ArticlePDF Available

Geographic range, population structure and conservation status of the green python (Morelia viridis), a popular snake in the captive pet trade


Abstract and Figures

Accurate knowledge of distribution and population size is required for effective conservation and management of wild species. Here we report on the first estimates of the distribution and density of the green python (Morelia viridis), an iconic rainforest species widely kept in captivity. We used climatic modelling to predict its distribution in Papua New Guinea, and both climate and vegetation mapping to predict its Australian distribution. We used mark-recapture methods to estimate the density and population structure of green pythons at Iron Range, northern Australia. Bioclimatic analyses suggested that there is extensive climatically suitable habitat in Papua New Guinea (= 200000 km(2)), but very little in Australia (similar to 300 km(2)). However, use of vegetation maps increases the predicted suitable area of occupancy in Australia to 3127 km(2), including nine regional ecosystems. Density estimates at Iron Range were 4-5 ha(-1) in the complex vine forest regional ecosystem; however, only half of these were mature adults. The large predicted area of occurrence and the high density in the one intensively studied area suggest that the species is not vulnerable to extinction in the short term. However, more studies are needed in both New Guinea and Australia, especially to quantify the impact of harvesting green pythons for the pet trade.
Content may be subject to copyright.
The green python (Morelia viridis Schlegel, 1872) is a small
(<1.5 m) python inhabiting a large part of New Guinea, includ-
ing some satellite islands, and Cape York Peninsula, Australia
(Barker and Barker 1994; O’Shea 1996). Although there are
records from a large geographic area, the true distribution and
estimates of population size of the green python are largely
unknown. Globally, the species is listed on Appendix 2 of the
Convention on International Trade in Endangered Species
(Inskipp and Gillett 2003), while the Australian population is
listed as ‘Rare or Insufficiently Known’ (Cogger et al. 1994).
Such conservation assessments are hampered by an inadequate
knowledge of the species’ biology. It is an iconic rainforest
species, and one of the most sought-after snake species in the
captive pet industry. This interest from the captive trade stems
from the snake’s striking colours and remarkable colour
change, with individuals hatching either bright yellow or brick
red and changing to bright green (Wilson et al. 2006b; Wilson
et al. 2007). Many are exported from Indonesia (West Papua)
each year to satisfy the captive pet trade (UNEP-WCMC
CITES trade database). Despite the high profile of green
pythons and the potential pressures on wild populations, infor-
mation for determining their conservation status is severely
Here we use data at multiple scales to address these knowl-
edge gaps, including two complementary methods to predict
their distribution, both globally, and specifically in Australia.
Climatically suitable habitat was predicted for Australia and
Papua New Guinea, whereas in the Australian region these
results were combined with the distribution of regional eco-
system known to be used by green pythons. Density and popu-
lation structure were then estimated using data from an
intensively studied population in northern Australia. We
combine all known data and discuss the likely conservation
status of this species in the wild following IUCN criteria (IUCN
Species localities for predictive distributions
Localities in Papua New Guinea were primarily derived from
museum specimens, the locality lists in O’Shea (1996), or our
fieldwork in Papua New Guinea conducted in late 2005. In
Australia localities were primarily based on our fieldwork, with
additional localities from personal communications with local
people (see Appendices 1 and 2 for site details). Museum speci-
mens were used only if their locality could be accurately deter-
mined. Locations consisted of a latitude, longitude and
elevation. Where elevations were not recorded as primary data
they were derived from topographic maps.
Distribution prediction
BIOCLIM is part of the ANUCLIM software package (Houlder
et al. 1999) and is used to predict the bioclimatic space occupied
by an organism and to make predictions on the geographic pres-
ence or absence of that organism in a defined area. The
BIOCLIM analysis procedure and general limitations are
Australian Journal of Zoology, 2007, 55, 147–154
10.1071/ZO06078 0004-959X/07/030147© CSIRO 2007
David WilsonA,B,C and Robert HeinsohnA
AFenner School of Environment and Society, Australian National University, Canberra, ACT 0200, Australia.
BPresent address: School for Tropical and Marine Biology, James Cook University, Townsville 4811, Australia.
CCorresponding author. Email:
Abstract. Accurate knowledge of distribution and population size is required for effective conservation and
management of wild species. Here we report on the first estimates of the distribution and density of the green python
(Morelia viridis), an iconic rainforest species widely kept in captivity. We used climatic modelling to predict its
distribution in Papua New Guinea, and both climate and vegetation mapping to predict its Australian distribution. We used
mark–recapture methods to estimate the density and population structure of green pythons at Iron Range, northern
Australia. Bioclimatic analyses suggested that there is extensive climatically suitable habitat in Papua New Guinea
(200000 km2), but very little in Australia (~300 km2). However, use of vegetation maps increases the predicted suitable
area of occupancy in Australia to 3127 km2, including nine regional ecosystems. Density estimates at Iron Range were
4–5 ha–1 in the complex vine forest regional ecosystem; however, only half of these were mature adults. The large predicted
area of occurrence and the high density in the one intensively studied area suggest that the species is not vulnerable to
extinction in the short term. However, more studies are needed in both New Guinea and Australia, especially to quantify
the impact of harvesting green pythons for the pet trade.
Geographic range, population structure and conservation
status of the green python (Morelia viridis), a popular
snake in the captive pet trade
D. Wilson and R. Heinsohn148 Australian Journal of Zoology
explained in detail elsewhere (Nix 1986; Lindenmayer et al.
1991; Nix and Switzer 1991; Houlder et al. 1999).
We derived two predicted distributions from the locality
data, the total range of the species based on minimum and
maximum predicted bioclimatic values, and the ‘core’ distri-
bution based on the 10–90 percentile levels of the multivariate
bioclimatic profile (Lindenmayer et al. 1991; Sumner and
Dickman 1998). Core areas represent those areas that have the
greatest conservation value for a species, and may act as refugia
under altered climatic conditions (Lindenmayer et al. 1991).
Locality data from Australia and Papua New Guinea were used
independently for their own region when running BIOCLIM.
Regional ecosystem use
As no detailed land cover classifications are available for New
Guinea this technique was used only in Australia. The vegeta-
tion in Queensland has been categorised into regional ecosys-
tems comprising a vegetation community that is consistently
associated with a particular combination of landform and soil
(Sattler and Williams 1999). Locations were overlain onto the
regional ecosystem map for Queensland in ArcView 3.1 (ESRI
1999). The area of each regional ecosystem where green
pythons were recorded was taken from Neldner (1999) and
ArcView map areas (see Table 1). Habitat preferences per se
were not examined as this requires comparison on used and
unused sites, and unused sites are difficult to comprehensively
determine for this species.
Density and abundance
The density and regional ecosystem use of green pythons was
studied in the Iron Range area, in north-eastern Australia
(12°45S, 143°17E). The climate there is strongly seasonal,
with most rain falling in a distinct ‘wet’ season between
December and May (for more details see Wilson et al. 2006b).
Spotlighting transects were established in the four most
common regional ecosystems in the Iron Range area:
Complex vine forest (regional ecosystem 3.3.1),
Transitional rainforest (regional ecosystem 3.12.8),
Dune rainforest (regional ecosystem 3.2.12), and
Woodland (regional ecosystem 3.3.31).
For further details of these ecosystems see Neldner (1999).
Due to logistic constraints, transect lengths varied between
regional ecosystems and were discontinuous. There were 17 km
of transect in complex vine forest, 3 km in woodland, 2.3 km in
transitional rainforest and 1 km in dune forest, with a transect
width of 30 m (15 m either side of the transect path). This width
was the furthest from the transect that an individual could be
reliably detected, as determined during two weeks of survey
work before the first field season. Transects were surveyed for
green pythons each fortnight for two consecutive wet seasons
(December 2002–April 2003 and December 2003–April 2004)
for a maximum of 21 surveys. Surveys commenced after
2000 hours and all sightings were made by hand-held spotlight
from a slow-moving (<10 km h–1) car or by foot. Green pythons
observed on transects were initially marked with a uniquely
coded passive integrated transponder tag (Gibbons and Andrews
2004), which was recorded on all subsequent survey encounters
to generate a recapture history for each individual. The location
of all sightings was recorded using a GPS (Garmin 12).
The open-population Jolly–Seber method in the program
MARK (White and Burnham 1999) was used to analyse the
recapture histories for this population. This method estimates
the population abundance at the start of the survey period for a
known area. For this model we assumed that survival was time
dependent, but the probability of recapture was constant
throughout the study.
Population demographic data were also collected from indi-
viduals during spotlight surveys. Morphological measurements
taken from recaptures of individuals were used to determine
growth rates, while the initial capture record for each individual
was used to determine the sex and age structure of the popu-
lation. Ages were determined from snout–vent lengths using the
regression equation of Wilson et al. (2006b). The movements,
home range and habitat use of individuals and of different sex
and age classes were determined using radio-tracking tech-
niques on 27 individuals that were followed for up to 451 days
over 18 months (see Wilson et al. 2006afor more details).
In Papua New Guinea BIOCLIM predicts 245535 km2of cli-
matically suitable habitat, with a ‘core’ area of 26 321 km2
Table 1. Regional ecosystems categories where green pythons were found in Australia, the dominant tree species and the extent of each ecosystem.
Regional Vegetation description Total extent in Total extent on
ecosystem protected areas (ha)ACape York (ha)A
3.2.7 Corymbia intermedia or C. clarksoniana woodland in wet coastal areas 2420 11300
3.2.12 Araucarian microphyll vine forest on coastal dunefields and beach ridges 1170 12000
3.3.1 Closed semideciduous mesophyll vine forest 12930 48850
3.3.31 Eucalyptus tetradonta ±Corymbia clarksoniana ±C. tessellaris woodland on coastal plains 9460 55000
3.5.5 Corymbia novoguinensis ±C. tessellaris woodland on northern Cape York Peninsula none 6250
3.5.13 Melaleuca viridifolia, Asteromyrtus brassii woodland on flat plains 770 8615
3.11.3 Simple evergreen notophyll vine forest on exposed metamorphic and granitic slopes 8300 79 000
3.12.3 Notophyll vine forest 3150 77600
3.12.8 Corymbia clarksoniana ±C. tessellaris open forest on coastal ranges and lowlands 2170 33400
Total for all vegetation types 40370 332015
AValues taken from Neldner (1999). Protected areas are those listed under the Nature Conservation Act 1992, and include national parks, conservation parks
and resource reserves (Neldner 1999).
Australian Journal of Zoology 149
(Fig. 1). Large core areas are predicted on the lower slopes of
the Huon Peninsula and southern portion of the central
cordillera, plus parts of the trans-Fly region. Green pythons
were predicted to be absent from the central highlands and
swamp areas of the trans-Fly and Sepik drainages. BIOCLIM
also predicts green pythons to be on New Britain, and many of
the smaller satellite islands of New Guinea. In Australia
BIOCLIM predicts there to be 292.82 km2of climatically suit-
able habitat, with a core area of 15.73 km2(Fig. 2a). This core
area is contained within the Iron Range area, with smaller frag-
ments of suitable habitat predicted further south in the
McIllwraith Range and in isolated pockets (of single grid cells)
further north along the coast. Interestingly, sites in Australia
predicted no climatically suitable habitat in Papua New Guinea
and vice versa.
Regional ecosystem preferences in Australia
In Australia green pythons were recorded from nine regional
ecosystems, totalling an area of 3127 km2(Fig. 2b, Table 1).
Most of the suitable regional ecosystems were concentrated
around the Iron and McIlwraith Ranges, with a smaller discrete
area at the Lockerbie Scrub, and a few isolated patches between
them (Fig. 2b). There were also suitable areas south of the Laura
divide, notably in the Cape Melville area, but these were
excluded from further analysis (see Discussion). Most records
were from the Iron Range area, with two records each from the
McIlwraith Range area and Lockerbie Scrub. Most records were
from complex vine forest (regional ecosystem 3.3.1). There are
488.5 km2of this regional ecosystem on Cape York Peninsula,
with 129.3 km2of this in protected areas (Neldner 1999).
Density, abundance and population structure
In total, 101 individuals were captured 147 times in complex
vine forest during the fortnightly surveys over the two wet
seasons. The total number of green pythons in the surveyed area
Distribution of green pythons
0 200 400 600 800100
Sepik drainage
Huon Peninsula
New Britain
Fly delta
Fig. 1. BIOCLIM prediction of climatically suitable areas for the green
python (Morelia viridis) in Papua New Guinea. Light grey represents total
range, while dark grey represents the predicted core range. Dots are the
sighting locations on which the prediction is based.
075 22530037.5
Lockerbie Scrub
Iron Range
McIlwraith Range
Laura Basin
0 75 150 225 30037.5
Lockerbie Scrub
Iron Range
McIlwraith Range
Laura Basin
Fig. 2. Predicted distribution of the green python (Morelia viridis) in Australia. (a) BIOCLIM prediction of the total range. The core area is hidden under
the clump of sightings at Iron Range. (b) Prediction based on regional ecosystem distributions matched with known sighting localities. In both cases dots are
sighting localities on which the predictions are based.
D. Wilson and R. Heinsohn150 Australian Journal of Zoology
of this regional ecosystem was estimated at 227 ± 85 (s.e.) using
the Jolly–Seber model in MARK (White and Burnham 1999).
Given a survey area of 50 ha (0.5 km2, as defined in the
Methods), this equates to ~4–5 ha–1 in this regional ecosystem.
Seven individuals were caught in transitional rainforest, but no
recaptures were made, hence mark–recapture analysis tech-
niques could not be used to estimate abundance in this regional
ecosystem. Juveniles were found only in canopy gaps within
complex vine forest and were never found in transitional rain-
forest. No green pythons were recorded from either the wood-
land or dune rainforests, and these transects were discontinued
after 10 repeats.
On the basis of the known-age structure of this population
(Wilson et al. 2006b) these 227 individuals comprise 49 adult
females and 65 adult males, 75 immature females and 14 imma-
ture males, and 14 juvenile females and 10 juvenile males
(Fig. 3a). The age structure of this population is positively
skewed, with a mean age of 3.4 years and a maximum predicted
age of at least 13 years (Fig. 3b).
This study is the first to estimate the density and structure of a
green python population and adds to our knowledge of this
unusual and charismatic species in the wild. On the basis of bio-
climatic data and the distribution of regional ecosystems used,
green pythons appear to have a large potential distribution in
Papua New Guinea, including inshore islands and much of the
lowlands and foothills. In Australia green pythons are restricted
to very small areas of suitable habitat on eastern Cape York
Peninsula; however, the one intensively sampled regional
ecosystem contained high densities of individuals.
Overall, the BIOCLIM analysis indicated that green pythons are
potentially widely distributed in Papua New Guinea, but
restricted to small areas of far northern Australia. BIOCLIM
was chosen because of its simplicity of use and the requirement
of presence-only data, rather than the presence or absence
required by more detailed models (Elith et al. 2006). It is also a
well established approach that has been used previously to
predict the potential distributions for a variety of plant and
animal species in Australia (Nix 1986; Lindenmayer et al. 1991,
1996; Olsen and Doran 2002). Climatic conditions in areas
where green pythons were predicted to occur are characterised
by hot and wet summers with cooler winters, typically corre-
sponding to areas of rainforest.
In Papua New Guinea BIOCLIM predicts both a large core
and total area of suitable habitat (Fig. 1). The distribution of
green pythons could not be predicted in West Papua (Indonesia),
as we could find no accurate climate models for this area,
although we believe that this area would contain extensive suit-
able habitat due to the similar climate and landforms between
Papua New Guinea and West Papua. Large areas of the central
highlands were excluded, as were the higher altitudes on the
Huon Peninsula and south along the central cordillera, presum-
ably due to low minimum temperatures. Portions of the Fly delta
(in the south-west), which are covered in low alluvial plains and
flats, and the Sepik drainage (in the north), which is dominated
by lowland freshwater swamps (Paijmans 1976), are also
excluded from the predicted habitat. BIOCLIM predicts sub-
stantial climatically suitable habitat on New Britain, and on
islands between New Britain and the mainland (Fig. 1), despite
there being no records from any of these islands. These islands
have never been connected to mainland New Guinea (Mayr and
Diamond 2001), and green pythons have apparently not been
able to colonise these islands. However, their presence on other
oceanic islands such as Biak, off the north coast of Irian Jaya,
shows that colonisation over water can occur. In contrast, the
islands at the south-eastern tip of New Guinea were previously
connected to the mainland (Mayr and Diamond 2001) and do
contain green pythons (O’Shea 1996), as predicted by the
BIOCLIM analyses.
The climatic analysis did not show a clear demarcation
between the northern and southern watersheds in Papua New
Guinea (Fig. 1). This is in contrast to the findings of Rawlings
and Donnellan (2003), who found clear genetic differences
between the northern and southern populations.
Juvenile Immature Adult
Age class
Number of individuals
Age (years)
Number of individuals
Fig. 3. Demographic composition of the green python (Morelia viridis)
population in the survey area. Total numbers are based on the population
estimate from a Jolly–Seber model in MARK, while the proportion in each
category is based on the size distribution of all captures during fieldwork
(Wilson et al. 2006b). Size class distributions (a) of females (black
columns) and males (grey columns); age distributions of all individuals (b).
Age categories listed are the upper bound of each range.
Australian Journal of Zoology 151
Australian distribution
In Australia, both bioclimatic analysis and the distribution of
regional ecosystems where green pythons were found were used
to predict their potential distribution. Climatic modelling for the
Cape York Peninsula area is poor due to the low number of data
points (H. Nix, pers. comm. 2006), suggesting that, rather than
using the regional ecosystems to further decrease the area pre-
dicted by climatic analysis (Meyer and Thuiller 2006), the two
methods should be used in a complementary fashion. The area
of regional ecosystems in Australia where green pythons have
been found is considerably larger than the climatically suitable
area as predicted by BIOCLIM (3127 km2compared with
293 km2). Given this disparity, which one is closer to the true
extent of occurrence of the green python? The distribution of
location records in Australia may reflect the true distribution
and density of green pythons, but may also represent the easiest
access points into suitable habitat. Iron Range, where most
records occur, is a popular area with both amateur naturalists
and scientists due to the diversity of animal species that occur
there (Kikkawa et al. 1981). In comparison, the McIlwraith
Range is relatively remote and comparatively difficult to access.
Areas of suitable regional ecosystems south of the Laura Basin
were excluded from ‘area of occurrence’ estimates as no green
pythons have been recorded from this area, and the Laura Basin
is an effective barrier restricting the southward spread of rain-
forest species (Lavarack and Godwin 1987). Interestingly,
BIOCLIM predicted no substantial areas of climatically suitable
habitat in Australia where green pythons had not been previ-
ously recorded (Fig. 2a). Climatically suitable habitat exists in a
few isolated locations between Iron Range and the northern tip
of Cape York Peninsula; however, these do not appear to contain
suitable vegetation (based on regional ecosystem mapping) and
have never been surveyed for green pythons. Even the
Lockerbie Scrub, where there were two green python records,
has little climatically suitable habitat (Fig. 2a).
This study highlights the markedly different conclusions that
may be drawn using different distribution prediction methods
(Loiselle et al. 2003). There are biases in both methods and
neither option should be viewed as more accurate. Both models
should be used as indicators of predicted distribution, and be
considered in conjunction with other traits that may limit distri-
bution . Importantly, both methods predict the Iron Range area
to be core habitat for green pythons in Australia, highlighting
the conservation significance of this area (Mackay and Nix
2001). The Iron–McIlwraith Range area forms the largest
remaining area of lowland tropical rainforest in Australia, and is
distinct from more southerly rainforests in its flora and fauna.
Both floristically and faunally, this area is more related to the
Melanesian lowland rainforests of New Guinea, with which it
shares a large number of genera and species (Kikkawa et al.
1981; Webb and Tracey 1981; Crisp et al. 2001).
Density and abundance
Our estimate of 4–5 ha–1 in complex vine forest is well within
the range of densities reported, both for snake species world-
wide and most other published studies on tropical snake species
(Parker and Plummer 1987; Brown and Shine 2002). The
density of green pythons was actually greater than expected.
Individuals were rarely encountered during fieldwork, with new
individuals encountered only every 2–3 h (authors’ unpub. data).
However, in support of our estimates, although we caught only
101 individuals during surveys, in total we caught 207 individ-
uals in the study area during all research, similar to the 227 indi-
viduals predicted in the same area using the Jolly–Seber model.
Radio-tracking shows that individuals do not always come to the
ground at night to hunt (Wilson 2007), and this behaviour may
give the impression of fewer individuals than suggested by the
density estimates. The age structure of the Iron Range popu-
lation appears to be healthy, with significant numbers of juve-
niles and a high proportion of recently mature adults.
There have been two other well studied Australian pythons,
the carpet python (M. spilota imbricata), which had a density of
~0.5 ha–1 in temperate south-western Australia (Pearson et al.
2005), and the water python (Liasis fuscus), which had
encounter rates of ~0.5 h–1 (Brown and Shine 2002), similar to
the encounter rate for green pythons (authors’ unpub. data). By
comparison, Schulte (1988) estimated the density of the
emerald tree boa (C. caninus) to be 0.004 ha–1 in the Peruvian
Amazon. This discrepancy is particularly noteworthy as green
pythons and emerald tree boas show strikingly convergent evo-
lution in many aspects of their ecology. Both species are tropi-
cal, arboreal specialists and show ontogenetic colour change
from yellow or red juveniles to green adults at ~55 cm (Stafford
and Henderson 1996; Wilson et al. 2006b).
Our density estimates suggest that there are substantial
numbers of green pythons in complex vine forest at Iron Range,
and this area represents the single largest known population in
Australia. Adult green pythons were also observed in areas of
transitional rainforest, but juveniles were never recorded in this
regional ecosystem. This suggests that transitional rainforest
can be recolonised, but may not be suitable for breeding in.
Green pythons were opportunistically recorded in a further three
regional ecosystems in the Iron Range area; however, density
estimates are not available for these areas. These ecosystems lie
primarily within the protected area of the Iron Range National
Green pythons were twice recorded from the McIlwraith
Range to the south of Iron Range (Christian 1997; K. McDonald,
pers. comm. 2002) and these two areas may be connected by
gallery rainforest (Legge et al. 2004). Although individual daily
movements are small, the fact that they are constantly active and
males have a roaming strategy (Wilson et al. 2006a) suggests
that the Iron Range and McIlwraith populations would be con-
nected wherever habitat corridors exist. There is a smaller area of
suitable regional ecosystem further north at the Lockerbie Scrub
where green pythons have twice been recorded (Waldren 1996;
S. Templeton, pers. comm. 2004). At most, this area might
contain a few hundred individuals and the Lockerbie Scrub may
be too small to support a green python population that is viable
in the long term.
Although the density of green pythons in the complex vine
forest seems high, we hesitate to extrapolate our estimates to the
entire possible range of green pythons. It would be problematic
to extrapolate the density estimates from the one regional
ecosystem at Iron Range to the McIlwraith Range for several
reasons. First, there have been only two reports of green pythons
from the McIlwraith Range in the last 150 years (authors’
Distribution of green pythons
D. Wilson and R. Heinsohn152 Australian Journal of Zoology
unpub. data) and this makes any extrapolation tenuous. Second,
although a density estimate is available for one of the regional
ecosystems that occurs in the McIlwraith Range, actual densi-
ties may vary between the two areas, as has been shown in other
snake species (Parker and Plummer 1987; Henderson 2002).
Third, there are substantial areas of regional ecosystems where,
although green pythons have been recorded, density estimates
are not available. Ideally, surveys should be carried out in all
regional ecosystems where green pythons are known to occur,
both at Iron Range and in the McIlwraith Range region to esti-
mate densities in these potentially important ecosystems.
Conservation status
Our data suggest that green pythons do not satisfy the IUCN cri-
teria (Table 2) for being listed as vulnerable, either globally or
within Australia. We must stress that important information is
still lacking, especially for the potentially larger New Guinea
population and, as such, there is a large element of uncertainty
in this conclusion. Although our bioclimatic analysis demon-
strated large areas of suitable habitat in Papua New Guinea, and
similarly large areas are probably also available in western New
Guinea, there are currently no data available covering the extent
to which local people hunt them for food. Further, although they
are known to be taken for the international pet trade in West
Papua, numbers taken and population densities for different
regions are not currently available.
In Australia we have a more detailed knowledge of their
distribution and population size. Green pythons do not appear to
be vulnerable under any of the IUCN criteria (Table 2); however,
further work needs to be undertaken to refine our assessment.
Long-term information on population trends is lacking, as are
density estimates from the other eight regional ecosystems
where green pythons were recorded. The population structure at
Iron Range also shows that mature individuals comprise less
than 50% of the total population, reducing the effective popu-
lation size. Surveys are needed in both the McIlwraith Range
and at Lockerbie Scrub to determine the density and distribution
of green pythons in these areas. Hence we conclude that there
are insufficient grounds to list green pythons as threatened,
either globally or in Australia, but reiterate the need for further
study, especially in New Guinea.
We thank all the people who helped search for specimens in the field,
especially K. and A. Goetze, S. Legge, S. Murphy, B. Robinson, E. Sobey
and K. Wilson. K. Nissen and J. Stein helped to generate maps. This research
was funded by the Australian Geographic Society, National Geographic
Society and the Hermon Slade Foundation. Appropriate permits were issued
by the Queensland Environmental Protection Agency (WITK00337502)
and the Australian National University (C.RE.24.02 and 27.02).
Barker, D., and Barker, T. (1994). ‘Pythons of the World. Volume 1.
Australia.’(Advanced Vivarium Systems Inc.: Escondido, CA.)
Brown, G., and Shine, R. (2002). Influence of weather condition on activity
of tropical snakes. Austral Ecology 27, 596–605. doi:10.1046/j.1442-
Christian, T. (1997). A field trip to the McIlwraith Range – Cape York
Peninsula. Monitor 8, 160–169.
Cogger, H., Cameron, E., Sadlier, R., and Eggler, P. (1994). ‘The Action
Plan for Australian Reptiles.’ (Australian Nature Conservation Agency:
Crisp, M., Laffan, S., Linder, H., and Monro, A. (2001). Endemism in the
Australian flora. Journal of Biogeography 28, 183–198. doi:10.1046/
Elith, J., Graham, C., Anderson, R., Dudík, M., and Fer rier, S., et al. (2006).
Novel methods improve prediction of species’ distributions from occur-
rence data. Ecography 29, 129–151. doi:10.1111/j.2006.0906-7590.
ESRI (1999). ‘ArcView GIS Version 3.1.1.’ (Environmental Systems
Research Institute, Inc.: Redlands, CA.)
Gibbons, J., and Andrews, K. (2004). PIT tagging: simple technology at its
best. Bioscience 54, 447–454. doi:10.1641/0006-3568(2004)054
Henderson, R. (2002). ‘Neotropical Treeboas. Natural History of the
Corallus hortolanus Complex.’ (Krieger Publishing Company: Malabar,
Houlder, D., Hutchinson, M., Nix, H., and McMahon, J. (1999).
‘ANUCLIM Version 5.0 User Guide.’ (Centre for Resources and
Environmental Studies, Australian National University: Canberra.)
Inskipp, T., and Gillett, H. (Eds) (2003). ‘Checklist of CITES Species.’
Compiled by UNEP-WCMC. (CITES Secretariat and UNEP-WCMC:
IUCN (2001). ‘IUCN Red List Categories and Criteria: Version 3.1.’(IUCN
Species Survival Commission: Cambridge, UK.)
Kikkawa, J., Monteith, G., and Ingram, G. (1981). Cape York Peninsula:
major region of faunal interchange. In ‘Ecological Biogeography of
Australia’. (Ed. A. Keast.) pp. 1695–1742. (Dr W. Junk: The Hague.)
Lavarack, P., and Godwin, M. (1987). Rainforests of northern Cape York
Peninsula. In ‘The Rainforest Legacy’. (Eds G. Werren and A. Kershaw.)
pp. 201–222. (Australian Government Publishing Service: Canberra.)
Legge, S., Heinsohn, R., and Garnett, S. (2004). Availability of nest hollows
and breeding population size of eclectus parrots, Eclectus roratus, on
Cape York Peninsula Australian Wildlife Research 31, 149–161.
Lindenmayer, D., Nix, H., McMahon, J., Hutchinson, M., and Tanton, M.
(1991). The conservation of Leadbeater’s possum, Gymnobelideus
Table 2. Assessment of the world and Australian populations of the green python (Morelia viridis) against the summarised IUCN Vulnerable
criteria (IUCN 2001)
IUCN Criteria World Australia
A. Has there been, or is there projected to be, a population decrease of more than 30% in 10 years? Not known Not known
B. Is the population extent of occurrence 20000 km2or the area of occupancy 2000 km2?NoNo
C. Is the population estimated at 10000 mature individuals? No No
D. Is the population size 1000 individuals or with a very restricted area, such that it is quickly capable of No No
becoming endangered?
E. Does quantitative analysis show the probability of extinction in the wild is at least 10% within 100 years? Not undertaken
Does the species qualify for Vulnerable status? No No
Australian Journal of Zoology 153
leadbeateri (McCoy): a case study of the use of bioclimatic modelling.
Journal of Biogeography 18, 371–383. doi:10.2307/2845479
Lindenmayer, D., Mackey, B., and Nix, H. (1996). The bioclimatic domains
of commercially important eucalypts from south-eastern Australia.
Australian Forestry 59, 74–89.
Loiselle, B., Howell, C., Graham, C., Goerck, J., Brooks, T., Smith, K., and
Williams, P. (2003). Avoiding pitfalls of using species distribution
models in conservation planning. Conservation Biology 17, 1591–1600.
Mackay, B., and Nix, H. (2001). ‘The Natural History Signif icance of Cape
York Peninsula. Report for the Queensland Environmental Protection
Agency.’(ANUTECH Pty Ltd: Canberra.)
Mayr, E., and Diamond, J. (2001). ‘The Birds of Northern Melanesia.’
(Oxford University Press: New York.)
Meyer, C., and Thuiller, W. (2006). Accuracy of resource selection functions
across spatial scales. Diversity & Distributions 12, 288–297.
Neldner, V. (1999). Cape York Peninsula. In ‘The Conservation Status of
Queensland’s Bioregional Ecosystems’. (Eds P. Sattler and R. Williams.)
pp. 3/1–3/85. (Environmental Protection Agency: Brisbane.)
Nix, H. (1986). Biogeographic analysis of the Australian elapid snakes. In
‘Atlas of Elapid Snakes of Australia’. (Ed. R. Longmore.) pp. 4–15.
(Australian Government Publishing Service: Canberra.)
Nix, H., and Switzer, M. (Eds) (1991). ‘Rainforest Animals: Atlas of
Vertebrates Endemic to Australia’s Wet Tropics. Kowari 1.’ (Australian
Parks and Wildlife Service: Canberra.)
Olsen, P., and Doran, B. (2002). Climatic modelling of the Australian distri-
bution of the grass owl (Tyto capensis): is there an inland population?
Wildlife Research 29, 117–125. doi:10.1071/WR01030
O’Shea, M. (1996). ‘A Guide to the Snakes of Papua New Guinea.’
(Independent Publishing: Port Moresby.)
Paijmans, K. (1976). Vegetation. In ‘New Guinea Vegetation’. (Ed.
K. Paijmans.) pp. 23–105. (Australian National University Press:
Parker, W., and Plummer, M. (1987). Population ecology. In ‘Snakes:
Ecology and Evolutionary Biology’. (Eds R. Seigel, J. Collins and
S. Novak.) pp. 253–301. (Macmillan Publishing Co.: New York.)
Pearson, D., Shine, R., and Williams, A. (2005). Spatial ecology of a threat-
ened python (Morelia spilota imbricata) and the effects of anthro-
pogenic habitat change. Austral Ecology 30, 261–274. doi:10.1111/
Rawlings, L., and Donnellan, S. (2003). Phylogeographic analysis of the
green python, Morelia viridis, reveals cryptic diversity. Molecular
Phylogenetics and Evolution 27, 36–44. doi:10.1016/S1055-7903(02)
Sattler, P., and Williams, R. (Eds) (1999). ‘The Conservation Status of
Queensland’s Bioregional Ecosystems.’ (Environmental Protection
Agency: Brisbane.)
Schulte, R. (1988). Observaciones sobre la boa verde, Corallus caninus, en
el Departmento San Martin-Peru. Boletín de Lima 55, 21–26.
Stafford, P., and Henderson, R. (1996). ‘Kaleidoscopic Tree Boas.’(Krieger
Publishing Company: Malabar, FL.)
Sumner, J., and Dickman, C. (1998). Distribution and identity of species in
the Antechinus stuartii–A. flavipes group (Marsupialia: Dasyuridae) in
south-eastern Australia. Australian Journal of Zoology 46, 27–41.
Waldren, I. (1996). Rainforests of northern Cape York Peninsula and their
exploitation by herptiles. Monitor 8, 5–14.
Webb, L., and Tracey, J. (1981). Australian rainforests: patterns and change.
In ‘Ecological Biogeography of Australia’. (Ed. A. Keast.) pp. 607–694.
(Dr W. Junk: The Hague.)
White, G., and Burnham, K. (1999). Program MARK: survival estimation
from populations of marked animals. Bird Study 46S, 120–138.
Wilson, D. (2007). Foraging ecology and diet of an ambush predator: the
green python Morelia viridis. In ‘Biology of the Boas and Pythons’.
(Eds R. Henderson and R. Powell.) pp. 141–150. (Eagle Mountain
Publishing: Eagle Mountain, UT.)
Wilson, D., Heinsohn, R., and Legge, S. (2006a). Spatial ecology of dichro-
matic green pythons (Morelia viridis) in Australian tropical rainforests.
Austral Ecology 31, 577–587. doi:10.1111/j.1442-9993.2006.01519.x
Wilson, D., Heinsohn, R., and Wood, J. (2006b). Life history traits and
colour change in the arboreal tropical python Morelia viridis. Journal of
Zoology 270, 399–407. doi:10.1111/j.1469-7998.2006.00190.x
Wilson, D., Heinsohn, R., and Endler, J. (2007). The adaptive signif icance
of ontogenetic colour change in a tropical python. Biology Letters 3,
40–43. doi:10.1098/rsbl.2006.0574
Manuscript received 12 September 2006, accepted 15 May 2007
Distribution of green pythons
D. Wilson and R. Heinsohn154 Australian Journal of Zoology
Appendix 1. Details of locations in Papua New Guinea used in the
BIOCLIM prediction of the distribution of the green python (Morelia
viridis) in Papua New Guinea
Location Latitude Longitude Elevation (m)
Abam 8.95 143.183 50
Aitape 3.13333 142.333 0
Aiyurafu 6.56667 145.333 1968
Aramia River 7.93333 143.367 1
Baiyer 5.53333 144.15 1170
Biniguni 9.66667 149.283 198
Bulolo 7.2 146.65 794
Chimbu River 6.05 144.967 871
Dede 8.3 142.883 1
Derongo 5.41667 141.1 314
Fergusson Island 9.55 150.667 0
Finschafen 6.56667 147.85 0
Garaina 7.88333 147.15 699
Goroka 6.06667 145.383 1524
Kainantu 6.28333 145.867 1553
Kapuma 7.58333 144.967 1
Karimui 6.5 144.85 983
Kebil 6.2 145.033 1840
Kerema 7.96667 145.75 1
Kunini 9.08333 143 0
Kwima 6.13333 144.967 1541
Lae 6.73333 146.983 1
Lake Murray 6.81667 141.383 59
Lufa 6.31667 145.317 1621
Mafulu 8.51667 147.033 1500
Maiwara 10.35 150.35 0
Mt Lamington 8.91667 148.167 1679
Nivi 6.2 145.333 1646
Nondugl 5.86667 144.767 1702
Normanby Island 10 151.167 0
Okapa 6.53333 145.617 1814
Omati 7.73333 144.183 0
Popondetta 8.76667 148.25 156
Simbai 5.28333 144.517 2009
Sinaeada 10.3167 150.317 48
Sturt Island 8.16667 142.25 0
Telefomin 5.13333 141.617 1240
Urapmin 5.15 141.5 1808
Waghi 5.83333 144.633 0
Wau 7.33333 146.717 1200
Woitape 8.55 147.283 1850
Wombon 5.63333 141.1 191
Wonenara 6.8 145.883 1559
Zim 8.78333 143.1 91
Appendix 2. Details of locations in Australia used in the BIOCLIM
prediction of the distribution of the green python (Morelia viridis)
in Australia
Location Latitude Longitude Elevation (m)
Chili Beach 1 12.6299 143.422 5
Chili Beach 2 12.629 143.425 5
Chili Beach 3 12.6293 143.427 5
Iron Range 1 12.741 143.285 20
Iron Range 2 12.7437 143.283 50
Iron Range 3 12.7104 143.293 80
Iron Range 4 12.7541 143.288 50
Iron Range 5 12.7143 143.319 70
Iron Range 6 12.7096 143.297 80
Iron Range 7 12.7644 143.287 92
Iron Range 8 12.78 143.309 104
Iron Range 9 12.7769 143.282 116
Iron Range 10 12.7061 143.297 57
Iron Range 11 12.6987 143.3 128
Iron Range 12 12.6993 143.303 140
Iron Range 13 12.7458 143.232 152
Iron Range 14 12.7136 143.3 164
Iron Range 15 12.7459 143.23 103
Lockerbie 1 142.58 10.78 5
Lockerbie 2 142.46 10.79 80
Peach Creek 1 13.7367 143.339 530
Peach Creek 2 13.7372 143.339 550
... This species may have a lifespan from 15 to 20 years [65], which also increases its invasive potential. Despite snakes from genus Morelia being popular in the pet trade [67] no record of this species outside its native range exists [12]. ...
Full-text available
Because biological invasions can cause many negative impacts, accurate predictions are necessary for implementing effective restrictions aimed at specific high-risk taxa. The pet trade in recent years became the most important pathway for the introduction of non-indigenous species of reptiles worldwide. Therefore, we decided to determine the most common species of lizards, snakes, and crocodiles traded as pets on the basis of market surveys in the Czech Republic, which is an export hub for ornamental animals in the European Union (EU). Subsequently, the establishment and invasion potential for the entire EU was determined for 308 species using proven risk assessment models (RAM, AS-ISK). Species with high establishment potential (determined by RAM) and at the same time with high potential to significantly harm native ecosystems (determined by AS-ISK) included the snakes Thamnophis sirtalis (Colubridae), Morelia spilota (Pythonidae) and also the lizards Tiliqua scincoides (Scincidae) and Intellagama lesueurii (Agamidae).
... Weigner and D. Wilson, pers. comm.;Wilson and Heinsohn, 2007), such statements do not automatically prove its absence, and may simply be an indication of the elusiveness of the species (Covacevich et al., 1982;Meiri, 2016). The scarcity of anecdotal reports surrounding the presence of V. prasinus on Cape York Peninsula may also be influenced by the arboreal nature of the species (but see, e.g. ...
Full-text available
We summarize and evaluate anecdotal information and observations about the potential occurrence of the New Guinea emerald tree monitor lizard, Varanus prasinus (Schlegel, 1839), on mainland Australia. Several independent but unconfirmed reports about sightings of large green lizards in the rainforests on Cape York Peninsula have been published in the last 40 years, but still no photographs or voucher specimens exist. The closest confirmed occurrence of V. prasinus to mainland Australia is on Moa (also known as Mua or Banks) Island, one of the islands in the Torres Strait that separates New Guinea from Australia. The shallow tropical waters of the Sahul Shelf surrounding these small islands were dry land during Pleistocene glacial periods and facilitated faunal exchanges between both huge landmasses in the past. Consequently, a natural occurrence of V. prasinus on Cape York, together with the endemic Varanus keithhornei (Wells and Wellington, 1985), the canopy goanna, seems plausible. Likewise, a possible polymorphism in the color pattern of the latter species as a putative result of repeated introgression and/or hybridisation events as source for the sightings of green tree monitors deserves further investigations. Therefore, we encourage future field work in this remote area to finally answer the question if the New Guinea emerald tree monitor is native to mainland Australia.
... However rainforest pockets were probably preserved at the very least in topographic refugia (e.g. the escarp- ment of the Great Dividing Range) due to reliable orographic rainfall, which is better reflected in our MAXENT predictions. Enough rainforest for Palm Cockatoos was probably maintained at the Iron and McIlwraith Ranges given the persistence of other large rainforest-dependent vertebrates that have disappeared from rainforests elsewhere in Australia (Eclectus Parrots (Eclectus roratus); Legge et al. 2004; Green Pythons (Morelia viridis); Wilson and Heinsohn 2007). Evolution of a unique vocal dialect at Iron Range may have occurred in isolation in the refugial population there, similar to isolation-recombination dynamics creating dialect boundaries in other Australian parrots (e.g. ...
Full-text available
Species persistence and maintenance of genetic diversity are strongly affected by dispersal and historical distribution, especially when species depend on habitat that is non-uniform or fluctuates dramatically with changing climate. Australo-Papuan rainforest has fluctuated dramatically since the last glacial maximum (around 20 kya). To understand how prehistoric climate fluctuation affected population connectivity and genetic diversity in a rainforest edge species, we screened 27 Palm Cockatoo samples from Cape York Peninsula (Australia) and southern Papua New Guinea (PNG) in 1132 single nucleotide polymorphisms in 342 nuclear loci and the mitochondrial ND2 gene. We also modelled the birds’ distribution at present, mid-Holocene (~6 kya) and the last glacial maximum (~21 kya). Population differentiation in nuclear genomic data among Australian populations aligns with distribution contraction to mountainous refugia at the mid-Holocene (~6 kya). Lack of nuclear divergence between PNG and Australia may reflect late-Holocene recolonisation, but different ND2 haplotypes suggest early stages of divergence. Although admixed individuals suggest some gene flow, recent movement restriction to/from Australian refugia is suggested by a unique ND2 haplotype, genomic divergence and a vocal dialect boundary shown previously. Our results show how prehistoric climate fluctuation affects present-day and future species conservation in dynamic rainforest edge ecosystems.
... Notwithstanding, the LB is considered to have been important in both limiting the modern distributions of southward-colonising mesic forest species and driving divergence among populations of species that span the barrier (Lavarack & Godwin, 1987). Southward-colonisers include palm cockatoos (Murphy, Double & Legge, 2007), cuscuses (Winter & Leung, 1995), rodents (Bryant et al., 2011), and pythons (Wilson & Heinsohn, 2007). Most such species apparently colonised Australia from New Guinea via the Torres Strait land bridge during Plio-Pleistocene glacial cycles before being isolated on Cape York by rising sea levels (see Bryant & Fuller, 2014). ...
Full-text available
The influence of climatic changes occurring since the late Miocene on Australia’s eastern mesic ecosystems has received significant attention over the past 20 years. In particular, the impact of the dramatic shift from widespread rainforest habitat to a much drier landscape in which closed forest refugia were dissected by open woodland/savannah ecosystems has long been a focal point in Australian ecology and biogeography. Several specific regions along the eastern coast have been identified previously as potentially representing major biogeographical disjunctions for closed forest taxa. Initially, evidence stemmed from recognition of common zones where avian species/subspecies distributions and/or floral communities were consistently separated, but the body of work has since grown significantly with the rise of molecular phylogeographic tools and there is now a significant literature base that discusses the drivers, processes and effects of these hypothesised major biogeographical junctions (termed barriers). Here, we review the literature concerning eight major barriers argued to have influenced closed forest taxa; namely, the Laura Basin, Black Mountain Corridor, Burdekin Gap, Saint Lawrence Gap, Brisbane Valley Barrier, Hunter Valley Barrier, Southern Transition Zone and East Gippsland Barrier. We synthesise reported phylogeographical patterns and the inferred timing of influence with current climatic, vegetation and geological characteristics for each barrier to provide insights into regional evolution and seek to elicit common trends. All eight putative biogeographical barriers are characterised currently by lowland zones of drier, warmer, more open woodland and savannah habitat, with adjacent closed forest habitats isolated to upland cool, wet refugia. Molecular divergence estimates suggest two pulses of divergence, one in the early Miocene (~20–15 Mya) and a later one from the Pliocene–Pleistocene (~6–0.04 Mya). We conclude with a prospectus for future research on the eastern Australian closed forests and highlight critical issues for ongoing studies of biogeographical barriers worldwide.
... Whilst studies of snake distributions employing novel methodologies such as ecological niche modelling (e.g. Wilson & Heinsohn, 2007;Roberio et al., 2009;Santos et al., 2009) have provided information on the potential distribution of understudied snakes, none have attempted to directly attribute predicted distributions to the fragmentation of a species' habitat. In fact, the lack of similar previous studies (but see Luiselli & Capizzi, 1997;Kjoss & Litvaitis, 2001) is indicative of the difficulty in conducting such work. ...
Full-text available
Coronella austriaca is the United Kingdom’s rarest snake, being confined to the lowland heathlands of Dorset, Hampshire and Surrey. As a result, it remains the least understood; despite being listed as a key biodiversity action plan species. Substantial loss and fragmentation of its primary UK habitat - lowland heathland - has occurred in recent times, and yet research examining the population ecology and conservation genetics of this species remains limited. As a result, this PhD research was developed to fill this need. Based on three years of data collection, a combination of field studies, laboratory experiments, mathematical modelling and genetic analyses, were employed in an attempt to answer questions of relevance to the future conservation management of this species. Modelling smooth snake occupancy of remnant heathland patches using an information-theoretic approach showed patch size and the percentage of grassland in surrounding matrix habitats to be the primary determinants of smooth snake presence. Field-based studies based on 27 arrays of artificial refugia showed the size of trees and prey abundance to be important in determining mean smooth snake capture rates at occupied sites. Eight previously described microsatellite markers were used to complete the first assessments of the genetic population structure of C. austriaca at two spatial scales. Initial fine-scale analysis of structuring based on 11 sampling localities within a heathland/coniferous forest mosaic found significant population structuring as a result of isolation-by-distance effects, in addition to evidence of male-biased dispersal. At the wider scale, analysis of seven distinct populations across Dorset also found small but significant differences in genetic diversity. The observed patterns were not consistent with isolation-by-distance effects, nor was there any evidence of them being a result of habitat patch size or isolation. Phylogenetic analysis of the coarse-scale microsatellite data showed some evidence of population clustering based on their geographic locality in relation to the historical extent of Dorset’s heathland, suggesting they represent distinct management units. The reproductive ecology of C. austriaca was also examined using a combination of field data and microsatellite analysis. In contrast to continental populations, there was no relationship between female body size and litter size. However, there was a negative relationship between relative clutch mass and female body size, suggesting that there may be a trade-off between female survival and reproductive output. Microsatellite based genotyping of neonates from 16 litters born in the laboratory provides the first evidence of multiple paternity occurring in C. austriaca.
Full-text available
Green tree python (Morelia viridis, Schlegel 1872) is a highly sought-after Indonesian/Papuan NG/Australian species in terms of the international trade in reptile pets. As the trade in wild animals is mostly prohibited nowadays, captive breeding supplies the international pet trade. There is evidence that captive breeding might be used as a cover for specimen’s illegally sourced from the wild, as there are very few possibilities of distinguishing wild from captive-bred animals. These rely on invasive sampling (cutting off the end of the tail in order to obtain a sample of blood/muscle/bone tissues) or presence of ecto- and/or endoparasites (method overcome by breeders housing animals in semi wild conditions). Therefore, we examined the possibility of using stable isotope analysis for determining: either the place of origin or diet as a means of defining whether they are captive bred or illegally sourced from the wild. We also review the use of non-invasive samples of shed (moulted) skins. We conclude that shed skins that are currently not used for identifying the source of green tree python could be used as forensic evidence, subject to the development of a viable method.
Populations of blue-tongue skinks inhabiting eastern Wallacea and New Guinea are traditionally assigned to either Tiliqua gigas, which is endemic to this area, or to Tiliqua scincoides that extends its range from the Australian continent. Despite a wide morphological variation among local populations, genetic data from non-Australian populations were almost absent. We examined 128 specimens and sequenced mitochondrial ND4, 12rRNA and nuclear cmos genes. A phylogenetic analysis revealed the presence of two main clades corresponding to species Tiliqua scincoides and Tiliqua gigas. We provide the first report of Tiliqua scincoides from the Aru Island and confirm that it is genetically related to the Tanimbar populations reported as Tiliqua scincoides chimaerea. The other samples belonged to Tiliqua gigas, which also shows a distinct phylogenetic structure congruent with the geographic origin of the samples. The main split conforms to the north-south pattern of genetic variation, which was also reported in other animal species in New Guinea. (1) Samples from the northern coast of Irian Jaya, Seram, and Kai Islands belong to a distinct clade, which further splits supporting the recognition of the Tiliqua gigas keyensis subspecies. (2) Samples from the Bird’s Head and southern coast of the Irian Jaya and Halmahera Island form the other clade within Tiliqua gigas, in which the Halmahera samples formed a shallow, but clearly distinct branch. The haplotype network analysis of mitochondrial ND4 gene in Tiliqua gigas samples suggests a strong differentiation among major population groups.
The green python (Morelia viridis) is an iconic snake species highly sought after in the pet trade and is the target of illegal collection. Despite their popularity, some important ecological attributes of green pythons remain unknown, making their effective conservation management difficult. Detection-only surveys were conducted throughout the potential range of the green python in Australia, and intensive mark-recapture surveys were conducted in the areas where there have been previous records. In total, 298 green pythons were located in the Iron, McIlwraith and Kawadji-Ngaachi Ranges of Cape York, distributed over an estimated area of 2289 km 2 , where they frequented rainforest habitats and adjacent vine thickets. They were not found in the Lockerbie Scrub or Jardine River Catchment, despite anecdotal records. Green python density was estimated to be 540 km-2 in the Iron Range and 200 km-2 in the McIlwraith Range, where the percentages of adults captured were 56% and 83%, respectively. The differences between abundance and population demographics in the Iron and McIlwraith ranges may be due to differences in prey abundance and the impacts of collection. The results of this study provide baseline data to conservation managers and policy makers for the future conservation management of this species in Australia.
Full-text available
The GABA-A receptor agonist midazolam is a compound widely used as a tranquilizer and sedative in mammals and reptiles. It is already known that this benzodiazepine produces small to intermediate heart rate (HR) alterations in mammals, however, its influence on reptiles’ HR remains unexplored. Thus, the present study sought to verify the effects of midazolam on HR and cardiac modulation in the snake Python molurus. To do so, the snakes’ HR, cardiac autonomic tones, and HR variability were evaluated during four different experimental stages. The first stage consisted on the data acquisition of animals under untreated conditions, in which were then administered atropine (2.5 mg kg-1; intraperitoneal), followed later by propranolol (3.5 mg kg-1; intraperitoneal) (cardiac double autonomic blockade). The second stage focused on the data acquisition of animals under midazolam effect (1.0 mg kg-1; intramuscular), which passed through the same autonomic blockade protocol of the first stage. The third and fourth stages consisted of the same protocol of stages one and two, respectively, with the exception that atropine and propranolol injections were reversed. By comparing the HR of animals that received midazolam (second and fourth stages) with those that did not (first and third stages), it could be observed that this benzodiazepine reduced the snakes’ HR by ~60%. The calculated autonomic tones showed that such cardiac depression was elicited by an ~80% decrease in cardiac adrenergic tone and an ~620% increase in cardiac cholinergic tone – a finding that was further supported by the results of HR variability analysis.
Full-text available
A complete guide to the snakes of the island of New Guinea (with particular emphasis on the eastern half constituting the sovereign state of Papua New Guinea) and the islands to the east, e.g. Bismarck, Admiralty, d'Entrecasteaux, Louisiade and North Solomons Archipelagoes. Although out of print and also now out of date, this is still the definitive and much sought guide to the snake fauna of this region. It includes a section of snakebite first aid and hospital treatment by Drs David A Warrell and David G Lalloo. A 2nd edition is currently in preparation. A pdf of the 1st edition is available to researchers who contact the author personally.
Full-text available
This study describes genetical differences between three morphologically similar species of Antechinus in south-eastern Australia, and uses the climatic model BIOCLIM to clarify their expected geographical distributions. Allozyme electrophoresis revealed Nei’s distances of >0.2 between A. flavipes and A. stuartii and A. flavipes and A. agilis, the latter a newly recognised species in south-eastern Australia. Fixed allele differences were determined in five proteins between A. stuartii and A. flavipes from an area of sympatry in northern New South Wales, confirming their genetic distinctness. A smaller distance (0.08) separated A. stuartii from A. agilis, but fixed allele differences in albumin and mannose phosphate isomerase distinguished these species clearly. Locality records for the three species were compiled from the electrophoretic results, museum specimen records and published data, and used to generate expected distributions for each species. A. flavipes is predicted to occur primarily in warm, inland areas of south- eastern Australia with a mean annual rainfall of 785 mm, but to occur along the coast in South Australia and southern Queensland. In contrast, the distributions of A. stuartii and A. agilis are predicted to be broadly coastal, with the former occurring in northern New South Wales and southern Queensland in areas with high mean annual rainfall (1430 mm) and temperature (16.0°C), the latter in southern New South Wales and Victoria in cooler areas (11.8°C annual mean) with intermediate rainfall (1071 mm). Sympatry appears to be limited between A. flavipes and its two congeners; A. stuartii and A. agilis are predicted to be parapatric with only two small areas of overlap being evident.
Full-text available
The Australian distribution of the grass owl (Tyto capensis) is poorly understood. It has been proposed that there are two centres of distribution: a resident coastal population in the north-east, and a less stable inland population from which there is Australia-wide dispersal when good seasons are followed by deteriorating conditions. We analysed records of the grass owl and modelled its bioclimatic profile and distribution, which was typically subtropical, warm to hot humid with no dry season or a dry winter. This predicted a north-east sub-coastal to coastal, permanently occupied, core distribution for the owl. We found no evidence for a permanent or isolated inland population, nor for inland populations being the sole source of dispersers, as has been suggested previously. Most inland and northern records were made in the 1970s when grass owls colonised the arid inland, the Kimberley and the far north of the country in association with events leading to the flooding of Lake Eyre. The data suggest that grass owls disperse from their core range after exceptionally good breeding seasons to areas made temporarily favourable by exceptional rainfall or flooding, only to disperse again when conditions become drier. These dispersal events are not tied uniquely to outbreaks of the long-haired rat (Rattus villosissimus), but to a variety of terrestrial prey with dynamic life histories driven by rainfall.
The distribution of the Australian mainland endemic subspecies of the eclectus parrot, Eclectus roratus macgillivrayi, is currently confined to the lowland rainforests of the Iron–McIlwraith Ranges of eastern Cape York Peninsula. Females breed in large hollows in emergent rainforest trees that are readily visible from above. Aerial surveys were used to sample 58% of the rainforest (454 km2) of the Iron Range region to estimate the density of these nest trees. Corrections for overcounting bias (not all observed emergent trees were active nest trees) and undercounting bias (not all active nest trees were visible from the air) were made by ground-truthing over 70 trees. The tree count data were treated in two different ways, producing estimates of 417 (s.e. = 25) and 462 (s.e. = 31) nest trees for the Iron Range region. Long-term observational data on the number of eclectus parrots associated with each nest tree were used to estimate the population size of eclectus parrots at Iron Range: 538–596 breeding females, and 1059–1173 males. These results have three implications. First, this relatively low population estimate suggests that the Australian subspecies of eclectus parrots should be considered vulnerable to habitat loss or perturbation, especially in light of their complex social system, male-biased adult sex ratio, low breeding success and high variance in reproductive success among females. Second, the low density of nest trees suggests that eclectus parrots are absent from the rainforests of Lockerbie Scrub and the Jardine dunefields because these areas are too small. Finally, if eclectus parrots persisted in the Iron–McIlwraith region during the rainforest contractions of Pleistocene glacial maxima (e.g. 14 000–17 000 years ago), the refugium in this region must have been fairly substantial in order to support a viable population – probably larger than previously assumed.
The occurrence of Leadbeater's possum, Gymnobelideus leadbeateri (McCoy) is restricted to a narrow set of climatic conditions which are confined to the Central Highlands of Victoria. This area encompasses the entire known range of the species. Bioclimatic analyses using the computer program BIOCLIM allowed the prediction and definition of the potential limits of the distribution of G. leadbeateri. This was used as a basis for selecting sites for surveys for G. leadbeateri outside the Victorian Central Highlands. G. leadbeateri was not detected in field surveys of these sites. The bioclimatic profile of G. leadbeateri was closely related to the bioclimatic profiles of several species of trees, including Eucalyptus regnans, Acacia obliquinervia and Acacia frigiscens. However, none of the plants that were examined in the study could be used as an indicator of the presence of G. leadbeateri. On the basis of its bioclimatic profile, the distribution of G. leadbeateri is likely to undergo a considerable contraction as a result of climatic changes associated with the 'greenhouse effect'. The importance of this result and other findings from the study are discussed in terms of the value of BIOCLIM and bioclimatic analyses in wildlife conservation.
A bioclimatic analysis was undertaken for four commercially valuable wood production trees from south-eastern Australia: - Eucalyptus regnans (mountain ash, swamp gum or stringy gum); Eucalyptus delegatensis (alpine ash, white-top stringybark or gum-top stringybark); Eucalyptus nitens (shining gum or silvertop) and Eucalyptus fastigata (brown barrel or cut-tail). A reference resource for forest managers is presented that includes summary information on a range of bioclimatic conditions which characterise the extant distribution of each tree species. These data were used to map the predicted potential bioclimatic domains of the four species. In each case, areas were identified with marginally suitable bioclimatic conditions that extended beyond the present known distributions of the species, while the core bioclimatic domain of each species often conformed more closely to its known range. The results of our investigation highlighted some clear differences in the bioclimatic regimes occupied by the four targeted taxa. The most marked contrasts were between E.fastigata and the other three species, particularly with respect to seasonal differences in rainfall patterns and annual variations in temperature regimes. E.fastigata typically occurs on sites characterised by lower cool season rainfall compared with the other tree species. Conversely, warm season rainfall was higher for E.fastigata although the differences between the four taxa were not large. Sites supporting E.fastigata experienced a greater range in temperature than those inhabited by the three other species of trees. The approach used in this study has potential value in a range of aspects of forest management including modelling site productivity, designing vegetation surveys, identifying areas to establish plantations, and determining provenances for plantation use.