Ecology, 89(5), 2008, pp. 1201–1207
? 2008 by the Ecological Society of America
ASSOCIATIONS OF VARIABLE COLORATION WITH NICHE BREADTH
AND CONSERVATION STATUS AMONG AUSTRALIAN REPTILES
ANDERS FORSMAN1AND VIKTOR A˚BERG
School of Pure and Applied Natural Sciences, University of Kalmar, SE-391 82 Kalmar, Sweden
quences of color polymorphisms. Previous endeavors have aimed at identifying conditions
that promote the evolution and maintenance within populations of alternative variants. But
the polymorphic condition may also influence important population processes. We consider
the prediction that populations that consist of alternative ‘‘ecomorphs’’ with coadapted gene
complexes will utilize more diverse resources and display higher rates of colonization success,
population persistence, and range expansions, while being less vulnerable to range
contractions and extinctions, compared with monomorphic populations. We perform pairwise
comparative analyses based on information for 323 species of Australian lizards and snakes.
We find that species with variable color patterns have larger ranges, utilize a greater diversity
of habitat types, and are underrepresented among species currently listed as threatened. These
results are consistent with the proposition that the co-occurrence of multiple color variants
may promote the ecological success of populations and species, but there are also alternative
We evaluate predictions concerning the evolutionary and ecological conse-
environmental variability; evolution; population persistence; range expansion; reptiles.
adaptive polymorphism; Australia; complex phenotypes; conservation; ecological diversity;
Animal color patterns have a long tradition in
generating and testing theories central to ecology and
evolutionary biology (Darwin 1859, Cott 1940). Color
pattern may influence performance and fitness of
individuals, for instance by serving as a visual signal in
mate choice (Andersson 1994) or by providing protec-
tion from predators (Ruxton et al. 2004). In ectothermic
animals, such as reptiles and insects, which rely on
external radiant energy for body temperature regulation,
coloration may influence heating rates and body
temperature, body temperature may influence energy
budgets and locomotion, and locomotion may influence
the capacity to escape predation or to compete with
conspecifics for access to mates (Peterson et al. 1993).
Color pattern may also influence fitness interactively
with other traits. Natural selection seems to favor
certain combinations of coloration and behavioral,
morphological, and physiological traits, at the expense
of other trait–value combinations, rather than operating
only on coloration per se (King 1987, Brodie 1992,
Forsman 1995, Forsman and Appelqvist 1998). Such
correlational selection will promote the evolution of
genetic rearrangements that result in physical integra-
tion and associations between different traits (Jackson et
al. 1976, Brodie 1989, Endler 1995, Forsman et al. 2002).
For reasons just outlined, animal color patterns offer
ideal model systems for investigations into ecological
and evolutionary causes and consequences of polymor-
phisms. Our aim here is to evaluate the proposition that
variable coloration may influence population processes
(Dobzhansky 1951). We address predictions from a
conceptual model of the ecological and evolutionary
consequences of color pattern polymorphism (Forsman
et al. 2008). The model begins with an evolutionary
branching event from mono- to polymorphic condition,
and posits that color pattern will become associated
genetically and developmentally with other kinds of
traits through the action of correlational selection. At
this stage, the polymorphism lends itself admirably to
the adaptive landscape metaphor (Arnold et al. 2001),
with the fitness peaks representing alternative trait–value
combinations displayed by different morphs. It has
therefore been proposed that a population that consists
of two or more alternative ecomorphs that occupy
different niches should utilize a greater diversity of
resources and enjoy enhanced population stability and
persistence, colonization success, range expansions,
evolutionary potential, and speciation. Conversely,
polymorphic populations are predicted to be less
vulnerable to environmental change and at lower risk
of range contractions and extinctions, compared with
non-variable populations (Forsman et al. 2008). Here we
report on a comparative analysis of reptiles which was
carried out to test for ecological associations of variable
Based on literature data for lizards and snakes of
Australia, we examine whether species with variable
Manuscript received 10 October 2007; revised 13 December
2007; accepted 19 December 2007. Corresponding Editor: T. J.
color pattern have larger range sizes, utilize a larger
number of habitat types, and if they are underrepre-
sented among species that are considered to be
endangered, compared with species with non-variable
coloration. We choose to evaluate these particular
predictions for the following reasons. That the coexis-
tence of alternative color- or ecomorphs will result in
utilization by the population as a whole of a greater
diversity of resources (i.e., broader niche) represents the
very mechanism envisaged to positively influence the
dynamics and persistence of polymorphic populations
(Forsman et al. 2008). Range size may be regarded a
composite measure of the performance of populations
because it is influenced by colonization success, ability to
cope with environmental change, and population
persistence, all of which have been predicted to be
positively influenced by color polymorphism (Forsman
et al. 2008). Finally, by testing for an association with
conservation status we evaluate the proposition that
color pattern polymorphism will reduce the probability
of population declines and extinction.
MATERIAL AND METHODS
Reptiles lend themselves admirably to comparative
investigations of ecological correlates of color-pattern
variation. Chromatic polymorphism has evolved and
been lost independently in a series of distantly related
lineages of lizards and snakes in various parts of the
world (Greer 1989, Shine 1991, Wolf and Werner 1994).
Because reptiles are ectothermic, coloration may not
only influence survival and reproduction directly;
indirect effects of coloration on fitness mediated via
body temperature may also be important (Peterson et al.
1993). Associations of coloration with morphology,
behavior, and life history have been documented in
several species of lizards and snakes (Jackson et al. 1976,
King 1987, Brodie 1989, Forsman 1995, Shine et al.
1998), suggesting that the sympatric occurrence of two
or more alternative ecomorphs is relatively common in
this group of animals. Extensive information on the
ecology and conservation status of lizards and snakes
emanating from the work of dedicated herpetologists
has been compiled in readily available sources.
Sources of information
Data on color patterns were obtained from Reptiles
and Amphibians of Australia (Cogger 1992). We classified
species as being either variable or non-variable with
regard to color pattern based on descriptions in the field
guides. Theoreticians have been concerned with identi-
fying conditions that may promote the long-term
maintenance in populations of two or more discrete
phenotypes (Ford 1945, Hedrick 1986, Leimar 2005).
But there is no reason to expect polymorphisms to
persist in evolutionary time; they may be transient and
change from a polymorphic to a monomorphic state or
constitute intermediate stages in the evolution of
geographic variation, sexual dimorphism, or reproduc-
tive isolation and speciation (Forsman et al. 2008). We
therefore made no distinction between species described
as exhibiting true within-population polymorphism and
species displaying geographic variation in color pattern;
both were classified as variable coloration. We classified
species that were sexually dichromatic as having variable
coloration only if it was evident from the description
that polymorphism or geographic variation occurred
within one or both sexes; otherwise they were considered
non-variable. Similarly, species exhibiting ontogenetic
change in coloration were classified as variable only if
polymorphism or geographic variation occurred among
Comparative analyses should ideally be performed
based on highly resolved phylogenies that contain
information on branch lengths, such that the direction-
ality and order of state transitions can be estimated
(Pagel and Meade 2006). Unfortunately, such phyloge-
nies are available only for a subset of Australian reptiles.
To evaluate the hypothesis that evolutionary transitions
between non-variable and variable coloration have been
associated with changes in the ecological variables under
investigation, we therefore used a modified version of
the pairwise comparative method (Møller and Birkhead
1992, Maddison 2000). The pairwise approach is
suitable for situations where a phylogenetic hypothesis
of the species under investigation is unavailable or
poorly resolved, and it does not rely heavily on
assumptions about ancestral states, branch lengths, or
elaborate models of evolution. We relied on the
taxonomy used in Cogger (1992). Because the taxonomy
and phylogenetic affinities of reptiles is undergoing
constant revision, our database does not include all
currently recognized species; compare Cogger (1992)
and Cogger (2000). The pairwise comparative approach
is based on comparisons between matched pairs of
congeneric species that differ with regard to the trait of
interest (variable vs. non-variable coloration) but are
otherwise similar (Møller and Birkhead 1992). It may be
difficult, however, to decide which species constitute the
most appropriate match, particularly for groups that
contain poorly studied species. To circumvent this
problem, we include in our analysis data for all species
that we could reliably classify as either variable or non-
Because our approach is based on matched pair
comparisons, we include in our analyses only those
genera that contain species classified as having variable
coloration and species in the same genera that were
classified as non-variable. Those genera in which all
species were either variable or non-variable were
omitted. A list of species included in our data set and
their classification with regard to variable or non-
variable color pattern is available in the Appendix.
Do species with variable color patterns have larger ranges?
To determine range sizes for the species in our data set
we first digitized the distribution maps in the field guide
ANDERS FORSMAN AND VIKTOR A˚BERG 1202Ecology, Vol. 89, No. 5
(Cogger 1992) using a scanner and then performed
measurements on the digitized images using an image
process and analysis program (Scion Image, Release
Beta 4.0.2.; Scion Corporation, Frederick, Maryland,
USA). We calibrated the program based on a known
distance between two fixed locations on the map such
that range sizes were expressed in square kilometers
To avoid pseudo-replication we compute for each
genus the mean (if more than one species) range size for
species with variable coloration and the mean range size
for species with non-variable coloration. We then
perform the statistical analysis based on paired com-
parisons of generic means, using the MANOVA method
for repeated measures analysis of variance, and treating
the mean range size of variable and non-variable species
within each genus as a repeated measure (Forsman et al.
2002). Because range size may differ between lizards and
snakes, we included order (Sauria or Serpentes) as an
independent class variable in the model. We implement-
ed the analyses using procedure GLM in SAS, and
sigma (type III) sums of squares were used to test
hypotheses (SAS 2004). A major benefit of this approach
is that it enables us to test for a difference in range size
between species with variable and non-variable colora-
tion in a single analysis (rather than performing two
separate analyses) while controlling statistically for
overall differences in range size between lizards vs.
Do species with variable color patterns utilize
a larger number of habitat types?
To test for an association between coloration and
habitat use we counted the number of habitat types used
by each species, following Cogger et al. (1983). For each
species, Cogger et al. (1983) listed each habitat type in
which individuals have been observed or collected, and
the habitat types were taken from a standardized list.
Examples of habitat types include closed forest, desert,
hummock grassland, open heath, tall shrubland, and
woodland (in total ;25 types). We performed an
analysis based on paired comparisons of generic means,
as described before.
Are species with variable color patterns
less vulnerable to extinction?
We obtained information on conservation status of
the species from the online edition of the Action Plan for
Australian Reptiles (Cogger et al. 1993; data available
online)2This listing is based on an extensive series of
surveys done by the herpetological community and
information from state and territorial agencies; final
conservation status was estimated by a panel of
herpetologists using the ranking system of Millsap et
al. (1990). To ensure that our data set was based on
recent identification and listing of species as threatened,
we consulted in November 2006 the online version of the
Environment Protection and Biodiversity Conservation
Act List of Threatened Fauna, which presents a
reconsidered status of the initial list as information has
been updated and made available for assessment (data
In keeping with the pairwise approach, we analyzed
data using a repeated measures approach regarding
conservation status of species within the same genus but
classified as having either variable or non-variable color
pattern as a repeated measure using logit-model analysis
implemented with procedure GENMOD (SAS 2004). To
control for a possible difference in conservation status
between lizards and snakes, we included order as an
explanatory variable in the model. We used the Wald
statistic to assess the statistical significance of explana-
tory variables included in the model (Collett 1991). A
common variance inflation factor for all observations
(computed as the square root of the Pearson’s chi-square
divided by the degrees of freedom) was used to account
for problems with overdispersion.
Variable coloration is associated with larger range sizes
We obtained data for 323 species distributed among
Australian lizards (275 species representing 14 genera)
and Australian snakes (48 species, 9 genera). Summary
statistics of the pairwise comparisons of range size based
on within-generic means is presented in Table 1. Overall,
our analyses uncovered large and statistically significant
differences in range size between species with variable
and non-variable color patterns (Table 1, Fig. 1). Species
with variable coloration have a mean range size of 2.83
106km2(least squares means as obtained from a two-
factor ANOVA with color and order as independent
class variables), which is more than two times larger
than the corresponding value for species with non-
variable coloration, 1.2 3 106km2. The difference in
range size between species with variable and non-
variable color patterns is highly significant (MANOVA
method for repeated measures analysis of variance,
effect of color pattern: F1,21¼ 8.52, P ¼ 0.0082; Fig. 1).
Our results revealed no overall difference in range size
between snakes and lizards (F1,21¼ 0.25, P ¼ 0.62; Fig.
1), and the association of variable coloration with larger
range size does not differ between lizards and snakes (as
evidenced by the nonsignificant color by order interac-
tion effect: F1,21¼ 0.24, P ¼ 0.63). Our findings are
robust to choice of statistical test; we arrive at the
conclusion that range size is associated with variable vs.
non-variable coloration also if we analyze our data using
a nonparametric Wilcoxon matched-pair signed ranks
test (range size was largest in the species with variable
May 2008 1203ASSOCIATIONS OF COLOR POLYMORPHISM
coloration in 11 of the 14 genera of Australian lizards
and in eight out of nine genera within the Australian
snakes; Table 1, both P , 0.05).
Variable coloration is associated with utilization
of a greater diversity of habitat types
The Zoological Record of Australian Amphibians and
Reptiles (Cogger et al. 1983) scored the number of
habitat types used for 268 of the species included in our
data set of lizards and snakes with variable or non-
variable color patterns. A pairwise comparison of the
mean number of habitat types used by variable and non-
variable species is presented in Table 1.
We find a large, statistically significant difference in
number of habitat types used by Australian species of
reptiles with variable vs. non-variable color patterns
(Table 1, Fig. 1). Species with variable coloration
utilized on average 6.9 (95% confidence interval, 5.7–
8.0) different habitat types (least squares means ob-
tained from a two-factor ANOVA with color and order
as independent class variables), which is .160% the
number of habitat types used by species with non-
variable coloration (least squares mean ¼ 4.1 habitat
types; 95% confidence interval, 2.9–5.3). The larger
number of habitat types exploited by species with
variable color patterns is significant (MANOVA method
for repeated measures analysis of variance, effect of
color pattern: F1,21¼ 16.45, P , 0.0006; Fig. 2) and
equally evident in lizards and in snakes (as evidenced by
the nonsignificant color by order interaction effect: F1,21
¼0.02, P¼0.89). Importantly, the utilization of a larger
number of habitat types by species with variable as
compared with non-variable coloration remains signif-
icant when we adjust statistically for the confounding
effect of range size (ANCOVA with coloration and
order, lizards vs. snakes, treated as independent class
variables and range size as a covariate; effect of
coloration, F1,42¼ 4.17, P , 0.05; effect of order, F1,42
¼ 1.16, P ¼ 0.29; effect of range size, F1,42¼ 85.2, P ,
0.0001). We arrive at the conclusion that habitat use is
associated with variable vs. non-variable coloration also
if we analyze our data using nonparametric tests (habitat
use was more diverse in species with variable coloration
in 11 of the 14 genera of Australian lizards and in eight
out of nine genera within each of Australian snakes, P ,
0.05 for both; Table 1).
Species with variable coloration are estimated
as less vulnerable to extinction
In The Action Plan for Australian Reptiles (Cogger et
al. 1993) and the Environment Protection and Biodi-
versity Conservation (EPBC) Act List of Threatened
variable or non-variable color patterns.
Pairwise comparison of range size, habitat use, and conservation status for Australian species of lizards and snakes with
Coloration variable Coloration non-variable
(total no. species)
Notes: Values for range size are means computed across N species within each genus. Values for number of habitat types used are
means computed across N species within each genus. Species were classified as threatened based on information in Cogger at al.
(1993) and the EPBC Act List of Threatened Fauna. Genera consisting of only variable or only non-variable species are not
ANDERS FORSMAN AND VIKTOR A˚BERG 1204Ecology, Vol. 89, No. 5
Fauna, 50 of the 323 species of Australian lizards (275
species) and snakes (48 species) included in our data set
are listed as either ‘‘endangered,’’ ‘‘vulnerable,’’ or ‘‘rare
or insufficiently known.’’ The distribution of such
threatened species among lizards and snakes with
variable vs. non-variable color patterns is shown in
Table 1. Only one of 12 snake species (8%) with variable
coloration included in our data set is listed as threatened
(the death adder Acanthophis antarcticus), compared
with almost 15% of the species with non-variable
coloration (Fig. 2). None of the lizards with variable
color pattern were classified as threatened, whereas 17%
of lizards with non-variable color patterns fell into this
category. Our data thus suggest that compared with
species possessing non-variable color patterns, species
with variable color patterns are less likely to be
threatened (repeated measures logit-model analysis
using GENMOD with coloration, species with variable
vs. non-variable coloration in the same genus, as a
repeated measure, and order, lizards or snakes, as a
factor; effect of color pattern variability: v2¼ 4.42, df ¼
1, P , 0.05; Fig. 2). There is no difference between the
lizards and snakes included in our data set with regard
to endangerment (v2¼0.03, df¼1, P¼0.85). The lower
endangerment of species with variable color patterns
remains significant when we adjust statistically for the
potentially confounding effect of range size (effect of
coloration, v2¼6.62, df¼1, P , 0.01; effect of order, v2
¼0.02, df¼1, P¼0.88; effect of range size, v2¼0.15, df
¼ 1, P ¼ 0.69.
Our comparative analyses of Australian lizards and
snakes showed species with variable color pattern to
have larger range sizes, utilize a larger number of habitat
types, and to be underrepresented among species listed
as threatened. Species with variable color patterns have
a mean range size almost three times larger than their
counterparts with non-variable coloration (Fig. 1). This
finding concurs with the results of earlier comparative
studies on range size and color variation in birds (Fowlie
and Kru ¨ ger 2003, Roulin and Wink 2004) and insects
(Brisson et al. 2006).
Dobzhansky (1951) proposed that polymorphism will
enable a species to increase the efficiency of the
exploitation of the resources of the environment and
can be regarded as an adaptation to varied circumstanc-
es and allow larger ranges. He reported that two widely
ranging Drosophila species (D. willistoni and D. paulis-
torum) are more polymorphic (with regard to chromo-
some inversions) compared with two other species (D.
equinoxialis and D. tropicalis) of narrow range. Cain and
(bottom panel) number of habitat types utilized for Australian
species of lizards and snakes with variable and non-variable
coloration. This figure is based on generic means of data
provided in Table 1. The solid and dotted lines within the box
indicate median and mean, the boundaries of the box indicate
25th and 75th percentiles, whiskers below and above indicate
10th and 90th percentiles, and dots indicate outlying observa-
tions. Numbers in parentheses below indicate number of
Comparison of (top panel) distributional range and
lizards and snakes in our data set with variable and non-
variable color pattern classified as threatened (endangered,
vulnerable, or rare). Information on conservation status was
obtained from the action plan for Australian reptiles (Cogger et
al. 1993) and the EPBC Act List of Threatened Fauna.
Numbers above bars indicate number of threatened species.
Numbers in parentheses below indicate number of species
included in the data set.
Comparison of percentage of species of Australian
May 2008 1205ASSOCIATIONS OF COLOR POLYMORPHISM
Sheppard (1954) asked for evidence for increased
efficiency of resource exploitation in polymorphic
populations and pointed out that it might equally be
that the wide range has been a factor favoring
polymorphism as that polymorphism should be respon-
sible for the wider range. For groups of organisms for
which a phylogeny with information on branch lengths
is available it may be possible to determine whether it is
more likely that polymorphism has promoted range
expansions or if variable color patterns have evolved
more readily in species that inhabit larger areas (Pagel
and Meade 2006), but this is not an option in our case.
The pattern of variability correlating with range size
may in part be a consequence of taxonomic practices
and incomplete knowledge of phylogenetic affinities.
Nominal species with large ranges are more likely to be
variable across the range. If such variable species
represent imperfectly resolved groups that have subse-
quently been redescribed as distinct, more range limited,
and less variable taxa, then this may have contributed to
monomorphism being associated with smaller ranges.
Unfortunately, we are unable to evaluate or control for
this potential artifact.
Our analyses uncovered that species with variable
color patterns use a greater diversity of habitat types
(Fig. 1). This is not an epiphenomenon caused by species
with variable coloration having larger ranges; the
utilization of a larger number of habitat types by species
with variable coloration is evident also when we control
statistically for variation in range size. This result thus
corroborates one of the underlying mechanisms (i.e.,
broader niches) envisaged by Forsman et al. (2008) to
positively influence the dynamics and persistence of
Our results point to some interesting implications for
conservation biology. That we found species of lizards
and snakes with variable coloration to be underrepre-
sented among species listed as threatened in The Action
Plan for Australian Reptiles by Cogger et al. (1993) and
the EPBC Act List of Threatened Fauna is of particular
interest in this respect. Previous analyses of a diverse
array of taxa have identified several ecological predic-
tors of vulnerability, including small local population
sizes, small geographic-range size, large body size,
mating system, low reproductive investment, specialized
niche requirements, insular endemism, and the tendency
to form aggregations (Gaston and Kunin 1997, Reed
and Shine 2002).
The association of color pattern with conservation
status among the Australian reptiles included in our
study was strong (Fig. 2). None of the lizards with
variable color pattern included in our analyses were
classified as threatened, whereas .15% of lizards with
non-variable color patterns fell into this category.
Among snakes, the only species with variable coloration
included in our data set that is listed as threatened is the
death adder Acanthophis antarcticus. This species is
distributed throughout eastern and southern Australia,
but was classified by Cogger et al. (1993) as ‘‘rare or
insufficiently known.’’ We do not know why this species
has become rare. In their investigation of ecological
correlates of conservation status among Australian
Elapid snakes, Reed and Shine (2002) found that
threatened species were more likely to be ambush
predators than expected by chance. These authors
suggest that ambushing (sit-and-wait foragers) snakes
require specific ground cover, such as thick leaf litter in
the case of Acanthophis, and that broad scale clearing
and frequent fire events have reduced ground cover over
much of semiarid Australia, and thus contributed to the
decline of ambushing snakes. Alternatively, the low rate
of food intake associated with an ambushing foraging
mode may make species that rely on this strategy
particularly susceptible to prey scarcity and infrequent
reproduction (Reed and Shine 2002).
Specialists are unlikely to be able to use alternative
resources in response to changing environmental condi-
tions (Gaston and Kunin 1997). Reliance on a small range
of habitat or food types may therefore be an important
contributor to endangerment. Our finding that Australian
lizards and snakes with variable coloration use more
diverse habitat types and were less frequently classified as
threatened, compared with their congenerics with non-
variable coloration, accords well with this general notion.
Habitat specialization also has been found to explain a
significant amountofthe variationinratesofextinction of
reptile species from islands in the Mediterranean (Foufo-
poulos and Ives 1999).
In summary, based on our comparative analyses of
lizards and snakes, we find that species with variable
color patterns have larger ranges, utilize a greater
diversity of habitat types, and are underrepresented
among species currently listed as threatened, compared
with species with non-variable color patterns. All these
associations were strong and conform to the proposition
(Forsman et al. 2008) that the coexistence of two or
more color morphs may positively influence important
population processes. Because we were unable to
determine causal relationships, however, our conclu-
sions must be tentative. Future comparative approaches
should focus on taxonomic groups for which molecular
phylogenies with information on branch lengths exists
such that the directionality and sequence of state
transitions can be determined, and also address the
challenge of examining the influence of color variability
on changes and processes directly, rather than estab-
lishing associations with states. Our present findings
suggest that such endeavors are likely to be worthwhile.
We are grateful to the many dedicated herpetologists who
have produced the data, and to Red List Database for
providing access to data. Many people have contributed with
inspiring discussions and/or provided useful suggestions and
comments on the manuscript and data set, including but not
limited to R. Shine, J. Ahnesjo ¨ , A. P. Møller, and the people at
‘‘the Shine lab’’ at The University of Sydney who provided a
friendly atmosphere during the stage of manuscript prepara-
ANDERS FORSMAN AND VIKTOR A˚BERG 1206 Ecology, Vol. 89, No. 5
tion. The Swedish Science Council (grant to A. Forsman) and
the University of Kalmar provided financial support.
Andersson, M. 1994. Sexual selection. Princeton University
Press, Princeton, New Jersey, USA.
Arnold, S. J., M. E. Pfrender, and A. G. Jones. 2001. The
adaptive landscape as a conceptual bridge between micro-
and macroevolution. Genetica 112–113:9–32.
Brisson, J. A., J. Wilder, and H. Hollocher. 2006. A
phylogenetic analysis of the Cardini group of Drosophila
with respect to changes in pigmentation. Evolution 60:1228–
Brodie, E. D. 1989. Genetic correlation between morphology
and antipredator behaviour in natural populations of the
garter snake Thamnophis ordinoides. Nature 342:542–543.
Brodie, E. D. 1992. Correlational selection for colour pattern
and antipredator behaviour in the garter snake Thamnophis
ordinoides. Evolution 46:1284–1298.
Cain, A. J., and P. M. Sheppard. 1954. The theory of adaptive
polymorphism. American Naturalist 88:321–326.
Cogger, H. G. 1992. Reptiles and amphibians of Australia.
Fifth edition. Reed Books, Sydney, Australia.
Cogger, H. G. 2000. Reptiles and amphibians of Australia.
Sixth edition. Reed New Holland, Sydney, Australia.
Cogger, H. G., E. E. Cameron, and H. M. Cogger. 1983.
Zoological catalogue of Australia. 1. Reptiles and amphib-
ians. Bureau of Flora and Fauna, Canberra, Australia.
Cogger, H. G., E. E. Cameron, R. A. Sadlier, and P. Eggler.
1993. The action plan for Australian reptiles. Australian
Nature Conservation Agency, Canberra, Australia.
Collett, D. 1991. Modelling binary data. Chapman and Hall,
Cott, H. B. 1940. Adaptive coloration in animals. Methuen and
Company, London, UK.
Darwin, C. 1859. On the origin of species by means of natural
selection. Murray, London, UK.
Dobzhansky, T. 1951. Genetics and the origin of species. Third
edition. Columbia University Press, New York, New York,
Endler, J. A. 1995. Multiple-trait coevolution and environmen-
tal gradients in guppies. Trends in Ecology and Evolution 10:
Ford, E. B. 1945. Polymorphism. Biological Review 20:73–88.
Forsman, A. 1995. Opposing fitness consequences of color
pattern in male and female snakes. Journal of Evolutionary
Forsman, A., J. Ahnesjo, S. Caesar, and M. Karlsson. 2008. A
model of ecological and evolutionary consequences of color
polymorphism. Ecology 89:34–40.
Forsman, A., and S. Appelqvist. 1998. Visual predators impose
correlational selection on prey color pattern and behavior.
Behavioral Ecology 9:409–413.
Forsman, A., K. Ringblom, E. Civantos, and J. Ahnesjo ¨ . 2002.
Coevolution of color pattern and thermoregulatory behavior
in polymorphic pygmy grasshoppers Tetrix undulata. Evolu-
Foufopoulos, J., and A. R. Ives. 1999. Reptile extinctions on
landbridge islands: life-history attributes and vulnerability to
extinction. American Naturalist 153:1–25.
Fowlie, M. K., and O. Kru ¨ ger. 2003. The evolution of plumage
polymorphism in birds of prey and owls: the apostatic
selection hypothesis revisited. Journal of Evolutionary
Gaston, K. J., and W. E. Kunin. 1997. Rare–common
differences: an overview. Pages 12–29 in W. E. Kunin and
K. J. Gaston, editors. The biology of rarity. Chapman and
Hall, London, UK.
Greer, A. E. 1989. The biology and evolution of Australian
lizards. Surrey Beatty and Sons, Chipping Norton, New
South Wales, Australia.
Hedrick, P. W. 1986. Genetic polymorphism in heterogeneous
environments: a decade later. Annual Review of Ecology and
Jackson, J. F., W. Ingram III, and H. W. Campbell. 1976. The
dorsal pigmentation pattern of snakes as an antipredator
strategy: a multivariate approach. American Naturalist 110:
King, R. B. 1987. Colour pattern polymorphism in the Lake
Erie water snake, Nerodia sipedon insularum. Evolution 41:
Leimar, O. 2005. The evolution of phenotypic polymorphism:
randomized strategies versus evolutionary branching. Amer-
ican Naturalist 165:669–681.
Maddison, W. P. 2000. Testing character correlation using
pairwise comparisons on a phylogeny. Journal of Theoretical
Millsap, B. A., J. A. Gore, D. E. Runde, and S. I. Cerulean.
1990. Setting priorities for the conservation of fish and
wildlife species in Florida. Wildlife Monographs 111:1–57.
Møller, A. P., and T. R. Birkhead. 1992. A pairwise
comparative method as illustrated by copulation frequency
in birds. American Naturalist 139:644–656.
Pagel, M., and A. Meade. 2006. Bayesian analysis of correlated
evolution of discrete characters by reversible-jump Markov
chain Monte Carlo. American Naturalist 167:808–825.
Peterson, C. R., A. R. Gibson, and M. E. Dorcas. 1993. Snake
thermal ecology: the causes and consequences of body-
temperature variation. Pages 241–314 in R. A. Seigel and
J. T. Collins, editors. Snakes—ecology and behavior.
McGraw Hill, New York, New York, USA.
Reed, R. N., and R. Shine. 2002. Lying in wait for extinction:
ecological correlates of conservation status among Australian
Elapid snakes. Conservation Biology 16:451–461.
Roulin, A., and M. Wink. 2004. Predator–prey relationships
and the evolution of colour polymorphism: a comparative
analysis in diurnal raptors. Biological Journal of the Linnean
Ruxton, G. D., T. N. Sherratt, and M. Speed. 2004. Avoiding
attack: the evolutionary ecology of crypsis, warning signals
and mimicry. Oxford University Press, Oxford, UK.
SAS. 2004. SAS/STAT 9.1 User’s Guide. SAS Institute, Cary,
North Carlina, USA.
Shine, R. 1991. Australian snakes. A natural history. Reed,
Shine, R., Ambariyanto, P. S. Harlow, and Mumpuni. 1998.
Ecological divergence among sympatric colour morphs in
blood pythons, Python brongersmai. Oecologia 116:113–119.
Wolf, M., and Y. L. Werner. 1994. The striped colour pattern
and striped/non-striped polymorphism in snakes (Reptilia:
Ophidia). Biological Review 69:599–610.
A table showing species of Australian lizards and snakes classified as being either variable or non-variable with regard to color
pattern (Ecological Archives E089-073-A1).
May 20081207 ASSOCIATIONS OF COLOR POLYMORPHISM