The comparative population ecology of two semi-aquatic varanids
James G. Smith 2007
The comparative population ecology of two semi
aquatic varanids
James Gordon Smith
B. Sc. (Hons) Melbourne
A thesis submitted to satisfy the requirements for the award of the degree of Doctor
of Philosophy in the Institute of Advanced Studies, School for Environmental
Research, Charles Darwin University, Darwin, Australia
January 2007
Declaration:
I hereby declare that the work herein, now submitted as a thesis for the degree of Doctor of Philosophy of the Charles Darwin University, is the result of my own investigations, and all references to ideas and work of other researchers have been specifically acknowledged. I herby certify that the work embodied in this thesis has not already been accepted in substance for any degree, and is not being currently submitted in candidature for any other degree. Signed:
Date:
Acknowledgments
I would like to thank the innumerable volunteers (particularly my friend “features”) I have had over the years for their assistance in the capture and recapture of goannas, most of the time in less than ideal circumstances.
I am indebted to Keith Newgrain for analysis of all isotopic samples, the staff at
Crocodylus Park, in particular Charlie Manolis and Yaakov Bar Lev who generously allowed me to use their crocodiles, and Eric Cox, Jason Stevens and Greg Brown at
Beatrice Hill Research Farm who provided logistical support. Peter Brocklehurst from the Northern Territory Government, Department of Natural Resources,
Environment and the Arts provided invaluable GIS base layers for use in many analyses. I would also like to thank Thomas Madsen for the generous use of his field laboratory, Paul Horner for assistance and access to museum specimens, Corey
Bradshaw for assistance in R programming, Jeremy Freeman for all things GIS,
Gavin Bedford for advice and Julian Gorman for all sorts of help, on and off the touch rugby field.
After much deliberation, Bob Richards let me continue to enter his barramundi farm to carry on catching V. indicus on his land, a decision that enabled us all to learn a little more about this amazing and reclusive creature. In addition, Rod and Lorraine kept me company and kept me laughing throughout my time in the mangroves. For this, and for all the after work beers, I thank the three of them.
Of my three supervisors, I thank Keith Christian for his unquestioning assistance,
Barry Brook for his immediate, concise and honest advice and most of all Tony
Griffiths, for this wonderful opportunity and his enduring patience.
There are so many other people in my life who deserve much kudos for their help in getting me through this, whether it was by distracting me with a frisbee, a guitar or a holiday, or by staving off my insanity by providing comprehensible discussions about statistics. Accordingly, all the people in this paragraph deserve equal ranking;
Lorrae McArthur, Wendy Telfer, Jennifer Koenig, Ben (I concur) Phillips, Ron Firth,
Chris Brady, Taegan (lil’ t) Calnan, Christopher (features) Schaefer, Brett Murphy,
James (Lyndley) Mckay, Cameron Fraser, Steve Comber, Steve (big hair) Sinclair,
Jeffrey Foucault, Tchavolo Schmitt, Florin Niculescu and Birelli Lagrene. To my parents, Jennifer and Gordon, my sisters Marinda and Samantha and to my beautiful
Miranda, I extend my deepest gratitude and love.
Jacinda Brown spent a joyous afternoon in the mangroves with me and I am grateful to her for some of the photos on the cover pages. Whoever reads this should head to
Mindil beach markets in Darwin and buy some of her pictures, there might even be a mertens or an indicus in there! Albert Koomen and the Australian Broadcasting
Commission also allowed me to use some stills that I had captured from the documentary “Goannas and Rubbish Frogs” that was made during this project. To the “other” indicus boys, Ron and Chris (and Geoffrey) I thank you mostly for keeping me laughing but also for keeping me cashed up.
And finally, I can’t finish without thanking the inimitable yet humble mosquito. My constant companion in the mangroves, my joints still ache just at the thought of you…
This work was funded by Charles Darwin University and a grant from the Natural
Heritage Trust (NHT). All research was conducted under Animal Ethics Permit No.
A00015 and Northern Territory Parks and Wildlife Permit No. 15254.
Table of contents
1. Introduction...... 1 Varanid phylogeny and taxonomy...... 1 Varanid habits and morphology...... 2 Varanid physiology...... 3 Significance of varanids in Australia ...... 4 Current threats to Australian varanid populations...... 5 Previous studies of varanid ecology...... 6 Seasonal responses by varanids in northern Australia...... 6 Varanids used in this study...... 8 Why use a comparative approach ? ...... 10 This study...... 10 2. The potential impact of cane toads on Australian reptiles ...... 13 Abstract ...... 13 Introduction...... 14 Methods ...... 19 Numbers of reptile species potentially at risk ...... 19 Reptiles’ tolerance to toad toxins...... 20 Results...... 27 Species potentially at risk from toads...... 27 Tolerance to toad toxins...... 27 Discussion ...... 33 Crocodiles ...... 34 Varanids ...... 35 Agamids...... 35 Freshwater turtles ...... 36 3. Study sites and general methods ...... 39 Abstract ...... 39 Study Sites ...... 41 Manton Dam (Varanus mertensi)...... 41 Adelaide River (Varanus indicus) ...... 41 Animal capture...... 44 Previous capture methods ...... 44
Capture methods used ...... 45 New pipe trap ...... 45 Effectiveness of design...... 46 Animal processing ...... 50 Radio transmitter implantation...... 50 Analytical framework...... 52 4. Using morphometrics to predict gender in varanids ...... 58 Abstract ...... 58 Introduction...... 59 Materials and methods...... 61 Model development...... 64 Model fitting ...... 65 Model selection...... 65 Characteristic body size ...... 68 Results...... 69 Combined adult and juvenile varanids...... 69 Adult varanids ...... 70 Juveniles ...... 71 Previous hypotheses ...... 71 Gender prediction in Varanus gouldii...... 72 Discussion ...... 74 5. Varanus indicus physiology...... 78 Abstract ...... 78 Introduction...... 79 Materials and methods...... 80 Radio telemetry...... 80 Thermoregulation...... 81 Field metabolism ...... 86 Energy budget calculations...... 86 Statistical analyses ...... 87 Results...... 88 Thermoregulation...... 88 Field metabolism, water flux and energy budget ...... 91 Discussion ...... 97
Thermoregulation...... 97 Water flux ...... 99 Field metabolic rates and seasonal energy budgets ...... 100 6. Home range and movements of V. indicus and V. mertensi ...... 104 Abstract ...... 104 Introduction...... 105 Materials and methods...... 107 Radio telemetry ...... 107 Home ranges of V. mertensi and V. indicus ...... 108 Analyses...... 109 Range overlap...... 110 Data analysis...... 111 Results...... 113 Home range ...... 113 Distances traveled ...... 121 Daily movements of V. indicus ...... 123 Overlap...... 123 Discussion ...... 126 Space use of semi aquatic varanids ...... 126 Distances traveled ...... 129 Varanus indicus daily activity ...... 130 Overlap...... 131 Home range estimation (V. indicus)...... 131 Conclusions ...... 134 7. Population dynamics of V. mertensi and V. indicus ...... 136 Abstract ...... 136 Introduction...... 137 Methods ...... 139 Regional climate...... 139 Study areas and capture details ...... 139 Model development and parameter estimation ...... 140 Survival model development...... 142 Survival estimation...... 142 Density estimation ...... 143
Results...... 144 Recapture...... 144 Survival...... 144 Density and sex ratios ...... 148 Discussion ...... 148 V. indicus recapture ...... 148 Survival...... 148 Which method best predicts varanid survival? ...... 152 Density...... 153 Sex ratios ...... 155 Conclusion...... 156 8. Synopsis ...... 158 Australian reptiles and the threat of cane toads ...... 158 The problem of sexing varanids in the field ...... 159 The physiology of Varanus indicus ...... 160 Intra and interspecific differences between terrestrial and semi aquatic varanids...... 160 Measuring varanid population dynamics...... 162 Conclusions ...... 162 9. References ...... 165
Table of Figures
Chapter 2 Cane toads and reptiles Figure 1. The approximate current 2004 and predicted distribution of the cane toad in Australia. ………………………………………………………….….16 Figure 2. Percentage reduction in speed as a consequence of toad toxin dose for 11 species of Australian reptile. ……………………………………………32 Chapter 3 Methods and study sites Figure 1. A map showing the location of the study region in Australia……………40 Figure 2. A map of Manton Dam, showing reeds and lilies (dark hatching) and rocky banks (no hatching), which provided access to the shoreline for noosing Varanus mertensi . ……………………………………………………….42 Figure 3. Aerial photograph showing discrete section of mangrove forest along the Adelaide River which was the study site for V .indicus...... 43 Figure 4. Number of captures of Varanus indicus using different baits in pipe traps (12 traps, 6 trap days per session). ……………………………………..63 Chapter 4 Gender prediction in varanids Figure 1. Morphometric measurements used in model construction. ……………63 Chapter 5 V. indicus physiology
Figure 1. Grand mean body temperatures (T b; as measured by radiotelemetry),
predicted T b (predT b), set point range, and operative temperatures (T emax
and T emin) as a function of time of day for free ranging V. indicus during the (a) wet season and (b) dry season…………………………….…….92 Chapter 6 Home range and movements Figure 1. Home ranges of Varanus indicus throughout 2002 and 2003, based on Minimum Convex Polygons clipped to remove unused habitat. ……………………………………………………………..…………..114
Figure 2. Schematic representation of the home ranges of V. mertensi around the bank of Manton Dam…………….……………………………………115 Figure 3. Relationship between range size and body mass for some medium sized varanids. ………………………………………………………………117 Figure 4. Relationship between home range and snout vent length (and 95% confidence intervals) in V. indicus as predicted by the model………..120 Figure 5. Mean distances (± SD) moved by V. indicus throughout the day during two weeks of radiotelemetry in the dry season of 2003……………………124 Figure 6. Mean (± SE) home range overlap between and within genders of a) V. indicus and b) V. mertensi . …………………….…………..………….125 Figure 7. Number of recaptures of V. indicus against maximum distance between the furthest two traps in which each individual was recaptured in. ……………………………………………….………………………..133 Chapter 7 Population dynamics Figure 1. Number of captures of V. indicus (bars) and proportion of captures/ recaptures (line) within each trapping period. ………………………..145 Figure 2. Male and female V. mertensi monthly survival probabilities (± 95% CI) based on model averaged values from all models in Table 2…………147 Figure 3: Relationship between age at maturity and known survival rates of lizards (closed circles) and snakes (open circles) (taken from Shine and Charnov 1992). ………………………………………………………………….151
Table of Tables
Chapter 3 Cane toads and reptiles Table 1. Species used for this study, collection localities, morphological statistics
and results of toxin trials.…………………..………..…..………………...21
Table 2. Australian reptile species potentially affected by the invasion of the cane
toad. …………………..…………………..…………………..…………..29
Chapter 4 Gender prediction in varanids Table 1. Number of specimens examined, mean snout vent lengths (SVL) ± SE and
subgeneric classification for six species of varanids measured…………..62
Table 2. Explanation of all derived parameters used in model construction………66
Table 3. Candidate model sets for predicting gender of each species, the hypothesis
examined under each model and references concerning sexual dimorphism
in varanids leading to each model. …………………..…………………..67
Table 4. Summary of second order Akaike’s information criterion (AIC c) and
associated statistics for all candidate models for the analysis of gender
prediction in adult and juvenile V. gouldii subsets. ……………………...73
Chapter 5 V. indicus physiology Table 1. Definitions of the thermoregulatory indices used. ………………………85
Table 2. The set point ranges and mean Tb (as measured in the lab) for V. indicus
and other varanids from the same region (data from Christian and Weavers
1996). …………………..…………………..…………………..………..89
Table 3. Summary of body temperature data and thermoregulatory indices of V.
indicus and three other varanids from the same region (Christian and
Weavers 1996) during the wet and dry seasons. ………………………….90
Table 4. Water flux rates, total body water (TBW = % body mass, derived from 18O
dilution), as determined from isotopic analysis, and water economy index
(WEI) for field active Varanus indicus during wet and dry seasons.
…………………..…………………..…………………….……………….93
Table 5. Field metabolic rate and water flux rates of V. indicus and another semi
aquatic varanid, V. mertensi (from Christian et al. 1996d) and two terrestrial
species from the same northern Australian region (from Christian et al.
1995). …………………..…………………..…………………..…………95
Table 6 . Carbon dioxide production, field metabolic rates (FMR), and mean body
mass of free ranging Varanus indicus during the wet and dry seasons…..96
Chapter 6 Home range and movements Table 1. Candidate model sets for predicting various spatial statistics in V. mertensi
and V. indicus , the hypothesis examined under each model and references
concerning space use in varanids leading to each model………………...112
Table 2 . Seasonal home ranges and summary data for Varanus indicus and Varanus
mertensi .……...……..……………..……………..………...…………….116
Table 3. Home range sizes and summary data for Varanus indicus and Varanus
mertensi . ……………..……………..……………..……………..………118
Table 4. Summary of model selection using AIC c for the analysis of home range size
in V. indicus and V. mertensi , using generalised linear modeling (family:
Gamma). ……………..……………..……………..……………..………119
Table 5. Summary of model selection using AIC c for the analyses of distances
moved in V. indicus and V. mertensi , using generalized linear modeling
(family: Gamma). ……………..……………..………………………….122
Chapter 7 Population dynamics Table 1 . Group and individual covariates used in the parameterisation of survival
and recapture probability models for Varanus mertensi and Varanus
indicus . ……………..……………..……………..………………………141
Table 2. Summary of Akaike’s information criterion (AIC c) and associated statistics
for candidate known fate models for the analysis of the survival probability
(S) of V. mertensi . ……………..……………..……………..…………...146
Table of Plates
Plate 1. An inaccessible section of bank within the Manton Dam study site, covered in Nymphaea violacea and Eleocharis dulcis...... 55 Plate 2. Section of the bank at Manton Dam (foreground) that is easily accessible by boat for spotting and catching Varanus mertensi ...... 55 Plate 3. A section of the Adelaide River V. indicus study site, showing mature Avicennia marina and Xylocarpus mekongensis along the creekline...... 56 Plate 4. A section of the V. indicus study site along the Adelaide River, dominated by Bruigeira gymnorhyza and Rhyzophora stylosa...... 56
Preface
The chapters presented in this thesis have been prepared as individual papers.
Consequently, some repetition of material is inevitable and unavoidable, however, attempts have been made to keep any repetition to a minimum. The inclusion of a co author on a paper reflects a supervisory role only, and the interpretation, ideas and conclusions as presented reflect solely my own original work, with the exception of
Chapter 2, where the designs of Phillips (et al. 2003) were expanded upon (with the primary author) to include reptiles other than snakes.
Abstract During the dry season in tropical northern Australia, when water and food are scarce but temperatures remain warm enough for activity, many terrestrial varanids lower their levels of activity (some to the point of inactivity). However, previous studies have shown that the semi aquatic varanid Varanus mertensi remains active throughout the year, maintaining a lower body temperature than terrestrial varanids.
Is the example found in V. mertensi indicative of other semi aquatic monitors? Here I describe, using a suite of novel and innovative sampling and analysis techniques, the population ecology of V. mertensi and another semi aquatic varanid species Varanus indicus . Specifically, I ask the questions of whether (i) V. indicus responds similarly to seasonality, and (ii) the home range and movement patterns of semi aquatic varanids also differ from fully terrestrial species. I also consider the likely impact of the invasive and toxic cane toad ( Bufo marinus ) on these species and other reptiles and to what degree it is possible to quantify the demography of free living varanids
(particularly when they are hard to sex in the field). I demonstrate that V. mertensi and V. indicus share broad similarities in life history traits with terrestrial varanids: males travel further in the breeding season; larger animals are more likely to move further; and home ranges overlap between conspecifics. However their semi aquatic lifestyle, while buffering them (both physiologically and ecologically) from seasonal water shortages, has implications for their use of space (in the case of V. indicus, leading to relatively small home ranges) and their population dynamics (high densities). In addition, I show that Bufo marinus poses a definite threat to these varanids along with many other reptile species. Lastly, I demonstrate that radiotracking is more robust than mark recapture for population monitoring of varanids.
Varanus indicus
Introduction
Varanus mertensi
1. Introduction 1
1. Introduction This thesis examines variation in the physiological, spatial, seasonal and population ecology of two semi aquatic varanids; Varanus mertensi and Varanus indicus . In contrast to semi aquatic varanids, terrestrial varanids are relatively well studied and show distinct intra and interspecific variation, including pronounced responses to seasonality. This chapter serves to review the current literature on both semi aquatic and terrestrial varanids, highlights the gaps in current knowledge that I will address, and culminates with the aims of the thesis research and a brief outline of each chapter.
Varanid phylogeny and taxonomy
All living varanid lizards (or ‘goannas’ as they are known in Australia) are members of the genus Varanus , the sole genus in the family Varanidae. Varanids are globally widespread with 54 species currently recognized; however, this number is constantly increasing as new species are described (e.g., Harvey and Barker 1998, Pianka et al.
2004). Varanids have probably retained the same basic body plan for a long time, as many fossils suggest a body plan similar to those of present day varanids (Pianka
1995).
The most recent work on the phylogeny of all varanids (Ast 2001) has recognised three major lineages within the genus Varanus : an African clade, which is basal to the rest of the group; an Indo Asian clade; and an Indo Australian clade. The monophyly of the Indo Australian clade, has been further supported by Fitch et al.
(2006). Within the Indo Australian clade the endemic Australian dwarf varanids
(Odatria ) are included as a sister group to the larger Australian varanids in the
1. Introduction 2 gouldii and varius groups (Fitch et al. 2006). Only four members of the Indo Asian group are extant in Australia ( V. doreanus , V. indicus , V. keithorni and V. finschi
(Pianka et al. 2004)).
Varanid habits and morphology
Possessing a generalised body plan and hampered by the putative constraints of ectothermy, one might conclude that varanids are too closely coupled to both physical forces and the spatial configuration of abiotic factors in their environment to allow for much adaptive radiation. Despite these apparent limitations, varanids have radiated extensively, particularly in Australia, and now persist in almost all climates and habitat types on the continent. Some have evolved a smaller body plan, and some gigantism; in fact the mass of the smallest and largest members of the genus varies by nearly five orders of magnitude (Pianka 1995). This enormous size range variation has enabled varanids to exploit many different habitats and climates. Such diversification includes species that are fossorial, arboreal, arenicolous or saxicoline in habit, whereas others have become semi aquatic and one has developed specialisations for coping with saline environments (Dunson 1974). A number of species utilise many of these strategies, to the extent that the now more geographically widespread species (e.g. V. gouldii ) have been able to exploit a wide variety of habitats and thermal environments.
Although many varanids are associated with aquatic habitats (riparian and floodplain associations), diversification by the genus into aquatic habitats has been limited. Of the 24 species of varanids in Australia (Cogger 2000), only four can be considered aquatic to any appreciable degree: Varanus mertensi, Varanus mitchelli, Varanus
1. Introduction 3 indicus and Varanus semiremex . This phenomenom is replicated in other squamate families, such as the agamids where there is only one (out of approximately 63 species) semi aquatic Australian species; the Eastern Water Dragon Physignathus lesueurii. In the varanids, the distributions of all the semi aquatic species lie in the tropics, where average year round temperatures are higher and hence restrictions to activity on ectothermic animals are reduced. Adaptations for an aquatic lifestyle include a laterally compressed tail, with a dorsal ‘fin’ in the final one third of the tail and elevated nostrils (Bedford and Christian 1996). Interestingly, all aquatic varanids in Australia have also evolved some extent of arboreality, although this trait is not restricted to aquatic varanids (e.g. V. scalaris , V. tristis , V. gilleni are non aquatic but arboreal).
The dietary habits of all varanids are fairly similar in that most are generalist carnivores, consuming high proportions of small invertebrate prey and a smaller proportion of larger vertebrate prey (Shine 1986, Bennett 1998). However, two species are known to be exclusively frugivorous (at certain times of the year, V. olivaceus and V. mabitang , Bennett 1998). Varanids activity varies from wide ranging active foragers, to some species that are considered almost purely ambush predators (Peters 1973, Auffenberg 1981, Dryden et al. 1990, Sweet 1999, Pianka et al. 2004).
Varanid physiology
The amount and predictability of energy (food and temperature) available to organisms is crucial both in influencing individuals’ ability for maintenance, growth, survival and reproduction, but these energetic factors may also influence the
1. Introduction 4 potential for adaptation and diversification (Huey et al. 2001). Therefore diversification in the ecology of a group of species that possess a relatively generalised body plan such as varanids may be associated with a suite of physiological and behavioural, as well as morphological adaptations, in response to different climates and resource conditions.
There have been various comparisons of variability in physiological traits for varanids including aerobic capacity (Christian and Conley 1994, Schultz 2002), heating and cooling rates (Wikramanayake and Dryden 1999) and standard metabolic rates (Thompson and Withers 1994). These studies indicate significant variation in physiological tolerances, which have been equated to differences in foraging mode.
They also indicate that many physiological parameters of varanids, such as standard metabolic rate and maximum oxygen consumption, are similar to other lizards. Thus, varanid lizards, may be as physiologically diverse as other lizard families.
Significance of varanids in Australia
The Australian continent contains most varanid species, with as many as six sympatric species in the arid regions of the interior (Pianka 1995) and up to eleven broadly sympatric species in tropical northern Australia (Cogger 2000). They are important components of the Australian biota, due to the roles they play in complex food webs and it has been suggested they are occupying the role of larger placental carnivores in some ecosystems (Pough 1973, Losos and Greene 1988). Varanids are also of enormous significance for Aboriginal people, both as a food source (Bomford and Caughley 1996, Vardon et al. 1998) and as cultural totems (Meehan 1982,
Altman 1987).
1. Introduction 5
Current threats to Australian varanid populations
Encroaching urbanisation and habitat alteration impact on Australian varanids.
Nevertheless, most varanid species in Australia have stable conservation status at present due to their large distributions and the lack of heavily trafficked roads across much of the continent (King and Green 1999).
Although no Australian varanids are listed as threatened under the Commonwealth
Environment Protection and Biodiversity Conservation Act 1999 (Commonwealth of
Australia 1999), nine species from the Northern Territory have been upgraded recently to threatened status because an introduced anuran potentially can have significant consequences for the future persistence or distributions of many species
(NRETA 2006). The cane toad ( Bufo marinus , Bufonidae) is a large (up to 230 mm body length) anuran native to South and Central America (Zug and Zug 1979) and represents a novel, abundant and highly toxic prey item to Australian predators, many of which are varanids. Doody et al. (2006) recently recorded declines in V. panoptes from the Daly River Region, immediately after the arrival of cane toads.
The habitat preferences of semi aquatic varanids in particular overlap greatly with those of cane toads, and it is expected that varanid populations will rapidly decline as a consequence of this immediate and continual sympatry. There is almost no baseline data on varanid population dynamics at present, and importantly, the opportunity to study large populations of varanids in the wet dry tropics of northern Australia is rapidly diminishing.
1. Introduction 6
Previous studies of varanid ecology
The majority of past autecological varanid studies have concentrated on a variety of aspects of varanid ecology: habitat selection and home range (Green and King 1978,
Gaulke 1992, Thompson 1992a, 1993, Weavers 1993, Philipps 1995, Traeholt 1995,
Sweet 1999, Thompson et al. 1999), field thermoregulatory and metabolic studies
(Stebbins and Barwick 1968, King 1980, King et al. 1989, Thompson 1992b,
Christian and Conley 1994, Traeholt 1995, Christian and Weavers 1996, Thompson and Withers 1998, Guarino et al. 2002), diet (Shine 1986, Gaulke 1991, James et al.
1992, King 1993, McCoid and Witteman 1993) and distribution patterns (Woinarski
1992, Woinarski and Gambold 1992). One longer term study of a varanid population has estimated density and sex ratios within a population (Auliya and Erdelen 1999), but there has been to date only one study (James 1996) that used mark recapture methods to quantify varanid population dynamics. A lack of population level studies prohibits well informed, long term management decisions for existing populations
(Caughley and Gunn 1996). One factor that has hindered such studies is the difficulty in obtaining high numbers of animals and subsequent recaptures from the field
(Shine 1986, Sweet 1999).
Seasonal responses by varanids in northern Australia
The tropical north of Australia has a monsoonal climate, which is characterised by annually dichotomous seasons (Bowman 2002). Very little rain falls throughout the dry season (May to October) whilst heavy rain is characteristic of the wet season
(November to April). During the dry season the average total precipitation is 122 mm, average maximum and minimum temperatures are 31 and 21ºC respectively, and average relative humidity ranges from 67 to 43% (Darwin Airport, Bureau of
1. Introduction 7
Meteorology 2004). These data are in striking contrast to the wet season, when the average total precipitation is 1592 mm, the average temperature range is between 32 and 24ºC and the average relative humidity ranges from 77 to 67% (Darwin Airport,
Bureau of Meteorology 2004). A further two transitional seasons can be discerned from within these (McDonald and McAlpine 1991), thus defining four seasons; early and late wet and early and late dry. The early wet and the late wet seasons are both characterised by high humidity and little rainfall.
The majority of terrestrial varanids in northern Australia occupy habitats that are influenced strongly by seasonality (Pianka et al. 2004). In response to dry season conditions, when food and water become scarce, these species show much reduced, or even no activity until the first rains of the wet season (Shine 1986, Christian et al.
1995, Christian and Bedford 1996, Christian et al. 1996b, Sweet 1999). Varanids in regions where water never disappears appear to be more buffered from these strong effects of seasonality (Shine 1986, Christian et al. 1996d).
Most of the river systems in Australia are comparatively short and coastal or ephemeral and many of these annually inundated regions dry out in the late dry season (October) leaving only remnant, deeper pools and billabongs still containing free water (Cowie et al. 2000). These deeper pools provide dry season refuges for many species including freshwater crocodiles Crocodylus johnstoni (Webb and
Manolis 1993), many species of birds (Corbett 1987, Marchant and Higgins 1990,
Whitehead et al. 1992), mammals (Corbett 1994), snakes (Madsen and Shine 1998,
Brown et al. 2002) and varanids (Shine 1986, Pianka et al. 2004).
1. Introduction 8
Another ecosystem that appears to be buffered from the high seasonality in northern
Australia are mangroves. Mangrove communities tend to be located within sheltered coastal areas, surrounding highly indented estuaries and offshore islands protected by
2 reefs and shoals (Tomlinson 1986). Over 4,000 km of mangroves are found along the 10,953 km Northern Territory coastline (Brocklehurst and Edmeades 1996).
Mangroves provide breeding, nesting, foraging and shelter sites for a variety of birds, reptiles, amphibians, and both terrestrial and aquatic mammals (Saenger et al. 1983,
Hamilton and Snedaker 1984, Noske 1996, Blamires and Nobbs 2000, Macintosh and Asthton 2002, Noske 2003, Pianka et al. 2004).
Examining the population ecology of V. indicus and V. mertensi , which both live in these seasonally buffered systems, forms the basis of this study. Importantly, both varanids could decline in numbers as a result of the cane toad invasion.
Varanids used in this study
Merten’s water monitor Varanus mertensi (Glauert 1951) is a semi aquatic varanid that is widely distributed across northern Australia, including Queensland, Western
Australia, and the Northern Territory. This species is highly adapted to freshwater streams and permanent water holes and hence its distribution is discontinuous throughout its range, particularly during the dry season (May October) when there is little water available (Shine 1986). This fact in particular highlights their potential vulnerability, as it is likely that cane toads will also congregate in these areas in the dry season. Some previous work on V. mertensi include a study of its energetics and waterflux of free living adults (Christian et al. 1996d, Thompson and Withers 1998), diet and foraging mode (Mayes et al. 2005b), and a comparative study of the diets,
1. Introduction 9 habits and reproductive biology of four species including V. mertensi (Shine 1986).
Varanus mertensi is unlike most varanid species in northern Australia in that they select body temperatures that are significantly lower than other terrestrial species
(Christian and Weavers 1996, Christian et al. 1996d) and can remain relatively active throughout the year, probably because their food supplies and habitat are not limited in the dry season (Christian et al. 1996d).
The mangrove monitor Varanus indicus (Daudin 1802) has a large distribution that extends throughout much of Micronesia with the southern extent of its range encompassing northern Queensland and the Northern Territory. Throughout its range, the size, pattern and scalation of populations are highly variable (Bennett
1998). In the Northern Territory V. indicus is associated almost entirely with the estuarine mangrove systems that line the major rivers. Many of these river systems in northern Australia have adjacent floodplains which are predicted to contain high numbers of cane toads in the near future (Freeland 1986), therefore the long term viability of many V. indicus populations remains unclear. There have been no long term studies of V. indicus , the majority of information in the literature consists of observations on their diet and reproductive habits on islands where they have been introduced (Dryden 1965, Wikramanayake and Dryden 1988, McCoid and Hensley
1991, Sprackland 1997). It is likely that V. indicus will show the same year round activity as V. mertensi , as mangroves are highly productive environments (Lugo and
Snedaker 1974, Clough and Attiwill 1982, Finlayson et al. 1988) and water is permanently present in the Northern Territory's large rivers where mangroves persist.
Thus, the metabolic and water turnover rates of V. indicu s are predicted to be more similar to those known for V. mertensi (Christian and Weavers 1996, Christian et al.
1. Introduction 10
1996d, Thompson and Withers 1998) than to other varanid species that decrease their activity in the dry season (Thompson 1992b, Christian and Bedford 1996, Christian et al. 1996b).
Why use a comparative approach ?
The apparent convergence of V. mertensi and V. indicus from two separate and distinct evolutionary clades into at least superficially similar semi aquatic habits in two vastly different habitat types warrants further attention. Thus, comparisons of the physiology and the resulting comparative ecology of these two species can reveal some insight into the breadth of adaptations shown by varanids in Australia.
This study
The main aims of this thesis are to: i) investigate whether the example of constant activity found in V. mertensi is replicated in V. indicus and; ii) to examine how different V. mertensi and V. indicus are to published information on terrestrial varanids, both ecologically and physiologically. Secondarily, I investigate the potential threat of Bufo marinus on varanids and other reptiles, and examine the use of morphometrics to predict gender in varanids.
This thesis is presented in 8 chapters. Chapters 2 and 4 6 are written as a series of papers that compare the population ecology and physiology of V. mertensi and V. indicus as well as highlighting the potential threat posed by Bufo marinus . A brief synopsis of each chapter is given below.
1. Introduction 11
Chapter 2 highlights the risks posed by cane toads to many of Australia’s reptile fauna, by examining the effects of toad toxin on numerous species from various taxa and by demonstrating the probable overlap of known reptiles’ distributions with the predicted distribution of the cane toad.
Chapter 3 describes the study sites and the general methods used in this study, including analytical methods, and a new trapping technique designed to increase capture rates of varanids in tidal, estuarine systems.
Chapter 4 is also a largely methodological description. It explains the development of models in an attempt to predict the gender of wild captured varanids based on various morphometric characters of x rayed and museum specimens (i.e. animals of known sex). Gender determination is important for field studies of varanids, as the results of other studies have been limited by the problems associated with accurate sex determination.
Chapter 5 explores the comparative physiology of V. mertensi and V. indicus . The thermoregulatory characteristics of V. indicus are examined in light of what is known of V. mertensi and other varanids. The field metabolic rates and rates of water turnover of V. indicus between seasons are then examined.
Chapters 6 and 7 examine the spatial and temporal population ecology of V. mertensi and V. indicus . In Chapter 6 the home ranges and movement patterns of each species are documented. In Chapter 7, the influence of season, gender and body size on the
1. Introduction 12 survival of these two species are examined and two techniques for measuring the survival of free living varanids are then compared.
Chapter 8 contains a synthesis of the physiology and ecology of the two species, a synopsis of conclusions presented in the previous chapters and suggestions for further research.
The potential impact of cane toads on Australian reptiles Lizard runway
Bufo marinus
Cr ocodile runway
Crocodile runway This chapter has been published as: Smith, J.G. and Phillips, B.J. (2006) Toxic tucker: the potential impact of cane toads on Australian reptiles. Pacific Conservation Biology 12:40 49
Varanus panoptes in lizard runway
2. The potential impact of cane toads on Australian reptiles 13
2. The potential impact of cane toads on Australian reptiles
Abstract
Cane toads are a highly successful invasive species, having invaded more than 20 countries in the last 150 years. In Australia, they currently occupy more than 1 million km 2. Toads are highly toxic and Australian predators have no evolutionary history with the cardiac toxins in toad skin. As such, toads constitute a novel and extremely toxic prey for Australia’s predators. Australia’s reptiles are perhaps the largest group likely to be affected by the invasion of the toad. By examining species distributions, I conclude that 59% of agamids, 85% of the varanids and all of
Australia’s crocodiles and freshwater turtles are potentially at risk from toads. I then assayed eleven species of reptile; one freshwater turtle (Chelidae), two crocodiles
(Crocodylidae), two dragons (Agamidae), one python (Pythonidae) and five species of varanid (Varanidae) for resistance to toad toxin. I found a high level of variation between species in resistance to toad toxin but in all cases (except for one species of crocodile) all species were easily capable of eating a toad large enough to kill them.
I conclude that toads pose a real and ongoing threat to the majority of Australian reptile species I examined.
2. The potential impact of cane toads on Australian reptiles 14
Introduction
The introduction (intentional and accidental) of new organisms into existing native habitats is potentially one of the most detrimental processes affecting biodiversity conservation (Diamond 1989, Mack et al. 2000, IUCN 2001). New organisms can alter the ecology of invaded habitats in many ways including introducing new diseases, altering the vegetative structure, or preying on native species (Williamson
1996, Sandlund et al. 1999). The outcomes of ecological invasion vary considerably among systems, but potentially one of the most powerful effects involves the invasion of a toxic species into the range of native predators that have had no previous exposure to such toxins (e.g. Brodie and Brodie 1999). In such cases native predators may be unable to tolerate the novel toxin and may die in large numbers as they first encounter the invader.
The cane toad, Bufo marinus (Bufonidae), is a large (up to 230 mm body length) anuran native to South and Central America (Zug and Zug 1979). The species is a highly effective invader of new ecosystems; its distribution now extends to more than twenty new countries throughout the Caribbean and Pacific (Lever 2001) and it has recently been listed as one of the world’s top 100 most invasive species (IUCN
2001). The success of toads as a feral invader can probably be attributed to both their high fecundity and the fact that all life history stages are toxic (Flier et al. 1980,
Lawler and Hero 1997, Crossland 1998, Crossland and Alford 1998). The active principles of the toxin – bufogenins – are extremely powerful (Chen and Kovarikova
1967) and, as a defensive toxin, unique to the Bufonidae (Daly and Witkop 1971).
2. The potential impact of cane toads on Australian reptiles 15
In Australia, toads were introduced in 1935 as an agent for biological control by the sugar industry (Lever 2001). They have spread from their initial release points in eastern Queensland (Qld) to encompass more than 863,000 km 2 (50% of Qld, Sabath et al. 1981, Sutherst et al. 1995). Cane toads now extend into northern New South
Wales (NSW) and the Northern Territory (NT) and are predicted to further increase their range, primarily throughout coastal and near coastal regions of tropical
Australia, to encompass an area of approximately 2 million km 2 (Sutherst et al. 1995;
Fig. 1). Cane toads can reach extremely high densities in suitable habitat (up to 2138 individuals ha 1, Freeland 1986). Prior to this invasion, Australia had no Bufonid taxa (Lutz 1971). Toads thus represent a novel, common and highly toxic prey item to Australian predators.
There has been limited study of the impacts of the cane toads on native fauna throughout its introduced distribution. In Australia, it was not until the early 1960s
(> 25 years after the initial introduction) that anecdotal reports of population declines in native species became apparent: Breeden (1963) reported observations of declines in snakes, varanids ( Varanus spp .), frilled lizards ( Chlamydosaurus kingii ) and quolls (a marsupial carnivore, Dasyurus spp .) following the appearance of toads.
This was followed by observations of declines in snakes, varanids and birds following the arrival of toads in south eastern Queensland and northern New South
Wales (Pockley 1965, Rayward 1974). Covacevich and Archer (1975) provided further anecdotal evidence for the potential impact of toads on predators by collecting numerous reports of terrestrial predators (snakes, varanids, and marsupial carnivores) dying as a consequence of attempting to ingest toads.
2. The potential impact of cane toads on Australian reptiles 16
Current cane toad distribution Predicted 2030 under current climate Predicted 2030 global warming Australia
0 500 1000 1500 Kilometers Kilometres N
Figure 1. The approximate current 2004 and predicted distribution of the
cane toad in Australia. Predicted distribution is shown under
current climatic and 2030 global warming scenarios (after
Sutherst et al. 1995).
2. The potential impact of cane toads on Australian reptiles 17
Large, mobile predators are typically rarely encountered, wary, and secretive. Owing to this, the impact of the toad invasion on terrestrial predators has proved difficult to quantify (e.g. Catling et al. 1999). Pre invasion estimates of abundance are difficult to obtain and consequently are often not available. In the face of such challenges researchers have tended to focus on the impacts of toads on smaller more abundant organisms – typically potential prey or competitors. Quantitative data on interactions between small, abundant native species (primarily fish, frogs, and aquatic invertebrates) and toads have become available in recent years (Freeland and Kerin
1990, Crossland 1998, Crossland and Alford 1998, Catling et al. 1999, Williamson
1999, Crossland 2000, 2001). Several of these studies have concluded that the ecological impact of toads may be less extreme than might be anticipated (e.g.,
Freeland and Kerin 1990, Catling et al. 1999, Williamson 1999). This maybe an erroneous conclusion to reach in the absence of information on the effects of toads on large predators, particularly given the potentially important role large predators play in regulating communities (Pace et al. 1999, Terborgh et al. 1999, Terborgh et al. 2001).
Despite the fact that terrestrial predators were identified more than 30 years ago as the group most likely to be at risk from toads (Pockley 1965, Rayward 1974,
Breeden 1963, Covacevich 1975), there has been little published quantitative analyses of the potential or realized effects of cane toads on these species. Even if competition between toads and small vertebrates is minor and their role as predators on invertebrates is modest, they might still impose a massive ecological impact if they kill a high proportion of the anurophagous predators that attempt to ingest them.
2. The potential impact of cane toads on Australian reptiles 18
Recently, Phillips et al. (2003) examined the potential effect of toads on Australian snakes, and concluded that approximately 30% of the terrestrial snake fauna was at risk. Indeed, many of Australia's reptiles are potentially at considerable risk from toads as many species prey upon frogs (Losos and Greene 1988, Shine 1991a) and, unlike birds or mammals, reptiles have few options for prey manipulation. Reptiles often use their mouths to consume prey items in their entirety and, hence, cannot avoid direct exposure to toxins in the toad’s body.
It can be seen that there is a critical need to evaluate the severity of the probable impact of cane toads on Australian reptiles other than snakes. To understand the potential impacts, information is required in two separate areas: (1) how many
Australian species are potentially vulnerable to toads, based on their geographic distributions and dietary habits, where known (i.e., how many species are likely to eat toads and live in areas that toads will occupy); and (2) how many species can tolerate a quantity of toxins equivalent to ingesting a toad (i.e. what is their sensitivity to cane toad toxin)?
To answer these questions I reviewed published information on distributions and dietary habits of Australian reptiles and tested the ability of 11 “theoretically at risk” reptile taxa to tolerate toad toxins. The potential effect of cane toads on Australian snake species has been investigated in an earlier paper (Phillips et al. 2003), so with the exception of water pythons Liasis fuscus (which were not previously assessed) this chapter focuses on other reptile taxa.
2. The potential impact of cane toads on Australian reptiles 19
Methods
Numbers of reptile species potentially at risk
To identify species that might be affected by the invasion of the toad, I used ecoclimatic predictions of the likely eventual distribution of toads within Australia
(Sutherst et al. 1995) to identify those species that will or have come into contact with toads. For each species I estimated the percentage of their range currently encompassed by toads and the percentage of their range that is likely to become affected as toads reach the full extent of their likely range. This approach allowed me to estimate the relative risk that toads could pose to each species based on the percentage of the species’ range that will eventually be encompassed by toads. I excluded skinks, geckos and pygopodids from the analysis because the small size and habitat preference of the majority of these species makes them unlikely to be affected. Nevertheless there are likely to be some species in these groups that are affected so this analysis is likely to underestimate the total number of species affected by toads.
Sutherst et al. (1995) generated two maps of the likely final distribution of cane toads in Australia, one under the present climate and one under a conservative 2030 climate change scenario. The latter method produced a slightly larger predicted distribution for toads. I used both maps for the analysis (Fig. 1) but note that even the larger potential range might be a conservative estimate owing to adaptation by toads or lack of competition from congeners increasing its range outside the ecoclimatic envelope of its native range (used by Sutherst et al. 1995 to generate predictions for the Australian invasion). Thus, these estimates of the reptile species potentially affected may be conservative.
2. The potential impact of cane toads on Australian reptiles 20
The taxonomic breadth, sheer number of species potentially affected and patchy literature precluded a detailed study of the likelihood that each species will include toads in their diet. Additionally, because toads have life history traits that make them unusually likely to become prey for diurnal movement oriented predators such as agamids (see discussion) one might find no published accounts of consumption of anurans by predators but still expect that toads may become prey for these species.
While diet and the likelihood of consuming a toad are clearly very important to the potential impacts on each species, a careful analysis of each group and their likelihood of encountering and consuming toads is addressed here (see discussion).
Reptiles’ tolerance to toad toxins
I tested 11 species of Australian reptile for their susceptibility to toad toxin. As populations that are currently sympatric with toads may have adapted to this novel prey type, animals were collected from areas where toads were absent. Table 1 lists the species studied, with information on their body sizes and localities of collection.
Other than the crocodiles (which came from a captive population) all animals were collected from the field and tested within 2 d of capture. The study taxa included one python (Pythonidae), one freshwater turtle (Chelidae), two crocodiles
(Crocodylidae), two dragons (Agamidae) and 5 varanid species (Varanidae). These species were chosen because they were all identified as “at risk” and were sufficiently common at the study sites to enable collection. Animals that were obviously ill or in poor condition where excluded from this study. All animals
(except crocodiles) were kept outside in dry cloth bags that were contained in plastic boxes, and therefore subject to ambient
Table 1 . Species used for this study, collection localities, morphological statistics and results of toxin trials. Total n refers to the number of individuals tested for toxin resistance. Collection localities were all in the Northern Territory, Australia. Numbers in parentheses represent standard errors. Gape width is the distance across the head at the hinge of the jaw. The ID 50 for each species is expressed as (1) the dose of toxin per body mass of individual (0.002ml/g) labelled as ID50 (conc), (2) The absolute dose based on the weight of an average individual, expressed in milligrams of dried toad skin equivalent ABS ID50 (mg) and (3) as the percentage of the average animal's gape width that a toad’s head width, whose size is sufficient to provide the absolute dose, represents ID50 % Gape (see text for details).
Species Family Location total SVL (mm) Weight (g) Gape width ID50 (conc) ABS ID50 (mg) ID50 % n (mm) Gape
Crocodylus porosus Crocodylidae Captive 14 205.9 (32.8) 1003.8 (120) 46.7 (1.5) - - - Crocodylus johnstoni Crocodylidae Captive 12 124.7 (21) 192.3 (7) 28.2 (1) 2.76 21.28 75.48 Chelodina rugosa Chelidae Fogg Dam 6 220 (18) 1519.8 (312) 35.8 (2) 0.18 18.38 51.29 Chlamydosaurus kingii Agamidae Gunn Point 9 245 (9) 394.8 (55.6) 48.3 (3.9) 0.36 16.02 33.16 Lophognathus temporalis Agamidae Darwin 10 103.7 (6) 30.3 (6) 30.3 (12.6) 0.81 11.03 36.36 Varanus indicus Varanidae Adelaide River 12 395.8 (22) 875 (160) 32.1 (2) 2.21 28.05 87.42 Varanus mertensi Varanidae Lake Bennett 2 525.0 1689 34.2 - - - Varanus mitchelli Varanidae Manton Dam 1 280 214 22.8 - - - Varanus panoptes Varanidae Beatrice Hill 8 500.4 (59) 2848.5 (704.6) 38.8 (4) 0.19 21.47 55.33 Varanus scalaris Varanidae Gunn Point 6 210 (7) 120.5 (12.8) 17.1 (0.6) 0.94 15.29 89.32 Liasis fuscus Boidae Fogg Dam 6 677.7 (277) 541 (178) 17.65 (2.6) 2.76 17.60 99.71 2. The potential impact of cane toads on Australian reptiles 22
temperature. Both species of crocodile were housed in a constant temperature room at 32ºC.
I obtained toad toxin from skins of 85 freshly killed cane toads collected from the
Katherine area (NT). Toads were killed by freezing. Single extractions of toad toxin were taken for the entire study to remove among toad variance in toxicity and accurately control dosing. Thirty five freshly killed toads were measured for snout vent length, head width, and mass. Dorsal skin was removed (from the back of the head to the knees) including the parotoid glands and skins and allowed to dry at room temperature over several days before being weighed. Dried skins were blended with 10x v/w of 40% ethanol, the mixture strained and the solids discarded. The resulting liquid was allowed to evaporate to 50% of its initial volume at room temperature. I recorded the final volume and dispensed the extract into 25 ml containers to be frozen. Bufogenins are stable, partially water soluble compounds with a very high evaporation temperature (Meyer and Linde 1971). It would be anticipated that the extract thus contains toad toxins, although it is possible that some were lost due to saturation (see discussion). Sub samples of the combined extract were used to dose the test animals to remove the effects of any among toad variance in toxic loadings.
I tested the resistance of individual animals to bufogenins using the decrement in locomotor speed following a dose of toxin (methodology modified from that of
Phillips et al. 2003). After measuring the animal’s mass, snout vent length (SVL) and gape width, each animal was subjected to a locomotor trial. Before dosing, I subjected each animal to two locomotor trials one hour apart. In each trial I recorded
2. The potential impact of cane toads on Australian reptiles 23 six measurements of locomotor speed. The fastest speed from each trial was taken and the resulting times averaged over the two trials. This yielded an estimate of maximum locomotor speed before dosing ( b). Most animals would run away from the holder as soon as released, the occasionally reluctant animal was encouraged to move by tapping on the tail.
The following day I gave each animal a specific dose of toxin through a feeding tube attached to a syringe or calibrated micropipette. The tube was inserted into the animal’s oesophagus to a depth of 30% of its SVL. Locomotor trials commenced one hour after dosing. As for before dosing, I ran two locomotor trials: one hour post dose and two hours post dose. Maximum locomotory performance (swim speed or sprint speed) was calculated as before to yield an estimate of maximum speed after dosing ( a). I then calculated the percent reduction in speed (%redn) following dosing for each animal (%redn = 100 x (1 b/a)). Previous experiments on Keelback snakes (Phillips et al. 2004) illustrate that reduction in locomotor performance following this methodology is due to the toad toxin and not the carrier fluid.
Given the taxonomic breadth and size range of the species tested, the same locomotor test could not be used for each species. The snake and turtle were subjected to swimming trials as per Phillips et al. (2003) in a 2.4 m diameter circular swimming pool with speeds taken for each quarter of the pool. Sub adult crocodiles
(<1 m total length) were swum in a 2.4 m trough divided into four 600 mm sections
(speeds taken for the two middle sections). Agamids and varanids were run along a
15 x 1 m runway with a concrete floor. Running speeds were taken from three 4 m sections of this run (allowing 1.5 m at each end for starting and stopping).
2. The potential impact of cane toads on Australian reptiles 24
Apart from the variations in testing maximum locomotory performance, all methodologies (number of trials, timing of trials) were carried out in the same way.
Given pairs of trials were conducted in the same manner I anticipated that the derived index of toxin resistance should be broadly comparable across taxa.
Because most of the data was collected in the outdoors, temperature could not be rigorously controlled across trials (with the exception of the crocodiles). For all other species I kept temperature differences between before/after trials within 2ºC by running the post dose trial at a time when the water or ambient temperature was similar to that of the pre dose trial. Although maximum speed may vary with temperature, the repeatability of speed assays in snakes, at least, has been shown to be consistent across temperatures (Brodie and Russell 1999). Thus, I expected the percentage reduction measure to be unaffected by temperature differences across sets of before/after trials. All animals were adjusted to trial temperature by being allowed to adjust to ambient or water temperature for a minimum of half an hour before each trial.
Each species was subjected to a range of toxin doses, with the exact range based on observed effects. To minimize mortality, I initially tested animals of each species on low doses: A weak or zero effect in a trial meant that the dose was doubled for the next animal, a lethal effect meant that the next dose was quartered. I tested each animal once only. Where sample size permitted, I tested multiple individuals at each dosage level. Dosage rates were calculated on a volume to mass ratio for each individual (0.002 ml/g of body mass). Different dosages were achieved by dilution
2. The potential impact of cane toads on Australian reptiles 25 of the original toxin extract with distilled water. Six initial dilution levels were used
(0.025x, 0.05x, 0.1x, 0.2x, 0.5x and 1x) with some species later given intermediate doses. Higher doses were achieved by successively increasing the dose per mass of undiluted extract (thus 2x = 0.004 ml/g, 4x = 0.008 ml/g).
This design yielded data on reduction in locomotor speed as a function of dose for each species, with the ultimate aim to determine a lethal dose for each species.
However traditional approaches to estimating lethal doses (i.e. LD 50 ’s) tend to result in high levels of mortality in study animals. Here I estimate the dose required to render 50% of animals incapable of movement (‘Immobility Dose’ ID 50 ). This dose tends to be associated with death of the study animal (Phillips, et al. 2003) and thus is probably a good surrogate for an LD 50 estimate.
In almost all cases there was a strong positive relationship between dose and percentage reduction in speed, within the range of doses that elicited an effect.
Percent reduction scores were transformed according to the following formula modified from that of Brodie et al. (2002):
y’=ln(2/ y 1), where y is the proportional reduction in locomotor speed (%redn/100). There were three instances where the proportion reduction was <0. Because these values do not transform correctly they were entered as a proportional reduction of 0.01 (following
Brodie et al. 2002). This transformation makes it simple to estimate the dose giving a 100% reduction in speed (The ID 50 , y = 1) when y = 1, y’ = 0. Thus the ID 50 is the
X intercept of the regression of y’ on dose which can be estimated as –α / β , where α
2. The potential impact of cane toads on Australian reptiles 26 is the intercept and β is the gradient of the line. Least squares regressions of y’ on dose were conducted for each species and ID 50 estimates made.
A species’ vulnerability to toads will be determined not only by the amount of toxin that it can tolerate, but also by the size of toads that it consumes relative to its own body mass. An animal that eats only very small toads might thus be able to survive ingestion, whereas one that takes larger prey relative to its own body size might exceed the lethal dose. Because reptiles are generally gape limited predators, an animal’s head size offers an index of the maximum size of prey that it can consume
(e.g. Shine 1991b). The head width of a toad large enough to contain a potentially lethal dose of toxin for an average sized specimen of each species was calculated, and that prey size was compared to the gape width of this average animal.
For species with sufficient data I calculated the ID 50 (in terms of absolute dose) for an animal of average body size. I then converted this dose into the equivalent mass of toad skin and used the toad morphology data (specifically, the relationship between toad body size and skin mass) to calculate the size of toad that would constitute this ID 50 . To compare this potentially lethal minimum toad size to the size of toad that a given species could physically ingest, I calculated the average mass and gape width for each species. I then divided the ID50 toad size (expressed as toad head width) by the mean gape width of each species to provide an index of lethal prey size relative to the animal's physical ability to ingest a prey item of that size.
That is, the head width of a toad of size sufficient to provide the ID 50 to an average sized animal was expressed as a percentage of mean gape width for each species.
Percentages of <100% indicated that the animal could easily ingest a lethal sized
2. The potential impact of cane toads on Australian reptiles 27 toad, whereas higher values make it increasingly unlikely that the animal could ingest a toad large enough to kill it.
Results Species potentially at risk from toads.
Analysis of the distribution of Australian lizards, crocodiles and freshwater turtles suggests that 75 species are potentially at risk from the invasion of the cane toad
(Table 2). These represent both species of crocodile, all 14 species of tortoise, 37 of
63 species of agamid (59%) and 22 of 26 species of varanid (85%). Of these 75 “at risk” species, 34 (45%) are likely to have their range totally encompassed by that of the toad (under predicted 2030 climate change) and 7 (9%) have already had their range totally encompassed. Sixteen of the 75 “at risk” species are already recognized as being threatened either at a federal or state level (Cogger et al. 1993).
Tolerance to toad toxins
For most species tested, the percent reduction in locomotor performance was highly associated with survival after ingestion of toxin: most animals with 100% reduction in speed died 1 2 h after dosing. Animals with <100% reduction generally recovered over the course of 8 24 h, however some (notably the agamids and C. johnstoni ) died
12 h subsequent to lower doses. For the purposes of this analysis, these individuals were scored as showing a 100% reduction in speed. After receiving high doses of toxin, three of the agamids (two L. temporalis and one C. kingii ) appeared dead (eyes closed, stiff and without movement) for many hours during and after trials, but after
24 h had recovered, appearing active and alert.
2. The potential impact of cane toads on Australian reptiles 28
In most species tested (8 of 11), a higher dose (ml toxin/g) resulted in a greater reduction in locomotor performance (Fig. 2). The exceptions to this were either due to very small sample sizes (e.g. V. mertensi ) or because a dose that caused an effect on locomotor performance (e.g. C. porosus ) was not reached.
Crocodilians were the least susceptible to toxin, so much so that an ID 50 estimate could not be generated for C. porosus as the sample size did not permit higher doses.
The estimated ID 50 for C. johnstoni was 15 times that of the lowest estimate ( C. rugosa ). Amongst the varanids, the estimated ID 50 of V. indicus was >10 times that of the lowest ( V. panoptes ), however these differences were not statistically significant ( F2, 11 =2.56, p=0.12).
Overall, I found high levels of variation between species in ID 50 estimates.
Nevertheless when I expressed ID 50 as a percentage of gape width (all relevant toad allometries had r > 0.96, data not shown), I found that all species tested (except C. porosus ) are capable of consuming a toad large enough to cause death (i.e. the percentage scores are less than 100%; Table 1). Functionally then, of those species for which I could generate an estimate, most species exhibited relatively low levels of resistance to toad toxin.
2. The potential impact of cane toads on Australian reptiles 29 Table 2. Australian reptile species potentially affected by the invasion of the cane toad. Columns 2 4 give the percentage of the species’ range encompassed by the toad currently and under the predicted distribution of toads (under present climate and a 2030 predicted climatic scenario).
Percent overlap Present Predicted Current 2030 Crocodiles Crocodylus porosus 57 100 100 Crocodylus johnstoni 65 100 100 Freshwater Turtles Chelodina expansa 40 47 53 Chelodina longicollis 38 38 41 Chelodina novaeguinae 91 100 100 Chelodina oblonga 0 14 60 Chelodina rugosa 71 100 100 Chelodina steindachneri 0 3 3 Elseya dentata 61 100 100 Elseya latisternum 88 100 100 Emydura kreftii 100 100 100 Emedura macquarii 25 25 30 Emydura subglobosa 100 100 100 Emydura victoriae 0 100 100 Pseudemydura umbrina 0 0 100 Rheodytes leukops 100 100 100 Agamids Amphibolurus muricatus 32 45 55 Amphibolurus nobbi 62 62 67 Amphibolurus norrisi 0 0 0 Caimanops amphiboluroides 0 0 0 Chelosania brunnea 33 100 100 Chlamydosaurus kingii 74 100 100 Cryptogama aurita 0 0 0 Ctenophorus caudicinctus 10 18 18 Ctenophorus clayi 0 0 0 Ctenophorus cristatus 0 0 0 Ctenophorus decresi 0 0 0 Ctenophorus femoralis 0 0 0 Ctenophorus fionii 0 0 0 Ctenophorus fordi 0 0 0 Ctenophorus gibba 0 0 0 Ctenophorus isolepis 0 0 0 Ctenophorus maculatus 0 19 34 Ctenophorus maculosus 0 0 0 Ctenophorus mckenziei 0 0 0
2. The potential impact of cane toads on Australian reptiles 30
TABLE 1. (continued) Percent overlap Predicted Present Current 2030 Ctenophorus ornatus 0 7 29 Ctenophorus pictus 0 0 0 Ctenophorus reticulatus 0 0 4 Ctenophorus rufescens 0 0 0 Ctenophorus salinarum 0 0 6 Ctenophorus scutulatus 0 15 15 Ctenophorus vadnappa 0 0 0 Ctenophorus yinniethara 0 0 0 Diporiphora albilabris 20 100 100 Diporiphora australis 100 100 100 Diporiphora bennettii 37 100 100 Diporiphora bilineata 14 100 100 Diporiphora convergens 0 100 100 Diporiphora lalliae 8 13 13 Diporiphora linga 0 0 0 Diporiphora magna 36 100 100 Diporiphora pindan 0 100 100 Diporiphora reginae 0 0 0 Diporiphora superba 0 100 100 Diporiphora valens 0 0 0 Diporiphora winneckei 0 4 4 Hypsilurus boydii 100 100 100 Hypsilurus spinipes 50 50 75 Lophognathus gilberti 24 41 41 Lophognathus longirostris 0 2 2 Lophognathus temporalis 43 100 100 Moloch horridus Physignathus lesueurii 100 94 100 Pogona barbata 48 50 60 Pogona microlepidota 0 100 100 Pogona minima 0 10 20 Pogona minor 0 3 3 Pogona mitchelli 0 7 7 Pogona nullarbor 0 0 0 Pogona vitticeps 0 0 0 Tympanocryptis adelaidensis 0 7 21 Tympanocryptis cephalus 0 0 0 Tympanocryptis diemensis 0 0 25 Tympanocryptis intima 0 0 0 Tympanocryptis lineata 6 11 13 Tympanocryptis parviceps 0 0 0 Tympanocryptis tetraporophora 12 14 16 Tympanocryptis uniformis 32 100 100 Varanids Varanus acanthurus 17 30 30 Varanus baritji 70 100 100 Varanus brevicauda 0 4 4 Varanus caudolineatus 0 4 5
2. The potential impact of cane toads on Australian reptiles 31
TABLE 1. (continued) Percent overlap Predicted Present Current 2030 Varanus eremius 0 0 0 Varanus giganteus 0 0 0 Varanus gilleni 0 1 1 Varanus glauerti 0 91 91 Varanus glebopalma 42 92 92 Varanus gouldii 20 22 25 Varanus indicus 75 100 100 Varanus kingorum 0 100 100 Varanus mertensi 78 86 86 Varanus mitchelli 39 100 100 Varanus panoptes 37 63 65 Varanus pilbarensis 0 0 0 Varanus keithorni 0 100 100 Varanus primordius 25 100 100 Varanus rosenbergi 0 6 33 Varanus semiremex 100 100 100 Varanus spenceri 64 50 50 Varanus storri 36 50 50 Varanus scalaris 16 100 100 Varanus tristis 21 33 34 Varanus varius 42 44 48
2. The potential impact of cane toads on Australian reptiles 32
100 A)
50 V. indicus
V. mertensi
0 V. mitchelli
V. panoptes
V. scalaris
50
100 B) Reduction in speed (%)
50 C. kingii
L. fuscus
C. johnstoni
0 L. temporalis
C. porosus
C. rugosa 50 1 2 3 4 5 6
Ln (100 x Dose)
Figure 2. Percentage reduction in speed as a consequence of toad toxin dose for 11
species of Australian reptile. The x axis, is Ln(100 x dose) where dose is
expressed as a concentration of toxin extract administered at a rate of
0.002 ml/g. Plotted points represent the mean value for all individuals tested at each dosage level (error bars included for two species, the rest
omitted for clarity). A) Shows all the varanid species tested, B) shows the
remaining taxa.
2. The potential impact of cane toads on Australian reptiles 33
Discussion
Cane toads are a major problem throughout the Caribbean and the Pacific having successfully invaded many countries previously naïve to toads. In Australia they have been spreading rapidly for 70 years, and warnings of their possible ecological impact on native fauna have been voiced throughout that period (Lever 2001). Most studies on toad impacts to date have examined the effects of toads on potential prey items, competitors and aquatic invertebrate predators. Several of these studies have concluded that toad impacts are likely to be minimal on this component of native fauna (Freeland and Kerin 1990). Unfortunately, a lack of effect at lower trophic levels reveals little about potential impacts on higher trophic level predators, a component of native fauna that may be affected.
Recent work by Phillips et al. (2003) using similar methodology to this study has shown that toads are likely to have a negative impact on greater than 30% of
Australia’s terrestrial snake fauna. These findings extend this result to encompass other native reptiles. Importantly, the results of the two studies are qualitatively similar – indicating most taxa tested are capable of ingesting a toad large enough to kill them. It is however important to note that these results are conservative for several reasons. First, I only extracted toxin from the dorsal skin of toads. Toxin that is present in the ventral skin and internal organs was not included in the extraction. Second, the extraction process is unlikely to have been 100% efficient and some toxin will have been lost (i.e. toxin may have reached saturation in solution). Therefore, the actual lethal dose in terms of toad size is likely to be even
2. The potential impact of cane toads on Australian reptiles 34 lower than those listed in Table 1. Third, many reptiles will take multiple prey items.
All calculations are limited to the effect of a single prey item. Fourth, low sample sizes do little to take into account the effects of intra specific variation in reposnse to toad toxin. Despite these conservative biases, they still suggest that many species and individuals tested can easily ingest a single toad large enough to be ultimately fatal
(Table 1).
Furthermore, my analysis of geographic distributions suggests that a high proportion of Australia’s varanids (85%) and agamids (59%) and all freshwater turtles and crocodiles will share at least a part of their future distribution with the toad. The exact level of impact for each species will depend upon the probability of encountering toads and the likelihood of consuming toads large enough to cause mortality. Unfortunately to date there are relatively few published data on the diet and foraging methods of many of the “at risk” species. Below I set out a brief discussion of foraging modes and diet for each of the sampled families (omitting the
Pythonidae which was dealt with by Phillips et al. 2003) commenting on potential risk to each family by ingesting toxic toads.
Crocodiles
Both species of Australian crocodiles are generally nocturnal foragers and live in aquatic habitats. Furthermore, amphibians have been found in the stomachs of both species (Webb and Manolis 1993) and even large crocodiles are known to eat small prey items (C. Manolis pers. com.). These factors suggest that crocodilians are at a high risk of encountering cane toads (breeding adults or tadpoles) and will probably attempt to eat them. Freshwater crocodiles C. johnstoni are at a higher risk given
2. The potential impact of cane toads on Australian reptiles 35 that the young individuals tested were susceptible to relatively small doses of toxin and that this species is more reliant on freshwater habitats than C. porosus . Larger specimens may be buffered in that they would need to consume much higher doses of toxin to kill them. As a crocodile grows it will certainly reach a size where ingesting even a large toad is unlikely to kill it. Therefore, small individuals are likely to be at higher risk, an outcome that is paralleled in snakes, albeit for different reasons (Phillips and Shine 2006). Saltwater crocodiles C. porosus are less at risk given that even comparatively large doses induced very little effect.
Varanids
Many varanids are active, far ranging, diurnal foragers and all have a well developed vomeronasal system for detecting prey (Greer 1989). Furthermore many species have a propensity to actively dig up potential prey items (King and Green 1999). Diurnal varanids may even encounter adult toads (which are nocturnal) in their search for food through digging them up. Given that all the varanids tested (3 out of 5) could easily ingest a toad of sufficient size to kill them, I predict that many species will be affected by the cane toad. At particular risk are large, wide ranging, and semi aquatic active foragers. These predictions are supported by anecdotal reports in
Burnett (1997) that document apparent declines in several species of varanids following the arrival of toads in an area.
Agamids
Australian agamids are primarily sit and wait predators (Greer 1989). They are extremely visually oriented and are believed to make limited use of chemosensory faculties in prey acquisition (Greer 1989). Probably because of this reliance on
2. The potential impact of cane toads on Australian reptiles 36 visual detection, prey acquisition is a strictly diurnal activity. A wide range of invertebrates has been found in the guts of Australian agamids but vertebrates also occasionally figure in their diet. As a group it seems highly likely that most agamids are likely to occasionally consume toads that are day active. Importantly, metamorph and juvenile toads are diurnally active (Freeland and Kerin 1991, Cohen and Alford 1993). Freeland and Kerin (1991) report juvenile toads (30 70 mm SVL) as being diurnally active (primarily in the morning). Metamorphs (<30 mm) were active throughout the day but were more restricted to the water’s edge. Importantly, the average size of diurnally active toads increased with distance from the breeding site, a finding confirmed by Cohen and Alford (1993). Based on these analyses, a toad of 30 mm SVL would be sufficient to kill most C. kingii (a relatively large agamid) less than 200 mm SVL, and all individuals of L. temporalis . A 70 mm toad would kill even the largest C. kingii . Toads thus pose a serious threat to agamids, particularly those in the vicinity of habitat that is suitable for toad breeding.
Freshwater turtles
Australian freshwater turtles forage exclusively in the water. Most species are omnivorous, subsisting on vegetable matter and carrion but taking animal prey when possible (Cann 1998). A few species appear to be almost exclusively carnivorous.
While it is unlikely that any Australian turtle could consume a mature adult toad
(although see Hamley and Georges 1985), the tadpoles, eggs, juveniles and metamorphs (all of which are toxic) are a likely prey item for many species, particularly those frequenting still water where toads are likely to breed. In these respects, turtles of the genus Chelodina are the most likely to be at risk from toads.
These long necked species are better equipped hunters and thus consume a higher
2. The potential impact of cane toads on Australian reptiles 37 proportion of animal prey (Cann 1998). Also they tend to inhabit slow moving to still waterbodies such as are favoured by toads as breeding sites. My data indicate that C. rugosa at least is reasonably sensitive to toad toxin. Interestingly, Hamley and Georges (1985) were able to maintain several Elseya latisternum on a diet of toads for several months with no apparent ill effects, suggesting that this species of turtle has a much higher resistance to toad toxin. Further work is obviously required to assess variation in resistance to toad toxin among Australian freshwater turtles.
This data suggest many species of Australian reptiles could be adversely affected by the continued invasion of the cane toad. The exact magnitude of the effect will depend on factors specific to each species (for example prey handling ability, prey choice) and whether or not populations can mount an effective adaptive response. It seems prudent however, to treat the invasion of the cane toad as a serious threat to
Australian reptiles, and wildlife managers should give serious consideration to the impact of the cane toad on these species.
Cane toads are a problem in more than twenty countries and research into the impacts of toads in these countries is likely to be hampered by the same difficulties as face research in Australia. In these countries, as with Australia, it is important to assess the impact of toads on large predators, despite the logistical difficulties.
38
Adelaide River study site
Study sites and general methods
Parts of this chapter have been published as; Smith J G (2003) Fish and company smell after three days: increasing capture rates of varanid lizards. Herpetological Review. 35(1): 41 43
Manton Dam study site
3. S tudy sites and general methods 39
3. Study sites and general methods
Abstract
This chapter presents the general methodologies used throughout this thesis. Included is a description of the broad habitat associations of the study species, selection of study sites, and the capture and processing methods for each species, including the development of a new passive trap for capturing goannas. The analytical framework described in this chapter forms the basis for all analyses in subsequent chapters, unless otherwise stated. This chapter serves as a methodological description for all other chapters, and only the variants in methodology used and the particulars of each statistical analyses pertinent to each chapter are described subsequently. All fieldwork was carried out between June 2001 and February 2004 in the tropical north of the Northern Territory, Australia (Fig. 1).
3. S tudy sites and general methods 40
¡ ¡ ¡¡ ¡ ¢ ¡ ¡¡ ¡ £ ¡ ¡¡ ¡ ¤ ¡ ¡¡ ¡
S tu a rt Adelaide
H ¥
§ ¨ ©
§ # ig §
§ h §
§ w §
© a River site § ¨ y
¥ Darwin
Arnhe m Hig
hway
¥
¨ §
§ §
§ §
§ §
§ §
§ ¨ ¥
Manton
Dam site
¥
§ ¦
§ ¥
§ §
§ §
¥ §
¦ § ¥
0 10 20Kilometres Kilometers
s N
¡ ¡ ¡¡ ¡ ¢ ¡ ¡¡ ¡ £ ¡ ¡¡ ¡ ¤ ¡ ¡¡ ¡
Figure 1. A map showing the location of the study region in Australia (bold square, left) and the loc ation of the two study sites in relation to the city of Darwin (above).
3. S tudy sites and general methods 41
Study Sites
Manton Dam ( Varanus mertensi )
A population of V. mertensi was investigated at Manton Dam, which lies 65 km south of Darwin, Northern Territory, Australia (12º 87' S, 131º 11' E; Fig. 2). It is a large (11,675 ha) reserve containing a man made freshwater body (330 ha).
Melaleuca leucadendra and Melaleuca argentea dominate the vegetation at the waters edge. Sections of the bank are lined with Nymphaea violacea and Eleocharis dulcis , for up to 10 m into the water (Plate 1), whereas in other sections the bank is rocky and sustains little vegetation cover (Plate 2). The quantity of emergent vegetation along the shoreline dictated to a large degree which sections of the bank were accessible by boat and therefore allowed access to animals.
Adelaide River ( Varanus indicus )
Lining the estuarine (northern) section of Adelaide River are many discrete patches of mangrove forest (12º 33’ S, 131º 22’ E). Adjacent to one section there is a fish farm with an associated access road that provides the only year round terrestrial access to mangroves of this type (Fig. 3). The majority of research conducted on V. indicus was at this site. Throughout the 27.2 ha site, the mangrove community along the creekline is dominated by Avicennia marina and Xylocarpus mekongensis (Plate
3) whilst adjacent to the river are predominantly Brugueira gymnorhiza and
Rhyzophora stylosa (Plate 4). Twice daily, the mangrove forest or a portion thereof
(dictated by tide height) is inundated to a maximum depth of 45 cm (on the landward edge). As there was a high potential of encountering estuarine crocodiles Crocodylus porosus , particularly adjacent to the creeklines, field work was limited to the low tide periods.
3. S tudy sites and general methods 42
Kilometres Rocky shore Reeds and Lilies
Figure 2. A map of Manton Dam, showing reeds and lilies (dark hatching) and
rocky banks (no hatching), which provided access to the shoreline for
noosing Varanus mertensi .
3. S tudy sites and general methods 43
Study area boundary
Kilometres
Figure 3. Aerial photograph showing the discrete section of mangrove forest along
the Adelaide River which was the study site for Varanus indicus . The
white line through the forest patch delineates the southern boundary of the
study area.
3. S tudy sites and general methods 44
Animal capture
Previous capture methods
Due to the difficulties associated with capturing free living varanids, many methods have been tried. Large varanid species that can be easily approached can be noosed, although this method can result in injury to the animal. Varanids can also be dug from or trapped outside their burrows (e.g., Varanus gouldii , King 1980) but this method is labor intensive and destructive to habitats. Some species ( V. tristis ) have been caught by following their tracks in the sand to their resting sites (Thompson et al. 1999). However, for large, wide ranging species and those using habitats where it is difficult for humans to move about freely, conventional active capture techniques cannot be employed.
Passive capture methods have proved effective for some smaller varanid species.
Pitfall traps have been used to capture V. brevicauda, V. eremius, V. gilleni, V. gouldii and V. tristis (Downey and Dickman 1993). Passive nooses (set in likely travel corridors) have been used for the more elusive rock dwelling species V. glebopalma and V . glauerti (Sweet 1999). While effective, these methods require constant monitoring (three times daily), a task unachievable over large areas, and are unsuitable in tidal environments such as mangrove systems.
To circumvent the problems posed by tides, Auliya and Erdelen (1999) devised a floating carrion trap (box trap) to capture the large water monitor V. salvator . These traps were baited with fish or chicken viscera and relied on a varanid entering the trap, grabbing the bait and triggering a mechanism to close the door behind it.
3. S tudy sites and general methods 45
Although effective, these traps were large, were erected on site and took 5 7 d to construct using local materials (bamboo and wood). Erecting large numbers of traps of this type is likely to prove prohibitively expensive. Importantly, throughout north
Australian estuarine environments the estuarine crocodile Crocodylus porosus may be attracted to and possibly interfere with carrion traps of this type, as well as posing a risk to field workers.
Capture methods used
Due to the differences in accessibility between each site, combinations of the following different methods were used to locate and capture each goanna species used in this study. At Manton Dam individual V. mertensi were located from the water using a small, motorised punt. The animal was then approached along the shoreline and captured using a noose attached to an extension pole, or less often, by hand. At Adelaide River V.indicus is quite abundant but difficult to sight and even more difficult to capture, due to its effective camouflage and tendency to climb high into trees (sometimes 10 m or more) when alarmed (pers obs). To overcome this problem, I developed a new passive trap which proved to be highly effective and easy to use.
New pipe trap
A specific baited arboreal pipe trap was devised that was suited to the mangrove environment and overcame the weaknesses of many previous trap designs. The traps were made from 1 m lengths of PVC stormwater pipe, sealed at one end with PVC push on end caps. Traps are placed vertically with the sealed end at the bottom and rely on gravity and the smooth sides of the pipe to prevent escape of varanids
3. S tudy sites and general methods 46 entering them. Traps could be made with practically any diameter, depending on the size of the target species. To allow drainage of water after heavy rains and facilitate spread of bait odors, three 6 mm holes were drilled through the end caps and three near the base (approximately 100 mm from the end cap) at equal intervals around the pipe. The lip of each trap was filed smooth to remove the sometimes sharp finish created when cutting from a larger length of pipe. Depending on the availability of materials (new vs. secondhand pipe), these traps were quite inexpensive to produce.
Traps were strapped vertically on trees using 10 cm wide adhesive tape. Attachment was quick, and the trap and animals readily removed by simply cutting the tape with a sharp knife. Brightly coloured tape was used to increase the visibility of traps when searching for them in dense forest. If the traps were placed high enough in trees they could be left set, undisturbed by the raised water levels caused by tidal or seasonal influences. However, low set traps present no drowning risk to these semi aquatic animals because they are able to escape the trap as water levels rise. High set traps were checked once each day with a small hand held mirror, used to peer into the entrance.
Effectiveness of design
To test the effectiveness of the design, 12 pipe traps of two different diameters were used (6 of 150 mm and 6 of 225 mm). These were set within a 2 ha patch in the
Adelaide River study site. Each trapping session took place over six consecutive days in early August 2001. Traps were baited with a small amount of meat (for 6 days) or fish (another 6 days) and placed (in the bottom of the trap), with the trap
3. S tudy sites and general methods 47 vertically on the main trunk of mature mangrove trees (predominantly Avicennia marina , Sonneratia lanceolata and Xylocarpus mekongensis ).
Pipe traps of both sizes were found to be highly effective for this species, with an overall trap success rate of 50% (36 captures in 72 trap days). Daily trap success rates (number of animals captured divided by total number of traps) ranged from 8%
(1 capture) to 75% (9 captures). These results compare very favourably with previous hand noosing results for V. indicus from the same area, which on average yielded less than one successful capture per day. Traps were found to be safe, with no trap mortality and no obvious trap related injuries. No trapped animals appeared to be agitated or exhausted on removal and release.
The effectiveness of these traps appears to be related to the pungency of baits used, as strong smelling baits (rotting fish) tended to catch more individuals (Fig. 4). Traps set with less “smelly” food items (pork or lamb off cuts, or recently bought fish) caught no animals. Animals were captured more frequently when the fresh fish had been in the traps for extended periods (i.e. once it had begun to decompose).
Recaptures occurred both at the same trap and at different trap locations. On two occasions, two animals were captured in the same trap.
As many other Australian varanids are attracted to carrion ( V. gouldii, V. panoptes,
V. rosenbergii , pers obs, King and Green 1999), these traps may prove to be useful in other habitats where conventional methods cannot be applied. Care should be taken however, to use the appropriate attractants. Decomposing fish was found to be most effective for this study, possibly due to the rancid odors that can be produced from
3. S tudy sites and general methods 48 small quantities of fish (Richards and Hultin 2001). While this trial used arboreal placement, the pipe trap design can readily be adapted for use in rocky environments or even in ground use when matched with an auger of appropriate diameter. This new pipe trap design provided an effective and affordable method for increasing capture rates of V. indicus . The method was tried for V. mertensi , but caught no animals over 5 d despite observations of many individuals throughout the area
(Manton Dam) in that time.
After determining that the trap design was effective, V. indicus was trapped quarterly between September 2001 and March 2004 using this method. Trapping was conducted within the same discrete 27.2 ha patch of mangrove forest along the
Adelaide River. Two parallel trap lines (24 per line) were erected throughout the 13.6 ha site (which encompassed beyond where radiotracking occurred, see chapter 6).
Traps were placed 50 m apart running parallel with the edges of the mangrove strip and Adelaide River, baited with fish and checked every day for six days. To minimize the effects of the large tidal fluctuations experienced throughout the region, trapping during each period was conducted around the spring tides for that month which ensured that the tides were highest at this time. This provided standardization of any possible differential behaviour of V. indicus during different tides and was chosen because V. indicus also appear to be very active during these times (pers obs).
3. S tudy sites and general methods 49
18 Raw pork or chicken 9 16 Fresh Fish 8 8 Rotting Fish 14 7 7 12 6 10 8 4 6 3 2
% of total captures 4 1 2 0 0 0 0 0 0 0 0 0 1 2 3 4 5 6 Trap day
Figure 4. Number of captures of Varanus indicus using different baits in pipe traps
(12 traps, 6 trap days per session). Numbers above bars indicate actual
numbers caught. Baits were retained unchanged for the 6 d sampling
period. Raw pork or chicken caught no V. indicus .
3. S tudy sites and general methods 50
Animal processing
Initially, all varanids captured were given an individual identification number with a transponder (Digivet) implanted subcutaneously under loose skin on the neck.
Subsequent captures allowed these individuals to be readily identified using a scanner (Destron Fearing, U.S.A.). At each capture, the following data were recorded: date, time, mass to the nearest gram using an electronic balance (Scaleman,
U.S.A.), snout vent length (SVL), head length (HL), head width (HW), tail length
(TL) where measured to the nearest cm, and gender (where possible, by manual eversion of hemipenes and/or by morphology), reproductive condition and any visible scars were noted. Varanids were released at the same point of capture within
1 hour.
Radio transmitter implantation At each site, a sub sample of varanids was implanted with radio transmitters
(Holohil SB 1 and SB 2, Canada). As stated above, gender was determined by palpating the genital region in an attempt to induce males to evert their hemipenes, but because the gender of many species of goanna is notoriously difficult to determine in the field (Shine 1986, Gaulke 1997), after June 2001, all varanids that were to undergo implantation surgery at Charles Darwin University were taken to
Parap Veterinary Clinic and the basal section of the tail and pelvis was radiographed to determine gender (Shea and Reddacliff 1986). For V. indicus individuals were determined to be males if the four anterior pointing spines typical of hemibacula of the indicus group (Ziegler and Böhme 1999) were evident. The gender of varanids with incompletely ossified spines could not be resolved. Adult V. indicus (SVL >320
3. S tudy sites and general methods 51 mm, Wikramanayake and Dryden 1988) displaying no ossification were classified as females.
Implantation surgery involved placing the head and anterior portion of each goanna in a clear perspex tube attached to a Fluotech 3 Anaesthetic Machine. A constant flow of oxygen nitrous oxide mixture was provided, and halothane slowly administered in 0.5% increments. Typically a 3% mixture was needed to anaesthetise goannas, but this varied between species. When the goanna was unable to right itself, and showed no reflex action when the tip of its tail was lightly pinched, the operation commenced. This process usually took 20 40 min.
All instruments and radio transmitters were cold sterilized for a minimum of 3 h prior to surgery. Sterile water was used to rinse the instruments before use and sterile gloves and face masks were worn during the operation. Prior to making any incision the goanna's scales were swabbed with 80% ethanol and a betadine solution and placed on a clean sheet.
Surgical techniques followed closely those used by Sweet (1999). A small incision, approximately 1 2 cm long, was made on the right hand side of the goanna, below the liver and anterior to the gonads. The incision was made on the animal’s side so that the resultant wound did not rub against the ground. Muscle layers were then gently teased apart with sterile tweezers. The radio transmitter (rinsed with sterile water) was gently inserted into the body cavity and massaged anteriorly so that it did not rest on the wound. The wound was sewn together using Ethicon 5 0 sutures.
Antiobiotic powder and a medical sealant (Histoacryl) were placed over the sutures to prevent any bacteria entering the wound or body cavity. Following surgery,
3. S tudy sites and general methods 52 goannas were placed in clean cloth bags and kept for 24 h and then released at the point of capture. Due to mortality, battery failure or movement out of telemetry range, animals were followed for varying periods throughout the study.
Analytical framework
All ecological systems are comprised of complex interactions between many elements and hence the questions ecologists can ask about individual units within these systems are hindered by either having to control for these elements and/or by restricting questions about such systems to simple dichotomies (Johnson 1999). This is particularly the case with the hypothesis testing (frequentist) based approach, whereby simple dichotomies (significant vs. non significant) are developed for an
“isolated” situation or question whose boundaries (for significance) are arbitrarily defined (Yoccoz 1991, Nester 1996). Recently however, more objective techniques such as the Information Theoretic paradigm (Burnham and Anderson 2002) have been developed, which allow more powerful inferences to be made.
Rather than isolating and asking individual questions about a system that is known a priori to be multifaceted, this approach advocates the development of multiple working hypotheses. These multiple hypotheses (the candidate set) are based on what is already known of the system, and developing a heuristic measure of “distance” between conceptual reality and each approximating model. This process has been termed Kullback Liebler (K L) information and is based on Boltzmann’s theory of entropy (Burnham and Anderson 2001).
3. S tudy sites and general methods 53
In this approach each model in the candidate set is compared to each other model, given the data that has been collected. One way of comparing these models is by using Akaike’s Information Criterion (AIC), which is a formal relationship between
K L information and the theory of maximum likelihood (Burnham and Anderson
2001). During the model ranking process, each model is subjected separately to the principle of parsimony, whereby models that contain the most parameters (and could potentially demonstrate spurious results) are most heavily penalized. The model (or models) for which AIC is smallest is then selected as most likely given the empirical data at hand.
This method is therefore not a test in any way; it is based on notions of inference and strength of evidence, and is a more objective method for asking more relevant questions about complex ecological systems. Because the ratio of sample size to parameters was always low in this study, the second order bias corrected form AIC c was used as the basis for model selection (Burnham and Anderson 2002). These
AIC c values can be rescaled as simple differences ( i) allowing a quick comparison of the ranking of candidate models (Burnham and Anderson 2002). Models were deemed to have substantial support if i fell within 1 2 of the best model (Burnham and Anderson 2002). Akaike weights ( wi) were calculated to provide a measure of the relative likelihood of each model in the candidate set, and in many cases further inference was drawn from the averaged parameter estimates from all models in the candidate set (weighted according to the Akaike weights). This averaging allows model selection uncertainty to be incorporated into parameter standard errors and reduces bias in parameter estimates, particularly when there are a number of models with similar AIC c (Burnham and Anderson 2002).
3. S tudy sites and general methods 54
Analyses throughout the remainder of this thesis (excluding Chapter 5) are all based on this approach. Burnham and Anderson (2002) and more recently Stephens et al.
(2005) have advocated that where orthogonally designed experiments are used, hypothesis testing is still valid and consequently this method is preserved in the chapter that examines the physiology of V. indicus (Chapter 5).
3. S tudy sites and general methods 55
Plate 1. An inaccessible section of bank within the Manton Dam study site, covered in Nymphaea violacea and Eleocharis dulcis.
Plate 2. Section of the bank at Manton Dam (foreground) that is easily accessible by boat for spotting and catching Varanus mertensi .
3. S tudy sites and general methods 56
Plate 3. A section of the Adelaide River V. indicus study site, showing mature Avicennia marina and Xylocarpus mekongensis along the creekline.
Plate 4. A section of the V. indicus study site along the Adelaide River, dominated by Bruigeira gymnorhyza and Rhyzophora stylosa.
57
Using morphometrics to predict gender in varanids
This chapter has been published as; Smith J G, B. W. Brook, A. D. Griffiths and G. G. Thompson. (2007) Can morphometrics and information theory predict gender in varanids? Journal of Herpetology 41: 133 140
4. Using morphometrics to predict gender in varanids 58
4. Using morphometrics to predict gender in varanids
Abstract
Varanid lizards are difficult to sex in the field because commonly used techniques are not completely reliable and definitive techniques are not logistically or economically feasible for many field based applications. Previous work has shown that variation in morphometric variables can be used to determine gender in some species of varanid. Here I build on these previous exploratory analyses by developing a set of a priori models (containing morphometric variables) to predict gender for six species of Australian varanid, and then examining their relative support under the
Information Theoretic framework. I then use cross validation procedures to determine the reliability of the best supported models’ predictive ability. This analysis suggests that a very large sample size is required for building models to predict gender in many species, to the extent that most models could not predict gender at all. The most important sexually diagnostic features for many species were a number of head variables and (to a lesser extent) scaling of limb proportions. This analysis provides some useful statistical tools for the field sexing of adult and juvenile V. gouldii with a known level of reliability, and it also serves to highlight the danger of extrapolating from potentially spurious results when using exploratory methods or null hypothesis testing.
4. Using morphometrics to predict gender in varanids 59
Introduction
Many aspects of an animal’s biology are sex related. Habitat preferences, home and activity range size, behavior and feeding strategies are some of the ecological parameters affected by gender differences (Calder 1984, Shine et al. 1998b). Despite this, little is known about the influence of sex on varanid life histories, primarily because it is very difficult to determine their sex in the field (Green and King 1978,
Shine 1986, Gaulke 1997, Sweet 1999).
A variety of methods have been utilised to determine sex in varanids (Pianka et al.
2004). Adults of some species can be sexed ex situ by the radiographic visualisation of hemipenial bones if present (Shea and Reddacliff 1986), but the absence of these bones does not always guarantee a female or small ossifications in some females can be confused with hemipenial bones. Fiberoptic laparoscopy (Davis and Phillips
1991) and coelioscopy (Schildger et al. 1999) can be used for the sexing of individuals of almost any age, however, logistical and financial constraints generally prohibit these techniques being readily available to researchers in the field. Adult males of many species in the Odatria subgroup can be identified by scalation differences adjacent to the vent (G. Husband pers. com., Gaulke 1997), whereas other species exhibit sexual dimorphism in head length (Pianka 1994), snout size
(Bennett 1998), and overall body size (Shine 1986, Auffenberg et al. 1991). Manual or auto eversion of hemipenes (Auffenberg et al. 1991, Thompson 1992a, Weavers
1993, Thompson et al. 1999), and probing of hemipenial pockets (Auliya and
Erdelen 1999, Gaulke et al. 1999) have also been used as sex determining techniques although females of some species are known to possess similarly eversible structures
4. Using morphometrics to predict gender in varanids 60 (Böhme 1991) and conscious wild varanids actively resist probing and attempts at manual eversion (Gaulke 1997). More recently, Mayes et al. (2005a) used a combination of hemipene eversion and sex hormone ratios with considerable success.
However, the cost and effort associated with the hormonal analysis excludes it as a useful tool in the field.
Although numerous techniques for determining varanid sex exist, all have pitfalls.
After reviewing these techniques, Gaulke (1997) cautioned that no field methods of sex determination in varanids is completely reliable; including behavioral observations such as combat (once presumed to be only males) and the visiting of egg laying sites following egg deposition. Importantly, the uncertainty inherent in estimating a varanids’ gender in the field is not amenable to direct quantification, and as such an unmeasured yet implicit level of uncertainty is always carried forward into further analyses, but hitherto never accounted for. This may be at least part of the reason why past attempts at determining gender with statistical methods have met with mixed success.
Thompson (2002) determined that body length to head length ratios could not be used to determine sex in many species of the Varanus subgenus, but could be useful
(when combined with other methods) in assessing gender in some members of the
Odatria . Exploratory analyses (using stepwise discriminant analysis) of various morphological characters in Australian varanids of known sex using museum specimens suggested that morphological characters could be used to predict gender in ten of the 18 Australian varanid species they studied (Thompson and Withers
4. Using morphometrics to predict gender in varanids 61 1997). However, they found no single morphological variable, or suite of variables that could discriminate reliably between sexes for this genus.
Our aim is to build on the exploratory analysis of Thompson and Withers (1997) by examining the morphometric variables that they noted for gender prediction (among others derived from the literature and other field workers) for six species of varanids from the tropical north of Australia, using the largest data set yet assembled for these species. Because exploratory analysis lacks any basis for strong inference, I used an
Information Theoretic framework (Burnham and Anderson 2001) with the goal of developing a suite of morphological variables that can be measured in the field and used to predict gender with a known level of reliability.
Materials and methods A review of the relevant literature and conversations with many varanid field biologists and captive breeders was carried out to generate a suite of morphological variables most likely to enable the determination of gender (Fig. 1). The only variables chosen were those that were unlikely to change following preservation of specimens. These morphological variables were measured in six species of goannas from the Museum of the Northern Territory, the Queensland Museum and the
Western Australian Museum (Table 1). Measurements were made to +1 mm with the body of each specimen placed in the approximate position shown in Figure 1. For some specimens, where the end of the tail had obviously broken off, the tail measurement was not used. The actual gender of each individual was determined by visual inspection of the gonads.
4. Using morphometrics to predict gender in varanids 62
Table 1. Number of specimens examined, mean snout vent lengths (SVL) + SE and subgeneric classification for six species of varanids measured.
Species Subgenus gender n SVL V. gouldii Varanus M 96 310.90 + 0.98 F 59 267.46+ 1.16 V. mertensi Varanus M 30 348.03 + 3.56 F 21 347.38+ 2.74 V. panoptes Varanus M 22 315.55 + 6.92 F 11 324.00+ 7.24 V. eremius Odatria M 40 148.43 + 0.43 F 23 134.04 + 1.06 V. glebopalma Odatria M 24 313.79 + 1.53 F 7 251.71+ 3.86 V. mitchelli Odatria M 25 206.04 + 2.31 F 19 214.58 + 1.59
4. Using morphometrics to predict gender in varanids 63
HW Measurements Key HL Head Length HW Head Width HH Head Height FL Forelimb Length UFL Upper Forelimb Length LFL Lower Forelimb Length FL HL Hindlimb Length UHL Upper Hindlimb Length UFL LHL Lower Hindlimb Length TA Thorax abdomen Length TL Tail Length SVL Snout Vent Length L TOTL Total Length (SVL + TL) SVL TA
UHL HL
LHL HH
HL
TL
Figure 1. Morphometric measurements used in model construction.
4. Using morphometrics to predict gender in varanids 64 Model development
Following the recommendations of Burnham and Anderson (2002), multiple working hypotheses were developed a priori in an attempt to explain which morphometric variables can best predict gender. These hypotheses were based on a review of the relevant literature, discussions with fellow varanid biologists and the authors’ experience.
Many parameters were derived from relationships between morphometric variables
(Table 2) and then a set of models (incorporating single derived variables, multiple variables or combinations of variables, Table 3) was developed to represent each hypothesis. In addition, a further model was generated separately for each species, incorporating the best supported parameters that Thompson and Withers (1997) found to predict gender in varanids.
Consistent with Burnham and Anderson (2001), the number of models (R) compared to sample size for each species was kept low (R = 9 for all species except V. mitchelli , where R = 8, see Table 3), by dismissing some hypotheses suspected to be unrealistic or uninformative for this dataset on a priori grounds. For example, many museum specimens had broken tails and therefore the hypothesis that males have longer tails in some species (Gaulke 1997, Shine et al. 1998a), could not be examined. Since very large adults of many species can be recognised as males (Shine
1986), the question of whether the gender of individuals of similar size can be discerned by morphometrics is more interesting and applicable to field researchers, therefore the hypothesis of total size was also excluded. Because the candidate set for each species had one model that differed from the other species, the resultant global
4. Using morphometrics to predict gender in varanids 65 models (which incorporate all parameters in every model) were correspondingly different. A null model (where no morphological variable can predict gender, with the prediction based simply on the observed proportion of male: female) was also incorporated into each candidate set, as this is another plausible hypothesis.
Model fitting
Each model in the candidate set was analyzed as a binomial generalized linear model with a logit link function using gender as the response variable, in the program R
(version 1.9.0, R Development Core Team 2004). Generalized linear models were fit first to the entire data set of each species. Then where sample size allowed (n >20), data sets were split into mature individuals (based on size at known maturity from the literature) and the same models were refit to each of the adult subset and the juvenile subset.
Model selection
Model selection was performed using Information Theoretic model selection methods based on Akaike’s Information Criterion (AIC, Burnham and Anderson
2002). This procedure uses Kullback Leibler information as an objective basis for selecting the model that explains the most substantial proportion of variance in the data, yet excludes unnecessary parameters that cannot be justified by the data (the most "parsimonious model", sensu Burnham and Anderson 2002).
4. Using morphometrics to predict gender in varanids 66
Table 2. Explanation of all derived parameters used in model construction.
Parameter Derivation Explanation HeadL HL / SVL Head Length divided by SVL HeadW HW / SVL Head Width divided by SVL HeadH HH / SVL Head Height divided by SVL AVL ((FL*2+HL*2) / 4) / SVL Average length of all limbs divided by SVL HV 648.05*exp (0.0093*SVL) The relationship between Head Volume ((Head Width * Head Length * Head Height)/ 2) and SVL HL_FL (HL / FL) / SVL (HindLimb Length divided by Forelimb Length), all divided by SVL UpperFL UFL / SVL Upper Forelimb Length divided by SVL UpperHL UHL / SVL Upper Hind Limb Length divided by SVL Tail TL / SVL Tail Length divided by SVL Ta TA / SVL Thorax Abdomen Length divided by SVL LowerHL LHL / SVL Lower hind Limb divided by SVL UpperFL UFL / SVL Upper Forelimb divided by SVL LowerFL LFL / SVL Lower Fore Limb divided by SVL
4. Using morphometrics to predict gender in varanids 67 Table 3. Candidate model sets for predicting gender of each species, the hypothesis examined under each model and references concerning sexual dimorphism in varanids leading to each model. The first seven models were used for every species. An eighth model for each species was developed, incorporating parameters suggested by Thompson and Withers (1997). The global models are therefore model nine with the addition of the new terms included in model eight for that species. Varanus mitchelli had no new parameters added and therefore its global model is model nine. No. Model. Hypothesis Reference 1 HL Hind limb length varies in relation to Gaulke (1997) bodysize (SVL) between sexes
2 HeadW Head width varies in relation to bodysize This study (SVL) between sexes
3 AVL Average Limb Length varies in relation to Gaulke (1997) bodysize (SVL) between sexes Bennett 4 HeadV Head Volume varies in relation to bodysize (1998) Thompson (SVL) between sexes (2002) 5 HeadV + AVL A combination of Average Limb Length and Gaulke (1997) Head Volume vary in relation to and this study body size (SVL) between sexes
6 HL_FL A ratio of Hindlimb length to Forelimb Gaulke (1997) Length, varies in relation to body size (SVL) between sexes 7 ~1 null model: No parameters measured are able to predict gender V. glebopalma A combination of Upper Forelimb Length, Thompson 8 UpperFL + UpperHL + Tail vary in relation to body size (SVL) between and Withers sexes (1997) V. gouldii A combination of Upper Hindlimb Length and 8 UpperHL + HeadH Head Height, all vary in relation to body size “ (SVL) between sexes V. mertensi A combination of Thorax Abdomen Length, 8 Ta + LowerHL + Tail Lower Hindlimb Length and Tail Length , all “ vary in relation to body size (SVL) between sexes V. eremius Upper Forelimb Length varies in relation to 8 UpperFL “ body size (SVL) between sexes V. panoptes A combination of Upper Hindlimb Length, Lower LowerHL + LowerFL 8 Forelimb Length and Upper Forelimb length, all “ +UpperFL vary in relation to body size (SVL) between sexes HeadL + AVL + HeadV Global model: All parameters together predict 9 + HeadW + HL_FL gender. *plus additional parameters in model 8 (above) for each species
4. Using morphometrics to predict gender in varanids 68 Because the ratio of sample size to parameters was always low, the second order bias corrected form AIC c was used as the basis for model selection (Burnham and
Anderson 2002). These AIC c values were then rescaled as simple differences ( i) allowing a quick comparison for the ranking of candidate models (Burnham and
Anderson 2001). Akaike weights ( wi) were calculated for each species’ candidate set of models, thus providing a measure of the relative likelihood of each model, given each data set and each candidate set. Models were deemed to have substantial support if i fell within 1 2 of the best model (Burnham and Anderson 2001).
To examine the utility of each model as a predictive tool, the K fold cross validation prediction error (Blum et al. 1999) was estimated for all models in each candidate set with a substantial level of support. In this procedure, data are divided randomly into
K groups. For each group the generalized linear model is fit to data omitting that group and the prediction error (interpreted as percentages) associated with predicting observed responses (i.e. those left out of the fitting) in each group from the model fit, is generated. In this case, leave one out cross validation was used (where K= number of individuals), allowing all possible splits of the data to be examined. These values were then expressed as levels of prediction reliability (1 prediction error).
Characteristic body size
Kratochvíl et al. (2003) recommend cautious inspection of body scaling parameters before analyses and interpretations of sexual dimorphism are made. Therefore to establish which of three highly correlated variables best characterized “body size” of a varanid (for use in model development), snout to vent length (SVL), total length
(TOTL) and thorax to abdomen length (TA) were fit as generalized linear models
4. Using morphometrics to predict gender in varanids 69 (with gender as the response variable) separately across all species. All three models had substantial support, all having i within 1 2 of the best model. They therefore could not be discerned from one another (Burnham and Anderson 2001). SVL is the most widely used in the literature (Wikelski et al. 1997, Shine et al. 2001, Webb et al. 2001, Perry and Garland 2002) and was therefore chosen as the characteristic body size determinant.
Many physiological variables do not covary in a linear fashion with body mass and the intercept of the regression line between them does not go through zero (Packard and Boardman 1999). To examine this potential bias in the data, I plotted all variables used against SVL and fit each intercept through zero. All variables had R squared values of over 0.8, with the only exception to this being the variable Head
Volume (HV) whose relationship was shown to vary exponentially with SVL.
Therefore each equation for this relationship was incorporated into all models containing the HV parameter.
Results
Combined adult and juvenile varanids
For three ( V. panoptes, V. mitchelli and V. mertensi ) of the six varanids studied, the most parsimonious model selected in each candidate set was the null model, indicating that none of those models examined were able to predict varanid gender substantially better than chance. All candidate sets that ranked the null as the most likely model also ranked models containing head variable parameters with substantial support. Amongst the most supported models in most other species, various head variables, either solely or (less often) in combination with limb length
4. Using morphometrics to predict gender in varanids 70 parameters, were more prevalent in higher ranked models than models containing limb length parameters only. For V. eremius , the model containing the single variable
HL_FL was clearly the most supported in the candidate set, explaining 61.5% of the total variance. The cross validation prediction reliability however was 63.0%, indicating that using this parameter alone to determine gender is again not much better than chance (50% : 50%).
The model set for Varanus gouldii contained three models that were ranked with substantial support. None of the models could be easily distinguished from each other, as indicated by the Akaike weights (Burnham and Anderson 2002). The resultant cross validation prediction reliability across the three best ranked models
(mean=57% + 0.05 SD) indicated that these parameters are of little use for determining gender in this species. Varanus glebopalma was the only species for which a model or combination of other parameters (HV and AVL+HV) were the best supported models. The global model fit for this species was 58% and the resultant cross validation prediction reliability estimations averaged 82%.
Adult varanids
For two of the three species where sample sizes of adults were large enough for analysis ( V. mertensi and V. mitchelli) , two closely ranked best models both contained elements of head and limb variables, suggesting that the morphological differences between males and females in these species is either very slight and / or are associated with a change in overall allometry, rather than one single attribute.
4. Using morphometrics to predict gender in varanids 71 The global model for adult V. gouldii was ranked as the best model, as it was eighteen times better than the next best model in the candidate set (based on their evidence ratios, Anderson and Burnham 2002). Cross validation prediction reliability was very high (84%) indicating that the global model for this species could prove useful for predicting gender in adults (Table 4). Standardized regression coefficients
(as used by Conroy and Brook 2003) for the global model showed that the head volume parameter (HV) accounted for three times the variation of the other parameters.
Juveniles
Only three species had sufficient sample size (n >20) of juveniles for this analysis. In juvenile V. panoptes the null model could not be discerned from the other highest ranked models model set indicating that no model could be used to predict gender.
The V. mitchelli juvenile dataset had HL_FL ranked as the best model, and its evidence ratio indicated it was 3.4 times better than the next best model. The V. mitchelli global model explained 41% of the total variation and the cross validation prediction reliability for this model was 70%. This model is potentially a useful tool for sexing V. mitchelli . The three best ranked models for sexing juvenile V. gouldii were the same best ranked models in the adult and all individuals combined datasets.
Average cross validation prediction reliability for these models within the juvenile subset was 68%.
Previous hypotheses
Across all species, in all expanded data sets and adult subsets, none of the models predicted by Thompson and Withers (1997) were ranked within i =1 2 of the best
4. Using morphometrics to predict gender in varanids 72 model, and only 6% of the time they were ranked amongst those models with a reasonable level of support ( i ≤ 10).
Gender prediction in Varanus gouldii
Because the V. gouldii dataset was the largest (n = 155) and the resultant global models for the adult subset had a reasonable level of explained deviance (64%) and high cross validation prediction reliability (84%), equations were derived from the global models to predict gender in this species.
Male and female V. gouldii appear to mature at different sizes (Shine 1986), and therefore the utility of the following equation is only practical when applied to specimens larger than the size at maturity for males. To solve for gender (male; y >
0.5 < y ;female) in an adult (SVL >320 mm) V. gouldii with 84% reliability;
1 y = 1+ e (Model) where Model = 14.9+(1*AVL)+(0.0037*HV)+(5125.9*HL_FL)+(11.34*UpperHL), the terms AVL, HL_FL and UpperHL are as defined in Table 2 and
HV=648.05*exp (0.0093*SVL) . To solve for gender ( y) in a juvenile (SVL <280 mm) V. gouldii with 71% reliability;
1 y = 1+ e (Model) where, Model = 6.9 +(19.3*AVL)+(0.00557*HV)+(325*HL_FL) (35.7*UpperHL), the terms AVL, HL_FL and UpperHL are as defined in Table 2 and
HV=648.05*exp (0.0093*SVL) .
4. Using morphometrics to predict gender in varanids 73
Table 4. Summary of second order Akaike’s information criterion (AIC c) and associated statistics for all candidate models for the analysis of gender prediction in adult and juvenile V. gouldii subsets. All models are ranked according to support, thus i = 0 for the best model (bold type). log( L) is the maximised log likelihood of the model, AIC c is the selection criterion, K is the number of estimated parameters,
i is the difference between the model’s AIC c value and the minimum AIC c value and wi is the Akaike weight. Leave one out cross validation prediction reliability for all models with substantial support ( i ≤ 2) are also shown. Only models displaying a reasonable level of support ( i <10) are shown.
cv. log( L) Model AIC c K i wi reliability (%) adults (n= 83) global 20.791 54.150 6 0.000 0.909 84 HV 26.782 59.722 3 5.572 0.056 AVL + HV 26.326 60.919 4 6.768 0.030
juveniles (n=69) global 33.706 79.980 6 0.000 0.401 71 HV 36.954 80.068 3 0.087 0.383 67 AVL + HV 36.721 81.709 4 1.729 0.168 67 HL_FL 39.572 85.304 3 5.324 0.028
4. Using morphometrics to predict gender in varanids 74
Discussion Sexual dimorphism is common in lizard species (Olsson et al. 2002), with the most consistently dimorphic traits being head size (males having larger heads) and trunk length (the distance between the front and hind legs) is greater in females. This analysis also suggests that various head variables (principally head volume) and (to a lesser extent) scaling of limb proportions can also be important diagnostic features for gender prediction in some species of varanids. Bennett (1998) stated that male
Varanus salvadorii and Varanus albigularis develop bulbous snouts with extreme old age. Investigation of sexually mature individuals of these and other species may show similar head volume dissimilarities.
Although sexual dimorphism has been recorded among neonate snakes (King et al.
1999), allometric sexual dimorphism may not be present in varanids until maturation.
In this analysis, the candidate sets fit to adult varanids showed a higher proportion of models that could be discerned from null models and higher levels of explained deviance. However, the degree of precision by which these parameters can be used to determine gender, based on the cross validation assessment, was in most cases not high enough to warrant their use as a predictive tool.
The most effective gender prediction tool was developed for V. gouldii , where the global model for adult varanids, which combined several head and limb proportion variables, was able to predict gender with a reasonably high level of accuracy. Given that gender prediction in V. gouldii appears much more reliable than in other species, with higher levels of variation explained by the global model and high percentage of correct predictions, either V. gouldii is more sexually dimorphic than other species,
4. Using morphometrics to predict gender in varanids 75 or larger sample sizes may be required for effective sex prediction models to be developed for other species.
None of the models Thompson and Withers (1997) found significant using null hypothesis testing were well supported for any of the species I analyzed. This striking difference between results of the two studies highlights the limitations of the null hypothesis approach and shows, as suggested by Anderson and Burnham (2002), how exploratory analysis can sometimes produce spurious results. Clearly, although data driven methods can be useful in the early stages of analysis, we now have the benefit of more objective techniques, such as the Information Theoretic paradigm
(Burnham and Anderson 2002), which allow more powerful inferences to be made.
Varanid lizards probably recognize the gender of conspecifics using behavioral cues and pheromones. As varanids can be extremely shy animals, behavioral observation studies can be very labor and time intensive, and discerning gender by behavioral observation is not possible for most research projects unless definite mating / copulatory behavior is observed (e.g. King and Green 1979, Auffenberg 1981, Carter
1990, McCoid and Hensley 1991). Long term studies of varanids could take advantage of the olfactory sensitivity of a dog trained in recognising different sexes of a certain species by their odor. Other predictive techniques that have been suggested by various authors, such as examination of scale rosettes cranial to the vent in males (Auffenberg 1981) and more prominent skin flaps and scale micropores (S. Sweet pers com., Gaulke 1997), should be investigated in more detail for other varanids.
4. Using morphometrics to predict gender in varanids 76 The most reliable (albeit expensive and time consuming) method of sexing varanid lizards is by direct examination of the gonads (e.g. Schildger et al. 1999). However for the field biologist, sex specific DNA probes, requiring a blood sample to be taken from each individual, would probably be less costly, both economically and logistically. Such probes are available for Varanus komodoensis (Murphy et al. 2002) and Varanus rosenbergii (W. Smith pers. com.) but further work needs to be done before they are available for other species.
Creating larger datasets of morphometrics for other species and analyzing them using an a priori multiple working hypothesis approach could provide more useful gender prediction models. When used in conjunction with other methods, these models will help expand the capabilities of researchers and others to predict the gender of varanids in the field, quickly, inexpensively, and with a known level of uncertainty that can be appropriately propagated in further statistical analyses.
77
Varanus indicus physiology
Remote datalogger for recording V. indicus body temperatures This chapter has been submitted for publication as;
Smith, J.G., Christian, K. & Green, B. Physiological ecology of the mangrove dwelling varanid, Varanus indicus . Journal of Zoology
5. Varanus indicus physiology 78
5. Varanus indicus physiology
Abstract
Some species of terrestrial lizards in wet dry tropical climates reduce their body temperatures and activity, and lower their metabolic rates during the dry season when food and water resources are scarce. At least one semi aquatic varanid ( Varanus mertensi ) in this region does not respond in this way, presumably because of year round access to food and water. Varanus mertensi also selects lower body temperatures (T b) than its terrestrial counterparts. To investigate whether there is a trend for reduced (but seasonally constant) T b selection and seasonally constant metabolic rates in other semi aquatic varanids, I studied the thermal biology, energetics and water flux of Varanus indicus, a semi aquatic, mangrove dwelling varanid in tropical northern Australia. Varanus indicus selects consistently lower T bs than terrestrial varanids, and although they remain active year round, they reduce their activity in the dry season, but not to the extent of terrestrial varanids. Thus, although food and water depletion are the driving forces behind decreases in dry season T b selection and energetics for many varanids, V. indicus appears not to be subject to these pressures to the same extent. Varanus indicus field metabolic rates decrease by 38% in the dry season, due mostly to a reduction in activity. However their field metabolic rate does not differ significantly from other terrestrial varanids or another semi aquatic varanid from the same region. In contrast, the water flux rates of V. indicus are lower than V. mertensi but not different from any similar sized terrestrial varanids studied to date.
5. Varanus indicus physiology 79 Introduction
Lizards worldwide inhabit a wide range of habitats and climates (Pianka and Vitt
2003) and are therefore subject to a variety of seasonal and thermal conditions.
Studies of seasonal energetics of lizards in the wet dry tropics have revealed varying responses to the slightly lower ambient temperatures and lower availability of food and water which characterise the dry season (Christian et al. 2003). All species have lower night time body temperatures (T b) in the dry season because of the prevailing environmental conditions (Christian et al. 1999b), some species actively select lower
Tbs during the day (Christian et al. 1983, Christian and Bedford 1995, 1996,
Christian et al. 1999a), some depress their metabolic rate (Christian et al. 1996b,
Christian et al. 1996c, Christian et al. 1999a), and all decrease their activity during the dry season. Consequently, the field metabolic rates (FMR) of these species are lower in the dry season than in other seasons (Christian and Green 1994, Christian et al. 1995, Christian et al. 1996b, Christian et al. 1996c, Christian et al. 1996d,
Christian et al. 1999a, Christian et al. 2003).
Although some species select lower T bs during the day, it is important to note that the thermal environment in the wet dry tropics permits for lizards to attain high body temperatures (T b) throughout the year (Christian and Bedford 1995, 1996, Christian and Weavers 1996). In contrast, the semi aquatic varanid Varanus mertensi, does not change its seasonal T bs, resting metabolic rates or activity levels considerably between seasons (Christian and Weavers 1996, Christian et al. 1996d). Although the field metabolic rates of V. mertensi are lower in the dry season than in the wet, this seasonal difference is largely attributable to the passive consequences of lower night
5. Varanus indicus physiology 80 time body temperatures (Christian et al. 1996d). Varanus mertensi may be less affected by seasonal changes in the wet dry tropics because the habitats it occupies provide it with year round access to water and food (Christian et al. 1996d). To further examine these apparent seasonal trends in semi aquatic, tropical zone lizards,
I investigated the thermoregulation and energetics of Varanus indicus (Daudin 1802), a similarly sized semi aquatic varanid that occupies mangrove systems in the wet dry tropics of northern Australia and other parts of Asia. Given that V. indicus also inhabits habitats that are ostensibly productive year round, I hypothesized that they would remain active throughout the year without the striking metabolic and behavioural adjustments of terrestrial species in response to the dry season conditions. Because other semi aquatic varanids appear to select lower T bs than terrestrial species of similar size (Wikramanayake and Green 1989, King 1991,
Wikramanayake and Green 1993, Christian and Weavers 1996), I also hypothesized that V. indicus would thermoregulate at relatively low T bs.
Materials and methods
Radio telemetry
Varanus indicus were captured at the Adelaide River study site using baited pipe traps as outlined in Chapter 3, placed in individual bags, and transported to Charles
Darwin University, Darwin. Within 2 d of capture, temperature sensitive radio transmitters (Holohil Systems, Canada, model SB 2) were implanted intraperitoneally. Beforehand, transmitters were calibrated at 5°C increments between 5°C and 40°C in a water bath (Grant, USA) using a mercury thermometer traceable to a standard. The final mass of transmitters ranged between 4 13 g and all
5. Varanus indicus physiology 81 implanted transmitters weighed <3% of an animals’ body mass. Transmitters were surgically implanted in 13 V. indicus (682 2173 g) between 2002 and 2004 using the method outlined in Sweet (1999) and Chapter 3.
Thermoregulation
I recorded body temperatures of free ranging V. indicus by using an automated system with a fixed antenna, receiver and digital processor (Telonics, USA) connected to a data logger (Campbell CR10X, USA) in 2003 and a coupled ATS
(Advanced Telemetry Systems, MN) datalogger (Model R2100) and receiver (Model
D5041A) in 2004. Both digital processors and data loggers were calibrated in the laboratory prior to taking measurements in the field. Where possible, body temperatures of multiple individuals were recorded at the same time at 15 min intervals. Subsequently, hourly means were calculated for each individual and grand means were calculated for all V. indicus sampled in a season for use in energy budget calculations.
A wooden thermal gradient was created in the laboratory (4 m x 1.5 m) in an air conditioned room. Two 60 watt heat lamps (0.5 m above the bottom of the container, and connected to timers) were spaced 0.5 m apart from one end of the container, thus providing a continuous gradient of substrate temperatures that ranged from 19 to
42 °C during the daytime hours (07:00 – 18:00). Individual V. indicus were placed one at a time inside the gradient with a thermistor inserted ~50 mm inside the cloaca.
The thermistor was attached to a HOBO XT temperature datalogger (Onset
Computer Corporation, U.S.A.), which was taped to the side of the tail using an
5. Varanus indicus physiology 82 adhesive bandage after being calibrated. The T b of each V. indicus was recorded every minute over a 2 d period.
The set point range has been defined as the ‘preferred’ T b of an animal, as measured in the laboratory without other ecological demands of the environment which may influence T b (Hertz et al. 1993, Christian and Weavers 1996). The set point range for
V. indicus was estimated using the central 50% of the T b from the thermal gradient to define the upper and lower limits (Hertz et al. 1993).
I recorded microclimate data as described in Christian and Weavers (1996). Various environmental measurements (mud, water, ambient, inside hollow temperatures) were measured every 15 min for two weeks each season. All measurements were recorded with Hobo Pro series dataloggers (Onset Computer Corporation, U.S.A.), with a 1.5 m thermistor attached, allowing simultaneous measurements of ambient and other substrates (mud, water, or hollow log) temperatures. Shaded air temperatures were measured at 2 m above the ground.
Operative temperatures (T e) represent the interaction between an animal and its environment by incorporating aspects of the physical characteristics of an animal
(e.g., absorptivity and surface areas) and the microhabitat in which it is situated
(Porter and Gates 1969, Bakken 1981, Bakken et al. 1985, Bakken 1992). The T emax and T emin were calculated from a steady state biophysical model based on equations in Tracy (1982) and similar to models used broadly (Porter and Gates 1969, Porter et al. 1973, Porter and James 1979, Porter and Tracy 1982, Waldschmidt and Tracy
5. Varanus indicus physiology 83 1983) including input data based on animal characteristics and microclimate data collected from the study site. An absorbtivity value of 87% was measured for V. indicus using methods outlined in Christian et al. (1996a).
Because of the relatively large size of V. indicus (mean mass =1265 g ± 43) the animals do not change T b instantly. Therefore T es alone are not sufficient as an index of body temperatures as the animals move through a thermally complex environment
(Christian et al. 2006). Thus, to develop thermoregulatory indices, I used the null model based approach of Christian et al. (2006), a synopsis of which is given below.
A null distribution of predicted T b (predT b) between the maximum (T emax) and minimum operative temperatures (T emin) was generated (in True BASIC). The predT b represents the body temperatures achievable by the animal based upon V. indicus ’ body size, T e, and the body temperature of the animal in the recent past (as the animal is warming or cooling). To calculate predT b a random starting place was assigned (corresponding to a site in which the microclimate at that time is between
Temax and T emin), and a random length of time for the modeled animal to remain in that spot was also assigned. The predicted body temperatures (predT b) was calculated for the animal in that place for each minute that the animal remained there.
This calculation was done using T e as the driving force, and the thermal time constant, (taken from regression equations generated from data in (Dzialowski and
O'Connor 2001) to incorporate body mass. The form of the equation is:
T = T0+ (T e T0)*e 1/ where T0 is the temperature at the time the animal first enters the new place.
Different time constants were used for warming and cooling. The program then
5. Varanus indicus physiology 84 assigned a new random place and time, calculated the corresponding predT b, and continued these steps over the course of the whole day (24 h), repeating this process for 1000 d to generate a null distribution of random predT b achievable throughout the day.
Using this null distribution, I determined a number of thermoregulatory indices for any time during the day (Table 1). I calculated the mean predT b for each hour of the day and then compared these values with the actual animal T b (as measured by telemetry) at each hour using the indices of Hertz et al. (1993). The mean deviation of an animal’s field active body temperatures from its set point (or preferred) range,
d b, as defined by Hertz et al. (1993) was calculated. This index is an indication of the accuracy of thermoregulation. A measure of the thermal quality of the animal’s environment ( d e), is the mean deviation of T e from the animal’s set point range
(Hertz et al. 1993) was also calculated. For large ectotherms, it is appropriate to substitute the mean achievable body temperature (predT b) for mean T e to calculate d e.
The “Effectiveness” of thermoregulation as defined by Hertz et al. (1993) as:
E = 1 – ( d b/ d e) was calculated, where a value for E of 0 indicates a random selection of thermal environments, and a value for E of 1 indicates careful thermoregulation. An alternative to this index, simply d e d b, was suggested by Blouin Demers and
Weatherhead (2001). The “Exploitation” of the thermal environment (Ex) was calculated as: Ex = (amount of time an animal has a body temperature within its set
5. Varanus indicus physiology 85
Table 1. Definitions of the parameters and indices used for quantifying
thermoregulation by V. indicus .
Index or symbol Definition Derivation
Tb Body temperature Measured remotely via radio telemetry
Te Operative temperature The equilibrium temperature that the lizards would obtain in a given microclimate.
Tset Set point range of T bs The central 50% of the Tbs selected in the thermal gradient.
db An index of field body Mean of the absolute temperatures relative to value of the deviations of the set point range field active T bs from T set .
de An index of average Mean of the deviations of thermal quality of a Te from T set . habitat
E Effectiveness of 1 (d b/d e) temperature regulation
Ex Exploitation of the The time in which an thermal environment animal’s T bs are within Tset , divided by the time available for the animal to have its T b within T set .
5. Varanus indicus physiology 86 point range) ÷ (amount of time during the day in which it is possible for the animal to achieve its set point range) (Christian and Weavers 1996).
Field metabolism
3 18 The doubly labeled water ( H H2 O) technique (Lifson and McClintock 1966) was used to measure field metabolic rates (FMR) and water flux of free ranging animals over a mean period of 21 d in the wet season and 31 d in the dry season. Blood samples (0.3 ml) were taken from the caudal vein before injections of 440 L (mean value) of 18 O and 1 ml of tritiated water, 8 to 12 h after injection, and upon recapture.
Isotopic samples were measured at the Division of Sustainable Ecosystems, CSIRO,
Canberra. The techniques for sample analysis to determine isotopic measurements of
FMR and water flux are described in detail elsewhere (Christian and Green 1994,
Christian et al. 1995, Christian et al. 1996b, Christian et al. 1996c, Christian and
Weavers 1996, Christian et al. 1996d).
For comparison, I used the empirical equation for reptiles (Nagy et al. 1999) to predict the FMR of V. indicus , using the mean mass of the animals I sampled in each of the two seasons. Allometric equations for reptiles from arid and tropical regions
(Nagy 1982) were used to calculate predicted water flux rates for comparison with measured values from V. indicus .
Energy budget calculations
Energy expenditure was derived from field CO 2 estimates using a thermal equivalent
1 of 26 kJ L CO 2. Field metabolic rate was partitioned into components due to T b and activity for each season by combining field and laboratory data. A regression
5. Varanus indicus physiology 87 equation was derived to relate body temperature to resting metabolism (as determined from laboratory measurements, Schultz 2002). This was then combined with field body temperatures separately for both seasons, during the day (09:00 –
18:00 h) and night (the remaining 14 h) to estimate the total resting metabolism
(TRM) under field body temperature conditions (Benabib and Congdon 1992). The oxygen consumption values were converted to units of energy using the energy
1 equivalent of 20.08 kJ L O 2 (Benabib and Congdon 1992).
Activity respiration (AR, Benabib and Congdon 1992) is calculated as the difference between the FMR and TRM and provides an index of the amount of energy expended in activities such as locomotion, digestion, and reproductive costs (van
Marken Lichtenbelt et al. 1993). The percentage of the total field metabolism allocated to activity (%AR) is calculated as AR/FMR X 100 (Anderson and Karasov
1981). The field maintenance scope is derived from the ratio of FMR/TRM
(Congdon and Tinkle 1982), and if the animal's body mass changes <1% d 1, this can be termed the sustained metabolic scope (SusMS, Peterson et al. 1990).
Statistical analyses
Seasonal means were compared using ANCOVA with body mass as a covariate.
Differences between means were considered statistically significant when P≤ 0.05 and means are presented ±1 SE. I report results of new statistical comparisons
(ANCOVA) with other species of Varanus of similar size from the same region using data collected in previous studies (Christian et al. 1995, Christian et al. 1996d).
5. Varanus indicus physiology 88 Results
Thermoregulation
The mean mass of the V. indicus used in the calculation of standard metabolic rate
(SMR, 997 ± 204, Schultz 2002) was slightly smaller than those of the animals used in the calculations of FMR (1265 ± 43). The set point range selected by V. indicus was 30.9 – 33.5ºC (Table 2). In the wet season, V. indicus grand mean T b overlaps the set point range for 4.5 hours (15:30 20:00) and in the dry season, the grand mean
Tb overlaps the set point range for 6.0 hours (13:00 19:00; Fig. 1). The daytime grand means in the wet (30.0 °C) versus the dry seasons (29.5 °C) were very similar, as were the night time grand means (29.5 and 28.0 °C respectively). Maximum temperatures selected by V. indicus during the dry and wet seasons were very similar
(33.1 and 32.1 °C respectively) and there was only 1.8 °C difference between the wet
(27.0 °C) and dry season (25.2 °C) minimum temperatures.
Table 3 shows various thermoregulatory indices and the number of hours per day that varanids can and do achieve their set point range, comparing V. indicus from this study with V. panoptes, V. gouldii and V. mertensi from Christian and Weavers
(1996). During the hours before and after sunset, in both the wet and dry seasons, the
Tbs of V. indicus are higher than either T e value, indicating that the animals are in tree hollows (a phenomenon also supported by radiotelemetry observations) which provide a warmer microhabitat than ambient external temperatures (Christian and
Weavers 1996).
5. Varanus indicus physiology 89
Table 2. The set point ranges and mean T b (as measured in the lab) for V. indicus and other varanids from the same region (data from Christian and Weavers 1996).
Lower set Upper set Mean T b in lab
Species point ( °°°C) point ( °°°C) (°°°C)
V. indicus 30.9 33.5 32.3
V. panoptes 35.8 37.6 36.7
V. gouldii 34.0 36.3 35.1
V. mertensi 33.1 35.5 34.2
5. Varanus indicus physiology 90 Table 3. Summary of body temperature data and thermoregulatory indices of V.
indicus (bold, this study) and three other varanids from the same region (Christian
and Weavers 1996) during the wet and dry seasons.
Time Time set Grand set point point Mean Midday range range T T b b possible exploited Ex Species (°C) (°C) de db de db E (h) (h) (%) Wet Season V. indicus 30.3 30.6 3.2 0.9 2.31 0.7 10.0 2.6 26 V. panoptes 36.4 35.2 7.0 1.5 5.5 0.8 8.7 1.0 12 V. gouldii 35.9 35.9 6.4 0.2 6.15 0.9 9.1 8.7 96 V. mertensi 34.0 34.0 8.9 0.5 8.45 0.9 11.3 8.1 72 Dry Season V. indicus 29.7 31.4 1.6 1.0 0.66 0.4 8.0 5.1 64 V. panoptes 36.2 36.2 2.1 0.9 1.28 0.6 7.3 5.0 68 V. gouldii 28.2 28.2 1.9 5.7 3.75 2.0 7.8 0.0 0 V. mertensi 33.4 33.4 1.4 1.0 0.48 0.3 8.1 6.0 74 Tb = Body temperature Te = Operative temperature db = Field T b relative to set point range de = Average thermal quality of habitat E = Effectiveness of temperature regulation Ex = Exploitation of the thermal environment
5. Varanus indicus physiology 91
During the dry season, the d e is approximately half what it was in the wet season, a result of suitable microclimatic conditions for V. indicus to stay within their set point range. According to Christian and Weavers (1996), V. mertensi also experiences low de during this season. During the wet season V. indicus is the least careful thermoregulator of all similar sized tropical Australian varanids studied to date (as indicated by index E, Table 3). Varanus indicus also does not thermoregulate as carefully in the wet as in the dry season (E = 0.40 versus E = 0.72 respectively).
Varanus indicus in the dry season exploits 64% of available time in its set point range, whereas in the wet season V. indicus spends very little time (Ex = 26%) within its set point range.
Field metabolism, water flux and energy budget
Six lizards were recaptured after injection with isotopes in the wet season and four were recaptured in the dry season. Recaptured wet season lizards gained a mean of
10% body mass, and dry season animals lost a mean of 1% body mass. Total body water estimates remained the same in the wet and dry seasons (Table 4). Water influx rates were significantly lower ( P = 0.0006) in the dry season, representing a 43% decrease in water flux and wet season water influx rates were much higher than those predicted for lizards in arid and tropical zones (Nagy 1982). However, in the dry season, the predicted water influx rates for tropical lizards exceeded those found in
V. indicus. During the dry season V. indicus water flux was significantly lower than
V. mertensi (P = 0.02) but not different from V. panoptes or V. gouldii . During the wet season, V. indicus water flux was significantly higher than V. panoptes ( P =
0.0007) but not different from V. mertensi or V. gouldii .
5. Varanus indicus physiology 92 Tb Wet Tb Wet a) 50 predTadjTb Wetb Wet
TTemax Wetemax Wet TTemin Wetemin Wet
u 40 set point range l
30 Temperature (°C) 20
10 0 2 4 6 8 10 12 14 16 18 20 22 24 Hour TTb Dryb Dry 50 b) predTadjTb Dryb Dry
Temax Dry Temax Dry Temin Dry Temin Dry 40 u set point range l
30
Temperature (°C) 20
10 0 2 4 6 8 10 12 14 16 18 20 22 24 Hour Figure 1. Grand mean body temperatures (T b; as measured by radiotelemetry),
predicted T b (predT b), set point range, and operative temperatures (T emax
and T emin) as a function of time of day for free ranging V. indicus during the (a) wet season and (b) dry season.
5. Varanus indicus physiology 93
Table 4. Water flux rates, total body water (TBW = %body mass, derived from 18 O dilution), as determined from isotopic analysis, and water economy index (WEI, see text for details) for field active Varanus indicus during wet and dry seasons. Sample sizes and masses are as in Table 1. Standard deviations are in parentheses. The rates of water flux predicted by the allometric equations for arid and semi arid zone reptiles and for tropical, subtropical zone reptiles (Nagy 1982) are also shown.
Wet season Dry season (n = 6) (n = 4) Mass (g) 1210 (580) 1242 (618)
Water influx (ml d 1) 83.4 (9.4) 35.8 (10.8)
Predicted: arid (ml d 1) 24.4 25.0
Predicted: tropical (ml d 1) 51.0 51.9
Water influx (ml kg 1 d 1) 81.3 (32.6) 32.2 (13.3)
TBW (%) 71.2 (3.0) 70.0 (2.7)
WEI (ml kJ 1) 0.5 (0.14) 0.3 (0.05)
5. Varanus indicus physiology 94
The water economy index (WEI) is the ratio of water flux to FMR with the units of ml kJ 1 (Nagy and Peterson 1988). During the wet season WEI was significantly higher (64%, P = 0.02) than the dry season value. During the dry season V. indicus’
WEI is lower than V. mertensi ( P = 0.009) but not V. gouldii (P= 0.66) or V. panoptes ( P = 0.70). During the wet season the WEI is higher in V. indicus than in V. gouldii ( P = 0.02) and V. panoptes ( P=0.04), but not different from V. mertensi ( P =
0.54).
The FMR was significantly lower in the dry season than in the wet (Table 5, P =
0.01), representing a 38% (69.6 kJ d 1) reduction in FMR. The decrease in dry season
FMR is due in part to lower night time temperatures (3%) but mostly to a decrease in activity (97%). The wet season FMR of V. indicus was not different from V. gouldii,
V. panoptes or V. mertensi , nor was the dry season FMR different. All other indices
(AR, SusMS and AS) were higher in the wet season than the dry (Table 6). The predicted values of FMR for a reptile (Nagy et al. 1999) were much lower in the wet but very similar to those of V. indicus in the dry season.
The density of V. indicus along Adelaide River is estimated as 10.9 adult animals’ ha 1 using recapture estimates and the number of captures per trapping session
(Chapter 8). Using this value, the estimated energy expenditure for the population is
1328 kJ ha 1 d 1 in the wet season and 826.7 kJ ha 1 d 1 in the dry season (62% of wet season expenditure).
5. Varanus indicus physiology 95
Table 5. Field metabolic rate and water flux rates of V. indicus (this study) and another semi aquatic varanid, V. mertensi (from Christian et al. 1996d) and two terrestrial species from the same northern Australian region (from Christian et al.
1995).
FMR Water flux Species Period (kJ kg 1 d 1) (ml kg 1 d 1) V. indicus wet season 161.4 81.3 dry season 98.7 32.2 V. mertensi wet season 120.7 63.1 dry season 81.1 66.6 V. gouldii active 196.4 50.7 inactive 66.5 19.5 V. panoptes active 143.2 41.4 inactive 56.3 21.0
5. Varanus indicus physiology 96 Table 6. Carbon dioxide production, field metabolic rates (FMR), and mean body mass of free ranging Varanus indicus during the wet and dry seasons. Standard deviations are shown in parentheses. Predicted FMR was calculated with the equation for reptiles from Nagy et al. (1999). The total resting metabolism (TRM) was calculated using field body temperatures in the laboratory defined equation relating body temperature to resting metabolism, summed over 24 h and can be divided into resting metabolism during active periods (RMA) and resting metabolism during inactive periods (RMI). The amount of energy expended in activity (AR) is estimated as the difference between the FMR and TRM. The sustained field metabolic scope (SusMS) is calculated as FMR/TRM. The percentage of the total field costs allocated to activity (%AR) is calculated by AR/FMR x 100. The field activity scope (AS) is = (FMR RMI)/RMA.
Wet season Dry season (n = 6) (n = 4) Mass (g) 1210 (580) 1242 (618)
CO2 (ml g 1 h 1) 0.36 (0.05) 0.16 (0.06)
FMR (kJ kg –1d–1) 161.4 (31.4) 98.7 (35.3)
FMR (kJ d –1) 184.4 (71.8) 114.8 (45.1)
Predicted FMR (kJ d –1) 107.8 110.4 RMA (kJ kg –1 d 1) 11.1 11.2 RMI (kJ kg –1 d 1) 14.9 13.0 TRM (kJ kg –1 d 1) 26.0 24.2 AR (kJ kg –1 d 1) 135.4 74.5 %AR 83.9 75.5 SusMS 6.2 4.1 AS 13.2 7.6
5. Varanus indicus physiology 97
Discussion Thermoregulation
Although no statistical comparisons were made, the T b’s selected by V. indicus were
4ºC lower (in both seasons) than the midday T b’s of the semi aquatic V. mertensi , which selects significantly lower T b’s than two terrestrial varanids ( V. panoptes and
V. gouldii ) from the same area (Christian and Weavers 1996). Varanus indicus also uses the widest range of temperatures of the northern Australian lizards studied to date.
If V. indicus were inactive throughout the dry season, in order to maximize their savings of energy and water, they could select deep shade for the entire day in order to achieve the coolest T bs possible. However, Fig. 1(b) shows that around 09:00 T bs are the same as T emin but rapidly increase from 11:00 until their T b is within their set point range, then T b gradually decreases from around 14:00 for the remainder of the day. This suggests that V. indicus remain active during the dry season, a trend that is supported by radiotracking data (Chapter 6). Furthermore, this suggests that these lizards are actively thermoregulating in the dry season, albeit not to the maximum extent possible in their thermal environment. They could emerge, bask and achieve their set point range 1.5 h earlier in the dry if they exploited their thermal environment to the fullest (Fig. 1b).
All three tropical varanids (including V. indicus ) that remain active during the dry season show a similar trend of exploiting their thermal environments to a greater extent during this time. Varanus indicus spend some days in one place (in tree hollows or in the same tree, Chapter 6). The surrounding flooded wetlands and high
5. Varanus indicus physiology 98 tide levels during the wet season within Adelaide River probably provide sufficient food (all animals gained mass during this time) such that animals would not have to forage each day. It is possible that removing these days from the analysis may increase the Ex values for both seasons. However, as T b was logged remotely for this study, days of inactivity for each individual could not be determined.
During the dry season, the d e is approximately half the wet season value, indicating more suitable microclimatic conditions for V. indicus to stay within their set point range. Varanus mertensi also experiences low d e during this season, with the habitat for V. panoptes and V. gouldii being slightly less thermally favourable relative to their set point ranges (Christian and Weavers 1996). It seems likely that V. indicus thermoregulate at T b’s that embody a trade off between saving energy and water, while being able to remain active (Huey and Stevenson 1979). If V. indicus remain active to forage, they would also need to be able to avoid predators (Christian and
Tracy 1981) and digest any food they obtain.
Mean midday (11:00 h 16:00 h) T bs were marginally lower than the set point range during both seasons and were consistently lower than the other three tropical species.
The mean lab T b selected by V. indicus was also lower than V. panoptes , V. gouldii and V. mertensi . Varanus indicus’ set point range and midday T bs are also lower than other tropical, terrestrial species but similar to the semi aquatic V. mertensi and other semiaquatic varanids (Wikramanayake and Green 1989, Traeholt 1995,
Wikramanayake and Dryden 1999). The mangrove forests that V. indicus inhabit are generally heavily shaded (mean canopy cover range= 51 75%), and although thermal microhabitats are present all year that would enable their set point temperatures to be
5. Varanus indicus physiology 99 attained, it is possible that V. indicus selects lower T b’s because the effort required to shuttle between these patches of sun may outweigh the advantages of higher T b selection (Hertz et al. 1993). In this way, V. indicus share some similarities with
Hypsilurus spinipes , an agamid that lives in deeply shaded rainforest and is a complete thermoconformer (Rummery et al. 1995).
Given that semi aquatic varanids evolved from different varanid lineages (Ast 2001), the thermal data presented here and elsewhere (Christian et al. 1996d) suggest that the relatively low T b’s selected by semi aquatic species are not related to phylogenetic history.
Water flux
Water influx rates of V. indicus were much higher in the wet season than those predicted for lizards in arid and tropical zones (Nagy 1982), but in the dry season, the influx rates were less than the predicted value. Varanus mertensi has greater rates of water flux than those predicted by Nagy (1982) in both the wet and dry seasons
(Christian et al. 1996d). However, the wet dry tropics of northern Australia do not fit neatly into these categories because although the wet season is humid and could be classed as ‘tropical’, the dry season has characteristics similar to those in the arid zone.
Given that V. indicus inhabit environments that abut floodplains and river systems year round, it is reasonable to assume that they have access to water throughout the year and therefore obtain at least part of their water from sources other than food. To cope with estuarine environments they may possess specialized salt secretory glands
5. Varanus indicus physiology 100 like another Australian varanid mangrove specialist, V. semiremex (Dunson 1974).
The dry season water flux rates of V. indicus are lower than those of V. mertensi but not different from the similar sized terrestrial species studied to date (Christian et al.
1995, Christian et al. 1996b). Varanus mertensi (which is active year round) showed no seasonal change in water flux rate, whereas, like V. indicus , the dry season water flux rates in V. panoptes and V. gouldii (which reduce their activity for some months) were half their wet season values.
The water economy index (WEI, Nagy and Peterson 1988) assesses the water fluxes of animals relative to their energy expenditure. The lower the ratio, the less water animals use for the same energy output, and thus desert animals tend to have lower
WEI than non desert species. The dry season WEI for V. indicus was lower than the semi aquatic V. salvator and the terrestrial V. bengalensis (calculated from Dryden et al. 1992, dry season values reported only). Both of these species inhabit areas in Sri
Lanka with abundant water supplies, and they obtain approximately 60% of their water from drinking and pulmocutaneous exchange. Although water is always available for V. indicus , food may be less abundant during monthly low tide phases in the dry season (when the adjacent floodplain is also dry), and hence V. indicus may forage less.
Field metabolic rates and seasonal energy budgets
The extent of seasonal metabolic reduction exhibited by V. indicus (38%) during the dry season is low compared to other varanid species in the region, except V. mertensi
(33%, Christian et al. 1996d) . In V. mertensi , the reduction is due to the passive consequence of lower night time temperatures, whereas this phenomenon only
5. Varanus indicus physiology 101 accounts for 3% of the dry season reduction in V. indicus , the remainder due to decreased activity. This observation is supportive of other radio tracking data
(Chapter 6) indicating that V. indicus is less responsive to seasons than the savannah species, but more so than V. mertensi . So far, all lizards studied in the Australian wet dry tropics exhibit lower FMRs in the dry season, but the relative contributions of the behavioral and physiological mechanisms to achieve this varies among species
(Christian et al. 2003).
The estimated energy expenditure for the V. indicus population in the wet season is
1,328 kJ ha 1 d 1 and in the dry season, 827 kJ ha 1 d 1. The energetic expenditure of an island population of the iguana Cyclura nubila was estimated as 4,800 kJ ha 1 d 1
(Christian et al. 1986). This high value resulted from the combination of large body sizes (up to 6,000g c.f. V. indicus max =3,750 g) and a high population density, over two times the maximum value estimated for V. indicus along Adelaide River ( C.
1 1 nubila 23 ha cf V. indicus 10.9 ha ). Cyclura nubila also select higher daytime T b’s
(30.5 38.6 °C) which would contribute to this higher value. [0]
A population of the small (mass= 30 – 40 g) insectivorous agamid Lophognathus temporalis , which lives in tropical savannahs of northern Australia, expends 113 kJ ha 1 d 1 in the dry season and 612 kJ ha 1 d 1 in the wet season (Christian et al. 1999a) even though they can persist at much higher population densities than V. indicus
(107 125 lizards ha 1). Similarly the densities of Anolis bonairensis (mean= 8 g) ,
Cnemidophorus murinus (mean= 77 g) and Gonatodes antillensis (mean= 1 g) in the
Caribbean were 1318, 561, and 4200 individuals ha 1 respectively (Bennett and
Gorman 1979) and their field active estimates of energy expenditure were 693, 2,510
5. Varanus indicus physiology 102 and 379 kJ ha 1 d 1, respectively. The herbivorous Galapagos marine iguana
Amblyrhynchus cristatus can reach densities of 2,000 individuals km 1 of coastline on some islands (Wikelski et al. 1997) and the daily energy expenditure of a representative 1 kg adult is 70 kJ d 1 (Nagy and Shoemaker 1984), which is approximately half the average value for V. indicus.
Clearly the amalgam of body size, density, foraging mode, seasonality and activity levels determine the energy flow through different lizard species in different systems.
Varanus indicus provides an interesting example because it shows generally intermediate physiological responses to seasonal changes, between the semi aquatic
V. mertensi that remains active year round and the terrestrial V. gouldii and V. panoptes that spend at least some time inactive during the driest part of the year. An investigation of the seasonal productivity of the mangrove system along Adelaide
River (particularly with regard to V. indicus ’ prey items) would help to determine the causes of their reduced activity and energy expenditure in the dry season.
103
X ray photograph of V. indicus showing location of implanted transmitter and antenna, as well as hemipeneal ossifications.
Home range and movements of
V. mertensi and V. indicus
6. Home range and movements 104
6. Home range and movements of V. indicus and V. mertensi
Abstract
Varanids are some of the world’s largest lizards with many large terrestrial species occupying large home ranges. Little is known, however, about space use by semi aquatic varanids. I used radio telemetry and mark recapture techniques to examine differences in space use between sexes, seasons, and body sizes in the two semi aquatic Australian varanids, Varanus mertensi and Varanus indicus . Both V. mertensi and V. indicus were active throughout the year, unlike other terrestrial varanids in
Australia that are limited by resources (particularly food and water) during the dry parts season. Varanus indicus has a surprisingly small home range when compared to that of other similar sized varanids, and the size of Varanus mertensi ’s home ranges vary substantially, the full extent of which may not have been detected despite the duration of this study. Body size was the only discernible influence on the home range of V. indicus (larger animals having larger home ranges), whereas no influences on the home range of V. mertensi could be detected. An interaction between body size and gender influenced the distances V. indicus moved, whereas interactions between body size, gender, and seasonality all influenced the distances
V. mertensi moved. Given the linear nature of the space utilised by V. mertensi , the calculation of home range asymptotes were not considered realistic for this species, because bootstrap and similar methods have no flexibility at present for removing sections of minimum convex polygons that are unlikely to be used. In contrast, the two year mark recapture data for V. indicus supported the precision of the home range estimates derived from radiotracking.
6. Home range and movements 105
Introduction
It is generally accepted that most animals limit their movements to a defined area over a period of time and this area is typically referred to as a ‘home range’
(Kernohan et al. 2001). Considerable work has been devoted to examining the factors that affect inter and intra species variation in home range size. Lizard home ranges increase with body mass and decrease with increased habitat productivity (Turner et al. 1969, Christian and Waldschmidt 1984, Perry and Garland 2002). Larger species with higher energetic requirements must occupy large areas to obtain sufficient food, but even small species occupy relatively large home ranges when the availability of food is low (Stanner and Mendelssohn 1987). Variation in home range size amongst lizards has also been attributed to differences in age and sex (Schoener and Schoener
1982), dominance status (Schoener and Schoener 1982), trophic level (Schoener
1968), and foraging mode (Rose 1982, Christian and Waldschmidt 1984, Perry and
Garland 2002).
Pronounced seasonal dichotomies in many climates leads to reduced activity or even inactivity by lizards during colder or less productive periods (Green and King 1978,
Christian et al. 1995, Griffiths 1999). Some species of terrestrial lizards in wet dry tropical climates reduce their body temperatures as well as their activity, and lower their metabolic rates during the season when food and water resources are scarce
(Christian and Weavers 1996, Christian et al. 1999a). Two semi aquatic lizards
(Varanus mertensi and V. indicus ) from tropical Australia do not respond in this way, ostensibly because of year round access to food and water. Chapter 5 and Christian et al. (1996d) have demonstrated that they can both species remain active throughout
6. Home range and movements 106 the year, but to what extent their use of space is effected by seasonality has not yet been examined.
Several studies have examined the use of space by varanids in various parts of the world, and in general, these studies have found that males tend to move further than females in the breeding season (Stanner and Mendelssohn 1987, Carter 1990,
Thompson et al. 1998, Guarino 2002, Ibrahim 2002, Perry and Garland 2002).
Varanids also show large home range overlap and little evidence for territoriality
(although see Sweet 1999 for an exception), and some studies over longer periods have demonstrated that some varanids do not form fixed home ranges at all
(Auffenberg 1981, 1988, 1994).
Northern Australia has a high diversity of varanids with as many as 11 species occurring in the same region (Sweet 1999). Clearly, varanids have been able to exploit a wide variety of niches and habitats: saxicoline ( V. glebopalma, V. acanthurus ), arboreal ( V. prasinus, V. glauerti ), semi fossorial ( V. primordius ) and semi aquatic ( V. indicus, V. mertensi, V. mitchelli, V. semiremex) (Pianka et al.
2004). The use of space by varanids in these structurally complex environments will no doubt be different to the ground dwelling species (e.g. V. gouldii and V. panoptes ).
There are four semi aquatic species of varanid in Australia, and the larger two,
Varanus mertensi and Varanus indicus , inhabit ostensibly different aquatic environments. The objective of this study was to examine the use of space by these two large, semi aquatic varanids in the tropical north of Australia. Specifically, I
6. Home range and movements 107 examine whether gender, body size and seasonality affect the space used by these species. In keeping with the findings of Chapter 5 and Christian et al. (1996d), I hypothesize that seasonal effects on V. indicus and V. mertensi’ s space use will be less pronounced than in other sympatric varanids and that body size will scale positively with space use. I also predict a male bias towards greater movements, in agreement with all other work on varanids to date.
Materials and methods
Radio telemetry
For study sites of both species and all capture methods employed refer to Chapter 3.
Individuals of both species were implanted intra peritoneally with VHF radio transmitters (Holohil SB 1 and SB 2, Canada). Prior to implantation, the 13 V. mertensi analysed in this study were sexed using external morphological features and hemipenial eversion (where possible), and all 11 V. indicus were x rayed to determine gender.
Varanus mertensi was radio tracked manually by boat at least once every month between August 2000 and March 2002. Individuals were located using an ATS
Receiver (Advanced Telemetry Systems, U.S.A.) and hand held Yagi antenna. All locations were recorded using a GPS (Garmin 12XL). The majority of fixes of V. mertensi were obtained during the wet seasons of 2000 and 2001, and seasonal comparisons of movements are limited to the fixes of only five individuals in the dry season of 2001 due to transmitter failures.
6. Home range and movements 108 For V. indicus , daily locations of each lizard were recorded within a two week period in both the wet season and dry seasons (see Chapter 3 for definitions) of 2002.
Because many V. indicus were located in the same place after many consecutive days, a more intensive period was undertaken in the 2003 dry season, incorporating four fixes d 1 (where possible, restricted by tides), three hours apart, to examine whether individuals were moving throughout the day but then returning to the same locations. The poor condition of the access track in 2003 wet season prevented a similarly intensive wet season sample. The closest adjacent mangrove strip to the V. indicus study site was located downstream on the eastern (opposite) bank; it was surveyed numerous times for missing radio tracked varanids to document possible dispersal events or larger home range areas; however none were ever recorded.
Home ranges of V. mertensi and V. indicus
For animals that occupy linear habitats such as river systems, conventional calculations of home range size are not applicable (Tucker et al. 1997, Kay 2004).
For animals whose movement patterns are geographically restricted, GIS software can be used to subtract unused habitat from initial calculations of home range area.
Linear home ranges have been used to quantify home ranges of freshwater crocodiles
(Tucker et al. 1997), river otters (Melquist and Hornocker 1983), and estuarine crocodiles (Kay 2004). Given any two points along the bank in a linear home range, the path the animal traversed is known at the broad scale. This definition becomes less precise when considering animals traversing a linear habitat around a polygon, such as a dam (i.e. animals could swim across, rather than travel around). Although many conventional methods of space use analysis are advantageous because of greater comparability between studies and the relatively objective criteria for
6. Home range and movements 109 determining ‘normal’ movements and activity centres, these methods were deemed inappropriate for V. mertensi . I therefore estimated the home range of V. mertensi in a linear fashion that I defined as the distance in metres along the bank between two most distant points recorded for each individual, delineated by a 10 m buffer either side of the bank edge. This buffer was chosen because V. mertensi seldom traveled further than 10 m away from the water’s edge even when being pursued (pers obs) and were often detected in reeds or lilies up to 10 m into the water. No observations were made of V. mertensi swimming across the dam, either directly or in successive daily telemetry locations. Therefore, these linear home ranges are considered a valid estimate of the space use for this species. For V. indicus , conventional Minimum
Convex Polygons (MCPs) were generated, and unused habitats (e.g. floodplains in the dry season) were removed from the resultant polygons.
Analyses
All spatial analyses were performed using ArcView GIS software (ver. 3.2a, ESRI,
Redlands, California) with the animal movement extension (AMAE ver. 2.04, Hooge and Eichenlaub 2001) and geo referenced base layer maps from the Northern
Territory Government, Department of Natural Resources, Environment and the Arts.
Common automated methods, such as bootstrapped area observation curves and incremental analysis were not appropriate for V. indicus because of a large number of fixes from the same location (see Results). Random selection of subsets of these points in many instances will not define an MCP. Incremental area curves were not constructed for V. mertensi because the home ranges of V. mertensi are highly linear; therefore, the only way that area curves would not asymptote is if each individual
6. Home range and movements 110 continuously traveled in the one direction around the bank of the dam and thus increasing its home range, a clearly untenable notion.
Although locations obtained over short intervals such as the two week periods examined for V. indicus are generally termed ‘activity ranges’ rather than home ranges (Thompson 1994, Thompson et al. 1998, Thompson et al. 1999), I examined the accuracy of the home range estimates by comparing them with data from a simultaneous mark recapture study (Chapter 7). Using these data, I calculated home ranges from the maximum straight line distances between the two furthest traps in which an individual was caught. In this analysis, I used only individuals that were captured more than three times and whose entire recapture record was contained within the inside of the trapping grid (i.e., with no records at the extreme ends of the site), were used.
Range overlap
Static interactions between individuals were stimated using percentage overlap of
100% MCPs for V. indicus, and LHR for V. mertensi :
R1,2 = A 1,2 /A 1 and R 2,1 = A 1,2 /A 2, where R 1,2 is the proportion of Animal 1’s range overlapped by Animal 2’s range,
R2,1 is the proportion of Animal 2’s range overlapped by Animal 1’s range, and A 1,2 is the area of overlap between R 1 and R 2 (Kernohan et al. 2001). Straight line distances traveled were calculated using the Animal Movement package for Arcview
3.2 (for V. indicus ), and the network analyst function (Arcview 3.2) for V. mertensi .
The latter model constrained all distance calculations within the linear home range of each animal.
6. Home range and movements 111
Data analysis
Three explanatory variables that have been shown to be influential on the space use of varanids were used to generate multiple working hypotheses to examine which variables best predict various spatial statistics in both species. I hypothesize that either gender (Stanner and Mendelssohn 1987, Carter 1990, Thompson et al. 1998,
Ibrahim 2002, Perry and Garland 2002), body size (Turner et al. 1969, Christian and
Waldschmidt 1984), season (Green and King 1978, Auffenberg 1988, Auffenberg et al. 1991, Ibrahim 2002), or combinations of these variables, would influence the movements and home ranges of V. indicus and V. mertensi . Accordingly, these were used as single variable models, or in combinations, to represent each hypothesis
(Table 1).
Each model was run as a generalised linear model (GLM) in the program R (ver
1.9.0, R Development Core Team 2004). Because the ratio of sample size to parameters was usually low, the second order bias corrected form of AIC (AIC c) was used as the basis for all model selection (Burnham and Anderson 2002). For calculations of the distance an animal is likely to move and the probability of an animal moving (as both of these datasets contained repeated measures of the same individuals), model selection was undertaken using mixed effects models that account for individuals as a random effect (Crawley 2003). Examination of subsequent interactions was carried out on the full effects models using the effects package in R (Fox 2003).
6. Home range and movements 112 Table 1. Candidate model sets for predicting various spatial statistics in V. mertensi and V. indicus , the hypothesis examined under each model and references concerning space use in varanids leading to each model.
No. Model. Hypothesis Reference 1 sex Male and female varanids Stanner and demonstrate differences in their Mendelssohn 1987, use of space. Carter 1990, Thompson et al. 1998, Ibrahim 2002, Perry and Garland 2002 2 SVL Different sized (SVL) varanids Turner et al. 1969, utilise space differently (e.g. larger Christian and animals travel further). Waldschmidt 1984 3 season Varanids’ movement patterns and Green and King 1978, home ranges vary according to Auffenberg 1988, seasonality. Auffenberg et al. 1991, Ibrahim 2002
6. Home range and movements 113
Results Home range
Within both species home range sizes varied greatly among individuals (Figs 1 and
2, Table 2), and V. indicus displayed home ranges that were more typical of smaller varanids (Fig. 3). Varanus mertensi mean linear home ranges (BA) were 13.6 ha and
6.3 ha for males and females, respectively. Mean home range size V. indicus was 4.7 ha for males and 0.9 ha for females (Table 3).
Body size (SVL) was the best ranked in the candidate set for modeling the home ranges of V. indicus (Table 4), having 94% likelihood of being the best model (19 times more likely than the next model) and explaining 28% of the total variation.
Figure 4 shows the fitted model curve (with 95% confidence intervals) over the actual range of values recorded, which illustrates that home ranges of V. indicus increase only slightly as animals increase in body size. The large home range outlier in Figure 4 is from one large male (2050 g, mean of 100 animals was 1102 g) whose movements were much greater in the dry season of 2002.
Differences in home ranges between seasons, genders or body sizes of V. mertensi could not be discerned definitively as the null model was the best supported (Table
4) and the univariate models sex, season and SVL explained very small amounts of the variance (mean=1.6%).
6. Home range and movements 114
Metres
Figure 1. Home ranges of Varanus indicus throughout 2002 and 2003, based on
Minimum Convex Polygons clipped to remove unused habitat. Open
polygons: males; stippled polygons with thick borders: females. Inset:
General location of study site in Adelaide River, Australia.
6. Home range and movements 115
Kilometres
Figure 2. Schematic representation of the home ranges of V. mertensi around the
bank of Manton Dam. Solid lines: males, dotted lines: females. In reality
these linear home ranges overlay on the bank of the dam. Inset: General
location of study site in Australia.
6. Home range and movements 116
Table 2 . Seasonal home ranges and summary data for Varanus indicus and Varanus mertensi . One fix per day was taken, except for individuals M1, F3, F4 and F5, where the intensive dry season fixes were recorded. Home range is estimated using
100% Minimum Convex Polygons (MCP) for V. indicus and linear home ranges
(LHR) for V. mertensi (see text for details). M= Male, F=Female.
Wet season Dry season max max Sex SVL Mass # MCP dist. # MCP dist. & ID (mm) (g) fixes (ha) (m) fixes (ha) (m) V. indicus M1 360 900 4* 3.0 238 45 0.43 104 M2 435 1200 12 0.4 123 13 0.12 160 M3 480 2050 13 2.8 157 12 12.00 667 M4 450 1550 12 0.7 97 13 2.70 321 M5 380 900 15 1.7 174 15 0.14 134 M6 440 1350 14 0.4 179 F1 390 900 12 0.3 87 22 N/A ** 44 F2 420 1000 12 0.4 111 14 0.92 192 F3 390 640 45 0.20 104 F4 360 500 45 0.27 93 F5 430 900 45 0.49 76 LHR LHR V. mertensi (ha) (ha) M1 420 950 23 2.1 558 27 16.8 4633 M2 500 1850 30 5.1 1620 30 1.9 727 M3 525 2350 59 3.1 964 M4 540 2500 38 7.7 M5 510 2250 30 2.7 1304 M6 520 2400 32 6.0 1990 M7 485 1650 40 6.3 2142 M8 415 1150 20 6.9 3345 F1 460 1750 32 3.1 852 29 5.2 2217 F2 510 2250 29 13.4 2982 14 4.0 969 F3 410 990 31 11.2 964 19 1.1 365 F4 430 1250 31 3.0 F5 485 1350 14 3.0 * Only four fixes could be obtained and then animal could not be followed onto floodplain in wet season ** Animal stayed in two trees for entire period and therefore an MCP could not be generated
6. Home range and movements 117
70
60 V. griseus
50
40
30 V. tristis 20 V. rosenbergi Mean range (ha) V. glauerti 10 V. gouldii V. mertensi V. glebopalma V. indicus V. olivaceous 0 0 500 1000 1500 2000 2500 3000 3500 4000 4500 Mean mass (g) Figure 3. Relationship between range size and body mass for selected medium sized
varanids (V. rosenbergi , Green and King 1978, V. olivaceous , Auffenberg
1988, V. indicus, V. mertensi , this study, V. gouldii , Thompson 1994, V.
glauerti , Sweet 1999, V. glebopalma , Sweet 1999, V. tristis , Thompson et
al. 1999, V. griseus , Ibrahim 2002). Closed diamonds are reported home
ranges, open squares are reported activity ranges (hence home ranges may
be larger).
6. Home range and movements 118 Table 3 . Home range sizes and summary data for Varanus indicus and Varanus mertensi . Note that the number of fixes for V. indicus M1 also incorporates the intensive dry season data (4 fixes per day, see text).
Sex SVL Mass # MCP # & ID (mm) (g) fixes (ha) zero moves V. indicus M1 360 900 49 4.9 0 M2 435 1200 25 1.1 13 M3 480 2050 25 12.0 3 M4 450 1550 25 3.1 14 M5 380 900 30 2.4 8 F1 390 900 34 0.3 14 F2 420 1000 26 1.6 8 LA V. mertensi (ha) M1 420 950 50 21.8 0 M2 500 1850 60 5.3 0 F1 460 1750 61 3.7 0 F2 510 2250 43 14.0 0 F3 410 990 50 1.1 1
6. Home range and movements 119
Table 4 . Summary of model selection using AIC c for the analysis of home range size in V. indicus and V. mertensi , using generalised linear modeling (family: Gamma). log( L) is the maximised log likelihood of the model, AICc is the selection criterion,
K is the number of estimated parameters, i is the difference between the model’s
AIC c value and the minimum AIC c value and wi is the Akaike weight. All models are ranked according to support, thus i = 0 for the best model (bold type). The term
%dev describes the amount of deviation each model explains (%). Only models displaying a reasonable level of support ( i<10) are shown.
Model log( L) AIC c K i wi %dev V. indicus SVL 15.615 36.153 2 0.000 0.940 28 sex 18.574 42.070 2 5.917 0.049 14 null 21.729 45.744 1 9.590 0.008 V. mertensi null 166.564 335.461 1 0.000 0.545 sex 166.346 337.782 2 2.321 0.171 3 season 166.491 338.073 2 2.612 0.148 1 SVL 166.563 338.216 2 2.755 0.137 0
6. Home range and movements 120
14
12
10
8
6 Activity range (ha)
Home range (ha) 4
2
0 360 380 400 420 440 460 480 Snout vent length (mm) Figure 4. Relationship between home range and snout vent length (and 95%
confidence intervals) in V. indicus as predicted by the SVL model in Table
4. Single dots represent observed individuals.
6. Home range and movements 121 Distances traveled
Mean daily and maximum distances traveled varied greatly both between seasons and between sexes (Table 2). Mean maximum distances traveled during the wet season were 147 m for V. indicus and 1672 m for V. mertensi . In the dry season these distances were 189 m and 1782 m for V. indicus and V. mertensi, respectively. Body size, gender and season all influenced the distances V. mertensi move, with the global model including all of these variables being the best supported in the candidate set (Table 5). These three variables also influence the distances V. indicus move, with the best model being the interaction between sex and body size but with substantial support for several other models.
In the dry season similar sized male and female V. indicus move similar distances, with larger individuals of both genders moving further. In the wet season, V. indicus males of all sizes are likely to move similar distances, whereas in the dry season, smaller females are likely to move further than larger ones. The longest movements recorded for V. indicus were less than V. mertensi , but male V. indicus moved greater distances in both the dry (447 m vs. 136 m) and wet seasons (340 m vs. 96 m).
Larger male V. mertensi in the dry season move less than smaller males, whereas females of all sizes appear likely to move similar distances. Both larger males and females move further in the wet season than smaller ones. In the wet season, the longest linear distance traveled by V. mertensi was 1240 m by a male and 2403 m by a female. During the dry season, the longest linear distances traveled were 3073 m by a male and 1276 m by a female.
6. Home range and movements 122 Table 5 . Summary of model selection using AIC c for the analyses of actual distances moved in V. indicus and V. mertensi , using mixed effects modeling (family: Gamma) incorporating individuals as a random effect. All models are ranked according to support, thus i = 0 for the best model (bold type). log( L) is the maximised log likelihood of the model, AIC c is the selection criterion, K is the number of estimated parameters, i is the difference between the model’s AIC c value and the minimum
AIC c value and wi is the Akaike weight. The term %dev describes the amount of deviation each model explains (%). Only models displaying a reasonable level of support ( i<10) are shown.
Model log( L) AIC c K i wi %dev V. indicus sex * svl 40.114 90.715 5 0.000 0.248 14 sex * season 40.258 91.004 5 0.289 0.215 13 season * svl 40.238 90.964 5 0.249 0.219 13 sex 42.673 91.539 3 0.824 0.164 8 svl 42.880 91.952 3 1.237 0.134 7 season 45.531 97.256 3 6.541 0.009 2 null 46.478 97.051 2 6.336 0.010 global model fit 36% V. mertensi global † 130.295 279.996 9 0.000 0.743 13 season * svl 136.362 283.178 5 3.182 0.151 9 sex * season 136.780 284.014 5 4.018 0.100 9 season 141.767 289.713 3 9.716 0.006 6 global model fit 13% † V. mertensi global model: sex*season*svl
6. Home range and movements 123
Daily movements of V. indicus
The dry season intensive radio tracking survey (4 fixes per day, 3 h apart) revealed that V. indicus individuals tend to move small distances more often than was detected by daily recordings of location, but they do not move far from their previous location. In 61% records, animals did not move (between consecutive fixes), and many individuals remained in the same tree for up to seven days. Others moved occasionally throughout the day, but generally only small distances (35–65 m), and most movement was in the middle of the day (Fig. 5).
Overlap
There was home range overlap between individuals of both species, both between and within sexes and mean overlaps in V. indicus home range were consistently greater than those of V. mertensi (Fig. 6). Intraspecifically, there were few obvious patterns except that the male to female overlap for V. indicus increased in the wet season. The lower female to female overlap in both species probably results from fewer radiotracked females.
6. Home range and movements 124
160
140
120
100
80
60
40 Mean distance moved (m) 20
0 09:00 12:00 15:00 18:00 Time Figure 5. Mean distances (± SD) moved by V. indicus throughout the day during two
weeks of radiotelemetry in the dry season of 2003. Values shown are means
of 39% of records, taken 4 times daily (where possible according to tide
height). In the remaining 61% of records, individuals did not move.
6. Home range and movements 125 a) 50
45 40
35
30 25
Overlap % 20 15
10 5 0 F to F M to F M to M
b) 100 90 80 70 60 50 40 Overlap % 30 20 10 0 F to F M to F M to M Gender and seasons Figure 6. Mean (± SE) home range overlap between and within genders of a) V.
indicus and b) V. mertensi . Open bars: dry season; filled bars: wet season;
striped bars: both seasons combined. The y axes are not to uniform scale.
6. Home range and movements 126
These estimates for both species are likely to be conservative because radio tracked
V. indicus were found in the same tree and some were occasionally in the same tree as other individuals that were not being radiotracked (5% of observations) and V. mertensi were often seen together, particularly when basking on large rocks.
Discussion
Space use of semi aquatic varanids
Typically, the areas occupied by terrestrial varanids of similar size to V. mertensi and
V. indicus are large (King 1980, Thompson 1992a, Weavers 1993, Christian et al.
1995, Ibrahim 2002), and many larger terrestrial varanids (>5000g) have home ranges exceeding 150 ha (Auffenberg 1981, Philipps 1995, Guarino 2002). Varanus indicus, however, has home ranges that are typical of smaller species (Thompson
1993, James 1996, Sweet 1999). Previous relationships between body size (mass) and home range size (based mostly on small species, Turner et al. 1969, Christian and Waldschmidt 1984), including the most recent by Guarino (2002, which takes into account larger monitors), also estimated much higher home range values than those recorded herein for V. mertensi and V. indicus . Caution in these comparisons must be stressed because all the home ranges of varanids reported to date vary in study duration and sample size, and few of these studies have examined home range asymptotes in their analysis.
One possible explanation for the smaller than expected home ranges of V. indicus is that they do not need to travel long distances to forage. Daily tidal fluctuations may
6. Home range and movements 127 bring in many of their prey items such as fish, mudskippers and crabs (Lugo and
Snedaker 1974, Losos and Greene 1988, Robertson and Daniel 1989, Saenger and
Snedaker 1993, Smith et. al. unpublished data) and their arboreal nature and the abundance of hollows may provide other known food sources such as birds eggs, small mammals, other reptiles and arthropods (pers obs, Losos and Greene 1988).
Similarly, the arboreal Varanus olivaceous has small home ranges at certain times of the year because they do not travel far from the fruiting trees upon which they feed
(Auffenberg 1988). For both V. indicus and V. olivaceous , the unrecorded but prominent vertical component to their space use would increase the volume of their home ranges (Perry and Garland 2002). A more informative measure may be to define three dimensional home ranges for these species (Jenssen and Nunez 1998,
Lovern 2000).
In contrast to V. indicus , V. mertensi tend not to form stable home ranges, at least not throughout the 3 years of this study. Some of the particularly large home ranges observed for V. mertensi were heavily influenced by single radio tracking fixes, occasionally large distances from where these animals were usually encountered. For example one animal traveled 5.47 km in 5 days. Rose (1982) postulated that lizards that are not territorial and continually forage in search of food may cumulatively increase their home range sizes. It is thus possible that V. mertensi similarly wander
(as they do in other man made water bodies, Christian 2004) as they forage along this linear strip. Similarly Auffenberg (1981, 1988, 1994) reports that some Varanus komodoensis and Varanus bengalensis are transient in this way.
6. Home range and movements 128 While I studied V. mertensi’ s use of space in a large man made waterbody, its home range dynamics may differ in other natural permanent water bodies such as billabongs which are comparatively small. In the wet season, V. mertensi can be found far from these water bodies when water is abundant in the landscape (Christian
2004). Whether these individuals disperse away from or return to the same locations each year remains to be investigated.
Home range sizes of Varanus bengalensis and Varanus albigularis have been found to vary according to the availability of food (Auffenberg et al. 1991, Philipps 1995).
However, for varanids in regions where temperature is more or less constant and where food is available year round (which may occur in permanent aquatic systems), less pronounced seasonal effects would be expected. Indeed this is the case throughout Indonesia for the semi aquatic Varanus salvator (Gaulke et al. 1999).
The most predominant prey item ( Holthuisana sp. ; freshwater crabs ) of V. mertensi in man made irrigation channels of Western Australia, were present during all seasons in both these studies (Shine 1986, Mayes et al. 2005b). Manton Dam may be a similar source of year round resources for V. mertensi .
Although mangrove systems are highly productive (Finlayson et al. 1988), worldwide they show distinct seasonal fluctuations in fish (Robertson and Duke
1990, Rooker and Dennis 1991), crab (Emmerson 1994) and arthropod abundance
(Lefebvre et al. 1994), with the highest abundances being associated with increased rainfall. Many predatory species respond to this seasonality in resources by being more active, and breeding, during these resource rich times (Lefebvre et al. 1994,
Noske 1996). Similarly, V. indicus could be responding to decreased productivity in
6. Home range and movements 129 the mangrove forest during the dry season by eating and drinking (and thus moving) less (Chapter 5). Even if the large seasonal differences in prey abundance usually experienced by varanids in terrestrial habitats are minimised by regular tidal influxes from the Adelaide River, larger individuals may still need to travel further, or more often, to meet their relatively larger energetic requirements (Garland Jr 2002).
Further research is required to determine to what degree food resources fluctuate throughout the year for V. indicus .
Distances traveled
Both species and individuals in this study showed great variation in the distances traveled, both within and between seasons, genders and throughout the range of body sizes. Previous studies have similarly shown that daily distances traveled by varanids can vary greatly between days, seasons and individuals (Auffenberg 1981, Stanner and Mendelssohn 1987, Auffenberg 1988, Auffenberg et al. 1991, Thompson 1992a,
Weavers 1993, Philipps 1995, Thompson et al. 1999). Auffenberg et al. (1991) found no significant relationship between the body size of adult V. bengalensis and the distances they moved each week, and no relationship between sex and distance moved, although some males moved more than females in some seasons.
One particularly large V. indicus (2050 g, mean of 104 animals 1102 g) showed movements during the dry season that greatly increased its home range and which were all made on consecutive days in June. These observations are consistent with other records of male varanids traveling large distances to visit females during their breeding season (Gaulke et al. 1999, Thompson et al. 1999). A pair of V. indicus were also seen mating on a tree branch (in a copulation position as described by King
6. Home range and movements 130 and Green 1999) during this time. Thus, it appears that large male V. indicus also travel greater distances during their breeding season (late June/early July), presumably to find mates.
Varanus indicus daily activity
Intensive (3 hourly) radio tracking (and trapping data) showed that V. indicus occasionally move short distances during the lower tides throughout each day, but never very far, and thus the daily measurements were not recording the small movements they make as they emerge (presumably to forage) for a small distance and then return to their original tree. Varanus indicus moved relatively little when compared to other similar sized, terrestrial varanids (Green and King 1978, King et al. 1989, Auffenberg et al. 1991, Ibrahim 2002), V. mertensi (this study) and the larger but also semi aquatic Varanus salvator (Gaulke et al. 1999). Again, by decreasing their activity (i.e., not moving for many days or moving very little) V. indicus would decrease their energy expenditure (as demonstrated in Chapter 5) and need to forage less. It is likely V. indicus are employing this sedentary strategy whilst occupying a productive system in order to enable them to attain such a comparatively large body size. These findings agree with Perry and Garland (2002), who have shown that the distance lizards move (hence home range sizes) are mediated by dietary needs and food availability.
The activity of Varanus indicus during high tide periods is unknown, although they are observed more commonly as the tide is rising (pers obs). Preliminary spool and line tracking suggested that they travel along small run off channels within the
6. Home range and movements 131 mangrove forest as the water recedes, to forage on prey traveling out with the tide
(pers obs), as has been observed for Crocodylus porosus (pers obs).
Overlap
The large overlap of home ranges between the sexes and throughout seasons for both
V. indicus and V. mertensi is consistent with other studies that show varanids are generally not territorial (Green and King 1978, Auffenberg 1981, Stanner and
Mendelssohn 1987, Thompson 1992a, Thompson 1994, although see Sweet 1999 for an exception). Further support for this result was provided by the V. indicus trapping data, where many individuals were captured simultaneously, including up to three individuals caught in the same trap (Chapter 3). This, together with the fact that other
V. indicus were commonly observed in the area, clearly shows high densities of animals with overlapping home ranges, further evidence for high productivity (at least for V. indicus ’ food) in this mangrove system.
Overlap values for V. mertensi were only derived from the edges of each animals home range (as they are linear around the bank edge), and these values are probably also underestimated since only 22% of the 57 individuals caught in the study period were radio tracked, and individuals were observed many times basking together on large rocks. Christian (2004) also records large overlapping home ranges for V. mertensi.
Home range estimation ( V. indicus )
Trapping results show that the home range estimates derived from radio tracked V. indicus are probably very close to the entire space they use . Eleven trapping sessions
6. Home range and movements 132 (conducted quarterly over three years), had an average maximum distance between the two furthest recaptures (up to 16 for one individual) of 152 m (Fig. 7). The mean maximum distance between all recapture points for each individual was 188 m, indicating that the combined wet and dry season activity ranges as determined by radio telemetry are probably representative of home ranges for this species.
Previous studies on small mammals have shown that trapping produces smaller home range estimates than radiotracking (Pavey et al. 2003), particularly at low densities
(Ribble et al. 2002). This phenomenon does not seem to be the case for V. indicus, and may be partly explained by the discrete nature of the mangrove forest and the effectiveness of the trap design. For instance, radiotelemetry revealed that very few animals (n=2 animals, 1.7% of total fixes) ventured out into the adjacent floodplain or fish farm during this study. Given that the recapture rates for V. indicus can be quite high (Smith 2004, Chapter 3), it is likely that sedentary individuals were being recaptured sufficiently frequently to estimate their home range to a similar precision as that detected by radiotracking.
6. Home range and movements 133
400 350 R2 = 0.53 300
250
200 150
100
50
Maximum distance between traps (m) 0 0 5 10 15 20 No. of recaptures Figure 7. Number of recaptures of V. indicus against maximum distance (m)
between the furthest two traps in which each individual was recaptured.
6. Home range and movements 134 Conclusions
Influences such as season and body size govern the spatial dynamics of V. indicus and V. mertensi , yet these effects are less influential for these two semi aquatic species than for terrestrial varanids, most likely because of the apparently high (and more seasonally constant) availability of food and water for these species.
Examination of the varanid literature reveals that much variation exists in both inter and intra specific spatial interactions and no single variable or small set of variables is likely to govern their space use. This is a reflection of varanid’s broad ecological variation and life history traits. While the use of GIS based software has increased our understanding of animals’ use of space, development of more sophisticated programs to examine complex systems (such as the three dimensional use of space by arboreal or semi aquatic species, Lovern 2000) would provide more realistic estimations of the space used by species such as V. indicus and V. mertensi .
135
Setting pipe traps in the mangroves A set pipe trap
Population dynamics of V. mertensi and V. indicus Searching for V. mertensi at Manton Dam
7. Population dynamics 136
7. Population dynamics of V. mertensi and V. indicus
Abstract
The population dynamics of Australia’s two largest semi aquatic varanids, Varanus indicus and Varanus mertensi , form the basis for this chapter. Estimation of V. indicus’ survival probability was attempted using mark recapture data from 3 years of trapping data, whilst V. mertensi survival probability was derived from known fate modeling of radiotracked individuals (2.6 years). Violation of the equal catchability assumption by V. indicus precluded any investigation into the probable causes of any variation in apparent survival probability for this species. However for V. mertensi, I demonstrate definitively what intuition suggests; that apparent survival probability in long lived lizards is high over short sampling periods, with body size, gender, and a linear trend all influencing these estimates. Both species, in particular V. indicus showed very high densities when compared to other varanids of their size, this high density likely results from the highly productive habitats with which they are associated. Radiotracking was found to be more effective for the long term monitoring of varanid populations; however, removal of all traps and old bait sources between mark recapture sampling periods is a possible method to reduce trap shyness in further studies of this type. For survival estimation in populations of species as long lived as varanids, longer term studies (perhaps spanning decades) are required.
7. Population dynamics 137
Introduction
Lizards are a common component of many faunal assemblages throughout the world
(Pianka and Vitt 2003) and have evolved a variety of life history strategies (Ballinger
1983, Dunham and Miles 1985, Shine 1985, Dunham et al. 1988, Vitt et al. 2003). At the population level, these variants in life history traits ultimately correspond to differences in longevity, fecundity, and abundance amongst the taxa (Pianka and Vitt
2003). Two major generalisations can be made about the population dynamics of lizards: 1) longevity is highly variable between taxa (Auffenberg 1981, Thompson et al. 1992, Bull 1995, Elmouden et al. 1997, Hutchinson et al. 2001, Slavens and
Slavens 2003); and 2) most species (of any size) have very high mortality of hatchlings and juveniles (Auffenberg 1981, Civantos and Forsman 2000, Pianka and
Vitt 2003). Survival of adult lizards can vary between sexes (Laurie and Brown
1990b), animals of different body sizes (Laurie and Brown 1990a), and with seasons
(Tinkle 1967, Ferguson et al. 1980, Bauwens 1981, Lebreton et al. 1992).
Importantly, introduced species of both predators (Berry and Gleeson 2005,
Schoener et al. 2005) and prey (Doody et al. 2006) can dramatically alter lizard populations.
Past studies of survival in lizards examined relatively short lived species (Turner et al. 1970, Ferguson et al. 1980, Andrews 1991, Diaz 1993, Tinkle et al. 1993,
Schoener et al. 2004) or estimated survival for individuals during the early life stages only (Andrews et al. 2000, Civantos and Forsman 2000, Fox and McCoy 2000,
Olsson and Madsen 2001, Hare et al. 2004, Le Galliard et al. 2005). Demographic
7. Population dynamics 138 studies of long lived species have shown that the probability of survival in adults is high between any given years (Auffenberg 1981, Bull 1995).
Although the demography and life history of most varanids in Australia is unknown
(Bennett 1998, although see James 1996) varanids in general are long lived (Flower
1933, 1937, Auffenberg 1981, Smirina and Tsellarius 1996, de Buffrénil and Hémery
2002). Many varanid populations in northern Australia are threatened by the introduced cane toad Bufo marinus (Doody et al. 2006, Chapter 2). Any large population reduction of long lived varanids could have strongly negative implications for their long term persistence and repercussions throughout food webs in which they are components, particularly because they are often the top predators
(Pough 1973, Losos and Greene 1988).
Here I estimate the survival probability of two northern Australian species of varanid, Varanus mertensi and Varanus indicus . Both are medium sized, semi aquatic varanids, which inhabit freshwater and estuarine environments (respectively) throughout tropical northern Australia (Pianka et al. 2004). This study represents the first attempt to quantify the possible effects of seasonality, gender and body size on the probability of survival in these species. I predict that survivorship is high in both species and (based on information in other chapters) that seasonality has little effect on this parameter. I also report on the sex ratios of V. mertensi and V. indicus , discuss the precision of these estimates, and investigate the implications for each species on their estimated densities.
7. Population dynamics 139
Methods
Regional climate
Although north Australia has a monsoonal climate characterised by distinct wet (high rainfall, high humidity) and dry (little rainfall, low humidity) seasons (Bowman
2002), the further two transitional periods of high humidity and little rainfall; late wet early dry and late dry early wet were also used. These periods coincide with the trapping periods for V. indicus (see below), and the wet dry dichotomy enables comparisons between other studies.
Study areas and capture details
Varanus indicus was trapped quarterly (48 traps, 6 days each period) between
September 2002 and March 2004 (see Chapter 3) in a patch of mangrove forest. This patch is bounded by the Adelaide River, its associated floodplains and a commercial fish farm, all of which V. indicus rarely enter (Chapter 6). Thus, emigration from the site was expected to be minimal during each trapping occasion, and only the end of the trap line facing the un trapped section of forest (200 m wide) constituted an area with a high probability of dispersal into or out of the trapping grid.
7. Population dynamics 140
Varanus mertensi were captured along the shoreline of Manton Dam (12º 87' S, 131º
11' E) using a small, motorized boat and noosing pole. Twenty two adult V. mertensi were implanted with radiotransmitters (Holohil SB 1 and SB 2, Canada), and an attempt at resighting was made each month between August 2000 and April 2002
(see Chapter 3 for details).
Model development and parameter estimation
Following the recommendations of Burnham and Anderson (2002), multiple working hypotheses were developed a priori to identify the variables (or combinations thereof) that were most likely to influence survival of varanids (Table 1). A set of additive capture mark recapture or known fate models was then developed to represent each hypothesis. Model selection was performed using Information
Theoretic model selection methods based on Akaike’s Information Criterion (AIC,
Burnham and Anderson 2002) and model averaged values were then used to investigate any change in survival probability over the course of the study.
7. Population dynamics 141
Table 1. Group and individual covariates used in the parameterisation of survival and recapture probability models for Varanus mertensi and Varanus indicus .
Covariates/ Factors Values Description Group covariate Sex (sex) * † F Male Gender Female Time specific group covariates Trend (tr) * † C As per number of Variable for linear trend capture intervals over the coarse of the experiment Season (seas) * † F Wet Two major seasonal periods Dry Total rainfall (mm) from the three * C Rainfall (r) months preceding each capture period (recorded 10 km away) Period (per) * † F Early wet, Late wet, Divides the year Early dry, Late dry into four 3 monthly periods Uniform probability across all * † Constant (.) No variation groups and time intervals Individual covariate Snout Vent Length (SVL) * † C Body size expressed as svl (mm) at first capture *= V. indicus F= Factor †= V. mertensi C= Continuous variable
7. Population dynamics 142
Survival model development
Both V. mertensi and V. indicus appear to be less responsive than terrestrial varanids
(sensu home ranges, movements and physiological responses) to the vast seasonal fluctuations of the region, because they both inhabit areas (mangroves and permanent water bodies) ostensibly with access to resources throughout the year, unlike the majority of other varanid species (Chapters 5 and 6, Christian et al. 1996d).
Nevertheless, the seasonal parameters (rainfall, season and its subset, period) were incorporated into the candidate model sets to investigate the potential effect of these phenomena. Because the body size and gender of lizards influences their home range size (Turner et al. 1969, Schoener and Schoener 1982, Christian and Waldschmidt
1984, Perry and Garland 2002) and their exposure to predators, these variables were also included in the survival models. A null model (with no fixed differences in apparent survival) and a linear trend in survival over time were also incorporated into each candidate set. Gender was incorporated as a group factor because individual covariates draw on more computer processing power but have no effect on model parameter estimates or model selection (Franklin 2000).
Survival estimation
The encounter histories developed by relocating radiotracked V. mertensi each month were implemented into a known fate analysis in Program MARK (White and
Burnham 1999). This analysis estimates the apparent survival parameter S when unhindered by recapture probability (Cooch and White 2006).
Survival of V. indicus was estimated using Cormack Jolly Seber (CJS) models, which consider survival and recapture probability separately, in Program MARK
7. Population dynamics 143
(Cooch and White 2006). Goodness of fit of the global model ф s+l+r+seas+per ps+l+r+seas+per (without the individual covariates) was assessed using the parametric bootstrap approach in Program MARK. One thousand bootstrap simulations indicated that the probability of obtaining a deviance as large, or larger, than that observed for the global CJS model, was 0.088, thus the model ф s+l+r+seas+per ps+l+r+seas+per showed no strong violation of the catchability assumption. However because the number of recaptures declined markedly after the first half of the study
(Fig. 1), trap ‘shyness’ (where among individuals captured at occasion i, the “old” individuals tend to be less recaptured than the “new” individuals) was tested for using the program U Care (Choquet et al. 2003).
Density estimation
The density at Manton Dam was calculated from the number of individual captures divided by the area that they are known to occupy (the shoreline incorporating 10 m into the water and 10 m on the bank, Chapter 6) around the entire bank. This value is the minimum density of animals known to be present at Manton Dam.
Estimates of the population density of V. indicus were calculated using:
N(D)= (N / P) pi (A) where N(D) is the density of animals (ha), N is the number of animals caught , P is the recapture rate and pi (A) is the proportion of the study area encompassed by the trapping grid (Matlock et al. 1996). The fourth trapping period was used to estimate density because high numbers were still being captured at that time, and the recapture rate had not yet started to decline (see Results).
7. Population dynamics 144 Results Recapture
Throughout the capture recapture study of V. indicus (3,168 trap days), a total of 100 individuals (93 adults) were captured, with over 76% (mean SVL= 445 mm, range=
310 590 mm) of these subsequently recaptured at least once. The maximum number of captures in one six day period (288 trap days) was 37 individuals, with recaptures increasing over the first half of the study and gradually declining in concert with the number of captures (Fig. 1). Trap ‘shyness’ was confirmed using U Care (χ 2= 29.4, two sided test, df= 31, P = 0.027), thus violating the assumption of equal catch ability amongst individuals and precluding any further analyses of recapture rate.
Survival
The average quaterly adult apparent survival probability of V. indicus was estimated from the global model as 0.70 and annual adult survival probability was estimated as
0.24. Because the assumption of equal catchability of V. indicus was violated, no model ranking or further inferences could be made about their apparent survival .
Over the two years of the study, only one V. mertensi (a female) was found dead, with the transmitter and some remains located at the base of a white bellied sea eagle
Haliaeetus leucogaster nest. Five survival models, incorporating sex, a linear trend, and the effect of periodicity for V. mertensi , could not be differentiated from one another (as the i were all less than 2; Table 2) and together comprised 66% of the model weights. Model averaged values indicate that average monthly survival probability was 0.99 for both males and females during half of the study period, with a gradual decline towards the conclusion of the study (Fig. 2), paralleling a decline in
7. Population dynamics 145
40 1
0.9 35 0.8 30 0.7 25 0.6
20 0.5
0.4 No. captures 15
0.3 Recaptures/ captures 10 0.2 5 0.1
0 0 Late Early Late Early Late Early Late Early Late Early Late Dry Wet Wet Dry Dry Wet Wet Dry Dry Wet Wet 2002 Trapping Period 2004
Figure 1. Number of captures of male and female (combined) V. indicus (bars) and
proportion of recaptures/ captures (line) within each trapping period.
7. Population dynamics 146
Table 2 . Summary of Akaike’s information criterion (AIC c) and associated statistics for candidate known fate models for the analysis of the survival probability ( S) of V. mertensi . All models are ranked according to support, thus i = 0 for the best model
(bold type). log( L) is the maximised log likelihood of the model, AICc is the selection criterion, K is the number of estimated parameters, i is the difference between the model’s AIC c value and the minimum AIC c value and wi is the Akaike weight. The term %dev describes the amount of deviation each model explains (%).
Only models displaying a reasonable level of support ( i <10) are shown.
Model log( L) AIC c K i wi %dev
Str 3.959 12.000 2 0.000 0.277 34 Ssex + tr 3.216 12.595 3 0.595 0.206 47 Ssvl + tr 3.653 13.469 3 1.469 0.133 39 S. 6.014 14.055 1 2.055 0.099 Ssex 5.182 14.445 2 2.445 0.082 14 Ssex + SVL + tr 3.142 14.558 4 2.558 0.077 48 Sseas 5.363 14.807 2 2.808 0.068 11 Ssex + SVL 5.015 16.194 3 4.194 0.034 17 Sper 4.882 18.037 4 6.037 0.014 19 Ssex + per 4.068 18.550 5 6.550 0.010 32 Subscripts: sex= gender tr= linear trend seas= season per= period .= constant SVL= snout vent length
7. Population dynamics 147
1
0.9 0.8 0.7
0.6
0.5
0.4 0.3 Survival probability ф 0.2
0.1
0 Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul 01 Aug Sep Oct Nov Dec Jan Feb Mar 00 00 00 00 00 01 01 01 01 01 01 01 01 01 01 01 02 02 02 Date
Figure 2. Male and female V. mertensi monthly survival probabilities (± 95% CI)
based on model averaged values from all models in Table 2. Males; solid
line, females: dotted line.
7. Population dynamics 148 functioning transmitters. Annual adult survival probability for V. mertensi was 0.89 for both sexes.
Density and sex ratios
A male to female sex ratio of 1.7:1 was recorded for V. mertensi and the minimum density of V. mertensi around Manton Dam was estimated as 4.6 adult varanids ha 1
(note: each hectare is designated as a 500 m long, 20 m wide linear trace of the shoreline). A male to female sex ratio of 2.2:1 was recorded for V. indicus and the density of V. indicus was estimated as 10.9 (95% CI upper=26.2, lower=10.1) adults ha 1.
Discussion
V. indicus recapture
Many varanids, including V. indicus , display movement and home range size disparities between genders across different body sizes (Stanner and Mendelssohn
1987, Carter 1990, Thompson et al. 1998, Ibrahim 2002, Perry and Garland 2002,
Chapter 6). In this study, males were captured and recaptured more often, lending further credence to the findings of Chapter 6 that males (particularly large ones) move more often than females.
Survival
The high survival probabilities estimated for V. mertensi throughout this study are not surprising because varanids generally are known to be long lived. For example, the time to maturation for many varanids is between 1 3 years (King and Rhodes
7. Population dynamics 149
1982, Auffenberg 1988, Bennett 1998, de Buffrénil and Rimblot Baly 1999) and V. komodensis take 5 6 years to reach maturity (Auffenberg 1981). The maximum age of free living V. griseus has been estimated at 12 14 years for males and 6 7 years for females (Smirina and Tsellarius 1996) and wild V. glebopalma can live for at least 8 years (pers obs). Varanus indicus in captivity has survived for over 17 years and V. mertensi over 20 years (Snider and Bowler 1992). Captive Varanus flavescens lives for at least 5 years and 8 months, captive V. gouldii and V. varius have survived for almost 7 years (Flower 1937) and a captive V. niloticus lived for 15 years (Flower
1933). Wild Varanus komodensis are reported to live for 20 years and have an estimated lifespan of fifty years (Auffenberg 1981). Thus, varanids are some of the longest lived squamates. In addition, mortality in lizards is generally at its highest among the hatchlings and juveniles (Auffenberg 1981, Civantos and Forsman 2000,
Pianka and Vitt 2003); survival of most adult varanids within any two year period is highly probable.
Although V. mertensi showed relatively constant survival probability, the decreasing survivorship detected in the last 5 months of the study were driven presumably by the decline of battery life in implanted transmitters. The effectiveness of using pipe traps to capture and (more importantly) recapture V. indicus declined after the first year as they became increasingly trap shy. Hence, violation of the assumption of equal catch ability in V. indicus is the probable cause of the lower (relative to V. mertensi ) survival probability estimate. Annually, the probability of an individual V. indicus surviving (as derived from this model) to the next year is very low (24%) for such an apparently long lived species. Clearly, this value is not accurate, and it is probably a much higher value, such as that found for V. mertensi.
7. Population dynamics 150
Using the estimated survival rates from this study, life expectancy (the mean number of years an adult individual could be expected to survive) was predicted by the equation:
−1 Life expectancy = ln(s) where s = mean annual survivorship (van der Toorn 1999). Using this equation, and assuming a similar annual survival rate for both species, life expectancy is 7.8 years.
This value is less than half the value of some captive V. mertensi and V. indicus specimens (Snider and Bowler 1992). Longevity records taken from captive individuals however, commonly signify upper age limits and do not represent the realistic longevity of an individual living under natural conditions (Foufopoulos and
Ives 1999).
Shine and Charnov (1992) demonstrated a curvilinear relationship between the age at maturity and the annual survival rate of lizards and snakes. Their model, however, only considered small lizard species. Although the age at maturity for most species
(including V. mertensi and V. indicus ) is unknown, it has been estimated at 1 3 years for many species of varanid (King and Rhodes 1982, Auffenberg 1988, Bennett
1998, de Buffrénil and Rimblot Baly 1999) and at between 5 6 years for V. komodoensis (Auffenberg 1981). The findings of this study (and others), of survival amongst larger species, concur with Shine and Charnov (1992, Fig. 3). Indeed, given the probable high longevity of many varanids, it is likely that most varanids would occur in this section of the curve.
7. Population dynamics 151
8 7 V. komodoensis 6 5 V. mertensi and V. indicus 4
3
Age at maturity (years) 2
1
0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1 Annual adult survival rate Figure 3: Relationship between age at maturity and known survival rates of lizards
(closed circles) and snakes (open circles) (taken from Shine and Charnov
1992). The dotted ovals represent the likely locations of V. mertensi and
V. indicus , and V. komodoensis in this relationship.
7. Population dynamics 152
Which method best predicts varanid survival?
Although emigration from a population cannot be separated from deaths in Cormack
Jolly Seber models (Cooch and White 2006), massive emigration (or deaths) as would be needed for the apparent survival probability of V. indicus reported here to be real, is highly unlikely, which calls into question the utility of the pipe trap method for long term studies of this type. Emigration on this scale can also be ruled out based on various quantitative data. For instance 11 individuals (11% of the marked population) were implanted with radio transmitters (Chapter 6) and no dispersal across the Adelaide River or out across the adjacent floodplain was detected during that study.
Further evidence for a sedentary V. indicus population (for at least the majority of individuals) is provided by the capture histories themselves. As trapping continued throughout the study, the numbers of animals captured decreased, but of those that were recaptured, many had been captured early on, or many times throughout the study. Of the last four trapping periods, 77% of captures were from the end of the trapping grid that faces the only edge of the forest where no traps were, i.e. these were new animals whose home range possibly covers only some of the trapping grid, not animals that had already been caught before. Similarly, James (1996) pitfall trapped 111 Varanus brevicauda but recaptured only 19 individuals over three years, thus, trapping varanids over the long term may not be a reliable way of collecting long term population data.
The pipe trap capture method was very useful in initially obtaining high numbers of
V. indicus; however, known fate modeling using radiotelemetry data appears to be a
7. Population dynamics 153 much more reliable method for the long term population modeling of varanids.
Radiotracking is less subject to biases in survival estimation because each individual can be sighted at every interval.
A possible modification to the pipe trapping method would be the removal of the traps at the end of each trapping period. During this study, traps were inverted after each period and the subsequent bait pile on the forest was eventually disposed of, presumably by the many scavengers (including V. indicus, pers obs) in the mangroves (Kathiresan and Bingham 2001) and the tide. Although at the beginning of each new trapping session (three months hence) no bait was evident. It is still possible that V. indicus could detect any odors present due to their enhanced vomeronasal system (Cooper 1997), and were habituated to the smell without the presence of bait, and/or, they associated the traps with human presence. Moving such a large number of traps into and out of the mangroves every three months was not done over the course of this study because of the logistical difficulties.
Density
Previous studies have estimated the density of varanids from riverside counts
(Erdelen 1991), observations (Stanner and Mendelssohn 1987, Auffenberg 1988), single searches (Bennett 2000) and mark recapture (James 1996, Bennett 2000). This study is the first to estimate the densities of varanids as derived from recapture probabilities. Caution should be stressed in the over interpretation of these results however, given that they are derived from a model whose main assumption was violated. Even with this caveat in mind, the simple number of varanids ha 1 estimation as calculated for V. mertensi leads to 7.4 for V. indicus , a similar value.
7. Population dynamics 154
The density of the adult V. indicus population along Adelaide River is high, especially when compared to other, similar sized varanids. For instance Varanus griseus densities have been estimated at between 0.01 0.02 animals ha 1 (Pianka et al.
2004) in various desert localities throughout its range, with differences in some instances attributed to environmental factors such as annual rainfall (Stanner and
Mendelssohn 1987) and habitat productivity (Auffenberg 1989). Mean adult V. komodoensis densities have been estimated at 0.03 animals ha 1 and these estimates varied greatly over five sites due to food and shelter resource availability
(Auffenberg 1981). Varanus olivaceous has been recorded at densities of 0.61 animals ha 1 (Auffenberg 1988). The minimum density of V. mertensi (derived from captures only) were lower than V. indicus , although still higher than other varanids recorded to date.
Semi aquatic varanids in tropical regions can sustain higher levels of activity throughout the year than terrestrial varanids (Christian et al. 1996d, Chapter 5) because of the apparently constant availability of food in these more productive environments (Mayes et al. 2005b). Varanus indicus in particular has small home ranges and does not move very far throughout the year (Chapter 6). The ostensibly greater concentration of resources available to semi aquatic varanids may allow for the high population densities seen in this study. Densities of the similarly semi aquatic Varanus salvator in Sumatra (Indonesia) have been estimated at between
0.002 0.01 animals ha 1 from riverside counts (Erdelen 1991), but their method only provides an index of abundance (Williams et al. 2002) and hence the reported
7. Population dynamics 155 densities were much lower than those estimated by experienced varanid skin suppliers from the same area (Erdelen 1991).
The productivity of the environment plays a crucial role in determining both prey abundance (Shine and Madsen 1997) and habitat quality (Greenberg 2001, Galán
2004), and hence effects the underlying population sizes of many taxa (Madsen and
Shine 2000). The resultant biomass in these large populations of semi aquatic varanids also surpasses those of terrestrial species. On Komodo Island, the biomass of V. komodoensis is 335.9 g ha 1, and on Padar it is 231.6 g ha 1, both from an average animal mass of 6000 g (Auffenberg 1981). The average mass of V. indicus is
1102 g and the biomass is therefore 7900 g ha 1. The biomass of V. mertensi is similarly high, at 6700 g ha 1 (mean mass = 1455 g). However, throughout much of
V. mertensi’s range, the density and resultant biomass may differ greatly from those reported here, because during the dry season they would have no choice but to migrate to residual waterbodies as the floodwaters from the wet season recede. These smaller waterbodies may not sustain such high densities of animals.
Sex ratios
The sex ratios presented here are based on genders determined in the field, on species that are notoriously difficult to sex (Chapter 4); thus these data are imprecise. The male to female sex ratio of 2:1, calculated for V. indicus , is lower than the 3:1 ratio recorded by Wikramanayake and Dryden (1988) based on 98 individual V. indicus from an introduced population on Guam. The sex ratio of all captured V. mertensi was similar (1.7:1). Most studies of varanid populations have recorded a male bias
(Auffenberg 1981, Shine 1986), even from the majority of preserved collections
7. Population dynamics 156
(Pianka 1969, 1971, King and Green 1979, Auffenberg 1981). A bias towards males in the majority of these studies is likely a consequence of males being more active
(more visible) and, hence, captured more often (King and Rhodes 1982). For example, a 22 month study of V. olivaceus recorded a 1:1 ratio, even though the males were more active and moved further (Auffenberg 1988). This was attributed to the use of hunting dogs to find animals, rather than relying on more passive
(trapping) or “incidental” (active searching) methods. A further study of V. acanthurus museum specimens that were collected whilst in refugia (i.e. not active) also recorded a 1:1 ratio (King and Rhodes 1982). Thus, it is likely that the sex ratios of free ranging populations of varanids are equal in most populations.
Conclusion
This chapter has demonstrated quantitatively that the short term survival probability of long lived varanids is high and that radiotracking is a much more robust, albeit sometimes less reliable method of survival estimation than mark recapture. Longer term studies, probably in the order of decades, would be required to definitively quantify differences in survival probability between male and female varanids of different sizes. The sex ratios presented here reflect the easier detectability of males
(as found in Chapter 6) and, therefore, as with many other field studies are probably not representative of the sex ratios in free ranging populations.
The densities of both V. mertensi and V. indicus surpass those of similar sized varanids in dissimilar habitats, leading again to the conclusion that high resource productivity plays an important in role in buffering semi aquatic varanids (from water and presumably prey abundance differences) between seasons.
157
Varanus mertensi and Bufo marinus
Synopsis
Varanus indicus
8. Synopsis 158
8. Synopsis Throughout this thesis I have examined variation in the physiological, spatial, seasonal and population ecology of V. mertensi and V. indicus . The two major questions addressed herein have both been answered to the affirmative; i) yes, V. indicus demonstrates year round activity like V. mertensi ; ii) yes, semi aquatic varanids are different from terrestrial varanids, physiologically, morphologically, in their use of space (both in relation to movements and in the densities they attain), and importantly, in how they respond to seasonality. Of the two semi aquatic varanids in this study, Varanus mertensi is the least responsive to seasonality. This, the final chapter will recapitulate the main findings of each chapter and highlight some avenues for future research focus.
Australian reptiles and the threat of cane toads
In chapter 2 I demonstrated that the cane toad is a serious threat to much of
Australia’s reptile fauna. This work and other recent studies (Phillips et al. 2003,
Oakwood 2004) support and reiterate the warnings about the possible effects of cane toads on large predators that have been voiced since their introduction in 1935
(Breeden 1963, Pockley 1965, Rayward 1974, Covacevich, 1975). Earlier works
(Freeland and Kerin 1990, Catling et al. 1999, Williamson 1999) suggesting that toad impacts are likely to be minimal (on all species) because of studies on small, abundant fauna, are now called into question.
Populations of predatory species will probably be affected to different degrees, depending on their prey choice and handling abilities, given that for many of these species, one predatory encounter is enough to kill them and many other species will
8. Synopsis 159 undoubtedly be affected, either directly or indirectly, from flow on effects down the food chain. Consequentially, the faunal composition of many ecosystems may be quite different to present day and this transition period, such as we are experiencing at present in northern Australia, provides many avenues for further research. In particular additional research opportunities exist for investigating whether V. mertensi , V. indicus and other varanid populations can mount an effective, long term, adaptive response to cane toads.
The problem of sexing varanids in the field
My analysis demonstrates that various head variables, principally head volume, and to a lesser extent, scaling of limb proportions can be important diagnostic features for gender prediction in some species of varanids. However definitively determining the sex of many varanid species in the field is likely to remain fraught with uncertainty.
Given that the dataset used in this thesis all but exhausts the supply of preserved specimens in most Australian museums, it is unlikely that this analytical method using morphometrics could be developed any further. In the absence of other, more definitive tools, such as DNA probing, further confidence in reliably field sexing varanids could be gained by developing a variety of diagnostic measures (both qualitative and quantitative). These techniques need to be species specific and the particulars of these for many species still need to be determined.
My analysis found that none of the models developed by Thompson and Withers
(1997) using null hypothesis testing were well supported for any of the species I analysed. This chapter highlights the limitations of the null hypothesis approach and in particular how exploratory analysis can sometimes produce spurious results
8. Synopsis 160
(Anderson and Burnham 2002). The use of more objective techniques such as the
Information Theoretic paradigm (Burnham and Anderson 2002) and cross validation procedures as applied here, allowed more powerful inferences to be made and more robust conclusions to be reached.
The physiology of Varanus indicus
Varanus indicus shows intermediate physiological responses to seasonal changes, between the semi aquatic V. mertensi that remains active year round (Christian et al.
1996), and the terrestrial V. gouldii and V. panoptes that spend at least some of their time inactive during the driest part of the year (Christian et al. 1995). Thus, V. indicus , like V. mertensi, is not subject to the pressures of dry season food and water depletion to the extent that terrestrial ones are. Although V. indicus remain active year round, the decrease in their dry season activity (and FMR) could be caused by reduced food resources at that time of year. An investigation into the seasonal productivity of the mangroves along Adelaide River (particularly with regard to V. indicus ’ prey items) would help to clarify this.
Intra and interspecific differences between terrestrial and semi aquatic varanids
Semi aquatic varanids share many life history characteristics with terrestrial varanids; males travel further in the breeding season, larger animals are more likely to move further, and there is large overlap between the home ranges of conspecifics.
In many other ways however, semi aquatic varanids (at least V. mertensi and V. indicus ) are different from terrestrial ones, the most important being that they can be active throughout periods of the year when terrestrial ones cannot. Ostensibly (and based on some indirect evidence, Christian et al. 1996d, Martin 2005, Mayes et al.
8. Synopsis 161
2005b, and this thesis), mangroves and freshwater bodies contain food items and water for V. mertensi and V. indicus year round, thereby enabling continuous activity.
Varanus mertensi has a catholic diet (Shine 1986, Losos and Greene 1988, Mayes et al. 2005b), and this broad diet enables it to cope with any seasonal and spatial differences in prey availability (Mayes et al. 2005b); however V. mertensi is unique in that its habitat is relatively linear. Thus, although there appears to be food present continuously, they must travel to forage. For many individuals this requirement leads to a large, linear home range. It is unclear if the highly variable home range sizes of
V. mertensi as found in this study are a consequence of resource variation around
Manton Dam.
In contrast to V. mertensi , V. indicus do not need to forage over large areas, probably because (in a broad sense) food is brought to them via the incoming tides, leading to very small home ranges, albeit with a strong and unreported vertical component up trees and very high densities. It is probable that northern Australian estuarine and mangrove macrofaunas also vary seasonally (Robertson and Duke 1990, Rooker and
Dennis 1991, Emmerson 1994, Lefebvre et al. 1994), but the broad diet of V. indicus
(Losos and Greene 1988) also allows them to stay in the same small area continuously.
8. Synopsis 162
Measuring varanid population dynamics
This study demonstrated that V. mertensi and V. indicus are long lived species as was indicated by longevity in captive specimens (Slavens and Slavens 2003). My research has also evaluated the efficacy of different methods for longer term monitoring. Although the pipe trap method can capture many V. indicus quickly, it may not be as robust a technique as radio tracking for the long term monitoring of varanid populations, although removing all traps between trapping periods may help alleviate trap shyness. The high densities recorded herein for both species contrast sharply to other studies of similar sized monitors (Auffenberg 1988, 1989, Pianka et al. 2004), again suggesting high resource availability.
Conclusions
The thesis illuminates many aspects of Australian semi aquatic varanid ecology, and demonstrates the broad ecological similarities between semi aquatic and terrestrial varanids and the reduced effect seasonality has on both semi aquatic species. Their relatively higher densities and small home ranges are more than likely a result of this lessened seasonal effect.
Many conclusions about the ecology of V. mertensi and V. indicus are predicated on the assumption that season is a useful surrogate for food availability. Although there is strong physiological and some other empirical evidence to suggest this is so, direct investigation into the seasonal availability of these varanids’ prey in these, and other northern Australian estuarine and freshwater systems, would complete this picture.
8. Synopsis 163
Many new logistical and analytical techniques have been developed or applied for the first time in this thesis. I have developed a new method for capturing elusive varanids in difficult habitats and have applied the null model approach of Christian et al. (2006) to modelling the thermoregulation of V. indicus. I have used the
Information Theoretic approach to predict varanid gender and model space use in varanids, and incorporated Cormack Jolly Seber and known fate modelling with model ranking procedures to estimate the survival of varanids. All of these methods and findings add substantially to our current knowledge of Australian varanids and provide useful techniques for future studies.
164
9. References 165
9. References Altman, J. C. 1987. Hunter gatherers today: an Aboriginal economy in north
Australia. Australian Institute of Aboriginal Studies, Canberra.
Anderson, D. R., and K. P. Burnham. 2002. Avoiding pitfalls when using
Information Theoretic methods. Journal of Wildlife Management 63 :912 928.
Anderson, R. A., and W. H. Karasov. 1981. Contrasts in energy intake and
expenditure in sit and wait and widely foraging lizards. Oecologia 49:67 72.
Andrews, R. M. 1991. Population stability of a tropical lizard. Ecology 72 :1204
1217.
Andrews, R. M., T. Mathies, and D. A. Warner. 2000. Effect of incubation
temperature on morphology, growth and survival of juvenile Sceloporus
undulatus . Herpetological Monographs 14 :420 431.
Ast, J. C. 2001. Mitochondrial DNA evidence and evolution in Varanoidea
(Squamata). Cladistics 17 :211 226.
Auffenberg, W. 1981. The behavioural ecology of the Komodo Monitor. University
of Florida Press.
Auffenberg, W. 1988. Gray's monitor lizard. University of Florida Press, Gainsville.
Auffenberg, W. 1989. Utilization of monitor lizards in Pakistan. Traffic Bulletin,
WWF, 11 :8 12.
Auffenberg, W. 1994. The Bengal monitor. University of Florida Press, Gainsville.
Auffenberg, W., Q. Nazar, and N. Khurshid. 1991. Preferred habitat, home range and
movement patterns of Varanus bengalensis in southern Pakistan. Mertensiella
2:7 28.
9. References 166
Auliya, M. A., and W. Erdelen. 1999. A field study of the water monitor lizard
(Varanus salvator ) in west Kalimantan, Indonesia. New Methods and Old
Problems. Mertensiella 11 :247 266.
Bakken, G. S. 1981. How many equivalent black body temperatures are there?
Journal of Thermal Biology 6:59 60.
Bakken, G. S. 1992. Measurement and application of operative and standard
operative temperatures in ecology. American Zoologist 32 :194 216.
Bakken, G. S., W. R. Santee, and D. J. Erskine. 1985. Operative and standard
operative temperatures: tools for thermal energetics studies. American
Zoologist 25 :933 943.
Ballinger, R. E. 1983. Life history variations. Pages 241 260, 465 473 in R. B. Huey,
E. R. Pianka, and T. W. Schoener, editors. Lizard Ecology, Studies of a
model organism. Harvard University Press, Cambridge.
Bauwens, D. 1981. Survivorship during hibernation in the European common lizard,
Lacerta vivipara . Copeia 1981 :741 744.
Bedford, S., and K. A. Christian. 1996. Tail morphology related to habitat of
varanids lizards and some other reptiles. Amphibia Reptilia 17 :131 140.
Benabib, M., and J. D. Congdon. 1992. Metabolic and water flux rates of free
ranging tropical lizards Sceloporus variabilis . Physiological Zoology 65 :788
802.
Bennett, A. F., and G. C. Gorman. 1979. Population density and energetics of lizards
on a tropical island. Oecologia 42 :1432 1939.
Bennett, D. 1998. Monitor lizards: Natural history, biology and husbandry. Frankfurt
am Main, Frankfurt.
9. References 167
Bennett, D. 2000. The density and abundance of juvenile Varanus exanthematicus
(Sauria: Varanidae) in the coastal plain of Ghana. Amphibia Reptilia 21 :301
306.
Berry, O., and D. M. Gleeson. 2005. Distinguishing historical fragmentation from a
recent population decline Shrinking or pre shrunk skink from New Zealand?
Biological Conservation 123 :197 210.
Blamires, S. J., and M. Nobbs. 2000. Observations of mangrove habitation by the
monitor lizard Varanus panoptes . Northern Territory Naturalist 16 :21 23.
Blouin Demers, G., and P. J. Weatherhead. 2001. Thermal ecology of black rat
snakes ( Elaphe obsoleta ) in a thermally challenging environment. Ecology
82 :3025 3043.
Blum, A., A. Kalai, and J. Langford. 1999. Beating the hold out: Bounds for K fold
and progressive cross validation. Pages 203 208 in Proceedings of the 10th
annual conference on computational theory.
Böhme, W. 1991. Hemiclitoris discovered: a fully differentiated erectile structure in
female monitor lizards ( Varanus spp .) (Reptilia: Varanidae). Journal of
Systematics and Evolution Research 33 :129 132.
Bomford, M., and J. Caughley, editors. 1996. Sustainable use of wildlife by
Aboriginal peoples and Torres Strait Islanders. Bureau of Resource Sciences,
Canberra.
Bowman, D. J. S. 2002. The Australian summer monsoon: a biogeographic
perspective. Australian Geographic Studies 40 :261 277.
Breeden, K. 1963. Cane toad ( Bufo marinus ). Wildlife in Australia 1:31.
9. References 168
Brocklehurst, P., and B. Edmeades. 1996. Mangrove survey of Darwin harbour,
Northern Territory. Northern Territory. Dept. of Lands, Planning and
Environment. Darwin.
Brodie, E. D. III , and E. D. Brodie, Jr. 1999. Predator prey arms races: asymmetrical
selection on predators and prey may be reduced when prey are dangerous.
BioScience 49 :557 568.
Brodie, E. D. III, and N. H. Russell. 1999. The consistency of individual differences
in behaviour: temperature effects on antipredator behaviour in garter snakes.
Animal Behaviour 57 :445 451.
Brodie, E. D., Jr., B. J. Ridenhour, and E. D. Brodie, III. 2002. The evolutionary
response of predators to dangerous prey: Hotspots and coldspots in the
geographic mosaic of coevolution between garter snakes and newts.
Evolution 56 :2067 2082.
Brown, G. P., R. Shine, and T. Madsen. 2002. Responses of three sympatric snake
species to tropical seasonality in northern Australia. Journal of Tropical
Ecology 18 :549 568.
Bull, C. M. 1995. Population ecology of the sleepy lizard, Tiliqua rugosa , at Mt
Mary, South Australia. Australian Journal of Ecology 20 :393 402.
Burnett, S. 1997. Colonizing cane toads cause population declines in native
predators: reliable anecdotal information and management implications.
Pacific Conservation Biology 3:65 72.
Burnham, K. P., and D. R. Anderson. 2001. Kullback Leibler information as a basis
for strong inference in ecological studies. Wildlife Research 28 :111 119.
9. References 169
Burnham, K. P., and D. R. Anderson. 2002. Model selection and multimodel
inference: A practical Information Theoretic approach, 2nd edition. Springer,
New York.
White, G. C. and K. P. Burnham. 1999. Program MARK: survival estimation from
populations of marked animals. Bird Study 46 : 120 139.
Calder, W. A. 1984. Size, function and life history. Harvard University Press,
Cambridge.
Cameron, E. E., and H. G. Cogger. 1992. The herpetofauna of the Weipa region,
Cape York Peninsula. Technical reports of the Australian Museum 7:1 200.
Cann, J. 1998. Australian freshwater turtles. Beaumont Publishing, Singapore.
Carter, D. B. 1990. Courtship and mating in wild Varanus varius (Varanidae:
Australia). Memoirs of the Queensland Museum 29 :333 338.
Catling, P. C., A. Hertog, R. J. Burt, J. C. Wombey, and R. I. Forrester. 1999. The
short term effect of cane toads ( Bufo marinus ) on native fauna in the Gulf
Country of the Northern Territory. Wildlife Research 26 :161 185.
Caughley, G. 1980. Analysis of vertebrate populations. Wiley, New York.
Caughley, G., and A. Gunn. 1996. Conservation biology in theory and practice.
Blackwell Science, Cambridge.
Chen, K. K., and A. Kovarikova. 1967. Pharmacology and toxicology of toad venom.
Journal of Pharmaceutical Sciences 56 :1535 1541.
Choquet, R., A. M. Reboulet, R. Pradel, O. Giminez, and J. D. Lebreton. 2003.
User's manual for U Care. Mimeographed document, CEFE/CNRS,
Montpellier.
9. References 170
Christian, A., G. S. Bedford, B. Green, A. D. Griffiths, K. Newgrain, and T. J.
Schultz. 1999a. Physiological ecology of a tropical dragon, Lophognathus
temporalis . Australian Journal of Ecology 24 :171 181.
Christian, A., I. E. Clavijo, N. Cordero Lopez, E. E. Ellias Maldonado, M. A.
Franco, M. V. Lugo Ramirez, and M. Marengo. 1986. Thermoregulation and
energetics of a population of Cuban iguanas ( Cyclura nubila ) on Isla
Magueyes, Puerto Rico. Copeia 1986 :65 69.
Christian, K. A. 2004. Varanus mertensi . Pages 410 415 in E. R. Pianka, D. R.
King, and R. A. King, editors. Varanoid lizards of the world. Indiana
University Press, Bloomington Indianapolis.
Christian, K. A., and G. S. Bedford. 1995. Seasonal changes in thermoregulation by
the frillnecked lizard, Chlamydosaurus kingii , in tropical Australia. Ecology
76 :124 132.
Christian, K. A., and G. S. Bedford. 1996. Thermoregulation by the spotted tree
monitor, Varanus scalaris , in the seasonal tropics of Australia. Journal of
Thermal Biology 21 :67 73.
Christian, K. A., G. S. Bedford, and T. J. Schultz. 1999b. Energetic consequences of
metabolic depression in tropical and temperate zone lizards. Australian
Journal of Zoology 47 :133 141.
Christian, K. A., G. S. Bedford, and S. T. Shannahan. 1996a. Solar absorbance of
some Australian lizards and its relationship to temperature. Australian Journal
of Zoology 44 :59 67.
Christian, K. A., and K. E. Conley. 1994. Activity and resting metabolism of varanid
lizards compared with 'typical' lizards. Australian Journal of Zoology 42 :185
193.
9. References 171
Christian, K. A., L. K. Corbett, B. Green, and B. W. Weavers. 1995. Seasonal
activity and energetics of two species of varanid lizards in tropical Australia.
Oecologia 103 :349 357.
Christian, K. A., and B. Green. 1994. Seasonal energetics and water turnover of the
frillneck lizard Chlamydosaurus kingii , in the wet dry tropics of Australia.
Herpetologica 50 :274 281.
Christian, K. A., B. Green, G. S. Bedford, and K. Newgrain. 1996b. Seasonal
metabolism of a small, arboreal monitor lizard, Varanus scalaris , in tropical
Australia. Journal of Zoology 240 :383 396.
Christian, K. A., A. D. Griffiths, and G. S. Bedford. 1996c. Physiological ecology of
frillneck lizards in a seasonal tropical environment. Oecologia 106 :49 56.
Christian, K. A., and C. R. Tracy. 1981. The effect of the thermal environment on the
ability of hatchling Galapagos land iguanas to avoid predation during
dispersal. Oecologia 49 :218 223.
Christian, K. A., C. R. Tracy, and W. P. Porter. 1983. Seasonal shifts in body
temperature and use of microhabitats by Galapagos land iguanas (C onolophus
pallidus ). Ecology 64 :463 468.
Christian, K. A., R. Tracy, and C. R. Tracy. 2006. Evaluating thermoregulation in
reptiles: an appropriate null model. The American Naturalist 198 :421 430.
Christian, K. A., and S. Waldschmidt. 1984. The relationship between lizard home
range and body size: A reanalysis of the data. Herpetologica 40 :68.
Christian, K. A., and B. W. Weavers. 1996. Thermoregulation of monitor lizards in
Australia: An evaluation of methods in thermal biology. Ecological
Monographs 66 :139 157.
9. References 172
Christian, K. A., B. W. Weavers, B. Green, and G. S. Bedford. 1996d. Energetics and
water flux in a semiaquatic lizard, Varanus mertensi . Copeia 1996 :354 362.
Christian, K. A., J. K. Webb, and T. J. Schultz. 2003. Energetics of bluetongue
lizards ( Tiliqua scincoides ) in a seasonal tropical environment. Oecologia
136 :515 523.
Civantos, E., and A. Forsman. 2000. Determinants of survival in juvenile
Psammodromus algirus lizards. Oecologia 124 :64 72.
Clough, B. F., and P. M. Attiwill. 1982. Primary productivity of mangroves. Pages
213 222 in B. F. Clough, editor. Mangrove ecosystems in Australia —
structure, function and management. Australian Institute of Marine Science,
Townsville.
Cogger, H. G. 2000. Reptiles and Amphibians of Australia. Reed New Holland,
Sydney, Australia.
Cogger, H. G., E. E. Cameron, R. A. Sadlier, and P. Eggler. 1993. The action plan
for Australian reptiles. Endangered Species Unit, Australian Nature
Conservancy.
Cohen, M. P., and R. A. Alford. 1993. Growth, survival and activity patterns of
recently metamorphosed Bufo marinus . Wildlife Research 20 :1 13.
Commonwealth of Australia (1999). Environmental Protection and Biodiversity
Conservation Act 1999. Commonwealth of Australia Act 91. Canberra.
Congdon, J. D., and D. W. Tinkle. 1982. Energy expenditure in free ranging
sagebrush lizards ( Sceloporus graciosus ). Canadian Journal of Zoology
60 :1412 1416.
9. References 173
Conroy, S. D. C., and B. W. Brook. 2003. Demographic sensitivity and persistence
of the threatened white and orange bellied frogs of Western Australia.
Population Ecology 45 :105 114.
Cooch, E., and G. White. 2006. Program MARK: A gentle introduction, 4th edition.
Colorado State University.
Cooper, W. E. 1997. Independent evolution of squamate olfaction and
vomerolfaction and correlated evolution of vomerolfaction and lingual
structure. Amphibia Reptilia 18 :85 105.
Corbett, L. 1994. The dingo in Australia and Asia. University of New South Wales
Press, Sydney.
Corbett, L. K., 1988. Fluctuations in the yearly numbers of magpie goose ( Anseranas
semipalmata ) nests at Kapalga in Kakadu National Park with emphasis on the
effects of feral buffalo and rainfall: Final Report to the Australian National
Parks and Wildlife Service from the CSIRO Division of Wildlife and
Ecology. 47 p. Winnellie, N.T. CSIRO Tropical Ecosystems Research Centre.
Reprint no. 1683.
Covacevich, J., and M. Archer. 1975. The distribution of the cane toad, Bufo
marinus , in Australia and its effects on indigenous vertebrates. Memoirs of
the Queensland Museum 17 :305 310.
Cowie, I. D., P. S. Short, and M. Osterkamp Madsen. 2000. Floodplain flora; a flora
of the coastal floodplains of the Northern Territory, Australia. Parks and
Wildlife Commission of the Northern Territory, Darwin.
Crawley, M. J. 2003. Statistical Computing:An introduction to R using S Plus. John
Wiley & Sons, West Sussex.
9. References 174
Crossland, M. R. 1998. Ontogenetic variation in toxicity of tadpoles of the
introduced toad Bufo marinus to native Australian aquatic invertebrate
predators. Herpetologica 54 :364 369.
Crossland, M. R. 2000. Direct and indirect effects of the introduced toad Bufo
marinus (Anura: Bufonidae) on populations of native anuran larvae in
Australia. Ecography 23 :283 290.
Crossland, M. R. 2001. Ability of predatory native Australian fishes to learn to avoid
toxic larvae of the introduced toad Bufo marinus . Journal of Fish Biology
59 :319 329.
Crossland, M. R., and R. A. Alford. 1998. Evaluation of the toxicity of eggs,
hatchlings and tadpoles of the introduced toad Bufo marinus (Anura,
Bufonidae) to native Australian aquatic predators. Australian Journal of
Ecology 23 :129 137.
Daley, J. W., C. W. Myers, and N. Whittaker. 1987. Further classification of skin
alkaloids from neotropical poison frogs (Dendrobatidae) with a general
survey of toxic/noxious substances in the amphibia. Toxicon 25 :1023 1095.
Daly, J. W., and B. Witkop. 1971. Chemistry and pharmacology of frog venoms.
Pages 497 519 in W. Bucherl and E. E. Buckley, editors. Venomous animals
and their venoms. Academic Press, New York.
Daudin, F. M. 1802. The Indian Tupinambis. Histoire Naturelle, Generale et
Particular des Reptiles. Volume III .
Davis, R. B., and L. G. Phillips. 1991. A method of sexing the Dumeril's monitor
Varanus dumerilii . Herpetological Review 22 :18 19.
9. References 175 de Buffrénil, V., and G. Hémery. 2002. Variation in longevity, growth, and
morphology in exploited Nile monitors ( Varanus niloticus ) from Sahelian
Africa. Journal of Herpetology 36 :419 426. de Buffrénil, V., and F. Rimblot Baly. 1999. Female reproductive output in exploited
Nile monitor lizard ( Varanus niloticus L.) populations in Sahelian Africa.
Canadian Journal of Zoology 77 :1530 1539.
Diamond, J. M. 1989. Overview of recent extinctions. Pages 37 41 in D. Western
and M. C. Pearl, editors. Conservation for the 21st century. Oxford
University Press, Oxford.
Diaz, J. A. 1993. Breeding coloration, mating opportunities, activity, and survival in
the lacertid lizard Psammodromus algirus . Canadian Journal of Zoology
71 :1104 1110.
Doody, J. S., B. Green, R. Sims, D. Rhind, P. West, and D. Steer. 2006. Indirect
impacts of invasive cane toads ( Bufo marinus ) on nest predation in pig nosed
turtles (C arettochelys insculpta ). Wildlife Research 33 :349 354.
Downey, F. J., and C. R. Dickman. 1993. Macro and microhabitat relationships
among lizards of sandridge deserts in central Australia. Pages 133 138 in D.
Lunney and D. Ayers, editors. Herpetology in Australia: A diverse discipline.
Royal Zoological Society of N.S.W., Mosman.
Dryden, G. L. 1965. The food and feeding habits of Varanus indicus on Guam.
Micronesica 2:73 76.
Dryden, G. L., B. Green, D. King, and J. Losos. 1990. Water and energy turnover in
a small monitor lizard, Varanus acanthurus . Australian Wildlife Research
17 :641 646.
9. References 176
Dryden, G. L., B. Green, E. D. Wikramanayake, and K. G. Dryden. 1992. Energy
and water turnover in two tropical varanid lizards, Varanus bengalensis and
V. salvator . Copeia 1992 :102 107.
Dunham, A. E., and D. B. Miles. 1985. Patterns of covariation in life history traits of
Squamate reptiles: The effects of size and phylogeny reconsidered. American
Naturalist 126 :231 257.
Dunham, A. E., D. B. Miles, and D. N. Reznick. 1988. Life history patterns of
squamate reptiles. Pages 441 522 in C. Gans and R. B. Huey, editors.
Biology of the Reptilia. John Wiley and Sons, New York.
Dunson, W. A. 1974. Salt gland secretion in a mangrove monitor lizard. Comparative
Biochemistry and Physiology 47 :1245 1255.
Dzialowski, E. M., and M. P. O'Connor. 2001. Thermal time constant estimation in
warming and cooling ectotherms. Journal of Thermal Biology 26 :231 245.
Elmouden, E., F. H., J. Castanet, and M. Znari. 1997. A skeletochronological study
of age, maturity, growth and longevity in the north African agamid, Agama
impalearis Boettger, 1874. Annales des Sciences Naturelles Zoologie et
Biologie Animale 18 :63 70.
Emmerson, W. D. 1994. Seasonal breeding cycles and sex ratios of eight species of
crabs from Mgazana, a mangrove estuary in Transkei, southern Africa.
Journal of Crustacean biology 14 :568 578.
Erdelen, W. 1991. Conservation and population ecology of monitor lizards: the water
monitor Varanus salvator (Laurenti, 1768) in south Sumatra. Mertensiella
2:120 135.
Estes, R. 1984. Fish, amphibians and reptiles from the Etadunna Formation, Miocene
of South Australia. Australian Zoologist 21 :335 343.
9. References 177
Ferguson, G. W., G. H. Bohler, and H. P. Wooley. 1980. Sceloporus undulatus :
Comparative life history and regulation of a Kansas population. Ecology
61 :313 322.
Finlayson, C. M., B. J. Bailey, W. J. Freeland, and M. R. Fleming. 1988. Wetlands of
the Northern Territory. in A. J. McComb and P. S. Lake, editors. The
conservation of Australian wetlands. Surrey Beatty & Sons, Sydney.
Fitch, A. J., A. E. Goodman, and S. C. Donnellan. 2006. A molecular phylogeny of
the Australian monitor lizards (Squamata: Varanidae) inferred from
mitochondrial DNA sequences. Australian Journal of Zoology 54 :253 269.
Flier, J., M. W. Edwards, J. W. Daley, and C. W. Myers. 1980. Widespread
occurrence in frogs and toads of skin compounds interacting with the ouabain
site of Na+, K+ ATPase. Science 208 :503 505.
Flower, S. S. 1933. Notes on some recent reptiles and amphibians of Egypt, with a
list of the species recorded from that Kingdom. Proceedings of the Zoological
Society (London) 1933 :735 851.
Flower, S. S. 1937. The duration of life in animals, further notes. III Reptiles.
Proceedings of the Zoological Society (London) 1937 :31 74.
Foufopoulos, J., and A. R. Ives. 1999. Reptile extinctions on land bridge islands:
Life history attributes and vulnerability to extinction. The American
Naturalist 153 :1 25.
Fox, J. 2003. Effect displays in R for generalised linear models. Journal of Statistical
Software 8:1 18.
Fox, S. F., and J. K. McCoy. 2000. The effects of tail loss on survival, growth,
reproduction, and sex ratio of offspring in the lizard Uta stansburiana in the
field. Oecologia 122 :327 334.
9. References 178
Freeland, W. J. 1986. Populations of cane toad Bufo marinus in relation to time since
colonization. Australian Wildlife Research 13 :321 330.
Freeland, W. J., and S. H. Kerin. 1990. Within habitat relationships between
invading Bufo marinus and Australian species of frog during the tropical dry
season. Australian Wildlife Research 15 :293 305.
Freeland, W. J., and S. H. Kerin. 1991. Ontogenetic alteration of activity and habitat
by Bufo marinus . Wildlife Research 18 :431 443.
Galán, P. 2004. Structure of a population of the lizard Podarcis bocagei in northwest
Spain: Variations in age distribution, size distribution and sex ratio. Animal
Biology 54 :57 75.
Garland Jr, T. 2002. Lizard home ranges revisited: Effects of sex, body size, diet,
habitat, and phylogeny. Ecology 83 :1870 1885.
Gaulke, M. 1991. On the diet of the water monitor, Varanus salvator , in the
Philippines. Mertensiella 2:143 153.
Gaulke, M. 1992. Distribution, population density, and exploitation of the water
monitor ( Varanus salvator ) in the Philippines. Hamadryad 17:21 27.
Gaulke, M. 1997. Sex determination of monitor lizards in the field A review of
methods. Hamadryad 22 :28 31.
Gaulke, M., W. Erdelen, and F. Abel. 1999. A radio telemetric study of the water
monitor lizard ( Varanus salvator ) in north Sumatra, Indonesia. Mertensiella
11 :63 78.
Glauert, L. 1951. A new Varanus from East Kimberley, Varanus mertensi sp. n.
West Australian Naturalist 3:14 16.
9. References 179
Green, B., and D. King. 1978. Home range and activity patterns of the sand goanna,
Varanus gouldii (Reptilia: Varanidae). Australian Journal of Wildlife
Research 5:417 424.
Greenberg, C. H. 2001. Response of reptile and amphibian communities to canopy
gaps created by wind disturbance in the Southern Appalachians. Forest
Ecology and Management 148 :135 144.
Greer, A. E. 1989. The biology and evolution of Australian lizards. Surrey Beatty
and Sons, Chipping Norton NSW.
Griffiths, A. D. 1999. Demography and home range of the frillneck lizard,
Chlamydosaurus kingii (Agamidae), in northern Australia. Copeia
1999 :1089 1096.
Guarino, F. 2002. Spatial ecology of a large carnivorous lizard, Varanus varius
(Squamata: Varanidae). Journal of Zoology 258 :449 457.
Guarino, F., A. Georges, and B. Green. 2002. Variation in energy metabolism and
water flux of free ranging male lace monitors, Varanus varius (Squamata:
Varanidae). Physiological and Biochemical Zoology 75 :294 304.
Hamilton, L., and S. Snedaker. 1984. Handbook for mangrove area management,
Commission of ecology. International Union for Conservation of Nature,
Gland, Switzerland.
Hamley, T., and A. Georges. 1985. The Australian snapping tortoise Elseya
latisternum : a successful predator on the introduced cane toad? Australian
Zoologist 21 :607 610.
Hare, K. M., C. G. Longson, S. Pledger, and C. H. Daugherty. 2004. Size, growth,
and survival are reduced at cool incubation temperatures in the temperate
lizard Oligosoma suteri (Lacertilia: Scincidae). Copeia 2004 :383 390.
9. References 180
Harvey, M. B., and D. G. Barker. 1998. A new species of blue tailed monitor lizard
(genus Varanus ) from Halmahera Island, Indonesia. Herpetologica 54 :34 44.
Hecht, M. K. 1975. The morphology and relationships of the largest known
terrestrial lizards Megalania prisca from the Pleistocene of Australia.
Proceedings of the Royal Society of Victoria 87 :239 250.
Hertz, P. E., R. Huey, and R. D. Stevenson. 1993. Evaluating temperature regulation
by field active endotherms: the fallacy of the inappropriate question.
American Naturalist 142 :796 818.
Hooge, P. N., and B. Eichenlaub. 2001. Animal movement extension to Arcview.
ver. 2.04. Alaska Biological Science Center, USGS, Anchorage.
Huey, R., E. R. Pianka, and L. J. Vitt. 2001. How often do lizards "Run on empty"?
Ecology 82 :1 7.
Huey, R., and R. D. Stevenson. 1979. Integrating thermal physiology and ecology of
ectotherms: A discussion of approaches. American Zoologist 19 :357 366.
Hutchinson, M. F., R. Swain, and Driessen. 2001. Snakes and lizards of Tasmania.
Nature conservation branch, Department of Primary Industries, Water and
Environment, Hobart.
Hutchinson, M. N., and S. C. Donnellan. 1993. Biogeography and phylogeny of the
Squamata. Pages 210 220 in G. J. B. Ross, editor. Fauna of Australia, Vol.
2A: Amphibia, Reptilia, Aves. Canberra: Australian Biological and
Environmental Survey.
Ibrahim, A. A. 2002. Activity area, movement patterns, and habitat use of the desert
monitor, Varanus griseus , in the Zaranik Protected Area, North Sinai, Egypt.
African Journal of Herpetology 51 :35 45.
9. References 181
IUCN. 2001. 100 of the world's worst invasive species: a selection from the global
invasive species database.
Jackson, R. 2005. The poorly known rusty monitor, Varanus semiremex : history,
natural history, captive breeding and husbandry. Herpetofauna 35 :15 24.
James, C. D. 1996. Ecology of the pygmy goanna ( Varanus brevicauda ) in spinifex
grasslands of central Australia. Australian Journal of Zoology 44 :177 192.
James, C. D., J. B. Losos, and D. R. King. 1992. Reproductive biology and diets of
goannas (Reptilia: Varanidae) from Australia. Journal of Herpetology 26 :128
136.
Jenssen, T. A., and S. C. Nunez. 1998. Spatial and breeding relationships of the
lizard, Anolis carolinensis : evidence of intrasexual selection. Behaviour
135 :981 1003.
Johnson, D. H. 1999. The insignificance of significance testing. Journal of Wildlife
Management 63 :763 772.
Kathiresan, K., and B. L. Bingham. 2001. Biology of mangroves and mangrove
ecosystems. Advances in marine biology 40 :81 251.
Kay, R. W. 2004. Movements and home ranges of radio tracked Crocodylus porosus
in the Cambridge Gulf region of Western Australia. Wildlife Research
31: 495 508.
Kernohan, B. J., R. A. Gitzen, and J. J. Millspaugh. 2001. Analysis of animal space
use and movements. Pages 125 166 in J. J. Millspaugh and J. M. Marzluff,
editors. Radio tracking and animal populations. Academic Press, San Diego,
CA.
King, D. 1980. The thermal biology of free living sand goannas ( Varanus gouldii ) in
Southern Australia. Copeia 1980 :755 767.
9. References 182
King, D. 1991. The effect of body size on the ecology of varanid lizards.
Mertensiella 2:204 210.
King, D. 1993. Diet and reproductive condition of free ranging Varanus timorensis .
The Western Australian Naturalist 19 :189 194.
King, D., and B. Green. 1979. Notes on diet and reproduction of the sand goanna,
Varanus gouldii rosenbergi . Copeia 1979 :64 70.
King, D., and B. Green. 1999. Goannas: The biology of varanid lizards, 2nd Edition.
UNSW Press, Sydney.
King, D., B. Green, and H. Butler. 1989. The activity pattern, temperature regulation
and diet of Varanus giganteus on Barrow Island, Western Australia.
Australian Wildlife Research 16 :41 47.
King, D., M. King, and P. R. Baverstock. 1991. A new phylogeny of the Varanidae.
Mertensiella 2:211 219.
King, D., and L. Rhodes. 1982. Sex ratio and breeding season of Varanus
acanthurus . Copeia 1982 :784 787.
King, R. B., T. D. Bittner, A. Queral Regil, and J. H. Cline. 1999. Sexual
dimorphism in neonate and adult snakes. Journal of the Zoological Society of
London 247 :19 28.
Kratochvíl, L., M. Fokt, I. Rehák, and D. Frynta. 2003. Misinterpretation of character
scaling: a tale of sexual dimorphism in body shape of common lizards.
Canadian Journal of Zoology 81 :1112 1117.
Laurie, W. A., and D. Brown. 1990a. Population biology of marine Iguanas
(Amblyrhynchus cristatus ). II. Changes in annual survival rates and the
effects of size, sex, age and fecundity in a population crash. Journal of
Animal Ecology 59 :529 544.
9. References 183
Laurie, W. A., and D. Brown. 1990b. Population biology of marine Iguanas
(Amblyrhynchus cristatus ). III. Factors affecting survival. Journal of Animal
Ecology 59 :545 568.
Lawler, K. L., and J. M. Hero. 1997. Palatability of Bufo marinus tadpoles to a
predatory fish decreases with development. Wildlife Research 24 :327 334.
Le Galliard, J. F., R. Ferrière, and J. Clobert. 2005. Juvenile growth and survival
under dietary restriction: are males and females equal? Oikos 111 :368 376.
Lebreton, J. D., K. P. Burnham, J. Clobert, and D. E. Anderson. 1992. Modeling
survival and testing biological hypotheses using marked animals: A unified
approach with case studies. Ecological Monographs 62 :67 118.
Lefebvre, G., B. Poulin, and R. McNeil. 1994. Temporal dynamics of mangrove bird
communities in Venezuela with special reference to migrant warblers. Auk
111 :405 415.
Lever, C. 2001. The cane toad. The history and ecology of a successful colonist.
Westbury Academic and Scientific Publishing, Yorkshire.
Lifson, N., and R. McClintock. 1966. Theory of the use of turnover rates of body
water for measuring energy and material balance. Journal of Theoretical
Biology 12 :46 74.
Losos, J. B., and H. W. Greene. 1988. Ecological and evolutionary implications of
diet in monitor lizards. Biological Journal of the Linnean Society 35 :379 487.
Lovern, M. B. 2000. Social and Hormonal Effects on the Ontogeny of Sex
Differences in Behavior in the Lizard, Anolis carolinensis . PhD Thesis.
Virginia State University, Blacksburg, Virginia.
Lugo, A. E., and S. Snedaker. 1974. The ecology of mangroves. Annual Review of
Ecology and Systematics 5:39 64.
9. References 184
Lutz, B. 1971. Venomous toads and frogs. Pages 423 473 in W. Bucherl and E. E.
Buckley, editors. Venomous animals and their venoms. Academic Press, New
York.
Macintosh, D. J., and E. C. Asthton. 2002. A review of mangrove biodiversity
conservation and management. Centre for Tropical Ecosystems Research,
University of Aarhus, Denmark.
Mack, R. N., D. Simberloff, W. M. Lonsdale, H. Evans, M. Clout, and F. Bazzaz.
2000. Biotic invasions: Causes, epidemiology, global consequences and
control. Issues in Ecology 5:1 20.
Madsen, T., and R. Shine. 1998. Spatial subdivision within a population of tropical
pythons ( Liasis fuscus ) in a superficially homogeneous habitat. Australian
Journal of Ecology 23 :340 348.
Madsen, T., and R. Shine. 2000. Silver spoons and snake sizes: prey availability
early in life influences long term growth rates of free ranging pythons.
Journal of Animal Ecology 69 :952 958.
Marchant, S., and P. J. Higgins, editors. 1990. Handbook of Australian, New Zealand
and Antarctic Birds. Oxford University Press, Melbourne.
Martin, J. 2005. The distribution, abundance and trophic ecology of the fishes of
Darwin Harbour mangrove habitats. PhD thesis, Darwin.
Matlock, R. B. J., J. B. Welch, and F. D. Parker. 1996. Estimating population density
per unit area from mark, release, recapture data. Ecological Applications
6:1241 1253.
Mayes, P. J., S. D. Bradshaw, and F. J. Bradshaw. 2005a. Successfully determining
the sex of adult Varanus mertensi (Reptilia: Varanidae) using a combination
9. References 185
of both hemipenile eversion and the ratio of androgens estradiol in plasma.
Annals New York Academy of Science 1040 :402 405.
Mayes, P. J., G. G. Thompson, and P. C. Withers. 2005b. Diet and foraging
behaviour of the semi aquatic Varanus mertensi (Reptilia: Varanidae).
Wildlife Research 32 :67 74.
McCoid, M., J., and R. A. Hensley. 1991. Mating and combat in Varanus indicus .
Herpetological Review 22 :16 17.
McCoid, M., J., and G. J. Witteman. 1993. Varanus indicus (Mangrove Monitor).
Diet. Herpetological Review 24 :105.
McDonald, N. S., and J. McAlpine. 1991. Floods and droughts: The northern
climate. in C. D. Haynes, M. G. Ridpath, and M. A. J. Williams, editors.
Monsoonal Australia, landscape, ecology and man in the northern lowlands.
AA Balkema, Rotterdam, The Netherlands.
Meehan, B. 1982. Shell bed to shell midden. Australian Institute of Aboriginal
Studies, Canberra.
Melquist, W. E., and M. G. Hornocker. 1983. Ecology of river otters in west central
Idaho. Wildlife Monographs 83 :1 60.
Meyer, K., and H. Linde. 1971. Collection of toad venoms and chemistry of toad
venom steroids. Pages 521 556 in W. Bucherl and E. E. Buckley, editors.
Venomous animals and their venoms. Academic Press, New York.
Murphy, J. B., C. Ciofi, C. de La Panouse, and T. Walsh, editors. 2002. Komodo
dragons: Biology and conservation. Smithsonian Institution Press.
Nagy, K. A. 1982. Field studies of water relations. Pages 483 501 in C. Gans and F.
H. Pough, editors. Biology of the Reptilia. Academic Press, New York.
9. References 186
Nagy, K. A., I. A. Girard, and T. K. Brown. 1999. Energetics of free ranging
mammals, reptiles and birds. Annual Review of Nutrition 19 :247 277.
Nagy, K. A., and C. C. Peterson. 1988. Scaling of water flux rate in animals.
University of California Publications in Zoology 120 :1 172.
Nagy, K. A., and V. H. Shoemaker. 1984. Field energetics and food consumption of
the Galapagos marine iguana, Amblyrhynchus cristatus . Physiological
Zoology 57 :281 290.
Nester, M. 1996. An applied statistician's creed. Applied Statistics 45 :401 410.
Noske, R. A. 1996. Abundance, zonation and foraging ecology of birds in mangroves
of Darwin Harbour, Northern Territory. Wildlife Research 23 :443 474.
Noske, R.A. 2003. The role of birds in mangroves pollination and insect predation.
in: Proceedings: Darwin Harbour region: Current knowledge and future
needs. Ed. Working Group for the Darwin Harbour Advisory Committee,
Department of Infrastructure, Planning and Environment, Darwin. pp 74 86.
NRETA. 2006. Classification status of animals of the Northern Territory. Schedule I;
Reptiles. Northern Territory Government, Department of Natural Resources,
Environment and the Arts.
Oakwood, M. 2004. The effect of cane toads on a marsupial carnivore, the northern
quoll, Dasyurus hallucatus . Report to Department of Environment and
Heritage, Darwin.
Olsson, M., and T. Madsen. 2001. Between year variation in determinants of
offspring survival in the sand lizard, Lacerta agilis . Functional Ecology
15 :443 450.
9. References 187
Olsson, M., R. Shine, E. Wapstra, B. Ujvari, and T. Madsen. 2002. Sexual
dimorphism in lizard body shape: the roles of sexual selection and fecundity
selection. Evolution 56 :1538 1542.
Pace, M. L., J. J. Cole, S. R. Carpenter, and J. F. Kitchell. 1999. Trophic cascades
revealed in diverse ecosystems. Trends in Ecology and Evolution 14 :483 488.
Packard, G. C., and T. J. Boardman. 1999. The use of percentages and size specific
indices to normalize physiological data for variation in body size: wasted
time, wasted effort? Comparative Biochemistry and Physiology A 122 :37 44.
Pavey, C. R., N. Goodship, and F. Geiser. 2003. Home range and spatial organisation
of rock dwelling carnivorous marsupial, Pseudantechinus macdonnellensis .
Wildlife Research 30 :135 142.
Perry, G., and T. Garland. 2002. Lizard home ranges revisited: effects of sex, body
size, diet, habitat, and phylogeny. Ecology 83 :1870 1885.
Peters, U. 1973. A contribution on the ecology of Varanus storri . Koolewong 2:12
13.
Peterson, C. C., K. A. Nagy, and J. Diamond. 1990. Sustained metabolic scope.
Proceedings of the National Academy of Science. USA 87 :2324 2328.
Philipps, J. A. 1995. Movement patterns and density of Varanus albigularis . Journal
of Herpetology 29 :407 416.
Phillips, B. L., G. P. Brown, and R. Shine. 2003. Asessing the potential impact of
cane toads on Australian snakes. Conservation Biology 17 :1738 1747.
Phillips, B. L., G. P. Brown, and R. Shine. 2004. Assessing the potential for an
evolutionary response to rapid environmental change: invasive toads and an
Australian snake. Evolutionary Ecology Research 6:799 811.
9. References 188
Phillips, B. L., and R. Shine. 2006. Allometry and selection in a novel predator/prey
system: Australian snakes and the invading cane toad. Oikos 112 :122 130.
Pianka, E. R. 1969. Habit specificity, speciation and species density in Australian
desert lizards. Ecology 50 :498 502.
Pianka, E. R. 1971. Notes on the biology of Varanus tristis . West Australian
Naturalist 11 :180 183.
Pianka, E. R. 1995. Evolution of body size: Varanids as a model system. American
Naturalist 146 :398 414.
Pianka, E. R., D. King, and R. A. King, editors. 2004. Varanoid lizards of the world.
Indiana University Press, Bloomington Indianapolis.
Pianka, E. R., and L. J. Vitt. 2003. Lizards: windows to the evolution of diversity.
University of California Press, London.
Pockley, D. 1965. The free and the caged. Blackwoods Magazine 298 :439 446.
Porter, W. P., and D. M. Gates. 1969. Thermodynamic equilibria of animals with
their environment. Ecological Monographs 39 :245 270.
Porter, W. P., and F. C. James. 1979. Behavioural implications of mechanistic
ecology II. The African rainbow lizard Agama agama . Copeia 1979 :594 619.
Porter, W. P., J. W. Mitchell, W. A. Beckman, and C. B. DeWitt. 1973. Behavioural
implications of mechanistic ecology. Oecologia 13 :1 54.
Porter, W. P., and C. R. Tracy. 1982. Biophysical analysis of energetics, time space
utilization, and distributional limits of lizards. in R. Huey, E. R. Pianka, and
T. Schoener, editors. Ecology of lizards. Harvard University Press,
Cambridge.
Pough, F. H. 1973. Lizard energetics and diet. Ecology 54 :837 844.
9. References 189
R Development Core Team. 2004. R: A language and environment for statistical
computing. in . R Foundation for statistical computing, Vienna, Austria.
Rayward, A. 1974. Giant toads a threat to Australian wildlife. Wildlife 17 :506 507.
Ribble, D. O., A. E. Wurtz, E. K. McConnell, J. J. Buegge, and K. C. Welch Jnr.
2002. A comparison of home ranges of two species of Peromyscus using
trapping and radiotelemetry data. Journal of Mammalogy 83 :260 266.
Richards, M. P., and H. O. Hultin. 2001. Rancidity development in a fish model
system as affected by phospholipids. Journal of Food Lipids 8:149 238.
Robertson, A. I., and P. A. Daniel. 1989. The influence of crabs on litter processing
in high intertidal mangrove forests in tropical Australia. Oecologia 78 :191
198.
Robertson, A. L., and N. C. Duke. 1990. Recruitment, growth and residence time of
fishes in a tropical Australian mangrove system. Estuarine, Coastal and Shelf
Science 31 :723 743.
Rooker, J. R., and G. D. Dennis. 1991. Diel, lunar and seasonal changes in a
mangrove fish assemblage off southwestern Puerto Rico. Bulletin of Marine
Science 49 :684 698.
Rose, B. 1982. Lizard home ranges: methodology and functions. Journal of
Herpetology 16 :253 269.
Rummery, C., R. Shine, D. L. Houston, and M. B. Thompson. 1995. Thermal
biology of the Australian forest dragon Hypsilurus spinipes (Agamidae).
Copeia 1995 :818 827.
Sabath, M. D., W. C. Boughton, and S. Easteal. 1981. Expansion of the range of the
introduced toad Bufo marinus in Australia 1935 1974. Copeia 1981 :676 680.
9. References 190
Saenger, P., E. J. Hegerl, and J. D. S. Davie. 1983. Global status of mangrove
ecosystems. Number 3, International Union for Conservation of Nature.
Saenger, P., and S. Snedaker. 1993. Pantropical trends in mangrove above ground
biomass and annual litterfall. Oecologia 96 :293 299.
Sandlund, O. T., P. J. Schei, and A. Viken, editors. 1999. Invasive species and
biodiversity management. Kluwer Academic, Boston.
Schildger, B., H. Tenhu, M. Kramer, M. Gerwing, G. Kuchling, G. Thompson, and
R. Wicker. 1999. Comparative diagnostic imaging of the reproductive tract in
monitors. Radiology Ultrasonography Coelioscopy. Mertensiella 11 :193
211.
Schoener, T. W. 1968. Sizes of feeding territories among birds. Ecology 49 :123 141.
Schoener, T. W., J. B. Losos, and D. A. Spiller. 2005. Island biogeography of
populations: An introduced species transforms survival patterns. Science
310 :1807 1809.
Schoener, T. W., and A. Schoener. 1982. Intraspecific variation in home range size
in some Anolis lizards. Ecology 63 :809 823.
Schoener, T. W., D. A. Spiller, and J. B. Losos. 2004. Variable ecological effects of
hurricanes: The importance of seasonal timing for survival of lizards on
Bahamian islands. Proceedings of the National Academy of Sciences of the
United States of America 101 :177 181.
Schultz, T. J. 2002. Oxygen transport in varanid lizards during exercise. PhD thesis.
Charles Darwin University, Darwin.
Serle, G., J. Ritchie, B. Nairn, and D. Pike. 2002. Australian dictionary of biography.
Cambridge University Press.
9. References 191
Shea, G., and G. L. Reddacliff. 1986. Ossifications in the hemipenes of Varanids.
Journal of Herpetology 20 :566 568.
Shine, R. 1985. The reproductive biology of Australian reptiles: a search for general
patterns. in G. C. Grigg, R. Shine, and H. Ehmann, editors. Biology of
Australasian frogs and reptiles. Surrey Beatty and Sons, Chipping Norton.
Shine, R. 1986. Food habits, habitats and reproductive biology of four sympatric
species of varanid lizards in tropical Australia. Herpetologica 42 :346 360.
Shine, R. 1991a. Australian snakes: a natural history. Reed Books, Sydney.
Shine, R. 1991b. Why do larger snakes eat larger prey items? Functional Ecology
5:493 502.
Shine, R., Ambariyanto, P. S. Harlow, and Mumpuni. 1998a. Ecological traits of
commercially harvested water monitors, Varanus salvator , in northern
Sumatra. Wildlife Research 25 :437 447.
Shine, R., and E. L. Charnov. 1992. Patterns of survival, growth, and maturation in
snakes and lizards. The American Naturalist 139 :1257 1269.
Shine, R., P. S. Harlow, J. S. Keogh, and Boeadi. 1996. Commercial harvesting of
giant lizards: the biology of water monitors, Varanus salvator , in southern
Sumatra. Biological Conservation 77 .
Shine, R., P. S. Harlow, J. S. Keogh, and Boeadi. 1998b. The influence of sex and
body size on food habits of a giant tropical snake, Python reticulatus .
Functional Ecology 12 :248 258.
Shine, R., and T. Madsen. 1997. Prey abundance and predator reproduction: rats and
pythons on a tropical Australian floodplain. Ecology 78 :1078 1086.
9. References 192
Shine, R., D. O'Connor, M. P. Lemaster, and R. T. Mason. 2001. Pick on someone
your own size: ontogenetic shifts in mate choice by male garter snakes result
in size assortative mating. Animal Behaviour 61 :1133 1141.
Slavens, F., and K. Slavens. 2003. Reptiles and amphibians in captivity Longevity
Home page. http://www.pondturtle.com/longev.html (October 2003).
Smirina, E. M., and A. Y. Tsellarius. 1996. Aging, longevity, and growth of the
desert monitor ( Varanus griseus Daud.). Russian Journal of Herpetology
3:(abstract only).
Smith, J. 2004. Fish and company smell after three days: Increasing capture rates of
carrion eating varanids. Herpetological Review 35 :41 43.
Smith, J.G. and Phillips, B.J. (2006) Toxic tucker: the potential impact of cane toads
on Australian reptiles. Pacific Conservation Biology 12 :40 49.
Snider, A. T., and J. K. Bowler. 1992. Longevity of reptiles and amphibians in north
American Collections. Herpetological Circular No. 21, 2nd edition. Society
for the study of reptiles and amphibians.
Sprackland, R. G. 1997. Mangrove monitor lizards. Reptiles 5:48 63.
Stanner, M., and M. Mendelssohn. 1987. Sex ratio, population density and home
range of the desert monitor ( Varanus griseus ) in the southern coastal plain of
Israel. Amphibia Reptilia 8:153 164.
Stebbins, R. C., and R. E. Barwick. 1968. Radiotelemetric study of thermoregulation
in a Lace Monitor. Copeia 1968 :541 547.
Stephens, P. A., S. W. Buskirk, G. D. Hayward, and C. Martínez Del Rio. 2005.
Information theory and hypothesis testing: a call for pluralism. Journal of
Applied Ecology 42 :4 12.
9. References 193
Sutherst, R. W., R. B. Floyd, and G. F. Maywald. 1995. The potential geographical
distribution of the cane toad, Bufo marinus L. in Australia. Conservation
Biology 10 :294 299.
Sweet, S. 1999. Spatial ecology of Varanus gauerti and V. glebopalma in northern
Australia. Mertensiella 11 :317 366.
Terborgh, J., J. A. Estes, P. Paquet, K. Ralls, D. Boyd Heger, B. J. Miller, and R. F.
Noss. 1999. The role of top carnivores in regulating terrestrial ecosystems. in
M. E. Soule and J. Terborgh, editors. Continental conservation: scientific
foundations of regional reserve networks. Island Press, Washington DC.
Terborgh, J., L. Lopez, P. Nunez, M. Rao, G. Shahabuddin, G. Orihuela, M. Riveros,
R. Ascanio, G. H. Adler, T. D. Lambert, and L. Balbas. 2001. Ecological
meltdown in predator free forest fragments. Science 294 :1923 1926.
Thompson, G. 1994. Activity area during the breeding season of Varanus gouldii in
an urban environment. Wildlife Research 21 :227 231.
Thompson, G. 2002. The feasibility of using body proportions in Western Australian
varanids ( Varanus ) as a method for determining a specimen's sex. Records of
the Western Australian Museum 20 :437 439.
Thompson, G., M. De Boer, and E. R. Pianka. 1998. Activity areas, daily
movements, ecology and thermoregulation of an arboreal monitor Varanus
tristis (Squamata: Varanidae) during the breeding season. Australian Journal
of Ecology 24 :117 122.
Thompson, G. G. 1992a. Daily distance travelled and foraging areas of Varanus
gouldii (Reptillia: Varanidae) in an urban environment. Wildlife Research
19 :743 753.
9. References 194
Thompson, G. G. 1992b. Effects of body mass and temperature on standard
metabolic rates for two Australian varanid lizards ( Varanus gouldii and V.
panoptes ). Copeia 1992 :343 350.
Thompson, G. G. 1993. Daily movement patterns and habitat preferences of Varanus
caudolineatus (Reptilia: Varanidae). Wildlife Research 20 :227 231.
Thompson, G. G., M. De Boer, and E. R. Pianka. 1999. Activity and daily
movements of an arboreal monitor lizard, Varanus tristis (Squamata:
Varanidae) during the breeding season. Australian Journal of Ecology
24 :117 122.
Thompson, G. G., and P. C. Withers. 1994. Standard metabolic rates of two small
Australian varanid lizards ( Varanus caudolineatus and V. acanthurus ).
Herpetologica 50 :494 502.
Thompson, G. G., and P. C. Withers. 1997. Comparative morphology of Western
Australian varanid lizards (Squamata: Varanidae). Journal of Morphology
233 :127 152.
Thompson, G. G., and P. C. Withers. 1998. Standard evaporative water loss and
metabolism of juvenile Varanus mertensi (Squamata: Varanidae). Copeia
1998 :1054 1059.
Thompson, M. B., C. H. Daugherty, A. Cree, D. C. French, J. C. Gillingham, and R.
E. Barwick. 1992. Status and longevity of the Tuatara, Sphenodon guntheri ,
and Duvaucel gecko, Hoplodactylus duvauceli , on North Brother Island, New
Zealand. Journal of the Royal Society of New Zealand 22 :132 130.
Tinkle, D. W. 1967. The life and demography of the side blotched lizard, Uta
stansburiana . Miscellaneous Publications, Museum of Zoology. University of
Michigan 132 :1 182.
9. References 195
Tinkle, D. W., A. E. Dunham, and J. D. Cogdon. 1993. Life history and demographic
variation in the lizard, Sceloporus graciosus : a long term study. Ecology
74 :2413 2429.
Tomlinson, P. B. 1986. The botany of mangroves. Cambridge University Press,
London.
Tracy, C. R. 1982. Biophysical modelling in reptilian physiology and ecology. Pages
275 321 in C. Gans and F. H. Pough, editors. Biology of the Reptilia.
Academic Press, New York.
Traeholt, C. 1995. A radiotelemetric study of the thermoregulation of free living
water monitor lizards, Varanus S. salvator . Journal of Comparative
Physiology 165 :125 131.
Tucker, A. D., C. J. Limpus, H. I. McCallum, and K. R. McDonald. 1997.
Movements and home ranges of Crocodylus johnstoni in the Lynd River,
Queensland. Wildlife Research 24 :379 396.
Turner, F. B., G. A. Hoddenback, P. A. Medica, and J. R. Lannon. 1970. The
demography of the lizard, Uta stansburiana Baird and Girard, in southern
Nevada. Journal of Animal Ecology 39 :505 519.
Turner, F. B., R. I. Jennrich, and J. D. Weintraub. 1969. Home ranges and body size
of lizards. Ecology 50 :1076 1081. van der Toorn, J. D. 1999. A survival guide to survival rates. Marine mammals:
Public display and research 3:27 38. van Marken Lichtenbelt, W. D., R. A. Weddelingh, J. T. Vogel, and K. B. M. Albers.
1993. Energy budgets in free living green iguanas in a seasonal environment.
Ecology 65 :575 581.
9. References 196
Vardon, M., S. Gaston, J. Niddrie, and G. Webb. 1998. Wildlife use at Momega,
north central Arnhem Land. Australian Biologist 12 :15 22.
Vitt, L. J., E. R. Pianka, W. E. Cooper, and K. Schwenk. 2003. History and the
global ecology of squamate reptiles. American Naturalist 162 :44 61.
Waldschmidt, S., and C. R. Tracy. 1983. Interactions between a lizard and its thermal
environment: Implications for sprint performance and space utilization in the
lizard Uta stansburiana . Ecology 64 :476 484.
Weavers, B. W. 1993. Home range of male lace monitors Varanus varius (Reptilia:
Varanidae), in south eastern Australia. Wildlife Research 20 :303 313.
Webb, G., and C. Manolis. 1993. Australian crocodiles: A natural history. Reed
Books.
Webb, J. K., G. P. Brown, and R. Shine. 2001. Body size, locomotor speed and
antipredator behaviour in a tropical snake ( Tropidonophis mairii ,
Colubridae): the influence of incubation environments and genetic factors.
Functional Ecology 15 :561 568.
Whitehead, P. J., B. A. Wilson, and K. Saalfeld. 1992. Managing the Magpie Goose
in the Northern Territory: approaches to conservation of mobile fauna in a
patchy environment. Pages 90 104 in I. Moffatt and A. Webb, editors.
Conservation and Development Issues in Northern Australia. North
Australian Research Unit, Australian National University, Darwin.
Wikelski, M., V. Carrillo, and F. Trillmich. 1997. Energy limits to body size in a
grazing reptile, the Galapagos Marine Iguana. Ecology 78 :2204 2217.
Wikramanayake, E. D., and G. L. Dryden. 1988. The reproductive ecology of
Varanus indicus on Guam. Herpetologica 44 :338 344.
9. References 197
Wikramanayake, E. D., and G. L. Dryden. 1999. Implications of heating and cooling
rates of Varanus bengalensis and V. salvator . Mertensiella 11 :149 156.
Wikramanayake, E. D., and B. Green. 1989. Thermoregulatory influences on the
ecology of two sympatric varanids in Sri Lanka. Biotropica 21 :74 79.
Wikramanayake, E. D., and B. Green. 1993. Thermal ecology of habitat and
microhabitat use by sympatric Varanus bengalensis and V. salvator in Sri
Lanka. Copeia 1993 :709 714.
Williams, B. K., J. D. Nichols, and M. J. Conroy. 2002. Analysis and management of
animal populations. Academic Press, San Diego.
Williamson, I. 1999. Competition between the larvae of the introduced cane toad
Bufo marinus (Anura: Bufonidae) and native anurans from the Darling
Downs area of southern Queensland. Australian Journal of Ecology 24 :636
643.
Williamson, M. 1996. Biological invasions. Chapman and Hall, London.
Woinarski, J. C. Z. 1992. Biogeography and conservation of reptiles, mammals and
birds across north western Australia: an inventory and base for planning an
ecological reserve system. Wildlife Research 19 :665 705.
Woinarski, J. C. Z., and N. Gambold. 1992. Gradient analysis of a tropical
herpetofauna: Distribution patterns of terrestrial reptiles and amphibians in
Stage III of Kakadu National Park. Wildlife Research 19 :105 127.
Yoccoz, N. G. 1991. Use, overuse, and misuse of significance tests in evolutionary
biology and ecology. Bulletin of the Ecological Society of America 72 :106
111.
9. References 198
Ziegler, T., and W. Böhme. 1999. Genital morphology and systematics of two
recently described monitor lizards of the Varanus (Euprepiosaurus) indicus
group. Mertensiella 11 :121 128.
Zug, G. R., and P. B. Zug. 1979. The marine toad, Bufo marinus : A natural history
resumé of native populations. Smithsonian Contributions to Zoology 284:58.
9. References 199