17.9 54.1 27.4 24.3 59.6 19.3 13.0 0 0

Coleoptera larvae 5.0 0 0 1.7 0 0 0.7 0 0 0.3 0 0 1.9 0 0 59 Coleoptera 0 0 0 0 62.3 0 0 54.1 0 0 49.4 0 0 0 0 (bombardier)

Collembola 0 0 0 0.2 0 0 1.6 0 0 0 0 0 0.9 0 0

Dermaptera 1.8 4.0 0 0.3 1.5 0 0 4.6 0 0.6 6.7 0 0.1 2.6 0

Diplopoda 0 0 0 0 1.5 0 0 2.8 0 0 3.4 0 0.9 0 0

Diptera 1.2 0 98.3 4.3 0 40.0 0.3 0 58.6 1.5 0 63.6 0.7 36.8 89.1

Diptera larvae 1.5 0 0 0.6 0 0 1.0 0 0 0 0 0 0 0 0

Formicidae 20.8 4.0 0 25.8 0 0 53.4 4.6 0 32.7 13.5 0 40.8 0 0

Gastropoda 0.8 0 0 0 0 0 6.3 0 0 0 0 0 0.4 0 0

Hemiptera 3.5 0.8 0 8.1 0 20.0 4.6 0 9.4 12.6 0 12.5 17.5 50.0 0

Hymenoptera 0.1 0 0 1.1 0 0 0.9 0 0 2.4 1.1 0 8.9 0 0

Order January February March April May

Stomach Pitfall Sweep Stom. Pitfall Sweep Stom. Pitfall Sweep Stom. Pitfall Sweep Stom. Pitfall Sweep

Isopoda 0 0 0 1.2 0 0 0 0 0 0 1.1 0 0.3 0 0

Isoptera 0 0 0 0 0 0 0 0 0 0.3 0 0 0.3 0 0

Lepidoptera 0.3 0 1.8 0 1.5 6.7 0 0 4.5 0 0 4.6 0.3 0 10.9

Lepidoptera larvae 25.7 0 0 0.9 0 0 0.4 0 0 0.3 0 0 1.9 0 0

Mantodea 0 0 0 0.3 0 0 0 0 0 0 0 0 0 0 0

Megaloptera 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Neuroptera 0 0 0 0 0 0 0 0 0 0 0 0 0.1 0 0

Odonata 0 0 0 0.5 0 0 0.1 0 0 0.3 0 0 0 0 0

Orthoptera 0 26.1 0 1.5 7.3 0 0.4 16.5 0 0 5.6 0 1.4 7.9 0 60

Plecoptera 0 0 0 0 0 0 0.6 0 0 1.5 0 0 2.2 0 0

Pseudoscorpiones 0 0 0 0 0 0 0 0 0 0 0 0 0.1 0 0

Pthitaptera 0 0 0 2.7 0 0 0.1 0 0 0 0 0 0 0 0

Scorpionidae 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Siphonaptera 0.1 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Thysanoptera 0 0 0 0.8 0 0 0 0 0 0 0 0 0 0

Vegetation 1.2 0 0 4.7 0 0 5.3 0 0 7.8 0 0 3.7 0 0

Unidentifiable 13.5 0 0 5.3 0 0 4.7 0 0 12.9 0 0 3.1 0 0

Table 14 Comparison of frequency of occurrence (%) of prey in L. fletcheri stomachs with relative abundance or prey collected in pitfall traps and sweep samples at Old Coree. Order January February March April May

Stomach Pitfall Sweep Stom. Pitfall Sweep Stom. Pitfall Sweep Stom. Pitfall Sweep Stom. Pitfall Sweep

Acarina 0 0 0 0 0 0 0 0 0 0 0 0 0.3 0 0

Aranae 0 14.5 0 3.7 13.0 0 0 17.4 0 3.5 9.0 0 3.8 2.6 0

Blattodea 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Chilopoda 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Cladocera 0 0 0 0 0 0 0 0 0 3.5 0 0 0 0 0

Coleoptera 0 50.6 0 48.1 13.0 33.3 7.1 2.8 27.4 24.6 10.1 19.3 29.7 0 0

Coleoptera larvae 0 0 0 3.7 0 0 0 0 0 0 0 0 2.6 0 0 61 Coleoptera 0 0 0 0 62.3 0 0 54.1 0 0 49.4 0 0 0 0 (Bombardier)

Collembola 0 0 0 0 0 0 0 0 0 0 0 0 0.9 0 0

Dermaptera 0 4.0 0 0 1.5 0 0 4.6 0 1.8 6.7 0 0 2.6 0

Diplopoda 0 0 0 0 1.5 0 0 2.8 0 0 3.4 0 2.9 0 0

Diptera 0 0 98.3 3.7 0 40.0 0 0 58.6 3.5 0 63.6 1.5 36.8 89.1

Diptera larvae 0 0 0 0 0 0 0 0 0 0 0 0 0.3 0 0

Formicidae 100.0 4.0 0 22.2 0 0 50.0 4.6 0 33.3 13.5 0 24.7 0 0

Gastropoda 0 0 0 0 0 0 7.1 0 0 0 0 0 0.3 0 0

Hemiptera 0 0.8 0 7.4 0 20.0 14.3 0 9.4 0 0 12.5 6.5 50.0 0

Hymenoptera 0 0 0 0 0 0 0 0 0 10.5 1.1 0 9.1 0 0

Order January February March April May

Stomach Pitfall Sweep Stom. Pitfall Sweep Stom. Pitfall Sweep Stom. Pitfall Sweep Stom. Pitfall Sweep

Isopoda 0 0 0 0 0 0 0 0 0 0 1.1 0 0.3 0 0

Isoptera 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Lepidoptera 0 0 1.8 0 1.5 6.7 0 0 4.5 0 0 4.5 0 0 10.9

Lepidoptera larvae 0 0 0 0 0 0 0 0 0 0 0 0 3.8 0 0

Mantodea 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Megaloptera 0 0 0 0 0 0 0 0 0 0 0 0 0.6 0 0

Neuroptera 0 0 0 0 0 0 0 0 0 0 0 0 0.6 0 0

Odonata 0 0 0 0 0 0 0 0 0 1.8 0 0 1.2 0 0

Orthoptera 0 26.1 0 7.4 7.3 0 0 16.5 0 0 5.6 0 3.2 7.9 0 62

Plecoptera 0 0 0 0 0 0 0 0 0 1.8 0 0 0 0 0

Pseudoscorpiones 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Pthitaptera 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Scorpionidae 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Siphonaptera 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Thysanoptera 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Table 15 Comparison of frequency of occurrence (%) of prey in L. fletcheri stomachs with relative abundance. Order January February March April May

Stomach Pitfall Sweep Stom. Pitfall Sweep Stom. Pitfall Sweep Stom. Pitfall Sweep Stom. Pitfall Sweep

Acarina 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Aranae 6.9 24.5 0.5 0 2.6 0 0 7.1 0 4.2 7.4 0 1.8 3.9 0

Blattodea 0 0 0 0 2.6 0 0 0.7 0 0 0 0 0 0 0

Chilopoda 0 0.3 0 0 0 0 0 0 0 0 1.6 0 0 0 0

Cladocera 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Coleoptera 55.2 41.1 20.0 21.1 18.8 0 11.8 3.3 27.4 16.7 10.1 19.3 18.2 0 16.1

Coleoptera larvae 0 0 0 0 0 0 11.8 0 0 4.2 0 0 0 0 0

63 Coleoptera 0 27.3 0 40.2 0 80.8 0 49.4 0 7.7 (bombardier)

Collembola 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Dermaptera 0 1.0 0 0 0 0 0 1.3 0 0 0.8 0 0 0 0

Diplopoda 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Diptera 6.9 0 92.6 0 0 99.7 0 0 58.6 0 0.8 63.6 16.4 0 81.5

Diptera larvae 0 0 0 21.1 0 0 5.9 0 0 0 0 0 0 0 0

Formicidae 6.9 9.6 0 36.8 26.5 0 47.1 2.0 0 20.8 6.6 0 38.2 86.5 0

Gastropoda 0 0 0 15.8 0 0 0 0 0 0 0 0 0 0 0

Hemiptera 3.4 0.3 1.0 0 0.9 0.2 11.8 0.3 9.4 33.3 0.8 12.5 3.6 0 0

Hymenoptera 0 0 0 0 0 0 0 0 0 8.3 0.8 0 7.3 0 0

Isopoda 0 0 0 0 0 0 0 0 0 0 0.8 0 0 0 0

Order January February March April May

Stomach Pitfall Sweep Stom. Pitfall Sweep Stom. Pitfall Sweep Stom. Pitfall Sweep Stom. Pitfall Sweep

Isoptera 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Lepidoptera 0 0 2.5 0 0 0.2 0 0.3 4.6 0 0 4.6 0 0 2.5

Lepidoptera larvae 20.7 0 0 5.3 0 0 0 0 0 0 0 0 0 0 0

Megaloptera 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Neuroptera 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Odonata 0 0 0.5 0 0.9 0 0 0 0 0 0.8 0 0 0 0

Orthoptera 0 21.0 0 0 6.8 0 5.9 3.6 0 0 5.7 0 0 1.9 0

Plecoptera 0 0 0 0 0 0 0 0 0 0 0 0 1.8 0 0

Pthiraptera 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 64

Pseudoscorpiones 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Scorpionidae 0 2.2 0 0 0.9 0 0 0.7 0 0 7.4 0 0 0 0

Vegetation 0 0 0 0 0 0 0 0 0 4.2 0 0 9.1 0 0

Unidentifiable 0 0 0 10.5 0 0 5.9 0 0 8.3 0 0 3.6 0 0

Carpet Python Study

Activity and Movements Throughout this study, radio-telemetered pythons spent long periods inactive, inside retreat sites and hence moved infrequently. More often than not, radio-telemetered pythons were not visible when located (73% of locations) and were only observed moving during 4% of the location events. Posture categories 2 were independent of sex (χ 10 = 29.26; P = 0.001); males were observed moving more frequently (54% of all observed movements) than females (25%) and juveniles (21%). Postures also differed significantly 2 between locations (χ 5 = 29.49; P < 0.001). Homestead snakes were less likely to be seen (79% of relocations = not visible, compared to 67% of relocations for Park snakes) and thus, were less likely to be observed in any of the other posture categories. There were no significant relationships between the number of fixes obtained and all but one of our 2 movement variables (movement frequency: r = 0.08; F 1, 16 = 1.22; P = 0.28; mean movement per move: 2 2 r = 0.01; F 1, 16 = 0.28; P = 0.60; mean daily displacements: r = 0.04; F 1, 16 = 0.67; P = 0.42. The 2 exception being total distance moved (r = 0.29; F 1, 16 = 6.2; P = 0.03). The total distance moved was greater with a greater number of fixes. Body size did not influence mean distance moved (SVL: r2 = 2 0.04; F 1, 16 = 0.65; P = 0.36; body mass: r = 0.04; F 1, 16 = 0.64; P = 0.44) or movement frequency 2 2 (SVL: r = 0.03; F 1, 16 = 0.40; P = 0.54; body mass: r = 0.004; F 1, 16 = 0.07; P = 0.80), but did 2 2 significantly influence total distance moved (SVL: r = 0.33; F 1, 16 = 7.51; P < 0.001; body mass: r = 0.32; F 1, 16 = 7.19; P = 0.02). Larger individuals moved greater distances.

Movement frequency was not influenced by location (F 1, 59 = 1.86; P = 0.182) or sex (F 2, 59 = 0.16; P = 0.855), but was significantly influenced by season (F 3, 59 = 25.72; P < 0.001) (Figure 32). There was a significant interaction between location and season (F 3, 59 = 3.40; P = 0.028), Park animals moved more frequently than Homestead animals during spring and summer. Total distances moved did not differ significantly between locations (location: F 1, 59 = 0.83; P = 0.368), but were significantly influenced by both sex (F 2, 59 = 3.71; P = 0.034) and season (F 3, 59 = 15.46; P < 0.001). Males moved further than females and juveniles (Table 16) and all animals moved further in spring and summer (Figures 33). Mean distances moved (per move) did not differ significantly between locations (F 1, 59 = 0.94; P = 0.339) or sexes (F 2, 59 = 1.45; P = 0.249), but differed significantly between seasons (F 3, 59 = 10.62; P < 0.001); pythons moved further in spring and summer (Figure 34). Mean and total distances moved were 2 2 not related to movement frequency (mean: r = 0.14; F 1, 16 = 0.31; P = 0.58; total: r = 0.13; F 1, 16 = 0.26; P = 0.61).

Home Range Estimates There was a negative relationship between the distance a snake was located from its point of release and 2 the number of days since release (r = 0.03; F 1, 250 = 5.49; P = 0.02), suggesting that pythons did not wander randomly, but restricted their movements to a distinct home range. No home range estimates 2 were correlated with the number of locations obtained: MCP: r = 0.06; F 1, 16 = 1.03; P = 0.37; 95% 2 2 kernel density estimate: r = 0.01; F 1, 16 = 0.17; P = 0.69; 50% kernel density estimate: r = 0.008; F 1, 16 = 0.13; P = 0.72.

Table 16 Movement and space use patterns in Morelia spilota. Note: values are means and standard errors are given in parentheses. Activity centre area Movement Total area usage Location usage

Mass Tracking Proportion SVL (cm) Mean 95% 50% (kg) period (days) of total area Total movement distance Movement Kernel Kernel MCP (ha) used as (m) moved (m/ frequency density density activity move) (ha) (ha) centre

Park

Juvenile 99.8 (5.7) 0.2 (0.1) 127 (82) 295.2 (65.1) 52.4 (12.3) 0.5 (0.14) 0.5 (0.3) 1.3 (0.6) 0.2 (0.1) 16.0 (1.7)

Female 166.2 (12.6) 1.3 (0.3) 292 (49) 1282.9 (447.9) 67.2 (13.6) 0.4 (0.01) 3.1 (0.5) 2.3 (0.9) 0.3 (0.1) 14.4 (2.3)

Male 157.8 (8.9) 1.1 (0.1) 91 (18) 1482.3 (364.0) 132.6 (19.5) 0.5 (0.14) 9.9 (6.2) 12.1 (9.3) 1.8 (1.5) 12.5 (1.7)

66 All 170 (50)a 1020.1 (248.6) 84.0 (14.5) 0.47 (0.06) 4.5 (2.3) 5.3 (3.2) 0.8 (0.5) 14.3 (1.1)

Homestead

Juvenile 106.7 (1.6) 0.3 (0) 253 (22) 789.5 (131.0) 93.2 (38.3) 0.3 (0.07) 3.0 (1.4) 5.2 (3.5) 1.0 (0.7) 17.8 (1.6)

Female 166.9 (14.4) 1.5 (0.4) 278 (48) 598.3 (184.2) 40.0 (13.8) 0.3 (0.04) 1.0 (0.7) 0.8 (0.6) 0.1 (0.1) 17.5 (3.6)

Male 173.4 (22.1) 2.1 (0.8) 352 (24) 2417.8 (1085.5) 83.1 (27.3) 0.4 (0.05) 9.6 (6.0) 3.6 (2.4) 0.7 (0.1) 17.3 (1.8)

all 300 (24) 1328.4 (482.7) 69.5 (15.1) 0.35 (0.03) 4.5 (2.3) 2.9 (1.2) 0.6 (0.3) 17.5 (1.4)

a Tracking periods for park snakes were shorter than those of homestead snakes owing to transmitter failure (one juvenile) and the death of several animals (one (female) as a result of predation and two (one juvenile and one male) from unknown causes – see Corey, 2007).

1.0 Park Homestead 0.8

0.6

0.4

0.2

Movement frequency

0.0 Spring Summer Autumn Winter

Season

Figure 31 Seasonal movement frequencies for M. spilota from the woodland of Willandra National Park and Willandra Homestead. Error bars are ± 1SE.

1.0 Park Homestead 0.8

0.6

0.4 Frequency

0.2

0.0 0-20 21-40 41-60 61-80 81-100 > 101 Movement interval (m)

Figure 32 Frequency distribution of movement distance intervals for M. spilota from the woodland of Willandra National Park and Willandra Homestead. Error bars are ± 1SE.

1200 1200 Spring Summer 1000 1000

800 800

600 600

400 400

200 200

0 0 Park Homestead Park Homestead 1200 1200 Autumn Winter 1000 1000

Distance moved (m) 800 800

600 600 mean distance total distance 400 400

200 200

0 0 Park Homestead Park Homestead Location

Figure 33 Seasonal distances moved by M. spilota from the woodland of Willandra National Park and Willandra Homestead. Mean distance is the mean distance moved per move, while total distance is the total distance moved (see materials and methods). Sample sizes refer to the number of snakes and error bars are ± 1SE.

Patterns of area usage were larger for park snakes and larger in males than females and juveniles (Table 16), but did not differ significantly between locations or sexes (two-factor ANOVA, MCP: location: F 1, 16 = 0.028; P = 0.869; sex: F 2, 16 = 2.46; P = 0.131; location × sex: F 2, 16 = 2.55; P = 0.123; 95% kernel: location: F 1, 16 = 0.68; P = 0.428; sex: F 2, 16 = 0.86; P = 0.450; location × sex: F 2, 16 = 1.83; P = 0.205; 50% kernel: location: F 1, 16 = 0.25; P = 0.628; sex: F 2, 16 = 0.74; P = 0.501; location × sex: F 2, 16 = 1.49; P = 0.268; Proportion of total area used as activity centre: location: F 1, 16 = 2.86; P = 0.119; sex: F 2, 16 = 0.34; P = 0.718; location × sex: F 2, 16 = 0.20; P = 0.826.

2 Minimum convex polygons were positively related to body size (SVL: r = 0.27; F 1, 16 = 5.62; P = 0.03; 2 mass: r = 0.25; F 1, 16 = 5.03; P = 0.04). However, both total and core activity area estimates were not 2 related to either SVL or body mass (SVL vs. 95% kernel: r = 0.06; F 1, 16 = 0.95; P = 0.34; mass vs. 2 2 95% kernel: r = 0.04; F 1, 16 = 0.69; P = 0.42; SVL vs. 50% kernel: r = 0.06; F 1, 16 = 0.91; P = 0.36; 2 mass vs. 50% kernel: r = 0.06; F 1, 16 = 0.93; P = 0.35). Both mean and total movements were significant 2 2 predictors of area usage (mean distance: MCP: r = 0.77; F 1, 16 = 21.78; p < 0.001; 95% kernel: r = 2 2 0.77; F 1, 16 = 21.36; P < 0.001; 50% kernel: r = 0.76; F 1, 16 = 20.28; P < 0.001; Total distance: MCP: r 2 2 = 0.91; F 1, 16 = 69.69; P < 0.001; 95% kernel: r = 0.63; F 1, 16 = 9.98; P = 0.006; 50% kernel: r = 0.62; 2 F 1, 16 = 9.52; P = 0.008). However, none of these were related to movement frequency (MCP: r = 0.27;

2 2 F 1, 16 = 1.17; P = 0.3; 95% kernel: r = 0.44; F 1, 16 = 3.65; P = 0.08; 50% kernel: r = 0.43; F 1, 16 = 3.42; P = 0.08). The areas occupied by the eight Homestead snakes exhibited considerable overlap and were centred on buildings (Figure 35). Park snakes showed less overlap, but this was partly due to animals being captured in different areas. In the park, five pythons were observed sheltering inside tree hollows that were within the home ranges of radio-telemetered animals, and towards the end of the study, another unmarked male was captured within the home range of a radio-tagged male. While not direct evidence, these observations suggest that park animals also display spatial and temporal overlap of home ranges. Nearly half of all radio-tracked animals (47%) returned to sites where they had previously been located. Around the homestead these sites were usually buildings (59% of revisited sites) or trees (31.8% of revisited sites). Homestead snakes tended to revisit previously used sites more than Park snakes; 15% of all movements were to previously occupied sites (compared to only 7% in the park). Park snakes returned to previously occupied tree hollows (75% of revisited sites) and logs (25% of revisited sites). Around the homestead, the number of sites revisited by the same individual ranged from 0–7, while in the park it ranged from 0–4. There was no relationship between any of the area use estimates and the 2 2 number of revisits to previously occupied sites (MCP: r = 0.09; F 1, 16 = 0.13; P = 0.7; 95% kernel: r = 2 0.22; F 1, 15 = 0.69; P = 0.42; 50% kernel: r = 0.20; F 1, 15 = 0.60; P = 0.45).

Patterns of Habitat Use Radio-tracked pythons spent much of their time in habitat types that provided hard cover and thus, were rarely visible when located. There were significant differences in the frequency of locations in the 2 2 different cover types, between both sex/age classes (χ 6 = 30.49, p < 0.001) and landscape types (χ 3 = 14.50, p = 0.001). Even though the use of different cover types differed between Homestead and Park animals, pythons from these areas displayed broadly similar patterns in their use of the different cover types. For example, males from both the homestead and park were observed basking more often (open cover type) than both females and juveniles, and females and juveniles spent more time in hard cover types such as tree hollows. The higher proportion of observations in dense cover types for males in the park reflects their less frequent use of tree hollows. Usage of the different cover types also varied 2 between seasons (χ 9 = 51.26, p < 0.001), with the proportion of locations in hard cover types highest in winter (98%, compared to 77% in spring, 67% in summer and 70% in autumn).

Radio-tracked pythons spent much of their time in arboreal locations (67% of all observations). The use 2 of terrestrial and arboreal habitats differed between both sex/age class (χ 2 = 7.81, p = 0.02) and 2 landscape types (χ 1 = 92.04, p < 0.001). Females and juveniles were more arboreal than males (both around the homestead and in the park) and Homestead pythons were more arboreal than park pythons. This increased tree usage by Homestead pythons reflects the fact that 76% of their ‘arboreal’ locations were inside roofs; whereas, all ‘arboreal’ locations for Park pythons were in trees. There was also a 2 significant seasonal difference in the proportion of arboreal and terrestrial locations (χ 3 = 73.44, p < 0.001). Arboreal relocations were more frequent in winter (91%, compared to 81% in autumn, 61% in spring and 46% in summer).

Patterns in Macro- and Microhabitat use

2 Macrohabitat use differed significantly between landscape types (χ 3 = 485.05, p < 0.001). Around the homestead, radio-tracked animals were located predominately in human-modified habitats (88% of locations); whilst in the park radio-tracked animals were rarely observed outside of the woodland (99% 2 of locations). Macrohabitat use differed significantly between the sex/age classes (χ 6 = 44.75, p < 0.001). Around the homestead, males and females used human-modified habitats in relatively equal proportions (92% and 91% of locations, respectively), but more than juveniles (74% of locations) who used woodland habitats (26% of locations) more than males and females did (6% and 9% respectively). Males used shrubland habitats (2% of locations), but females and juveniles did not. No Homestead animals were located in open habitats despite the extensive availability of these habitats in that area. In

the park, males, females and juveniles used woodland habitats in relatively equal proportions (98%, 99% and 98% of locations, respectively), but females were occasionally found in shrubland habitat (1% of locations). Males and juveniles were infrequently encountered in open habitats (2% and 2% of locations, respectively), and all such observations were when these animals were observed crossing patches of open habitat inside of the woodland (i.e., clearings) or roads. Macrohabitat use did not differ 2 significantly between seasons (χ 9 = 15.99, p = 0.07). 2 Microhabitat use differed significantly between locations (χ 6 = 309.28, p < 0.001). The primary differences were that Homestead pythons spent more of their time inside buildings (68% vs. 0%), and hence were less often observed in thickets (6% vs. 22%), logs (5% vs. 20%) and trees (18% vs. 49%). Of the relocations where animals inhabited buildings, 88% of these were in the roof cavity and 13% were from underneath buildings. The use of buildings by Homestead pythons differed between the sex/age 2 classes (χ 2 = 29.98, p < 0.001). Homestead males tended to use buildings more often (36% of all relocations) than females (25%) and juveniles (8%). Even though males used buildings more than females and juveniles, they were less arboreal because they spent more time underneath buildings (23% of all ‘building’ locations) than females (1%) and juveniles (0%).

2 Microhabitat use also differed significantly between sex/age classes (χ 18 = 130.28, p < 0.001). Juveniles spent more time in trees (59% of all locations) than females (36%) and males (9%). All sex/age classes were found in underground microhabitats; although, in these instances males and females were located in rabbit burrows, whereas juveniles were located inside of cracks in the dry creek bed. Males were observed more often in thickets (17% of all relocations) than females (11%) and juveniles (5%), and more often in logs (15% of all relocations) than females (8%) and juveniles (10%). There were 2 significant seasonal differences in microhabitat use (χ 18 = 130.28, p < 0.001). These differences tended to reflect the shift from terrestrial to arboreal microhabitat types as the weather got colder. Around the homestead, this shift was from trees and terrestrial locations to roof cavities, whilst in the park, the shift was from logs and thickets to tree hollows.

Microhabitat Selection We sampled habitat characteristics at 72 random locations, 24 locations for males, 25 locations for females, and 23 locations for juveniles. Pythons often remained in the same location for many days, and frequently returned to previously used locations. Because of this, and the fact that pythons spent considerable amounts of time in locations that could not be captured adequately with our sampling scheme (see methods), the number of characterised locations was much less than the actual number of relocations. This also meant that no comparisons could be made between Park and Homestead animals. In fact, 69% of the sample was comprised of pythons from the park, so there is a bias toward habitat selection in more ‘natural’ habitats. The mean scores for each variable for the random locations and the locations of the three groups of pythons are presented in Table 17. The overall MANOVA indicated that the habitat characteristics of the three groups of pythons and the randomly sampled points were significantly different (Wilk’s Λ = 0.26, F 36, 382 = 4.18, p < 0.0001). Distances between group centroids in the discriminant space showed that males, females and juveniles all used habitat non-randomly. Among the three classes of pythons, males, females and juveniles did not differ significantly from one another (Table 18). The discriminant function analysis derived three discriminant functions that summarised multivariate differences among the four types of locations. However, only the first discriminant function accounted for a significant amount of the total variation between group centroids, explaining 86% of the total variation between groups (Table 19). The variable HUNDER was highly correlated with %UNDER (r = 0.72) and so was removed from the analysis. The pooled within-group correlations of habitat variables indicated that %UVEG, %LEAF, LOG and PDBH contributed strongly and positively, whereas %GRND contributed strongly but negatively, to the first function.

Table 17 Means (± SE) of variables used in the analysis of microhabitat selection for the random locations and locations of male, female and juvenile M. spilota at Willandra. Random Male Female Juvenile Variablea (n = 72) (n = 24) (n = 25) (n = 23)

PDBH 7.52 (1.15) 18.22 (2.30) 21.61 (2.80) 35.39 (5.15)

NDBH 23.86 (1.17) 30.33 (2.27) 31.51 (2.00) 29.70 (1.73)

DTREE 14.27 (3.52) 3.24 (0.53) 5.34 (1.28) 2.61 (0.73)

DH20 67.65 (6.38) 17.81 (7.27) 25.28 (3.60) 10.82 (1.11)

HOVER 5.84 (0.34) 9.22 (0.87) 10.01 (0.68) 8.43 0.67)

HUNDER 0.55 (0.05) 1.34 (0.17) 0.93 (0.11) 1.22 (0.11)

%GRND 51.35 (3.86) 12.14 (4.05) 28.40 (5.80) 28.85 (8.23)

%GVEG 25.00 (3.40) 23.21 (6.01) 19.80 (4.72) 26.54 (6.68)

%UVEG 8.27 (1.01) 53.21 (9.43) 39.00 (7.12) 31.92 (7.90)

%LEAF 7.02 (0.72) 31.79 (7.84) 28.60 (4.99) 24.62 (5.44)

LOG 0.70 (0.09) 2.85 (0.34) 2.96 (0.31) 3.20 (0.55)

TEMP 25.60 (0.77) 26.85 (1.44) 23.22 (0.89) 23.79 (0.96)

a For definitions of variables see Table 1.

Table 18 Distances between the four group centroids in the discriminant space and their statistical significance for the analysis of habitat selection by M. spilota. The F statistic with 12 and 129 degrees of freedom is given in parentheses. Group

Group Random Male Female

Distance (F) p Distance (F) p Distance (F) p

Male 2.98 (7.28) < 0.0001

Female 2.76 (9.56) < 0.0001 1.38 (1.27) 0.25

Juvenile 2.60 (5.20) < 0.0001 1.81 (1.63) 0.10 1.10 (0.77) 0.68

Table 19 Summary statistics for the three discriminant functions and their pooled within-groups correlations (r) with the discriminating variables used in the analysis of habitat selection by M. spilota. Statistic Function 1 Function 2 Function 3

Eigenvalue 1.87 0.22 0.08

F 36, 382 = 4.18 F 22, 260 = 1.24 F 10, 131 = 0.67

p <0.0001 p = 0.22 p = 0.67

% of variance explained 85.9 10.3 3.8

r, PDBHa 0.38 -0.61 -0.29

r, NDBH 0.35 -0.07 0.10

r, DTREE -0.22 0.06 0.19

r, DH20 -0.35 0.12 0.33

r, HOVER 0.36 0.07 0.09

r, %GRNDa -0.40 -0.35 0.29

r, %GVEG -0.01 0.14 -0.40

r, %UVEGa 0.54 0.47 -0.16

r, %LEAFa 0.51 0.10 0.02

r, LOGa 0.49 -0.22 0.36

r, TEMP -0.09 0.43 -0.31

a Denotes five most significant values.

This function can be interpreted as a gradient from sites with smaller trees, less hollow logs, understorey vegetation and leaf litter, and more bare ground to sites with larger trees, more hollow logs, understorey vegetation and leaf litter, and as such, less bare ground. Separation of the random group from the three python groups along the first discriminant function reflects the fact that the available habitat (often open shrubland) offered little in the way of cover (more bare ground, less understory vegetation and leaf litter/fallen debris, with fewer hollow logs) and had fewer larger trees (Figure 36).

Patterns in Tree Use

Sixty-two trees in total were used by the 17 telemetered M. spilota as retreat sites throughout the study (Table 20). The mean number of trees used per individual was 3.6 and ranged from 1 to 10. There was no significant difference in the number of trees used by Homestead and Park pythons or between sex/age classes (two-factor ANOVA, location: F 1, 16 = 0.82, p = 0.38; sex: F 2, 16 = 2.81, p = 0.10). In addition, there was no significant difference in the mean minimum or maximum residence times between Park and Homestead snakes or sex/age classes (min. residence time: location: F 1, 16 = 3.03, p = 0.11; sex: F 2, 16 = 0.12, p = 0.86; max. residence time: location: F 1, 16 = 1.87, p = 0.2; sex: F 2, 16 = 1.57, p = 0.25). Seven pythons (41% of all radio-tracked animals) returned to trees that they had previously occupied, and of

these, three animals returned to the same tree on three different occasions. Only four pythons used trees that were used by other radio-tracked pythons. Two males from around the homestead used the same sugar gum (Eucalyptus cladocalyx), but at different times of the year.

1.5

1.0 Male

0.5

Random Non-gravid female 0.0

-0.5 Juvenile

-1.0

Second discriminant function discriminant Second -1.5 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 Smaller trees Larger trees Less understorey vegetation and leaf litter More understorey vegetation and leaf litter Fewer hollow logs More hollow logs More bare ground Less bare ground First discriminant function

Figure 34 Positions (± SE) of the group centroids of random locations and locations of male, non- gravid female and juvenile M. spilota on the two discriminant axes in the analysis of habitat selection.

Table 20 Summary of tree use by radio-tracked M. spilota. Homestead Woodland

Male Female Juvenile Male Female Juvenile

Number of pythons

Radio-tracked 3 3 2 3 3 3

Using tree hollows 3 3 2 3 3 3

Returning to previously used trees 1 2 0 0 2 2

Days in same tree

Minimum residence timea

Mean (± SD) 1.2 (0.4) 1.9 (0.8) 5.5 (2.8) 7.0 (6.6) 6.5 (3.1) 4.8 (4.6)

Range 1-2 1-3 1-21 1-14 2-21 1-21

Maximum residence timeb

Mean (± SD) 2.1 (1.2) 6.1 (5.2) 12.9 (12.6) 7.0 (6.6) 18.7 (11.7) 14.2 (13.8)

Range 1-5 2-20 2-75 1-14 2-75 1-75

Number of trees used

Total trees 7 15 9 3 15 13

Mean (± SD) 2 (1) 5 (4) 5 (1) 1 (0) 5 (5) 4 (4)

Range 1-3 2-10 4-9 0 1-10 1-9

a Minimum residence time refers to the number of days that snakes were actually observed in the same tree. b Maximum residence time is the number of days that snakes were potentially in the same tree.

While in the park, a male and female were observed together inside the same hollow of a black box (E. largiflorens) at the same time for several days in late spring.

Characteristics of Trees

Radio-tracked M. spilota did not use trees randomly throughout the study; instead, they were highly selective with respect to several characteristics. One surprising finding however was that they did not appear to prefer dead trees (unlike other semi-arboreal species of snakes (see Webb and Shine, 1997b). Dead standing trees constitute 13% of the available trees at Willandra (both around the homestead and in the park), but only 12% of the trees used by M. spilota. Dead trees at Willandra did not always contain hollows, thus, we classified them accordingly: ‘stags’ were dead standing trees containing hollows (see Webb and Shine, 1997b) and ‘dead’ trees were dead standing trees that did not contain hollows. Around

the homestead, stags constituted < 1% of the available trees, but 3% of the trees used by pythons, while dead trees comprised 4% of available trees and 3% of trees used by pythons (Table 21). In the park, stags constituted 5% of available trees but 17% of the trees used by pythons, while dead trees comprised 12% of the available trees but 0% of the trees used by pythons. The pythons’ use of dead rather than live trees was not significantly different from that expected under the null hypothesis that they use dead trees in 2 2 proportion to their availability, either around the homestead (χ 1 = 0.07, p = 0.93) or in the park (χ 1 = 0.21, p = 0.65). That is, pythons from around the homestead and in the park did not prefer dead rather than live trees.

Six species of trees were used by M. spilota at Willandra. Of live trees, Homestead pythons used peppercorn trees (Schinus areira) and black box (E. largiflorens) most often, while Park pythons used E. largiflorens most frequently. Pythons were highly selective in their usage of tree species, both around 2 2 the homestead (χ 8 = 29.90, p < 0.001) and in the park (χ 1 = 37.11, p < 0.001). Around the homestead, pythons used E. largiflorens and S. areira more often than would be expected by chance, based on the availability of these species. In the park, pythons used E. largiflorens much more often than would be expected by chance, based on the availability of this species. Morelia spilota was rarely observed using A. stenophylla, the most common tree species at Willandra.

Most of the trees at Willandra were considerably smaller than trees used by pythons throughout the study. This pattern was not only evident around the homestead and in the park, but also between dead and live trees. We analysed this data using 3-way ANOVA’s. The factors were location (homestead vs. park), tree status (live vs. dead) and usage (whether or not radio-tracked pythons used the tree), and the dependent variables were measures of the size of the trees (i.e., height or diameter at breast height [dbh]). Trees used by pythons were significantly larger than random trees (height: F 1, 983 = 16.82, p < 0.001; dbh: F 1, 983 = 57.14, p < 0.001) and live trees were significantly taller than dead trees (F 1, 983 = 5.05, p = 0.03), but not necessarily larger in diameter than dead trees (F 1, 983 = 1.10, p = 0.29).

Table 11 Abundance and availability of trees compared with usage of trees by M. spilota. Homestead Parka

Common name Species Available Used Available Used

N % N % n N % N % n

River cooba Acacia stenophylla 85 30.2 0 0 0 338 51.7 2 5.7 2

Black box Eucalyptus largiflorens 50 17.8 5 16.1 5 181 27.7 27 77.1 10

Pepper corn Schinus areira 73 26.0 18 58.1 6 0 0 0 0 0

Sugar gum Eucalyptus cladocalyx 6 2.1 3 9.7 3 0 0 0 0 0

Olive Olea europaea 9 3.2 2 6.5 2 0 0 0 0 0

76 Golden ash Fraxinus excelsior aurea 2 0.7 0 0 0 0 0 0 0 0

Belah Casuarina cristata 2 0.7 0 0 0 0 0 0 0 0

Date palm Phoenix dactylifera 1 0.4 1 3.2 1 0 0 0 0 0

Fruit tree Plum, apricot, etc. 18 6.4 0 0 0 0 0 0 0 0

Unidentified 22 7.8 0 0 0 0 0 0 0 0

Stag Dead standing trees containing hollows 1 0.4 1 3.2 1 34 5.2 6 17.1 5

Dead Dead standing trees without hollows 12 4.3 1 3.2 1 84 12.8 0 0.0 0

Total 281 31 637 35

N number of each species recorded; % percentage occurrence; n number of individual snakes using each tree species a Data includes trees that non-radio-tracked M. spilota were observed in (n = 5 individuals) – see methods.

14 (a) used 100 (b) random 12 80 10

8 60

6 40 DBH (cm) Height (m) 4 20 2

0 0

8 (c) 3 (d)

77 6 2

4 1 2 holes stem of No. No. of hollow branches hollow of No. 0 0

dead alive dead alive dead alive dead alive Homestead Park Homestead Park

Figure 35 Characteristics of trees selected by radio-tracked M. spilota compared with available trees in the same area. The histograms show mean values (± 1 SE) for dead standing trees and live trees, in both Willandra National Park and around Willandra Homestead.

There was no significant difference in the height of trees used by homestead and park pythons (F 1, 983 = 1.69, p = 0.19), but a significant difference in dbh (F 1, 983 = 7.06, p = 0.008) was detected. This difference in dbh between homestead and park trees reflects the tendency for E. cladocalyx and S. areira (common around the homestead) to be larger in dbh than E. largiflorens and A. stenophylla (2- way ANOVA, location and species as factors: p < 0.05 for all Holm-Sidak multiple pairwise comparisons). For the dependant variable dbh, there was a significant interaction between tree status and usage (F 1, 983 = 6.20, p = 0.013). Around the homestead, live trees were larger than dead trees and live trees around the homestead were larger than live trees in the park. Again, these differences reflect the fact that live trees around the homestead were usually S. areira and hence larger than dead trees in that area (Holm-Sidak multiple comparisons: p < 0.05).

Pythons occupied trees with high numbers of both stem and branch hollows. A three-factor ANOVA (factors: usage, status and location – see above) showed that M. spilota preferred trees with more branch hollows (F 1, 983 = 94.64, p < 0.001) and that there was no difference in the number of branch hollows between live and dead trees (F 1, 983 = 0.67, p = 0.41). In addition, there was a significant difference in the number of branch hollows between trees in human-modified and woodland environments (F 1, 983 = 9.45, p = 0.002) with trees in the park containing more branch hollows (on average) than trees around the homestead. There were significant interactions between location and usage (F 1, 983 = 7.76, p = 0.005), location and status (F 1, 983 = 14.80, p < 0.001) and location, usage and status (F 1, 983 = 4.92, p = 0.027).

The interactions reflect the fact that both used and available dead trees in the park contained more hollows compared to those around the homestead. Dead used trees around the homestead had more hollows than available dead trees, but this reflects that fact that there were only two dead trees around the homestead that were used by pythons, and these both happened to contain hollows. Live trees from both the park and around the homestead had more hollows than random trees, while random trees in the park contained more hollows than random trees around the homestead. The latter reflects the fact that many random trees around the homestead were usually fruit trees or exotic species (unidentified) that did not contain hollows. The only two species of tree around the homestead that did not contain hollows, but were used by pythons, were the olive trees (Olea europaea) and one python used the single date palm (Phoenix dactylifera). Both of these species produce extensive amounts of fruit which birds feed on, and have dense foliage that provides excellent sites from which to ambush prey. Live trees in the park that were used by pythons contained more hollows than available trees and finally, live trees around the homestead that were used by pythons had more hollows than dead trees.

Stem hollows followed a similar pattern as that described above for branch hollows in that trees used by pythons. These contained more stem hollows (F 1, 983 = 87.80, p < 0.001). There was no difference in the number of stem hollows between live and dead trees (F 1, 983 = 7.32, p = 0.55). However, there was no difference in the number of stem hollows between trees in human-modified and those in woodland environments (F 1, 983 = 0.31, p = 0.58). This time there were significant interactions between location and status (F 1, 983 = 6.69, p = 0.01) and usage and status (F 1, 983 = 5.39, p = 0.02).

Diet and Relative Prey Density

Thirty-four faecal samples were obtained from radio-tracked M. spilota and 12 direct observations of foraging encounters were made. Prey species composition of python diets differed significantly 2 between landscape types (χ 2 = 13.45, p = 0.001) (Table 22). Among the 28 prey items recorded for homestead pythons, 93% were mammals (of which rabbits comprised 54% and mice 43%) and 7% were birds. While among the 18 prey items recorded for park pythons, only 44% were mammals (of which rabbits comprised 88%), 50% were birds and 6% were reptiles. Mobile prey items comprised the majority of all prey (85%), but the presence of stationary items like nestlings and eggs (15%) suggests that radio-tracked pythons actively searched for prey.

Table 22 Prey items of M. spilota from Willandra Homestead and Willandra National Park, as determined from faecal examination and direct observation. Number of records Prey item Homestead Park

Mammals

Rabbit (Oryctolagus cuniculus) 14m, f 7m, f

Mouse (Mus musculus) 12m, f, J 1

Birds

Welcome swallow (Hirundo neoxena) 0 1f

Willie wagtail (Rhipidura leucophrys) 2f 0

Honeyeater (Lichenostomus sp.) 0 1j

Unidentified nestlings 0 5j

Egg shells 0 2f

Reptiles

Shingleback (Tiliqua rugosa) 0 1

N 28 18 m = prey items recorded from males; f = prey items recorded from females; j = prey items recorded from juveniles.

The majority of the prey items recorded (74%) were not native species; instead, they were exotic species like rabbits (Oryctolagus cuniculus) and mice (Mus musculus). A higher numbers of mice were observed around the homestead (n = 88) than in the park (n = 0). Interestingly, mice only appeared in one faecal sample from all park pythons. Mice comprised 39% of the prey records for Homestead pythons. Mouse abundance fluctuated seasonally around the homestead and was higher in spring and summer than it was in autumn and winter.

0.6

0.5 Park Homestead 0.4

0.3

trap night trap 0.2

0.1 No. of individuals per per individuals of No.

0 spring summer autumn winter Season

Figure 36 Relative abundance (number of individuals per trap night) of small mammals in Willandra National Park and around Willandra Homestead.

Discussion

Vertebrate Surveys

Low Detectability for Vertebrate Surveys There were no detectable differences between the number and abundance of vertebrate species utilising revegetated and non-revegetated sites. Differences may not have been detected due to low sample sizes. In addition, low detectability may have resulted from: 1) drought conditions, 2) the sampling methods used, and/or 3) that many species were not present. Animals often restrict their movements during dry, drought conditions when resources are scares in order to conserve water and energy. Moreover, many species seek shelter deep within the cracks in the earth where there is greater moisture present. Cracks also provide refuges from high air and soil surface temperatures. Restricted movements and cryptic behaviour make some species less visible. Animals may have been less detectable with the sampling methods used in this study compared to others. As shown, VES’ had a greater success rate in sampling vertebrates compared to Artificial Cover Surveys and PVC tube traps. It is recommended that during periods of drought VES’ be employed as the primary survey method. The results may also reflect that few vertebrate species were present at the study locations, which would indicate a depauperate fauna most likely resulting from significant land-use changes. It is recommended that reintroduction programs for native fauna be considered in association with revegetation activities in order to increase the speed at which different species can recolonise restored habitats.

Differences in Bird Species Between Sites Several and similar common birds species were observed in revegetated and non-revegetated sites. This may have resulted because both sites had similar vegetation structures despite one having plantings. Many of the plantings were unsuccessful and much of the existing vegetation died during the monitoring due to the drought conditions. When water is scares, it may require a greater period of time for plantings to become established and for birds to utilise restored habitats. In the current study, the data suggests that >5 years may be required for this to occur. However, there is some indication that the species that were observed only in the revegetated sites were those that require some understorey, or thick bush.

Frog Study

Species Abundance

More than 6000 L. tasmaniensis and L. fletcheri were captured during this study. Based on the production rate of approximately three juvenile frogs per m2, it can be estimated that nearly 40 million frogs hatched in rice bays between January 1 and May 30, 2007. This estimate is based on the figure that 12,000 ha of rice were sown this year. This figure however is less than 80% of the amount sown in non-drought years. In non-drought years about 120,000 ha of rice is sown; therefore, the number of frogs produced in bays may be as high as several billion in some years. The estimated numbers of frogs produced during this study supports the original findings of Doody et al. (2006), who estimated that nearly 500 million frogs were produced in the Riverina in 2003.

The number of captures in pitfall traps varied significantly between months. An increase in captures during certain periods may be correlated with increased rainfall as it has been reported that frogs usually increase their movements during precipitation (Lamoureux et al., 2002). The increased

capture rate for juveniles in February correlates with the timing of metamorphosis from the tadpole to juvenile stage. This, in association with a rainfall event, likely triggered the dispersal of juvenile frogs away from the rice bays and into the surrounding habitats. Limnodynastes tasmaniensis tadpoles may take up to 250 days to metamorphose. In contrast, L. fletcheri usual develop in a significantly shorter time period (approximately 60 days) (Barker et al., 1996; Lane and Mahoney, 2002). The first egg masses observed in the rice bays were in mid-November when the bays were filled; therefore, the development time for tadpoles from these masses was between 60 and 100 days. The low number of adults captured at Old Coree is likely due to the rice bays being encircled by silt fencing in mid- December. This fencing prevented frogs from entering the bays after this time.

The number of frogs observed in the rice bays at Old Coree and Leeton suggests that rice bays provide ideal habitat for frogs during various life stages. For example, rice bays may be considered as semi- permanent wetlands that supply water even after ephemeral wetlands have dried for the year (Richardson et al., 2001). This constant supply of water provides additional breeding opportunities, and a stable environment for developing tadpoles and juvenile frogs. Usually breeding in L. tasmaniensis and L. fletcheri is usually triggered by rainfall and females subsequently lay eggs in puddles (Barker et al., 1996). However, in these situations desiccation is a significant source of mortality for tadpoles (Lane and Mahony, 2002). The constant supply of water in rice bays during the growing season (approximately seven months) may reduce the risk of desiccation, thus ensuring that a very high proportion of tadpoles complete development.

Another likely reason for the production of very high numbers of frogs in rice bays is the relative lack of vertebrate predators. Fish, dragonfly larvae, water beetles and birds are the main predators of tadpoles (Lane and Mahoney, 2002). The mosquito fish (Gambusia holbrooki) is a major predator of tadpoles; they attack their tail fins, thereby immobilising them (Lane and Mahoney, 2001). Richardson et al. (2001) reported that fish were not present in some rice bays in the Murrumbidgee Irrigation District. Therefore, the risk of predation by fish may not be a significant in this ecosystem. Egrets (Ardea alba and A. intermedia) have been reported to consume tadpoles from rice bays in the Riverina (Richardson et al., 2001); although, the impact of water birds on tadpole survival is not clear.

Rice plants and other aquatic vegetation provide tadpoles with important microhabitats. Warkentin (1992) observed that aquatic vegetation offered food and sheltering sites to pond dwelling tadpoles, and that the use of these microhabitats was influenced by temperature, oxygen availability and predation risk. Aquatic vegetation is also important for the foamy egg masses that Limnodynastes produce. Females attach their egg masses to vegetation in order to receive protection and shelter until the eggs hatch (Lane and Mahoney, 2002). Although microhabitat use was not examined in the present study, tadpoles were observed sheltering among the rice plants during the day and swimming in the open water in the tofos in the evening.

Dietary Analysis

Both L. tasmaniensis and L. fletcheri consumed a wide variety of invertebrate species, which suggests that they are generalist predators. The diet of a generalist predator is strongly influenced by the availability of prey in the environment (Jenssen and Klimstra, 1966; Parker and Goldstein, 2004). Ants (Formicidae), beetles (Coleoptera) and true bugs (Hemiptera) were consumed most frequently by L. tasmaniensis and L. fletcheri. These invertebrates are important dietary components of many species of frogs in natural and agricultural ecosystems (Hirai and Matsui, 1999, 2001; Parker and Goldstein, 2004; Attademo et al., 2005; Siqueria et al., 2006).

The total volumetric consumption of L. tasmaniensis increased significantly in May. The total volume of prey consumed by L. fletcheri in May was not significantly greater than in other months but it represented a much larger range. These patterns in food consumption may be associated with preparations for hibernation. Parker and Goldstein (2004) observed a significant increase in the

number of prey items consumed by the Leopard Frog (Rana pipiens) in autumn when compared to other seasons.

Body Size Versus Prey Size Body size is expected to correlate with the size of the prey items consumed by an individual (Cogalniceanu et al., 2000). This general pattern was evident for L. tasmaniensis and L. fletcheri. The relationship between frog body size and prey size may be explained by the Optimal Foraging Theory, which suggests that it is beneficial for large-bodied frogs to consume larger prey items as they require more energy and nutrition than smaller individuals (Lima, 1998; Siqueria et al., 2006).

The relationship between body size and prey size may also be influenced by foraging strategy (Simon and Toft, 1991). Large prey items are typically captured by frogs that employ a sit-and-wait foraging strategy (Toft, 1981). On the other hand, small prey items are typically captured by species that implement an active foraging strategy (Toft, 1981). Therefore, variation in the size of prey may be expected to vary with stages of growth and size of an individual, which may result in shifts in foraging strategies (Lima, 1998). The similarity between the sizes of prey items in the diet in relation to body size of both species suggests that there is no difference in foraging mechanisms.

Ontogenetic Dietary Shifts As body size or mouth width increases, the size of the prey that an individual can consume might also increase. This may result in a dietary shift that is associated with growth. There is no evidence of ontogenetic shifts in diet between juvenile and adult L. tasmaniensis or L. fletcheri. However, shifts have been found in other species (Clarke, 1974; Christian, 1982; Hirai, 2002).

The lack of evidence for ontogenetic dietary shifts in L. tasmaniensis and L. fletcheri could have resulted from capturing many more juveniles than adults throughout the study. The possibility of this is supported by the significant, but relatively weak, regression analyses performed on the volume of the smallest and largest prey items consumed by the frogs. A lack of ontogenetic dietary shifts could also be due to a lack of abundance of large prey items occurring in the environment. A lack of large prey items would force large individuals to consume smaller prey items (Labanick, 1976).

Sex Related Differences in Diet A difference in diet may be expected in species that are sexually dimorphic. For example, in the Horned Leaf Frog (Proceratophrys appendiculata) larger bodied females consume larger prey items than small males (Boquimpani-Freitas et al., 2002). Male and female L. tasmaniensis and L. fletcheri are similar sized; therefore, it is not surprising that dietary differences were not detected between the sexes.

Dietary Differences Between Species

Sympatric species like Limnodynastes tasmaniensis and L. fletcheri are thought to be able to coexist because they utilise resources differently. Resources that are commonly partitioned include food and microhabitat (Toft, 1985). There was no significant difference in the diet of L. tasmaniensis and L. fletcheri: they consumed the same prey orders in similar proportions. Such extensive overlap between congeners has been reported in other species (Inger and Greenberg, 1966). The lack of difference in diet between L. tasmaniensis and L. fletcheri may be explained by the species similarities. In general, they occupy the same geographic distribution, utilise the same habitats and microhabitats, breed at the same time each year and appear to be morphologically similar (Duellman and Treub, 1985; Barker et al., 1996). This suggests that they will come into contact with the same invertebrate species and are therefore likely to have a similar diet.

Seasonal Availability of Prey There was a significant correlation between the availability of terrestrial invertebrates and the types of prey identified in the stomach contents of L. tasmaniensis and L. fletcheri, which indicates that their diet is influenced by the abundance of invertebrates in the environment. The significant correlations between the invertebrates captured in pitfall traps, indicates that both species consume terrestrial species predominately. This is expected for generalist predators (Labanick, 1976). The lack of correlation between the availability of aerial invertebrates and prey items consumed suggests that aerial invertebrates are not an important component of their diets. In rice fields in Japan, the banks between the rice bays were found to be important foraging grounds for the endangered Daruma Pond Frog (Rana porosa brevipoda) that also feeds mainly terrestrial invertebrates (Hirai and Matsui, 2001). This suggests that the banks of the bays offer important foraging habitat for these species.

The effectiveness of pitfall traps to sample the abundance of invertebrates in an ecosystem is dependant on the activity and abundance of the invertebrates (Cogalniceanu et al., 2004). For this reason pitfall trap sampling is biased towards predators of invertebrates, which may be regarded as sit- and-wait predators (Cogalniceanu et al., 2004). The positive correlation between the invertebrate species collected in pitfall traps and the species identified in the stomach contents suggests that L. tasmaniensis and L. fletcheri employ a sit-and-wait foraging method.

Both species consumed aquatic invertebrates like other frog species (Stewart and Sandison, 1972). However, L. tasmaniensis and L. fletcheri consumed aquatic invertebrates in very small proportions compared to terrestrial invertebrates. The aquatic sweep sampling indicated that there is a large quantity of potential aquatic prey available that was not consumed by either species. The consumption of aquatic invertebrates by L. tasmaniensis and L. fletcheri could become more important with changes in prey availability during different seasons as has been showed for other frog species (Jenssen and Klimstra, 1966).

Rice Pest Consumption

Limnodynastes tasmaniensis and L. fletcheri consumed rice pest species, which comprised approximately 5% of their diet. Likewise, frog species in Japan, South America and the United States are also know to consume pests of various agricultural crops (Brown, 1974; Hirai and Matsui, 1999; Attademo et al., 2005).

It was estimated that over four billion invertebrates were consumed during this study; therefore, approximately 220 million rice pests were consumed. These estimates are based on L. tasmaniensis feeding once a day and consuming an average of eight invertebrates during the feeding period. There is a considerable amount of variation in the number of invertebrates consumed by different species however. For instance, Fowler’s Toad (Bufo woodhousie fowleri), which is about the same size as the study species, has been reported to consume an average of 52 prey items per day, while the Green Tree Frog consumes an average of only 4.5 items per day (Brown, 1974).

There is a possibility that the number of pest species consumed by L. tasmaniensis and L. fletcheri could be higher than reported here. The sole focus in this study was rice pests and not other crop pests. Most rice farms will use one of four cropping systems in which winter or summer crops, or pasture are grown in rotation with rice. Therefore, the potential number of pest species available for consumption in the rice ecosystem is likely to be greater than on farms that only grow rice.

Carpet Python Study This study is one of few to provide detailed information on the spatial ecology of snakes inhabiting semi-arid regions of Australia and highlights several patterns in the spatial ecology of M. spilota from semi-arid NSW. Some of these patterns have implications for the conservation of this species. Firstly, there was considerable individual variation in all estimates of activity, movement and space use, both within sex/age classes and within human-modified and woodland environments. Secondly, despite this variation, common patterns did exist regarding seasonality of activity and movement. Thirdly, home ranges of M. spilota showed both spatial and temporal overlap. While broad patterns of activity and movement were relatively similar between landscapes with high and low human-modification, snakes inhabiting human-modified environments were generally more sedentary than those inhabiting woodland environments.

This study also described patterns of resource and macro- and microhabitat use in terms of sex, age, season and landscape type and examined structural characteristics associated with microhabitat selection. There were clear differences in habitat and resource use between radio-tagged animals inhabiting natural, woodland environments and those inhabiting human-modified environments. There were also broad similarities in the way that they utilised these resources. These broad patterns in macro- and microhabitat use were similar to those reported for M. spilota in other areas of Australia. However, M. spilota in western NSW were much more arboreal than M. spilota in the northeast and southeastern parts of the state. They also used tree hollows much more extensively. Furthermore, the examination of specific fine scale habitat characteristics provided the first quantitative data on microhabitat selection for this species.

Activity Patterns

Radio-tagged M. spilota were highly sedentary and spent long periods of time inactive in retreat sites throughout this study. The high proportion of time spent in such situations is most likely important in snakes that ambush prey (Slip and Shine, 1988c; Shine and Fitzgerald, 1996). Others have observed similar patterns in activity for species that utilise such foraging strategies (e.g., Hoplocephalus bungaroides, Webb and Shine 1997a; Gloydius shedaoensis, Shine et al., 2003) and for other populations of M. spilota (e.g., Slip and Shine, 1988b; Shine and Fitzgerald, 1996). This foraging strategy allows pythons to meet their modest energy requirements (Bedford and Christian, 1998), while remaining concealed from both potential prey and predators.

Male M. spilota were observed moving much more than both females and juveniles. This pattern was heavily influenced by the frequent movements of males in late winter and early spring. Both Slip and Shine (1988b) and Shine and Fitzgerald (1996) reported similar behaviour for male M. spilota in spring in eastern Australia. This increase in movement by males at these times of the year most likely reflects their search for reproductive females. All but one of the radio-tagged animals in the present study were inactive over the cooler months (mid April to late August) and were observed sheltering in either tree hollows or roof cavities.

Activity patterns also differed between landscapes. Pythons radio-tracked in woodland environments were more likely to be observed moving than those around the homestead and this may reflect the fact that homestead environments provided thermally suitable shelter sites (i.e., roofs) that contained a steady supply of prey. Although no previous studies have directly compared activity patterns of pythons in natural and heavily modified environments, other authors have noted the sedentary nature of pythons that inhabit such areas (e.g., Shine and Fitzgerald, 1996; Bedford, 2003; Heard et al., 2004). A sedentary life-style in snakes is often linked to a plentiful food supply (Gregory et al., 1987).

The increase in nocturnal activity during the warmer months may be to avoid the high daytime temperatures experienced at the study location. These temperatures (regularly > 35°C and often >

42°C) were probably high enough to suppress all diurnal activity. Similar shifts in activity have been reported in other snakes (e.g., Agkistrodon contortrix, Sanders and Jacob, 1981) and in other populations of M. spilota (Slip and Shine, 1988b; Shine and Fitzgerald, 1996), but not to the extent observed throughout this study.

The reproductive radio-tagged female was completely sedentary throughout her brooding period. Other authors have observed similar patterns in reproductive female pythons (e.g., Pearson et al., 2005), whereas Slip and Shine (1988d) reported that female M. spilota in south eastern Australia left their eggs on clear mornings to bask briefly before returning to them to continue incubation. Ambient temperatures in southeastern NSW are less extreme than those experienced at Willandra during the period when females incubate their eggs. Female M. spilota in southeastern NSW construct mounds of loose vegetation where they deposit and subsequently brood their eggs (Slip and Shine, 1988d). At Willandra, females chose well-insulated nest sites (tree hollows with a mean environmental temperature of 29.8°C (range = 16.9 – 42.2°C), or rabbit warren at 29.9°C (range = 20.3 – 40.9°C), and an old farm incinerator at 30.1°C (range = 18.6 – 41.6°C)). Differences in ambient air temperatures may account for the differences in behaviour between females from southeastern and inland Australia.

Movement Patterns Larger animals did not necessarily move further when they chose to move, nor did they move more frequently. However, body size was correlated with the total distance moved. This contradictory finding reflects the fact that larger animals were tracked for longer periods than smaller individuals (owing to the battery life of transmitters), and that the total distance moved was correlated with the number of relocations. Few studies have examined possible correlations between body size and movement patterns (Macartney et al., 1988); however, Clark (1974) reported that mean distance moved per day was correlated with body size in Thamnophis proximus, indicating that adults of that species were more vagile than juvenile snakes.

The radio-tagged female, which brooded eggs over the summer, continued to move and feed throughout the cooler months (late April through July). Incubating eggs is metabolically costly (Harlow and Grigg, 1984; Slip and Shine, 1988d; Bedford, 2003). Females may lose between 37% and 48% of their pre-oviposition mass, of which a third is attributed to metabolic costs associated with shivering thermogenesis (Slip and Shine, 1988a). The reproductive female in this study lost 49% of her pre-oviposition mass (Corey, 2007) and possibly lacked the body reserves to hibernate successfully. This individual was also the only animal observed moving, basking and feeding throughout the winter. Increased activity in snakes is thought to increase their vulnerability to anthropogenic mortality (Bonnet et al., 1999). This increased activity throughout the winter likely contributed to her death when she was later killed by a motor vehicle (Corey, 2007).

Radio-tagged pythons exhibited distinct seasonal patterns in their movement frequencies. Collectively, M. spilota moved more frequently in summer than they did in spring and not surprisingly, more frequently in both spring and summer than they did in winter. Although the two studies of M. spilota from eastern Australia (Slip and Shine, 1988b; Shine and Fitzgerald, 1996) did not specifically examine propensity to move, they both noted that movements decreased noticeably during the cooler months, more so in the southern most populations (Slip and Shine, 1988b). Distinct seasonal trends in movements have been observed for many species of snakes inhabiting both temperate (e.g., Shine, 1987; Brito, 2003; Heard et al., 2004) and arid regions (e.g., Secor, 1994; Beck, 1995). Temperature is considered one of the most important factors driving these patterns of movement (Huey, 1982). Ambient air temperatures at Willandra were warmer than those experienced in southeastern NSW (Corey, 2007; Bureau of Meteorology, 1901-2006 climate averages). Because all physiological processes are temperature dependent (McCue, 2004), these higher temperatures may lead to higher metabolic rates (Huey, 1982; 1991), which in turn requires more energy input. Thus, pythons at Willandra may need to move to new ambush sites more often in order to obtain food to

meet the metabolic costs associated with living in an environment with higher ambient air temperatures.

Pythons inhabiting woodland and human-modified environments differed in their movement frequencies. During the peak activity times (spring and summer), park snakes moved more often than their homestead counterparts. This may be due to relative prey abundance, which was higher in the human-modified environment than in the park. Macartney et al. (1988) proposed that snake movements should be related to spatio-temporal variation in abundance of resources, which often vary among environments. Thus, movement frequency by M. spilota at Willandra was most likely influenced by the relative abundance of suitable prey and its seasonal availability.

Despite the fact that males tended to move further than both females and juveniles, there were no overall sex or age related effects on either mean or total distances moved. Neither Slip and Shine (1988b), or Shine and Fitzgerald (1996), examined mean or total distances moved in their respective studies, but Pearson et al. (2005) reported no obvious sex or age related effects on movements in M. imbricata from southwestern WA. Mean daily displacement distances were much larger in this study than those reported for M. spilota in eastern Australia by Shine and Fitzgerald (1996). Like M. spilota in eastern Australia, pythons at Willandra also exhibited strong seasonal shifts in their distances moved. Radio-tagged M. spilota moved further during spring and summer than they did in autumn and winter.

Home Range Home ranges varied considerably among individuals. Minimum Convex Polygons (100%) estimates were considerably larger for males than those of females and juveniles. Overall, these estimates were considerably smaller than those reported for M. spilota in eastern Australia. The high variability among individuals could be attributed to differences in body size, because 100% MCP estimates were correlated with body size in the present study. Slip and Shine (1988b) also reported a correlation between body size and home range size, but Shine and Fitzgerald (1996) did not examine this relationship for M. spilota in northeastern NSW. Other studies examining the spatial ecology of snakes have also reported correlations between body size and home range size (e.g., Tiebout and Cary, 1987; Webb and Shine, 1997a; Wunderle et al., 2004), while others reported no correlation (e.g., Fitzgerald et al., 2002; Shine et al., 2003; Roth and Greene, 2006). Intuition suggests that foraging mode may influence home range size, with sit-and-wait foragers using smaller areas (e.g., Diffendorfer et al., 2005; Roth, 2005; Moore and Gillingham, 2006); however, many studies have reported very large home ranges for species, which display such foraging modes (e.g., Reinert and Zappalorti, 1988; Beck, 1995).

None of the home range estimates differed significantly between landscape types, despite the fact that pythons inhabiting woodland environments moved more frequently than those inhabiting human- modified environments. Shine and Fitzgerald (1996) also found that mean home ranges were similar in their two study areas, which experience different levels of human modification. All home range estimates were correlated with mean and total distances moved and thus, animals that moved further had larger home ranges. However, animals that moved more frequently did not necessarily have larger home ranges.

The degree of overlap between home ranges of different individuals can give insights into the mechanisms determining space use (Webb and Shine, 1997a). If the presence of one snake in an area makes the site less attractive to another snake, one would expect to see low overlap between individuals in space and/or time. This was not the case however as radio-telemetered M. spilota inhabiting both woodland and human-modified environments showed overlap and sympatric space use. Both Slip and Shine (1988b) and Shine and Fitzgerald (1996) reported similar overlap for M. spilota in their respective study areas. Other authors have reported spatial and temporal overlap in home ranges (e.g., Secor, 1994; Fitzgerald et al., 2002; Diffendorfer et al., 2005), while others have

found evidence to suggest mutually exclusive home ranges (e.g. Webb and Shine, 1997a; Whitaker and Shine, 2003). Nonetheless, the well-defined home ranges of radio-telemetered snakes and their tendency to revisit specific shelter sites (namely tree hollows and buildings) reflects a pattern common among many other species of snake (e.g., Weatherhead and Hoysak, 1989; Durner and Gates, 1993; Webb and Shine, 1997a).

Overall Patterns in Habitat Use Throughout this study all radio-tagged individuals spent much of their time in habitat types that provided full coverage and as a consequence were rarely seen when located. Both Slip and Shine (1988b) and Shine and Fitzgerald (1996) proposed that this inconspicuous behaviour might promote the survival of M. spilota in highly modified environments.

There were significant differences in the frequency of locations in the different cover types between sex and age classes and landscapes. In both human-modified and woodland environments, males were usually located more often in ‘open’ and ‘dense’ cover types than both females and juveniles. Moreover, individuals that inhabited woodland environments generally used ‘dense’ cover types more frequently than pythons, which inhabited human-modified environments. There were also significant seasonal differences in the use of different cover types. Individuals frequented ‘hard’ cover types during the cooler months and ‘dense’ and ‘filtered’ cover types during the warmer parts of the year.

Pythons may use hard cover types more than those that provide filtered, dense or no cover for many reasons. Firstly, habitat types that provide ‘hard’ cover may offer shelter sites with a less variable or somewhat ‘buffered’ thermal regime (Huey, 1991; Webb and Shine, 1997b, 1998b; Heard et al., 2004). In harsh environments, which may experience extremes in temperature like semi-arid Australia, ectothermic species like snakes may select habitat types that allow them to regulate their body temperatures more easily. Secondly, hard cover types offer increased protection from potential predators (Fitzgerald et al., 2002a). Because of their large body size and slow locomotion, pythons may be particularly obvious to potential predators, even more so in open habitat. Two of the radio- tagged individuals were killed when crossing roads with little cover during this study. Shine and Fitzgerald (1996) and others (e.g., Bonnet et al., 1999; Reinert and Rupert, 1999; Whitaker and Shine, 2000) have reported similar cases of mortality.

All of the radio-tagged pythons were highly arboreal. It has been reported that M. spilota spends less time in trees at other locations (Shine and Fitzgerald, 1996; Slip and Shine, 1988; Pearson et al., 2005). However, Bedford (2003) reported that M. spilota in Darwin spent about 90% of the time in arboreal sites. At Willandra, females and juveniles were significantly more arboreal than males. The trend for females and juveniles to use arboreal habitats more frequently than males may be related to subtle differences in their foraging ecology. For instance, the diets of male M. spilota at Willandra were mainly comprised of terrestrial prey items such as rabbits (Oryctolagus cuniculus), while those of juveniles and females were mainly composed of birds and their eggs.

Pythons inhabiting human-modified environments were more arboreal than those inhabiting woodland environments. This finding may reflect that pythons inhabiting human-modified environments were often located in the roofs of buildings. There were significant seasonal differences in the proportion of usage for arboreal locations. The use of these sites increased as environmental temperatures decreased. With the exception of the post-reproductive female that continued to move throughout winter, all animals over-wintered in arboreal locations.

Patterns in Macro- and Microhabitat Use Macrohabitat use differed significantly between landscapes. Not surprisingly, radio-tagged individuals that inhabited woodland environments were located predominately in woodland macrohabitats. However, individuals that inhabited human-modified environments were most commonly located in human-modified macrohabitats, despite having had access to nearby woodlands. This was largely due to the frequent usage of buildings.

Macrohabitat use also differed significantly between the sex/age classes. In human-modified environments, males and females used human-modified macrohabitats more than juveniles, which used woodland habitats more frequently. In human-modified environments, males occupied rabbit warrens. In northeastern Victoria, Heard et al. (2004) found that the distribution of M. spilota was correlated to the distribution of rabbits. Males were also observed crossing tracks and clearings in the woodland, or in grasslands adjacent to woodlands, while juveniles occupied cracks in the dry creek bed. These cracks were often quite deep and contained small lizards (e.g., Tessellated Gecko, (Diplodactylus tessellates) and skinks (Menetia sp.)). Juvenile pythons are known to consume small reptiles (Shine, 1998; Pearson et al., 2002a) and the abundance of this resource in this habitat may account for the occurrence of these individuals. Unlike M. spilota in southeastern NSW (Slip and Shine, 1988b) and in northeastern Victoria (Heard et al., 2004), radio-tagged M. spilota at Willandra did not display any seasonal variation in macrohabitat use. This is not surprising, given that M. spilota did not undertake any seasonal migrations to distant hibernacula during this study.

Microhabitat use differed significantly between landscapes. The primary differences were that Homestead pythons spent most of their time inside buildings; whereas Park pythons were mostly located in trees, logs or thickets. Shine and Fitzgerald (1996) reported similar differences in microhabitat propensities for M. spilota inhabiting areas with different levels of human modification.

Microhabitat use also differed significantly between the sex/age classes. Within human-modified habitats, males used buildings more than females and juveniles. Despite this finding, males were less arboreal because they often used the crawl spaces beneath buildings. These spaces contained high numbers of rabbits. Males were often observed with freshly captured rabbits or rabbit-sized boluses in their stomachs. Snakes may experience reduced locomotor capabilities after consuming a large prey item (Ford and Shuttlesworth, 1986); hence, these sheltered crawl spaces may offer increased protection whilst pythons are less mobile and more vulnerable to predators.

Unlike macrohabitat use, there were significant seasonal differences in the use of microhabitats. In both woodland and human-modified environments, these differences reflected a shift from terrestrial microhabitats (usually logs and thickets) used during warmer months to arboreal microhabitats (tree hollows and roofs) in the winter. During the warmer seasons, terrestrial microhabitats like thickets may provide areas of filtered and dense cover which have certain thermal benefits (Slip and Shine, 1988). These sites may also provide cover from which to ambush potential terrestrial prey (Tsairi and Bouskila, 2004). However, microhabitats that provide less cover may not provide adequate insulation against low temperatures during the cooler seasons. Shine and Fitzgerald (1996) reported similar shifts in microhabitat use for this species in northeastern NSW; whereas, Slip and Shine (1988b) reported shifts to terrestrial, rocky habitats during the colder months by M. spilota in southeastern NSW. It appears that microhabitat use by pythons at Willandra reflects both dietary and thermoregulatory requirements.

Microhabitat Selection

Morelia spilota is a sedentary ambush predator that is reliant on habitat structure for concealment from both prey and predators (Slip and Shine, 1988b, 1988a; Shine and Fitzgerald, 1996; Heard et al., 2004; Pearson et al., 2005). Despite this apparent reliance on habitat structure, only one previous study has examined specific structural and compositional habitat attributes associated with habitat use in this species (see Heard et al., 2004). The present study was the first to provide a quantitative assessment of microhabitat selection in M. spilota, and one of only a few such studies on Australian snakes in general.

Throughout this study, radio-tagged M. spilota selected sites that differed significantly from randomly sampled locations. Specifically, M. spilota selected sites that had more hollow logs, a higher coverage of leaf litter and fallen debris and understorey vegetation and were in closer proximity to larger trees than random sites. Heard et al. (2004) reported similar tendencies for M. spilota in northeastern VIC; however, microhabitats used were not compared to random sites. Thus, these authors were unable to make inferences about selection.

There are several reasons why M. spilota may select microhabitat sites with attributes such as those described above. Hollow logs provide shelter and sites for ambushing prey. Areas with more leaf litter and fallen debris may also provide suitable habitat for prey species (Bos and Carthew, 2003; Graham et al., 2005) and may have also provide increased terrestrial camouflage for pythons (Reinert, 1984b). The proximity to larger trees reflects the fact that they were more likely to contain hollow branches than smaller trees (Webb and Shine, 1997b; Fitzgerald et al., 2002a). These trees were important shelter sites for radio-tagged M. spilota.

There were no sex/age related differences in microhabitat selection; although, sample sizes for each group were relatively small. Other studies that have examined microhabitat selection in sit-and-wait ambush predators have reported similar findings (Reinert, 1984a, 1984b; Harvey and Weatherhead, 2006).

Patterns in Tree Use All animals used trees throughout this study and radio-telemetered pythons often spent long periods of time sequestered in tree hollows, particularly during the winter. Nearly half of all radio-telemetered animals returned to trees, which they had previously used and some animals returned to previously used trees up to three times. The frequent re-use of specific trees by individual snakes has been reported in other arboreal species of snakes (e.g., Hoplocephalus bungaroides, Webb and Shine, 1997b; Hoplocephalus stephensii, Fitzgerald et al., 2002a). In northeastern NSW, M. spilota used trees with a dense covering of vines, but did not use tree hollows (Shine and Fitzgerald, 1996).

Characteristics of Trees Used

Radio-tagged M. spilota consistently displayed non-random patterns of tree selection. Morelia spilota at Willandra used trees that provided many retreat sites (i.e., large trees, trees of particular species and those with high numbers of hollows). These patterns have also been documented in other arboreal snakes (e.g., H. bungaroides, Webb and Shine, 1997b; H. stephensii, Fitzgerald et al., 2002a).

The preference of Morelia spliota for hollow-bearing trees may explain their avoidance of the most common tree species at Willandra called the River Cooba (A. stenophylla) because they rarely contained hollows. In the two instances where this tree species was utilised, small juveniles were sequestered inside of cracks in the bark. The fact that general patterns in tree use showed few differences between woodland and human-modified environments suggests that M. spilota will use any tree containing suitable hollows. Thus, other species of trees not found at Willandra may be used

by M. spilota in areas with different vegetation types. Any subtle differences between tree usage in these landscapes may be associated with species-specific attributes of the different trees themselves, such as the larger diameters of Peppercorns (S. areira) and Sugar Gums (E. cladocalyx), which are common around the homestead, compared to that of Black Box (E. largiflorens) and A. stenophylla that are common in the woodlands. The two species of tree used by M. spilota in the human-modified environment that did not contain hollows were the Date Palm (P. dactylifera) and the Olive Tree (O. europaea). These two species were unique in that they both had a dense cover of foliage and large amounts of fruit that made them attractive to birds. Since pythons only frequented them when the fruit was ripening and birds were actively feeding, suggests that these trees may be important foraging sites.

Diet and Relative Prey Density

A variety of prey items were recorded for M. spilota during this study. These items are similar to those reported in other studies (e.g., Slip and Shine, 1988a; Shine and Fitzgerald, 1996; Fearn et al., 2001; Pearson et al., 2002a; Heard et al., 2004). Although the diets of M. spilota at Willandra were not as varied. This may reflect that many of the potential prey species in the semi-arid zone of NSW have undergone considerable declines (Lunney et al., 1996; Short, 1998). Two other trends observed in this study, the presence of high numbers of exotic prey species and stationary prey items, reiterate that M. spilota is an opportunistic predator utilising both a sit-and-wait and active foraging strategies.

The divergent diets between Homestead and Park pythons may be explained by differences in habitat use, in particular the increased use of trees by snakes inhabiting woodland environments. Differences in dietary composition among populations are widespread in snakes and are generally attributable to differences in the availability of different kinds of prey (Mushinsky, 1987). Some of the species that were commonly eaten by pythons in one landscape were almost absent from others. For example, mice (Mus musculus) were found around the homestead but not in the woodland. The fact that prey items like mice were more abundant in the human modified landscape during spring and summer may help to explain the more sedentary nature of animals inhabiting this landscape.

Implications for Conservation There is clear geographic variation in carpet python home range size. Such regional variation in home range size likely reflects local habitat quality, prey availability and perhaps body size (McNab, 1963; Harestad and Bunnell, 1979; Gregory et al., 1987). The relationships between these factors and the home range sizes observed in this study are not clear. Yet, the smaller home range size of populations in semi-arid NSW compared to those along the east- coast of the country precludes the identification of large home range size as an attribute explaining carpet python imperilment throughout western NSW. Animals with larger area requirements may be particularly susceptible to population declines as a result of habitat alteration and fragmentation. Small home ranges sizes suggest that pythons may be able to survive in relatively smaller patches of habitat provided they offer specific habitat characteristics. However, the high variability in home range size suggests that land managers should use the largest reported home range sizes, rather than the average, as a minimum space requirement for M. spilota in the semi-arid zone of NSW.

Similarly, there is geographic variation in carpet python daily movements. Daily distances moved by M. spilota at Willandra were up to four times greater than those of M. spilota in northeastern NSW. While pythons at Willandra moved further than individuals in northeastern NSW, these movements did not translate into larger home ranges. Shine and Fitzgerald (1996) did not examine movement frequency in their study, making comparisons difficult. For example, pythons at Willandra appear to have moved more frequently within smaller areas, while pythons in northeastern NSW moved less frequently, but further when they did move.

There are considerable challenges and risks associated with increased movements, especially in areas that have been heavily modified by humans (Roe et al., 2006). For instance, increased movement may expose individuals to natural or exotic predators, motor vehicles and encounters with humans (Bonnet et al., 1999). Little is known about the consequences of human landscape modifications for snake communities (Kjoss and Litvaitis, 2001). However, previous studies have demonstrated that the extent of a snake’s movements comprises one of the most important determinants of the risk of anthropogenic induced mortality (Bonnet et al., 1999; Roe et al., 2006).

Since 1788 at least 61% of the original native vegetation in NSW has been cleared, thinned or substantially disturbed (Environmental Protection Authority, 1997). The proportion of land cleared varies between region and community type (National Vegetation Advisory Council, 1999), but in some areas of semi-arid NSW has exceeded 90% (e.g., Bauer and Goldney, 2000; Driscoll, 2004). Morelia spilota in the semi-arid zone of NSW appear particularly reliant on tree hollows and a structurally diverse understory. Habitat alteration may simplify vegetation structure, thereby reducing the number of suitable shelter sites. Habitat modification and fragmentation associated with land clearing in semi-arid NSW has undoubtedly resulted in an increase in the proportion of edge habitat throughout the landscape, which in turn affects ecosystems by modifying ecological relationships such as predator-prey interactions (Donovan et al., 1995; Robinson et al., 1995; May and Norton, 1996). Morelia spilota may be particularly vulnerable to exotic predators as they are slow moving, non- venomous and inhabit regions of Australia where introduced predators are often abundant and widespread (Heard et al., 2006). Both Shine and Fitzgerald (1996) and Heard et al. (2006) reported high levels of predation on M. spilota (41% and 37% respectively) by the introduced red fox (Vulpes vulpes) and one animal was apparently killed by V. vulpes in the current study.

Pythons display several life history traits associated with a sit-and-wait foraging strategy (Webb et al., 2003). This may render them more susceptible to reductions in adult survivorship as a result of anthropogenic mortality. They have low rates of food intake, are slow growing (and hence take longer to reach maturity) and females reproduce on a less than annual basis (Slip and Shine, 1988a; Greer, 1997; Shine, 1998). While small, relic populations of M. spilota may be particularly susceptible to mortality events as a consequence of their small size, the role that predators like V. vulpes play in regulating population dynamics is unknown and may warrant further study. This is especially true given that habitat fragmentation throughout western NSW is likely to continue and thus increase the likelihood of predator-prey interactions.

Recommendations

Monitoring On-farm Vertebrate Wildlife Our program began successfully with considerable revegetation efforts by farmers associated with the Environmental Champions Program. However, our intention to establish a vertebrate monitoring program was undermined by the effects of drought on planting success, detectability, and resources (funding). The majority of plantings failed, and low captures of most vertebrates during visual searches and artificial shelter surveys precluded useful results. Only bird surveys produced quantitative results (but see below). As such our key recommendation is to establish a monitoring program only during reasonable (non-drought) conditions. In addition to the obvious drought impacts mentioned above, environmental conditions must be considered when comparing biodiversity indices across years.

The Importance of Frogs in Controlling Rice Pests Findings from both the present study and previous research suggest that up to hundreds of millions of frogs of one species (spotted grass frogs) are produced in rice bays in an average production year. Frogs thus likely comprise the highest biomass of any vertebrate on rice farms. Subsequently, their generalist dietary habits might be expected to have a significant impact on invertebrates inhabiting rice crops, including pests of rice plants. As predicted frogs consumed several species of rice pests, and as such their presence reflected natural pest control. Any reduction of frog abundance would accordingly result in increases in the abundance of rice pests and thus likely increase pest damage to rice crops.

We therefore recommend that:

• Frog abundance should be promoted on farms because of its role in natural pest control for rice crops, in addition to its importance within the food web.

• A passive approach is advisable given the profound numbers currently breeding on rice farms. This approach would (continued) restricted use of pesticides and herbicides that could reduce frog numbers. Although there are numerous other environmental incentives to reduce water use on farms, such reduction would be expected to reduce frog abundance

• Further studies be carried out on how frogs use rice farms, and what factors underpin frog abundance (other than water, as found in previous studies). For example, where do frogs hibernate, and what microhabitats do they utilize when rice bays are dry?

• Knowledge from that research could be used to suggest to ricegrowers what on-farm measures might promote frog abundance. For example, revegetation around dams could provide frogs with increased cover and food, especially when rice bays are not flooded

Ecology of Carpet Pythons in an Agricultural Landscape The inland carpet python is not only a species of special concern (having apparently declined in the Riverina Bioregion), but also represents a focal species for habitat conservation because it needs relatively intact woodland for its survival. The present study determined several aspects of the ecology of these snakes in an agricultural landscape. The major findings were that snakes (1) inhabited heavily human-modified areas (homestead buildings) when in proximity to intact woodland; (2) moved less frequently in these areas than in woodlands without buildings, probably because of the

higher concentration of food there; (3) relied on more exotic species for food than snakes inhabiting woodland; and (4) used tree hollows more often than coastal populations of carpet pythons.

Our recommendations for farmers with carpet pythons on their property who wish to conserve snakes and other woodland-dependent animals are:

• Retention of remnant vegetation, especially around creeks, rivers, and (black box) depressions

• Improvement of riparian remnants through revegetation efforts, including an understorey for protective cover

• Wherever possible link vegetation remnants with revegetation

• Retain trees and woodland habitat close to buildings

• Be aware of the role snakes play in helping to control invasive species such as the house mouse and European rabbit. Excessive poisoning of these animals around buildings could lead to detrimental effects on snakes

These recommendations are largely consistent with the Biodiversity Strategy and Plan developed for the Australian Rice Industry (Freudenberger and Stol, 2002).

References

Adcock, L. 2003. EMS activity update. Pages 11-14 in the Proceedings of the 3rd National Conference on Environmental Management Systems in Agriculture, Business and the Environment, November 10-13, 2003, Barossa Valley, South Australia.

Anonymous. (1989). Willandra National Park plan of management. Griffith: New South Wales National Parks and Wildlife Service. [Online URL] http://www.environment.nsw.gov.au/resources/parks/pomfinalWillandra.pdf

Attademo, A., Peltzer, P.M. and Lajmanovich, R.C. (2005): Amphibians occurring in soybean and implications for biological control in Argentina, Agriculture. Ecosystems and Environment (106): 389-394.

Australian Bureau of Agricultural and Resource Economics (ABARE), (2006), Australian crop report, no. 140, ABARE. at : http://www.abareconomics.com/publications_html/news/news.html downloaded 1st November 2007

Avery, D. T. 1996. Low on farm science, high on pledges. Des Moines Registar (9th June).

Barker, J., Grigg, G. C. and Tyler, M. J., (1996): A field guide to Australian frogs, Surrey Beatty and Sons Pty Limited, Chipping Norton, NSW, Australia

Batra, S. W. T. (1982), Biological control in agroecosystems, Science, (215):134-139.

Bauer, J. & Goldney, D. (2000). Extinction processes in a transitional agricultural landscape system. In Temperate eucalypt woodlands in Australia: Biology, conservation, management and restoration: 107-126. Hobbs, R.J. & Yates, C.J. (Eds.). Sydney: Surrey Beatty & Sons.

Beck, D.D. (1995). Ecology and energetics of three sympatric rattlesnake species in the Sonoran Desert. J Herpetol 29, 211-223.

Bedford, G.S. (2003). Ecology and physiology of pythons from the Northern Territory of Australia. PhD Thesis, Northern Territory University, Australia.

Bedford, G.S. & Christian, K.A. (1998). Standard metabolic rate and preferred body temperatures in some Australian pythons. Aust J Zool 46, 317-328.

Bennett, A. F, S. L. Kimber, and P. A. Ryan. 2000. Revegetation and wildlife: a guide to enhancing revegetated habitats for wildlife conservation in rural environments. Bushcare National Projects and Development Program Research Report 2/00.

Benson, J. (1999). Setting the scene: the native vegetation of New South Wales. Background paper (1) Sydney: Native Vegetation Advisory Council of New South Wales.

Bonnet, X., Naulleau, G. & Shine, R. (1999). The dangers of leaving home: Dispersal and mortality in snakes. Biol Conserv 89, 39-50.

Boquimpani-Freitas, L., Rocha Carlos F. D. and Van Sluys, M. (2002): Ecology of the Horned Leaf-Frog, Proceratophrys appendiculata (leptodactylidae), in an insular Atlantic Rain-Forest Area of Southeastern Brazil. Journal of Herpetology (36):318-322.

Bowmer, K. H., Bales, M., and J. Roberts. 1994. Potential use of irrigation drains as wetlands. Water Science Technology 29: 151-158.

Branch. W. R. (1976), Two exceptional food records for the african bullfrog Pyxicephalus adspersus (Amphibia, Anuran, Ranidae), Journal of Herpetology (10): 266-268.

Brito, J.C. (2003). Seasonal variation in movements, home range, and habitat use by male Vipera latastei in northern Portugal. J Herpetol 37, 155-160.

Brouder, S. M., and J. E. Hill. 1995. Winter flooding of rice-lands provides waterfowl habitat. California Agriculture 49: 58-64.

Brown, R. L. (1974), Diets and habitat preferences of selected anurans in southeast Arkansas, American midland naturalist (91): 468-473.

Brown, J.H. & Maurer, B.A. (1989). Macroecology: The division of food and space among species on continents. Science 243, 1145-1150.

Brown, G.P. & Weatherhead, P.J. (2000). Thermal ecology and sexual size dimorphism in northern water snakes, Nerodia sipedon. Ecol Monogr 70, 311-330.

Brown, G.P., Shine, R. & Madsen, T. (2005). Spatial ecology of slatey-grey snakes (Stegonotus cucullatus, Colubridae) on a tropical Australian floodplain. J Trop Ecol 21, 605-612.

Bureau of Meteorology. 2007. Rainfall averages for Australian sites (online). http://www.bom.gov.au/climate/averages downloaded 30th May 2007

Burt, W.H. (1943). Territoriality and home range concepts as applied to mammals. J Mammal 24, 346-352.

Butler, H., Malone, B. & Clemann, N. (2005). The effects of translocation on the spatial ecology of tiger snakes (Notechis scutatus) in a suburban landscape. Wildlife Res 32, 165-171.

Cambell, D. (1994). Clearing and cultivation in the Western Division. In Future of the fauna of western New South Wales: 201-205. Lunney, D., Hand, S., Reed, P. & Butcher, D. (Eds.). Sydney: Royal Zoological Society of New South Wales.

Campbell, H. W., and S. P. Christman. 1982. Field techniques for herpetofaunal community analysis. Pages 193-200 In N.J. Scott Jr. (ed) Herpetological Communities. U.S. Fisheries and Wildlife Service, Wildlife Research Report 13.

Charland, B.M. & Gregory, P.T. (1995). Movements and habitat use in gravid and nongravid female garter snakes (Colubridae: Thamnophis). J Zool 236, 543-561.

Caughley, G. & Sinclair, A.R.E. (1994). Wildlife ecology and management. Boston: Blackwell Scientific Publications.

Christian. K. A. (1982): Changes in the food niche during postmetamorphic ontogeny of the frog Pseudacris triseriata, Copeia, 73-80.

Clarke, R. D. (1974): Food habits of Toads, Genus Bufo (Amphibia: Bufonidae), American Midland Naturalist (91): 140-147.

Clark, D.R. (1974). The western ribbon snake (Thamnophis proximus): Ecology of a Texas population. Herpetologica 30, 372-379.

Cogalniceanu, D., Palmer, M.W. and Ciubuc, C., (2000): Feeding in anuran communities on islands in the Danube floodplain, Amphibia-Reptilia(22): pp. 1-19.

Cogger, H.G. 2000. Reptiles and amphibians of Australia. 6th edn. Sydney: Reed New Holland.

Cooke, S.J., Hinch, S.G., Wikelski, M., Andrews, R.D., Kuchel, L.J., Wolcott, T.G. & Butler, P.J. (2004). Biotelemetry: A mechanistic approach to ecology. Trends Ecol Evol 19, 334-343.

Cogger, H.G. (2000): Reptiles and amphibians of australia, New Holland Publishers Pty ltd, Sydney, Australia.

Coppel, H.C. and Mertins, J.W. (1977): Biological insect suppression, Springer-Valag Berlin, Heidleberg.

Corey, B. (2007). Spatial ecology and habitat use of carpet pythons (Morelia spilota) from semi-arid New South Wales. Honours Thesis, University of Canberra, Australia. [Online URL] http://aerg.canberra.edu.au/cgi-bin/biblio.cgi?action=Search

Corn, P. S. and R. B. Bury. 1990. Sampling Methods for terrestrial amphibians and reptiles. General Technical Report, US Agriculture and Forest Service, Pacific Northwest Research Station.

Covacevich, J. and Archer, M. (1970): The distribution of the cane toads Bufo marinus in Australia and effects on indigenous vertebrates, Queensland Museum, Australia.

Crump, M. L., and N. J. Scott. 1994. Visual encounter surveys. Pages 84-92 In W.R. Heyer, M. A. Donnelly, R. W. McDiarmid, L. C. Hayek, and M. S. Foster (eds) Measuring and Monitoring Biological Diversity. Standard Methods for Amphibians. Smithsonian Intsitution Press, Washington, D.C.

CSIRO, (1991): The insects of australia: at textbook for students and research workers, Melbourne University Press, Second edition.

Cunningham, G. M., Mulham, W. E., Milthorpe, P. L., and J. H. Leigh. 1981. Plants of Western New South Wales. Sydney: Government Printer.

Dargin, H. J. 1941. Rice growing in New South Wales. New South Wales Department of Agriculture, Sydney.

Darst, C.R., Menedez-Guerrero Pablo A., Coloma, L.A., and Cannatella, D. C. (2005): Evolution of dietary specialisation and chemical defence in poison frogs (Dendrobatidae): a comparative analysis, The American Naturalist (165): 56-69.

Davidson, B. R. 1988. Historical aspects of irrigation agriculture and river regulation in the Murray- Darling Basin, Australia. Regulated Rivers: Research and Management 2: 131-140.

Department of Environment and Water Resources, (1997): The Action Plan for Australian Frogs, Wildlife Australia.

Dickman, C.R., Pressey, R.L., Lim, L. & Parnaby, H.E. (1993). Mammals of particular conservation concern in the Western Division of New South Wales. Biol Conserv 65, 219-248.

Diffendorfer, J.E., Rochester, C., Fisher, R.N. & Brown, T.K. (2005). Movement and space use by coastal rosy boas (Lichanura trivirgata roseofusca) in coastal southern California. J Herpetol 39, 24- 36.

Dobson, A. P., A. D. Bradshaw, and A. J. M. Baker. 1997. Hopes for the future: Restoration ecology and conservation ecology. Science 277: 515-521.

Donald, P. F., Green, R. E., and M. F. Heath. 2001. Agricultural intensification and the collapse of Europe’s farmland bird populations. Proceedings of the Royal Society of London 268: 25-29.

Donnelly, M. A., (1991), Feeding patterns of the strawberry poison frog, Dendrobates pumilio (Anura: Dendrobatidae), Copeia(3):723-730.

Donovan, T.M., Lamberson, R.H., Kimber, A., Thompson, I.F.R. & Faaborg, J. (1995). Modelling the effects of habitat fragmentation on source and sink demography of Neotropical migrant birds. Conserv Biol 9, 1396-1407.

Doody, J.S., Osborne, W., Bourne, D. and Sims, R.A. (2004): Vertebrate biodiversity on Australian rice farms: An inventory of species, variation among farms, and proximate factors explaining that variation. Final report, project No. RIC04-10, Rural Industries Research and Development Corporation, Barton, Canberra.

Doody, J.S., Young, J. E. & Georges, A. (2002). Sex differences in activity and movements in the pig- nosed turtle, Carettochelys insculpta, in the wet-dry tropics of Australia. Copeia 2002, 93-103.

Driscoll, D A. (2004). Extinction and outbreaks accompany fragmentation of a reptile community. Ecol Appl 14, 220-240.

Duellman, W.E., Trueb, L. (1985): Biology of amphibians, McGraw-Hill, Inc. USA.

Durner, G.M. & Gates, J. E. (1993). Spatial ecology of black rat snakes on Remington Farms, Maryland. J Wildlife Manage 57, 812-826.

Duvall, D. & Schuett, G.W. (1997). Straight-line movement and competitive mate searching in prairie rattlesnakes, Crotalus viridis viridis. Anim Behav 54, 329-334.

Eardley, K. 1999. A foundation for conservation of the Riverina Bioregion. Sydney: NSW National Parks and Wildlife Service.

Eastburn, D. R. 1990. The River Murray: History at a glance. Murray-Darling Basin Commission, Canberra, ACT.

Elphick, C. S. 2000. Functional equivalency between rice fields and semi-natural wetland habitats. Conservation Biology 14: 181-191.

Elphick, C. S. 2004. Assessing conservation trade-offs: identifying the effects of flooding rice fields for waterbirds on non-target species. Biological Conservation 117:105-110.

Elphick, C. S., and L. W. Oring. 1998. Winter management of California rice fields for waterbirds. Journal of Applied Ecology 35: 95-108.

Elphick, C. S., and L. W. Oring. 2003. Conservation implications of flooding rice fields on winter waterbird communities. Agriculture, Ecosystems, and Environment 94: 17-29.

Environment Protection Authority. 1997. NSW State of the Environment 1997. Sydney: New South Wales Environment Protection Authority. [Online URL] https://www.dec.nsw.gov.au/soe/97/index.htm

Etkin, W. E. (1932): Growth and resorption phenomena in anuran metamorphosis, I., Physiology Zoology(5):275-300. as cited in Duellman and Trueb (1985).

Fasola, M. 1986. Resource use of foraging herons in agricultural and nonagricultural habitats in Italy. Colonial Waterbirds 9: 139-148.

Fasola, M., Canova, L., and N. Saino. 1996. Rice fields support a large portion of herons breeding in the Mediterranean Region. Colonial Waterbirds 19: 129-134.

Fasola, M., and X. Ruiz. 1996. The value of rice fields as substitutes for natural wetlands for waterbirds in the Mediterranean Region. Colonial Waterbirds 19: 122-128.

Fearn, S., Robinson, B., Sambono, J. & Shine, R. (2001). Pythons in the pergola: The ecology of ‘nuisance’ carpet pythons (Morelia spilota) from suburban habitats in south-eastern Queensland. Wildlife Res 28, 573-579.

Fischer, J., D. B. Lindenmeyer, and A. Cowling. 2004. The challenge of managing multiple species at muliple scales: reptiles in an Australian grazing landscape. Journal of Applied Ecology 41:32-44.

Fitzgerald, M., Shine, R. & Lemckert, F. (2002). Spatial ecology of arboreal snakes (Hoplocephalus stephensii, Elapidae) in an eastern Australian forest. Austral Ecol 27, 537-545.

Fores, E., and F. Comin. 1992. Ricefields, a limnological perspective. Limnetica 8: 101- 109.

Forstner, J.M., Forstner, M.R.J., and Dixon, J.R.(1998): Ontogenetic effects on prey selection and food habit of two sympatric east texas ranids: the southern leopard frog, Rana sphenocephala, and the bronze frog, Rana clamintas clamitans, Herpetological Review (29): 208.

Franca, L. F., Facure, K.G., and Giaretta, A.A. (2004): Trophic and spatial niches of two large-sized species of Leptodactylous (Anura) in south-eastern Brazil, Studies on Neotropical Fauna and Environment (39): 243-248.

Freed, A. N., (1980), Prey Selection and feeding behaviour of the green tree frog (Hyla cinerea), Ecology (3): 461-465.

Fujioka, M., and S. J. Lane. (1997). The impact of changing irrigation practices in rice fields on frog populations of the Kanto Plain, central Japan. Ecological Research 12: 101-108.

Glanzing, A. 1995. Native vegetation clearance, habitat loss biodiversity decline. Biodiversity Series Paper No. 6. Biodiversity Unit, Department of Environment, Sport and Territories, Canberra.

Gerakis, A., and K. Kalburtji. 1998. Agricultural activity affecting the functions and values of Ramsar wetland sites of Greece. Agriculture, Ecosystems and Environment 70: 119-128.

Greer, A.E. (1997). The biology and evolution of Australian snakes. Sydney: Surrey Beatty & Sons.

Gregory, P.T., Macartney, M.J. & Larsen, K.W. (1987). Spatial patterns and movements. In Snakes: Ecology and evolutionary biology: 366-395. Seigel, R.A., Collins, J.T. & Novak, S.S. (Eds.). New Jersey: The Blackburn Press.

Guarino, F. (2002). Spatial ecology of a large carnivorous lizard, Varanus varius (Squamata: Varanidae). J Zool 258, 449-457.

Gullan, P.J., and Cranston, P.S., (1994), The insects: An outline of Entomology, Chapman and Hall, Australia. 12-18.

Haddlington, P.W., and Johnston, J.A., (1982), An introduction to Australian insects, New South Wales University Press, Kensington, NSW.

Hale, M.A. and Elliott, D. (2003): Successes and challenges in augmentative biological control in outdoor agricultural applications: a producer’s perspective, 1st International Symposium on Biological Control of Arthropods.

Harestad, A.S. & Bunnell, F.L. (1979). Home range and body weight - a reevaluation. Ecology 60, 389-402.

Harlow, P. & Grigg, G. (1984). Shivering thermogenesis in a brooding diamond python, Python spilotes spilotes. Copeia 1985, 959-965.

Harris, S., Cresswell, W.J., Forde, P.G., Trewhella, W.J., Woollard, T. & Wray, S. (1990). Home- range analysis using radio-tracking data - a review of problems and techniques particularly as applied to the study of mammals. Mammal Rev 20, 97-123.

Hajek, C. H., and H. Johnson. 2002. Managing remnant vegetation. Landcare Notes LC0134.

Heard, G.W., Black, D. & Robertson, P. (2004). Habitat use by the inland carpet python (Morelia spilota metcalfei: Pythonidae): Seasonal relationships with habitat structure and prey distribution in a rural landscape. Austral Ecol 29, 446-460.

Heard, G.W., Robertson, P., Black, D., Barrow, G., Johnson, P., Hurley, V. & Allen, G. (2006). Canid predation: A potentially significant threat to relic populations of the inland carpet python Morelia spilota metcalfei (Pythonidae). Victorian Naturalist 123, 68-74.

Heatwole H. and Heatwole A. (1968): Motivational aspects of feeding behaviour in toads, Copeia (4):692-698.

Heitmeyer, M. E., Connelly, D. P., and R. L. Pederson. 1989. Habitat management for miugrating and wintering waterfowl in North America. Pages 475-505 In L. M. Smith, R. L. Pederson, and R. H. Kaminski (eds.) The Central, Imperial, and Coachella valleys of California. Texas Tech University Press, Lubbock.

Hengeveld, R. 1996. Measuring ecological biodiversity. Biodiversity Letters 3: 58-65.

Hirai, T. (2002): Ontogenetic change in the diet of the pond frog, Rana nigromaculata Ecological Research (17):639-644.

Hirai, T., and Matsui, M., (1999): Feeding habits of the Pond frog, Rana nigromaculata, inhabiting rice fields in Kyoto, Japan, Copeia (4);940-947

Hirai, T and Matsui, M. (2000): Myrmecophagy in a ranid frog Rana rugosa: Specilization or weak avoidance to ant eating? Zoological Science (17):459-466.

Hirai, T and Matsui, M. (2001): Ecological Food habits of an endangered Japanese frog, Rana porosa brevipoda. Research (16):737-743.

Hirai, T., Matsui, M. (2002): Feeding relationships between Hyla japonica and Rana nigromaculata in Rice Fields of Japan, Journal of Herpetology (36):662-667.

Hoddle, M.S. (2003): Classical control of arthropods in the 21st century, 1st International Symposium on Biological Control of Arthropods.

Hooge, P.N. & Eichenlaub, B. (2000). Animal Movement extension to ArcView. Version 2.0. Alaska Biological Science Centre, US Geological Survey, Anchorage, AK, USA. [Online URL] http://www.absc.usgs.gov/glba/gistools/

Howard, W. E. (1950): Birds as bullfrog food, Copeia (2):152

Huey, R.B. and Pianka, E R.(1981): Ecological consequences of foraging mode, Ecology (62)991-999.

100

Huey, R.B. (1982). Temperature, physiology and the ecology of reptiles. In Biology of the reptilia: Ecology and behaviour: 25-96. Gans, C. & Pough, F.H. (Eds.). London: Academic Press.

Huey, R.B. (1991). Physiological consequences of habitat selection. Am Nat 137, S91-S115.

Huey, R.B. & Pianka, E.R. (1981). Ecological consequences of foraging mode. Ecology 62, 991-999.

Inger, R. F. and Greenberg, B. (1966): Ecological and competitive relations among three species of frogs (Genus Rana), Ecology (47)746-759.

Jameson, D. L. and Myers, R. M. (1956): Food habits in juvenile frogs, Copeia(4):261.

Jenssen, T. A. and Klimstra, W. D. (1966): Food habits of the green frog, rana clamitans, in Southern Illinois, American Midland Naturalist(76)169-182.

Johnson, B.K., Christiansen, J.L., (1976): The food and food habit’s of Blanchard’s cricket frog, Acris crepitans blanchardi (amphibian, Anura, Hylidae, In Iowa, Journal of herpetology(10) 63-74

Jones, R.N, and Pittcock, B. (1997): Assessing the impacts of climate changes: the challenge for ecology. Pp. 311-322 in: frontiers of ecology. Klomp N. and Lunt, I. (eds.) Elsiever Science, Oxford.

Kernohan, B.J., Gitzen, R.A. & Millspaugh, J.J. (2001). Analysis of animal space use and movements. In Radio tracking and animal populations: 125-166. Millspaugh, J.J. & Marzluff, J.M. (Eds). San Diego: Academic Press.

Kimber, S. L., A. F. Bennett, and P. A. Ryan. 1999. Revegetation and wildlife: what do we know about revegetation and wildlife conservation in Australia? Report to Environment Australia.

King, M.B. & Duvall, D. (1990). Prairie rattlesnake seasonal migrations: Episodes of movement, vernal foraging and sex differences. Anim Behav 39, 924-935.

Kjoss, V.A. & Litvaitis, J.A. (2001). Community structure of snakes in a human-dominated landscape. Biol Conserv 98, 285-292.

Knight, R. L. 1999. Private lands: the neglected geography. Conservation Biology 13: 223-224.

Krebs, C.K. (2001): Ecology; the experimental analysis of distribution and abundance, fifth edition, Benjamin Cummins, an imprint of Addison Wesley Longman, Inc., San Francisco, US

Labanick, G. M. (1976): Prey availability, consumption and selection in the cricket frog, Acris crepitans. Journal of Herpetology (10): 293-298.

Lamoureux, V.S, Maerz, J.C. and Madison, D.M. (2002): Premigratory autumn foraging forays in the green frog, Rana clamitans Journal of Herpetology (36):245-254.

Lampo, M and De Leo, G.A. (1998): The invasion ecology of the toad Bufo marinus: From South America to Australia. Ecological Applications (8):388-396.

Lane, S. J., and M. Fujioka. 1998. The impact of changes in irrigation practices on the distribution of foraging egrets and herons (Ardeidae) in the rice fields of central Japan. Biological Conservation 83: 221-230.

Lane, S.J., and Mahony, M.J. (2002): Larval anurans with synchronous and asynchronous development periods: contrasting responses to water reduction and predator presence. Journal of Animal Ecology (71):780-792.

101

Lawrence, J.F. and Britton, E.B. (1994): Australian beetles, CSIRO, Melbourne University Press.

Legler, J.M. and Sullivan, L.J. (1979): The application of stomach-flushing to lizards and anurans. Herpetologica (35):107-110.

Leif, A.P. (2005). Spatial ecology and habitat selection of breeding male pheasants. Wildlife Soc B 33, 130-141.

Lemly, A. D., Kingsford, R. T., and J. R. Thompson. 2000. Irrigated agriculture and wildlife conservation: Conflict on a global scale. Environmental Management 25: 485-512.

Lima, A. P. (1998): The effects of size on the diets of six sympatric species of postmetamorphic litter anurans in central Amazonia, Journal of Herpetology (32): 392-399.

Lima, A. P, Magnusson, W.E. and Williams, D.G. (2000): Differences in diet among frogs and lizards coexisting in subtropical forests of Australia Journal of Herpetology (34): 40-46.

Lindenmayer, D., Claridge, A., Hazell, D., Michael, D., Crane, M., MacGregor, C., and Cunningham, R. (2003): Wildlife on farms- how to conserve native animals, CSIRO publishing, Melbourne, Australia.

Lindenmayer, D., and Burgman, M. (2005): Practical conservation biology, CSIRO Publishing, VIC, Australia.

Littlejohn, M J. and Martin A A. (1967): The rediscovery of Heleioporus australiacus (Shaw)(Anura: Leptodactylidae) in eastern Victoria, Proceedings of the Royal Zoological Society of Victoria.

Lovejoy, T. E. 1994. The quantification of biodiversity: An esoteric quest or a vital component of sustainable development. Philosophical Transactions: Biological Sciences 345: 81-87.

Loyn, R. H. 1986. The 20-minute search- a simple method for counting forest birds. Corella 10: 58- 60.

Macartney, M.J., Gregory, P.T. & Larsen, K.W. (1988). A tabular survey of data on movements and home ranges of snakes. J Herpetology 22, 61-73.

Maeda, T. 2001. Patterns of bird abundance and habitata use in rice fields of the Kanto Plain, central Japan. Ecological Research 16:569-585.

Maeda, N., and M. Matsui. 1989. Frogs and Toads of Japan. Bun-Ichi Sogo Shuppan, Tokyo.

Madsen, T. & Shine, R. (1996). Seasonal migration of predators and prey - a study of pythons and rats in tropical Australia. Ecology 77, 149-156.

Mahan, R.D. and Johnson, J.R. (2007): Diet of the gray treefrog (Hyla versicolor) in relation to foraging site location, Journal of Herpetology (41):16-23.

Maher, C.R. & Lott, D.F. (2001). A review of ecological determinants of territoriality within vertebrate species. Am Midl Nat 143, 1-29.

Manfredi, C., Soler, L., Lucherini, M. & Casanave, E.B. (2006). Home range and habitat use by Geoffroy's cat (Oncifelis geoffroyi) in a wet grassland in Argentina. J Zool 268, 381-387.

Margules, C. R, and R. L. Pressey. 2000. Systematic conservation planning. Nature 405: 243-253.

102

Martin, A.A. (1970): Parallel evolution in the adaptive ecology of Leptodactylid frogs of South America and Australia, Evolution (24): 643-644.

May, S.A. & Norton, T.W. (1996). Influence of fragmentation and disturbance on the potential impact of feral predators on native fauna in Australian forest ecosystems. Aust Wildlife Res 23, 387-400.

McCue, M.D. (2004). General effects of temperature on animal biology. In Temperature-dependant sex determination in vertebrates: 71-79. Valenzuela, N. & Lance, V.A. (Eds.). Washington: Smithsonian Books.

McNab, B.K. (1963). Bioenergetics and the determination of home range size. Am Nat 97, 133-140.

Miller, M. R., Sharp, D. E., Glime, D. S., and W. R. Mulvaney. 1989. Rice available to waterfowl in harvested fields of in Sacramento Valley, California. California Fish and Game 75: 113-123.

Minton, J. E. (1949): Coral Snake Preyed Upon by the Bullfrog, Copeia (4):288.

Moore, J.A. & Gillingham, J.C. (2006). Spatial ecology and multi-scale habitat selection by a threatened rattlesnake: The eastern massasauga (Sistrurus catenatus catenatus). Copeia 2006, 742- 751.

Morton, S.R. (1986). Land of uncertainty: The Australian arid zone. In A natural legacy: Ecology in Australia: 122-144. Recher, H.F., Lunney, D. & Dunn, I.A. (Eds.). Sydney: Pergamon Press.

Mullins, B. J. D., and L. J. Sutherland. 2001. Draft scoping report for the development of a biodiversity strategy and Plan for the Rice Industry. Report No. 11: Johnstone Centre: Research in Natural Resources and Society. Charles Sturt University.

National Vegetation Advisory Council. (1999). Setting the scene: The native vegetation of NSW. Sydney: Department of Land and Water Conservation. [Online URL] www.rbgsyd.nsw.gov.au/__data/assets/pdf_file/51480/settingthescene.pdf

NSW National Parks and Wildlife Service (2003): The Bioregions of New South Wales: their biodiversity, conservation and history, NSW National Parks and Wildlife Service, Hurstville

Odum, E. P. 1993. Ecology on our endangered life-support systems. Sinauer, Sunderland.

Pain, D. 1994. Case Studies of Farming and Birds in Europe: Rice Farming in Italy. Pages 45-53 In Studies in European Agriculture and Environmental Policy, R.S.P.D, Sandy, Bedfordshire.

Parker, M. L. and Goldstein, M. I. (2004): Diet of the Rio Grande leopard Frog (Rana berlandieri) in Texas, Journal of Herpetology (38):127-130.

Pearson, D., Shine, R. & Williams, A. (2005). Spatial ecology of a threatened python (Morelia spilota imbricata) and the effects of anthropogenic habitat change. Austral Ecol 30, 261-274.

Pearson, D., Shine, R. & How, R. (2002). Sex-specific niche partitioning and sexual size dimorphism in Australian pythons (Morelia spilota imbricata). Biol J Linn Soc 77, 113-125.

Peters, R.H. 1983. The ecological implications of body size. Cambridge: Cambridge University Press.

Petersen, C.R., Gibson, R.A. & Dorcas, M. E. (1993). Snake thermal ecology: The causes and consequences of body-temperature variation. In Snakes: Ecology and behaviour: 241-314. Seigel, R.A. & Collins, J.T. (Eds.). New Jersey: The Blackburn Press.

103

Pimentel, D., U. Stachow, D. A. Takacs, H. W. Brubaker, A. R. Dumas, J. J. Meaney, J. A. S. O’Neil, D. E. Onsi, and D. B. Corzilius. 1992. Conserving biodiversity in agriculture/forestry systems. Bioscience 42: 354-362.

Plant Health Australia (2005): National Rice Industry Biosecurity Plan, Plant health Australia, Canberra, ACT.

Plummer, M.V. (1981). Habitat utilisation, diet and movements of a temperate arboreal snake (Opheodrys aestivus). J Herpetol 15, 425-432.

Plummer, M.V. & Mills, N.E. (2000). Spatial ecology and survivorship of resident and translocated hognose snakes (Heterodon platirhinos). J Herpetol 34, 565-575.

Porteners, M.F. (1993). The natural vegetation of the Hay Plain: Booligal-Hay and Deniliquin- Bendigo. Cunninghamia 3, 1-122.

Porter, K. R. (1972): Herpetology, Saunders, Philadelphia, USA

Pulliam, H.R. (1975): Diet Optimization with nutrient constraints. American Naturalist (109):765-768.

Ramirez-Bautista, A and Lemos-Espinal, J. A. (2004): Diets of two syntopic populations of frogs, Rana vaillanti and Rana brownorum, From a tropical rain forest in Southern Veracruz, Mexico, The Southwestern Naturalist (49):316-320.

Redford, K.H., and Richter, B.D. (1999): Conservation of biodiversity in a world of use. Conservation Biology (13):1246-1256.

Reinert, H.K. & Rupert, R. (1999). Impacts of translocation on behaviour and survival of timber rattlesnakes, Crotalus horridus. J Herpetol 33, 45-61.

Reinert, H.K. & Kodrich, W.R. (1982). Movements and habitat utilisation by the massasauga, Sistrurus catenatus catenatus. J Herpetol 16, 162-171.

Reinert, H.K. & Zappalorti, R.T. (1988). Timber rattlesnakes (Crotalus horridus) of the Pine Barrens: Their movement patterns and habitat preference. Copeia 1988, 964-978.

Remsen, J. V., Swan, M. M., Cardiff, S. W., and K. V. Rosenberg. 1991. The importance of the rice- growing region of south central Louisiana to winter populations of raptors, waders and other birds. Journal of Louisiana Ornithology 1: 35-47.

Richardson, A.J., Taylor, I.R., and Growns, J.E. (2001): The Foraging Ecology of Egrets in Rice Field in Southern New South Wales, Australia, Waterbirds: The International Journal of Waterbird Biology (24): 255-264.

Roberts, J.D., Seymour, R.S. (1989): Non-foamy egg masses in Limnodynastes tasmaniensis from South Australia, Copeia (2): 488-492

Robinson, S.K., Thompson, F.R., Donovan, T.M., Whitehead, D.R. & Faaborg, J. (1995). Regional forest fragmentation and the nesting success of migratory birds. Science 267, 1987-1990.

Rodrigues, A. S. L., and K. A. Gaston. 2001. How large do resevre networks need to be? Ecology Letters 4: 602-609.

Roe, J.H., Gibson, J. & Kingsbury, B.A. (2006). Beyond the wetland border: Estimating the impact of roads for two species of water snakes. Biol Conserv 130, 161-168.

104

Roe, J.H., Kingsbury, B.A. & Herbert, N.R. (2004). Comparative water snake ecology: Conservation of mobile animals that use temporally dynamic resources. Biol Conserv 118, 79-89.

Rosenweiz, M. L. 2003. Win-win Ecology: How the Earth’s Species can Survive in the Midst of Human Enterprise. Oxford University Press, New York.

Roth, E. D. (2005). Spatial ecology of a cottonmouth (Agkistrodon piscivorus) population in east Texas. J Herpetology 39, 308-312.

Roth, T.C.I. & Greene, B.D. (2006). Movement patterns and home range use of the northern watersnake (Nerodia sipedon). Copeia 2006, 544-551.

Ruckstuhl, K. & Neuhaus, P. (2002). Sexual segregation in ungulates: A comparative test of three hypotheses. Biol Rev 77, 77-96.

Rural Industries Research and Development Corporation (RIRDC) (2000): Production of Quality Rice in South eastern Australia, Rural Industries Research and Development Corporation (RIRDC) and Irrigation Research and Extension Committee (IREC), NSW.

Sadlier, R.A. & Pressey, R.L. (1994). Reptiles and amphibians of particular conservation concern in the Western Division of New South Wales: A preliminary review. Biol Conserv 69, 41-54.

Sanders, J.S. & Jacob, J.S. (1981). Thermal ecology of the copperhead (Agkistrodon contortrix). Herpetologica 37: 264-270.

Saunders, D. A., Hobbs, R. J., and P. R. Erlich. 1993. Reconstruction of Fragmented Ecosystems: Global and Regional Perspectives. Surrey Beatty and Sons Pty. Ltd. Chipping Norton. 326 pp.

Sazima, I. (1971): The occurrence of marine invertebrates in the stomach contents of the frog Thoropa miliaris, Ciencia e Cultura (23):647-648.

Schoener, T.W. (1971). Theory of feeding strategies. Ann Rev Ecol Syst 11, 369-404.

Secor, S.M. (1994). Ecological significance of movements and activity range for the sidewinder, Crotalus cerastes. Copeia 1994, 631-645.

Shine, R. (1987). Intraspecific variation in thermoregulation, movements and habitat use by Australian blacksnakes, Pseudechis porphyriacus (Elapidae). J Herpetology 21, 165-177.

Shine, R. (1994). The biology and management of the diamond python (Morelia spilota spilota) and carpet python (M. s. variegata) in NSW. Species management report no. 15, Hurstville: New South Wales National Parks and Wildlife Service. [Online URL] http://www.bio.usyd.edu.au/Shinelab/publications/semipopular/Moreliaspilotadata.pdf

Shine, R. (1998). Australian snakes: A natural history. Revised edn. Sydney: Reed New Holland.

Shine, R. & Lambeck, R. (1985). A radiotelemetric study of movements, thermoregulation and habitat utilisation of Arafura filesnakes (Serpentes: Acrochordidae). Herpetologica 41, 351-361.

Shine, R. & Madsen, T. (1996). Is thermoregulation unimportant for most reptiles? An example using water pythons (Liasis fuscus) in tropical Australia. Physiol Zool 69, 252-269.

Shine, R. & Fitzgerald, M. (1996). Large snakes in a mosaic rural landscape: The ecology of carpet pythons Morelia spilota (Serpentes: Pythonidae) in coastal eastern Australia. Biol Conserv 76, 113- 122.

105

Shine, R. Sun, L., Fitzgerald, M. & Kearney, M. (2003). A radiotelemetric study of movements and thermal biology of insular Chinese pit-vipers (Gloydius shedaoensis, Viperidae). Oikos 342, 342-352.

Simon, M. P. and Toft, C. A. (1991): Diet Specialisation in Small vertebrates: Mite eating (acariphagy) in Frogs. Oikas (61): 263-278.

Siqueira, C. C., Van Sluys, M., Ariani, C. V. and Rocha, C. F. D. (2006): Feeding ecology of Thoropa miliaris in four areas of Atlantic Rainforest, Southeastern Brazil, Journal of Herpetology (40):520-525

Slip, D.J. & Shine, R. (1988a). The reproductive biology and mating system of diamond pythons, Morelia spilota (Serpentes: Boidae). Herpetologica 44, 396-404.

Slip, D.J. & Shine, R. (1988b). Habitat use, movements and activity patterns of free-ranging diamond pythons, Morelia spilota (Serpentes: Boidae): A radiotelemetric study. Aust Wildlife Res 15, 515-531.

Slip, D.J. & Shine, R. (1988c). Feeding habitats of the diamond python, Morelia s. spilota: Ambush predation by a boid snake. J Herpetol 22, 323-330.

Slip, D.J. & Shine, R. (1988d). Reptilian endothermy: A field study of thermoregulation by brooding diamond pythons. J Zool 216, 367-378.

Smith, P.J., Pressey, R.L. & Smith, J.E. (1994). Birds of particular conservation concern in the Western Division of New South Wales. Biol Conserv 69, 315-338.

Soule, M.. E. and M. A. Sanjayan. 1998. Conservation targets: do they help? Science 279: 2060-2061.

Sole, M,. Beckmann, O., Pelz, B., Kwet, A. and Engels, W. (2005): Stomach-flushing for diet analysis in anurans: an improved protocol evaluated in a case study in Araucaria forests, southern Brazil, Studies on Neotropical fauna and Environment (40): 23-28.

Sokal, R.R. & Rohlf, F.J. (1995). Biometry: The principles and practice of statistics in biological research. 3rd edn. New York: W.H. Freeman & Company.

Srivastva, J. P., Smith, H. N., and D. A. Forno. (1996). Biodiversity and agricultural intensification, partners for development and conservation. Washington D.C.

Stewart, M. M, and Sandison, P. (1972): Comparative food habits of sympatric Mink frogs, Bullfrogs and Green Frogs, Journal of Herpetology (6): 241-244

Thackway, R., and I. D. Creswell. 1995. An interim biogeographic regionalization for Australia: a framework for establishing the national system of reserves. Canberra: Australian Nature Conservation Agency.

Tiebout, H.M. & Cary, J.R. (1987). Dynamic spatial ecology of the water snake, Nerodia sipedon. Copeia 1987: 1-18.

Toft, C. A. (1980): Feeding ecology of thirteen syntopic species of anurans in a seasonal tropical environment, Oecologia (45):131-141.

Toft, C. A. (1981): Feeding ecology of Panamanian Litter Anurans: Patterns in diet and foraging mode, Journal of Herpetology (15):139-144.

Toft, C. A. (1985): Resource Partitioning in Amphibians and Reptiles, Copeia (1):1-21.

Tyler, M. J. (1976): Frogs, Collins, Sydney, Australia

106

Tyler, M.J. 1994. Australian Frogs A Natural History (Revised Edition). Reed, Sydney, Australia

Tyler, M.J., and Davies, M. (1979): Foam nest construction by Australian Leptodactylid frogs, Journal of Herpetology (13):509-510.

Van Sluys, M. and Rocha, C. F. D.,(1998): Feeing habitats and microhabitat utilization by 2 syntopic Brazilian Amazonian frogs (Hyla minuta and Pseudopaldicula sp. Rev. Brasil. Biology (58):559-562.

Vitousek, P.M., Mooney, H.A., Lubchenco, J. & Melillo, J.M. (1997). Human domination of earth’s ecosystems. Science 277, 494-499

Waddle, J.H. (2006): Use of amphibians as ecosystem indicator species, Unpublished thesis, University of Florida

Waldron, J.L., Bennett, S.H., Welch, S.M., Dorcas, M.E., Lanham, J.D. & Kalinowsky, W. (2006). Habitat specificity and home-range size as attributes of species vulnerability to extinction: A case study using sympatric rattlesnakes. Anim Conserv 9, 414-420.

Warkentin, K.M. (1992): Microhabitat Use and Feeding Rate Variation in Green frog Tadpoles (Rana clamitans), Copeia ( 3):721-740.

Waser, P.M. & Wiley, R.H. (1979). Mechanisms and evolution of spacing in animals. In Handbook of Behavioural Neurobiology, Volume 3: Social Behaviour and Communication: 159-224. Vandenberg, J. & Marler, P. (Eds.). New York: Plenum Press.

Weatherhead, P.J. & Hoysak, D.J. (1989). Spatial and activity patterns of black rat snakes (Elaphe obsoleta) from radiotelemetry and recapture data. Can J Zool 67, 463-468.

Webb, J.K. & Shine, R. (1997a). A field study of spatial ecology and movements of a threatened snake species, Hoplocephalus bungaroides. Biol Conserv 82, 203-217.

Webb, J.K. & Shine, R. (1997b). Out on a limb: Conservation implications of tree-hollow use by a threatened snake species (Hoplocephalus bungaroides: Serpentes, Elapidae). Biol Conserv 81, 21-33.

Webb, J.K., Brook, B.W. & Shine, R. (2003). Does foraging mode influence life history traits? A comparative study of growth, maturation and survival of two species of sympatric snakes from southeastern Australia. Austral Ecol 28, 601-610.

Weber, C.G. (1981): The bombardier beetle myth exploded, Creation/evolution (2):1-5.

Weigert, R.G. (1965): Intraspecfic variation in calories/g of meadow spittlebugs (Philaenus spumarius L.) BioScience (15):543-545.

Western, D. 2001. Human-modified ecosystems and future evolution. Proceedings of the National Academy of Sciences of the United States of America 98:5458-5465.

Western, D., and M. C. Pearl. (1989). Conservation in the 21st Century. Oxford University Press, New York.

Whitaker, P.B. & Shine, R. (2003). A radiotelemetric study of movements and shelter-site selection by free-ranging brownsnakes (Pseudonaja textilis, Elapidae). Herpetol Monogr 17, 130-144.

Wilson, D., Heinsohn, R. & Legge, S. (2006). Age- and sex-related differences in the spatial ecology of a dichromatic tropical python (Morelia viridis). Austral Ecol 31, 577-587.

Wilson, S. & Swan, G. (2003). A complete guide to reptiles of Australia. Sydney: Reed New Holland.

107

Wisler, C., Hofer, U. & Arlettaz, R. (2008). Snakes and monocultures: Habitat selection and movements of female grass snakes (Natrix natrix L.) in an agricultural landscape. J Herpetol 42, 337- 346.

Wood, S., Sebastian, K., and S. J. Scherr. 2000. Pilot Analysis of Global Ecosystems; Agroecosystems. Washington D.C.: International Food Policy Research Institute and World Resources Institute.

Woods, A. (1974): Pest control: A survey, McGraw- Hill Book Company (UK) Limited.

Worton, B.J. (1995). Using Monte Carlo simulation to evaluate kernel-based home range estimators. J Wildlife Manage 59, 794-800.

Wunderle, J.M.J., Mercado, J.E., Parresol, B. & Terranova, E. (2004). Spatial ecology of Puerto Rican boas (Epicrates inornatus) in a hurricane impacted forest. Biotropica 36, 555-571.

Zanatta, C. A., King, P., Axton, A., Sloane, P., S. C., Mathers, K., O'Connell, N., Thompson, P., Watts, D., Edwards, D., Heath, T., Asservatham, E., Stubley, D., Telford, M., McLeod, G., and A. Smith. 2001. Murray Irrigation Limited- Environment Report 2000/2001 Summary. Murray Irrigation Limited, NSW.

108 Reconciling Farming with Wildlife —Managing diversity in the Riverina rice fields—

by J. Sean Doody, Christina M. Castellano, Will Osborne, Ben Corey and Sarah Ross

Publication No. 10/007

Although the establishment of the Australian rice industry has RIRDC is a partnership between government and industry led to an altered pastoral landscape, it has also resulted in the to invest in R&D for more productive and sustainable rural seasonal availability of extensive modified aquatic habitats in industries. We invest in new and emerging rural industries, a a previously drier landscape. Consequently, rice-farming areas suite of established rural industries and national rural issues. are of considerable importance to a range of vertebrate fauna. Most of the information we produce can be downloaded for free This publication details findings from baseline monitoring of or purchased from our website . vertebrates on rice farms; a study of frogs as natural pest control for rice crops; and a study of how an iconic snake species RIRDC books can also be purchased by phoning utilises farms. The project concludes by integrating knowledge 1300 634 313 for a local call fee. gained in these studies with that of other studies to formulate management strategies for rice farmers and other stakeholders in the region.

Contact RIRDC: Level 2 15 National Circuit Ph: 02 6271 4100 Most RIRDC publications can be viewed and purchased at Barton ACT 2600 Fax: 02 6271 4199 our website: Email: [email protected] PO Box 4776 web: www.rirdc.gov.au Kingston ACT 2604 Bookshop: 1300 634 313 RIRDCInnovation for rural Australia