O R I G I NA L A RT I C L E
doi:10.1111/j.1558-5646.2009.00694.x
BODY SIZE DIVERSIFICATION IN ANOLIS:
NOVEL ENVIRONMENT AND ISLAND EFFECTS
Gavin H. Thomas,1,2 Shai Meiri,1,3 and Albert B. Phillimore1,4
1
NERC Centre for Population Biology & Division of Biology, Imperial College London, Silwood Park, Ascot, Berkshire, SL5
7PY, United Kingdom
2
E-mail: [email protected]
3
E-mail: [email protected]
4
E-mail: [email protected]
Received August 8, 2008
Accepted March 3, 2009
Extreme morphologies of many insular taxa suggest that islands have unusual properties that influence the tempo and mode of
evolution. Yet whether insularity per se promotes rapid phenotypic evolution remains largely untested. We extend a phylogenetic
comparative approach to test the influence of novel environments versus insularity on rates of body size and sexual size dimorphism
diversification in Anolis. Rates of body size diversification among small-island and mainland species were similar to those of anole
species on the Greater Antilles. However, the Greater Antilles taxa that colonized small islands and the mainland are ecologically
nonrandom: rates of body size diversification among small-island and mainland species are high compared to their large-island
sister taxa. Furthermore, rates of diversification in sexual size dimorphism on small islands are high compared to all large-island
and mainland lineages. We suggest that elevated diversifying selection, particularly as a result of ecological release, may drive high
rates of body size diversification in both small-island and mainland novel environments. In contrast, high abundance (prevalent
among small-island lizard communities) mediating intraspecific resource competition and male–male competition may explain why
sexual size dimorphism diversifies faster among small-island lineages than among their mainland and large-island relatives.
KEY WORDS:
Anolis lizards, body size, ecomorphs, islands, morphological diversification rates, novel environments, phylogeny,
sexual size dimorphism.
The extremes and unusual diversity of morphological forms found
on islands (Sondaar 1977; Case 1978), including dwarf and giant
morphs of many taxa (Russell 1877; Hooijer 1967; Keogh et al.
2005; Hedges 2008), have prompted comparisons of the rate of
trait evolution between insular and mainland taxa (Millien 2006;
Harmon et al. 2008; Pinto et al. 2008). High rates of trait evolution on islands are commonly attributed to ecological release
in which species’ expand their resource use or habitat primarily
because of a reduction in the number of competitors (Grant 1972).
Rapid trait change driven by ecological release is expected to occur following colonization of a novel environment that has fewer
potential competitors than the source (Grant 1972; Losos and De
Queiroz 1997). This scenario is likely to be particularly preva
C
2017
lent for island colonization (Lister 1989; Dayan and Simberloff
1998; Meiri et al. 2005) because islands are often species poor. If
ecological release is widespread among species following island
colonization then we might expect morphological and ecological
traits to diversify more rapidly among island species than their
mainland counterparts.
High rates of evolution may also occur when colonizing a
novel environment as a result of shifts in selection pressures driven
by, for example, differences in climate, vegetation, resource base,
competitors, or predators (Blondel 2000). In principle, this explanation is applicable to both novel island and novel continental
habitats (Campbell and Echternacht 2003). However, Price (2008)
suggests that the effects of differences in selection pressure will
C 2009 The Society for the Study of Evolution.
2009 The Author(s). Journal compilation Evolution 63-8: 2017–2030
G AV I N H . T H O M A S E T A L .
be more pronounced where there are multiple colonization events
onto different islands rather than into different novel continental
habitats. This is because the composition of island communities
(in terms of species identity) is likely to be more heterogeneous
(both between islands and through time) than the composition of
novel continental communities (Price et al. 2009). Consequently,
there should be greater variation in selection pressure between
species that have colonized multiple islands than between species
that have colonized a similar number of new areas of the mainland. This model therefore implies that there is greater potential
for rapid trait divergence among species that have colonized islands than among species that have colonized mainland novel environments. Rates of phenotypic diversification, however, could
be higher among species in both forms of novel environment than
among the source pool of species.
Recent studies of Australasian birds, Caribbean anoles, and
African chameleons have highlighted several systems in which islands are the source for mainland colonization (Raxworthy et al.
2002; Filardi and Moyle 2005; Nicholson et al. 2005; reviewed
in Bellemain and Ricklefs 2008). The biogeographic history of
Caribbean Anolis lizards (Nicholson et al. 2005) is well suited
to the study of morphological evolution in novel environments
versus islands per se. From a mainland South or Central American source, anoles diverged and speciated in situ and by dispersal
between the islands of the Greater Antilles (Cuba, Hispaniola, Jamaica, and Puerto Rico). In turn, the adaptive radiation of Greater
Antillean anoles has been the source of multiple colonization
events onto smaller islands throughout the Caribbean, and of recolonization of the mainland (Schoener 1969; Glor et al. 2005;
Nicholson et al. 2005). Anole communities on the Greater Antilles are species-rich and complex (Williams 1983; Losos et al.
2003) with as many as 14 or 15 species known to occur in sympatry in parts of Cuba (Diaz et al. 1998; Garrido and Hedges
2001). However, communities on small islands tend to be speciespoor (with a maximum of four anole species) and consequently
ecological opportunity is expected to be high for new colonizers. In contrast, mainland communities are more species-rich and
contain many potential competitors including the sister-clade of
Caribbean anoles (sometimes referred to as Dactyloa), and consequently have low expected ecological opportunity.
Island colonizers are expected to encounter low interspecific competition, but they may be subject to increased intraspecific competition due to density compensation (MacArthur et al.
1972). Density compensation describes the association between
low species richness and increased population density and seems
to be a common feature of insular lizard communities (Case 1975;
Buckley and Jetz 2007). If increased population density elevates
intraspecific competition, then there may be divergence in resource use within populations. This may lead to increased sexual
dimorphism, particularly in body size or in the trophic apparatus
2018
EVOLUTION AUGUST 2009
(Selander 1966). If the population density varies among islands
then we predict that sexual dimorphism should diversify more
rapidly among islands (competitor-poor, both high and low abundance) than among novel mainland areas (competitor-rich, usually
low abundance).
Our primary objective here is to compare rates of diversification in body size and sexual size dimorphism between Greater
Antillean (large island source pool), small-island colonizing, and
mainland colonizing anoles. However, several studies have suggested that anole lineages that dispersed away from the Greater
Antilles are an ecologically and morphologically nonrandom set
of species (Losos and De Queiroz 1997; Poe et al. 2007). This may
be important in interpreting any differences in rates of morphological diversification. Greater Antillean anoles have been classified
into six clearly defined ecological and morphological groupings
or “ecomorphs” (Williams 1972, 1983; Losos 1994). Both body
sizes (Schoener 1969; Williams 1983) and sexual dimorphism
(Butler et al. 2000; Butler et al. 2007) differ substantially between
ecomorphs. For example, “twig” anoles are typically small bodied and sexually monomorphic species, whereas “trunk-ground”
and “trunk-crown” species tend to have intermediate body sizes
and strong male-biased sexual size dimorphism. Most mainland
species have not been assigned to ecomorphs (Irschick et al. 1997)
but solitary species on small islands often resemble the “trunkcrown” or “trunk-ground” ecomorphs (Williams 1969; Losos and
De Queiroz 1997). This may be because small-island (or mainland) colonizers are derived from the “trunk-crown” or “trunkground” ecomorphs. Alternatively, colonizing species may have
converged on these two ecomorphs. If it is the former (as inferred
by Poe et al. 2007) then it is interesting to ask whether rates
of morphological diversification among small-island or mainland
species exceed rates among large-island “trunk-crown” or “trunkground” species even if rates are not greater than all large-island
species together. Therefore, morphological divergence of “trunkcrown” and “trunk-ground” species on the Greater Antilles may
be constrained by competition with other anoles that would be
absent from small islands or the mainland.
Here, we test whether colonizing lineages are nonrandom
with respect to the ecomorph of the likely founding lineage. We
then extend and apply a recent phylogenetic method (O’Meara
et al. 2006; Thomas et al. 2006) to examine the influence of
novel environments (mainland recolonizers) versus insularity per
se (small-island colonizers) on rates of body size diversification
in anoles.
Methods
NONRANDOM COLONIZATION AMONG ECOMORPHS
We tested for bias in the ecomorphs of anole lineages that have
colonized small islands or the mainland by reconstructing the
S I Z E D I V E R S I F I C AT I O N I N A N O L E S
ancestral ecomorph states on a recent phylogeny of anoles
(Nicholson et al. 2005). We classified each Greater Antillean anole
species using the ecomorph (sensu Williams 1972) designations
of Losos et al. (2006). Ecomorphs are named for the microhabitat
they occupy: grass-bush, trunk, trunk-ground, trunk-crown, twig,
and crown-giant. Some species do not fit into any of these six
categories and are classified as unique (Supplementary Appendix
S1 & S2). Two studies of 76 species in total (Supplementary Appendix S3) have shown that the six Greater Antilles ecomorphs
form distinct clusters in morphospace (Losos et al. 1998; Beuttell
and Losos 1999). Some species have not been subject to morphometric analyses but our main interest is in the ecological definition of ecomorph: definitions in Losos et al. (2006) were based on
qualitative observations in the field and descriptions of species’
habitat use from the literature (J. Losos, pers. comm.).
We used an ultrametric version of Nicholson et al’s
(2005) phylogeny with branch lengths proportional to time
based on penalized likelihood downloaded from http://
biosgi.wustl.edu/∼lososlab/anolis_mbg_2005/. We pruned the
phylogeny to include only Greater Antillean species (that is, only
the source pool species for which ecomorphs have been assigned;
Supplementary Appendix S4). Ancestral ecomorphs were
inferred using the maximum-likelihood Mk1 model in Mesquite
version 2.0 (Maddison and Maddison 2006, 2007). This analysis
confirmed that both small-island and mainland anoles are most
likely derived from species of the trunk-crown and trunk-ground
ecomorphs (Fig. 1 and Supplementary Appendix S4).
DATA
We categorized Anolis species as mainland, large-island, or smallisland species (Supplementary Appendix S1) following Nicholson
et al. (2005). Large islands (Cuba, Jamaica, Hispaniola, and Puerto
Rico) are all > 9000 km2 in area. Small islands are all < 3500 km2 .
There are no Caribbean islands of intermediate area. Small-island
status was only assigned to species endemic to small islands.
Because small-island and mainland species are all derived from
large-island lineages of the trunk-crown and trunk-ground ecomorphs (see above), we further divided large-island species into
two ecomorph categories: species of the trunk-crown and trunkground ecomorphs, and species that are unique or fit one of the
four remaining ecomorphs. Thus, we placed each species into
one of four geographical and ecomorph categories: small-island
species (scored as 0); large-island trunk-crown and trunk-ground
species (1); large-island other ecomorph species (2); and, mainland species (3).
Lizards continue growing after reaching sexual maturity and
the maximum, rather than mean, body size of a sample is often a
more appropriate estimator of age-independent adult size (Stamps
and Andrews 1992). Although maximum body size is likely to increase with sample size, around 20 individuals are considered
sufficient to provide a reliable estimate of asymptotic body size
with 25 individuals considered “adequate for most applications”
(Stamps and Andrews 1992). We compiled sex-specific data on
maximum snout vent length (SVL) of Anolis lizards from the literature and recorded sample sizes when available. All body size
data, including sample sizes and sources, are provided in Supplementary Appendices S1 and S2. Our focus is on anoles including
all Caribbean island species and their descendents that recolonized the mainland. We excluded species of Dactyloa, the mostly
mainland-dwelling South American sister group of Greater Antilles anoles (Nicholson et al. 2005) and note that this group is
extremely undersampled both morphologically and phylogenetically (Pinto et al. 2008). Although phylogenetic sampling of the
species that have reinvaded the mainland is not complete, the
sampled species are an unbiased representation of the diversity of
body sizes found in this clade (see data in Meiri 2008).
Low intraspecific sampling can inflate variance across
species and may influence estimates of relative morphological
diversification rates. This is particularly important if sampling
effort is inconsistent across groups. We used a chi-square test to
examine sampling bias for male and female size across the four
geographic and ecomorph categories. We divided species into
those with good (n ≥ 20) and poor (n < 20) sampling (following
Stamps and Andrews 1992) and assumed that species with no
sample sizes reported were poorly sampled (n < 20). We found
no evidence for differences in the quality of sampling between the
geographic and ecomorph classes (male SVL: χ2 = 2.524, df =
3, P = 0.471; female SVL: χ2 = 1.795, df = 3, P = 0.616). Using
more stringent definitions for good sampling quality (minimum
sample of 25, 30, 40, and 50 individuals), we still found no evidence for sampling bias. Nonetheless, we repeated all our main
analyses on a subset of the data that included only species with
maximum SVL based on at least 20 individuals (see below).
PHENOTYPIC DIVERSIFICATION RATES
The Brownian motion model of trait evolution describes a linear
increase in phenotypic variance with distance from the root of the
tree. The expected covariance among species can be described
by the variance–covariance matrix (V) representation of the phylogenetic tree. The Brownian model is a suitable model of trait
evolution under random genetic drift and also shares comparable expected covariance structures with directional, fluctuating,
and punctuated evolution (Hansen and Martins 1996). Following
Freckleton et al. (2002) the unbiased Brownian variance (σ2 ) is
given by
σ2 =
1
(y − α̂X)T V−1 (y − α̂X),
(n − 1)
(1)
where n is the number of tips, y is an n × 1 vector of trait
values at the tips, α is an n × 1 vector of the phylogenetic mean
EVOLUTION AUGUST 2009
2019
G AV I N H . T H O M A S E T A L .
Figure 1.
Anolis phylogeny. Includes all species used in this study, after Nicholson et al. (2005). Lineages are colored according to
geographic and ecomorph category. Asterisks indicate species that have been subject to morphometric analyses of ecomorphs (see
Supplementary Appendix S3 for further details).
for the trait, X is an n × 1 design matrix in which all entries
are set to one, and the superscript T shows that the transpose is
calculated. The Brownian variance is an estimate of the minimum
rate of evolutionary change (Garland 1992) and can therefore be
considered a measure of the rate of phenotypic diversification.
However, the Brownian model may incorrectly estimate the rate
of evolution (distinct from the rate of diversification) if traits have
evolved, for example, by directional, fluctuating, or punctuated
evolution.
If the rate of phenotypic diversification is heterogeneous then
the covariance among species may deviate from expectation de-
2020
EVOLUTION AUGUST 2009
rived from the phylogeny. Several methods have been proposed to
test for rate heterogeneity among lineages (Garland 1992; McPeek
1995; Mooers et al. 1999; O’Meara et al. 2006; Thomas et al.
2006). The maximum-likelihood method proposed by Thomas
et al. (2006) describes the expected covariance among species as
the entry-wise sum of two matrices, V 0 and V 1 , where V 0 refers
to branches of the phylogeny that share a binary character in state
0 and V 1 refers to the branches the character in state 1. To derive
the expected variance–covariance matrix, a scalar, θ, is applied to
one of the two matrices such that V = V 0 + θV 1 (note that the θ
parameter in our model is not the same as the mean θ in Butler
S I Z E D I V E R S I F I C AT I O N I N A N O L E S
and King’s (2004) Ornstein–Uhlenbeck model). The maximumlikelihood value of θ is then estimated where deviation from θ =
1 indicates rate heterogeneity. Here we extend the Thomas et al.
(2006) model to allow for multiple rate parameters such that V =
V 0 + θ 1 V 1 + ··· + θ k −1 V k −1 where k is the number of different
parts of the tree such that the estimate of the Brownian variance
(σ2 ) is given by
σ2 =
1
(y − âX)T (V0 + θ1 V1
(n − k)
+ · · · + θk−1 Vk−1 )−1 (y − α̂X).
(2)
In contrast to equation (1), here X is an n × k design matrix describing a multilevel factor. Our approach differs from the
“noncensored” method of O’Meara et al. (2006) because by including X as a design matrix, we allow a different phylogenetic
mean (as well as a different rate) in each of the k parts of the
tree rather than assuming a single phylogenetic mean across the
tree. Because multiple means are estimated, the denominator n −
1 in equation (1) is replaced by n − k in equation (2) (differing
from O’Meara et al who use n in their noncensored method). The
full derivation of the maximum-likelihood model is described in
detail by Freckleton et al. (2002).
Although the inclusion of different means has been questioned (Revell 2008), we argue that most hypotheses postulating
different rates imply different evolutionary regimes such that a
difference in mean is also a likely outcome. Means could differ
if trait evolution in one group is parallel (e.g., consistent shifts to
small body size in elephant species on islands compared to their
mainland sister species, Roth 1992; or the evolution of flightlessness in endemic rails, Trewick 1997), or if there is a single
shift in trait values at the base of a clade (e.g., the clade-wide
increase in bill length in Hawaiian honeycreepers, Lovette et al.
2001). A difference in means due to a single rapid change at the
base of clade is a form of rate shift. However, although it may
be possible to show that such a rate shift has occurred, it may
not be possible to identify which group increased or decreased in
rate. It is therefore informative to distinguish between a rate shift
that is due to a change in mean and one that is due to differences
in rates across all species in each group of interest. We show by
simulation that models assuming a common mean can indicate a
rate shift if the means of each group differ even if the Brownian
variances within each group do not (see Supplementary Appendix
S5). If the relevant hypothesis refers to differences in rates across
all species in each group of interest then the inference of a rate
shift due to differences in mean should be regarded as a type I error. Our model allows each group to effectively jump to different
means but within each group the trait follows a Brownian model.
Consequently, shifts in mean, but not in rates of whole groups,
are not inferred as rate shifts (Supplementary Appendix S5).
RATES MODELS
We used the phenotypic diversification rate tests described above
to compare rates of diversification in male maximum SVL, female maximum SVL, and sexual size dimorphism (SSD) across
the four island type/ecomorph categories. We log 10 transformed
male (162 species) and female (163 species) SVLs prior to analysis and calculated sexual size dimorphism (n = 160 species) as
log 10 (male SVL / female SVL) following the recommendations
of Smith (1999). Branches in the phylogeny were assigned to one
of the four island type/ecomorph categories (Fig. 1) based on the
ancestral state reconstruction described above and on Nicholson
et al. (2005). The phylogeny with branch assignments as node
labels is available in Supplementary Appendix S6. The most
complex model of phenotypic diversification rates has four rates,
one each for small-island lineages, large island trunk-ground and
trunk-crown lineages, large island “other" lineages, and mainland
lineages. In all models the parameter estimates were rescaled so
that θ = 1 for the small-island group to allow model averaging
(see below). The simplest model is the null constant-rate Brownian model. We fitted each of the 12 possible models to male SVL,
female SVL, and SSD in turn. We ranked models using the smallsample Akaike Information Criterion (AICc) and calculated both
delta AICc and Akaike weights (Burnham and Anderson 2002).
We used the Akaike weights to estimate model-averaged parameter estimates. We ran each set of 12 models four times using:
(1) the full dataset and allowing different means in each group;
(2) the full dataset and assuming a common mean; (3) the full
dataset with different means in each group but after transforming
the phylogeny according to the maximum-likelihood estimate of
the branch length transformation kappa (see below); and (4) a
reduced dataset including only species with SVL estimates based
on samples of at least 20 individuals and allowing different means
in each group. In the main text, we present only the first set of
models and the results of the remaining three sets of models are
available as Supporting Information (Supplementary Appendix
S7). R code for the phenotypic diversification rate tests and an
example analysis is available in Supplementary Appendices S8
and S9.
We also compared the maximum likelihood of each model
with the likelihood of the constant-rate model using the likelihoodratio statistic. This statistic is assumed to be asymptotically
chi-square distributed with degrees of freedom equal to the
difference in the number of parameters between the models (Edwards 1972). Previous studies based primarily on tworate models indicate appropriate type I errors and that parameter estimates are unbiased (O’Meara et al. 2006; Thomas
et al. 2006; Revell 2008), however, multiple parameter models have not previously been tested. We therefore simulated
the evolution of a trait along the anole phylogeny with a
single rate 10,000 times for each of the 12 models. We
EVOLUTION AUGUST 2009
2021
G AV I N H . T H O M A S E T A L .
compared each model with the null to estimate type I error
rates.
KAPPA TRANSFORMATION
If a trait evolves in a punctuated rather than gradual fashion
(Eldredge and Gould 1972) then there could be a bias toward
higher rates in one group if it has a predominance of short branches
relative to the groups with which it is being compared. This is relevant here because short branches separate many small-island
lineages and consequently high rates among these lineages could
be an artifact of a speciational evolutionary process rather than a
reflection of elevated rates of trait diversification on islands. We
therefore tested for speciational evolution in our data by estimating the parameter κ (Pagel 1997) on the phylogeny for each of the
three traits (male and female maximum SVL, and SSD) where
κ = 1 indicates evolutionary change consistent with a Brownian model, κ < 1 indicates that there is evolutionary stasis in
long branches, and κ > 1 indicates accelerated evolution in long
branches. The maximum-likelihood estimate of κ can be compared with a model with κ = 1 using the likelihood-ratio statistic
assuming a chi-square distribution with one degree of freedom.
Results
SIMULATIONS
Based on 10,000 simulations, we found very slightly elevated
type I errors for most models (Supplementary Appendix S10).
The maximum type I error rate across the full set of models was
0.058. Consequently, we also checked that models found to differ significantly from the null (constant rates) model using the
likelihood-ratio tests were also significant based on the simulated
distribution of the likelihood-ratio statistic. The qualitative interpretations of our results are not affected. However, we suggest that
simulations should be used a matter of course when using the rates
test, particularly when multiple rates are estimated.
MALE AND FEMALE SVL
The model-averaged parameter estimates for both male and female SVL show that the rate of phenotypic diversification is lower
among large island trunk-ground and trunk-crown species than in
the three other categories, which do not differ from one another
(male SVL, Table 1; female SVL, Table 2). This is consistent
with the single best-fitting model and the parameter estimates
in the four-rate model (Table 1 and Fig. 2A; Table 2 and Fig.
2B). Models in which the rates of phenotypic diversification were
equal for both large island categories but allowed to differ for
mainland and/or small-island lineages were substantially worse
than the best-fitting model (male SVL: AICc > 8; female SVL
AICc > 11). This suggests that rates among small-island or
mainland lineages do not exceed those of all large-island taxa but
2022
EVOLUTION AUGUST 2009
are higher than those of the large-island lineages from which they
are derived.
Models in which we assumed a common mean (see Supplementary Appendix S7) typically have slightly lower AICc values
than the equivalent multiple-means models, indicating that neither male SVL nor female SVL differs between groups. This is
evident from the phylogenetically corrected 95% confidence intervals (based on model averaged variances) for male SVL in millimeters from the multiple means models: small-island species =
63.0–74.9; large-island trunk-crown and trunk-ground species =
67.9–73.4; large-island other ecomorph species = 59.3–71.3; and
mainland species = 58.5–69.1. The equivalent 95% confidence
intervals for female SVL are: small-island species = 48.8–56.6;
large-island trunk-crown and trunk-ground species = 55.6–59.6;
large-island other ecomorph species = 51.4–64.0; and mainland
species = 52.0–61.9. The model averaged parameter estimates for
the common mean models were consistent with the multimeans
analyses (Supplementary Appendix S7). We found no evidence
for long-branch stasis (male SVL: κ = 0.841; comparison with
κ = 1: χ2 = 1.820, P = 0.177; female SVL: κ = 1.054; comparison with κ = 1: χ2 = 0.219, P = 0.640) and the model averaged
parameter estimates were again similar when we first transformed
the phylogeny according to the maximum-likelihood estimate of
kappa (see Supplementary Appendix S7). This was also the case
when we used the reduced dataset (see Supplementary Appendix
S7). Overall, and regardless of the choice of analysis, rates among
small-island or mainland lineages do not differ from those of all
large-island taxa but are higher than those of the large-island
lineages from which they are derived.
SEXUAL SIZE DIMORPHISM
The model-averaged parameter estimates for SSD show that the
rate of phenotypic diversification is highest among small-island
species with the lowest rates among both large-island other ecomorph species and mainland species. There is some evidence
for intermediate rates among large-island trunk-crown and trunkground species (Table 3). This is consistent with the single bestfitting model and the parameter estimates in the four-rate model
(Table 3; Fig. 2C).
Models in which we assumed a common mean (see
Supplementary Appendix S7) typically have higher AICc values, indicating that SSD differs substantially between groups
(phylogenetically corrected mean SSD ± 95% confidence
intervals based on model averaged variance: small-island
species = 0.097–0.130; large-island trunk-crown and trunkground species = 0.077–0.094; large-island other ecomorph
species = 0.047–0.058; mainland species = 0.030–0.040).
However, the model-averaged parameter estimates assuming
a common mean were consistent with those allowing multiple means (Supplementary Appendix S7). The model-averaged
Small islands
1.000
1.000
1.000
1.000 (0.533–2.008)
1.000 (0.542–2.043)
1.000
1.000
1.000 (0.828–3.152)
1.000
1.000 (0.658–2.520)
1.000
1.000 (0.673–2.574)
1.000
Model
M5
M3
M4
M9
M1
M11
All equal
M7
M2
M8
M10
M6
Model average
0.409 (0.243–0.731)
0.379 (0.226–0.673)
0.445 (0.266–0.789)
0.415 (0.242–0.728)
0.409 (0.244–0.725)
1.000
1.000
0.651
1.000
0.817
1.000
0.800
0.447
Large island TC & TG
1.000
1.000
1.214 (0.758–1.954)
1.018
1.115 (0.696–1.795)
1.455 (0.901–2.362)
1.000
1.059 (1.009–2.636)
1.741 (1.080–2.820)
0.817
1.000
0.800
1.062
Large island other
1.000
0.809 (0.500–1.367)
1.000
1.018
0.879 (0.542–1.484)
1.000
1.000
0.651
1.421 (0.877–2.400)
0.817
1.041 (0.642–1.761)
0.865 (0.667–1.828)
0.961
Mainland
0.000
1.598
1.625
2.183
3.724
6.233
6.430
6.846
6.961
8.217
8.562
10.312
deltaAICc
0.389
0.175
0.173
0.131
0.060
0.017
0.016
0.013
0.012
0.006
0.005
0.002
wtAIC
and ∗∗∗ denotes P<0.001. Models in which at least two rate categories are specified are named M1–M11 and the constant rate model is named All equal.
122.539
122.833
122.820
122.541
122.877
119.423
118.246
120.209
120.152
118.431
118.258
118.476
Maximum
likelihood
∗
(6)
(7)
∗
(7)
∗
(7)
∗
(8)
NS (6)
(5)
NS (7)
NS (7)
NS (6)
NS (6)
NS (7)
∗∗
P (k)
according to the small-sample Akaike Information Criteria (AICc): delta AICc shows the difference in AICc between the candidate model and the best-fitting model and wtAIC refers
to the Akaike weights. The maximum likelihood of each model is also compared with the constant-rates Brownian model (which has five parameters: four means and one rate)
using the likelihood-ratio statistic (χ2 ) with degrees of freedom equal to the difference in the number of estimated parameters (k) and where ∗ denotes P<0.05, ∗∗ denotes P<0.01,
Table 1. Rates of diversification in male snout vent length. The maximum-likelihood estimates of θ with approximate 95% confidence intervals in parentheses for male snout
vent lengths. Estimates of θ are shown for small-island, large-island trunk-ground and trunk-crown, large-island “other" ecomorph and mainland lineages. The models are ranked
S I Z E D I V E R S I F I C AT I O N I N A N O L E S
EVOLUTION AUGUST 2009
2023
2024
1.000
1.000
1.000 (0.337–1.308)
1.000 (0.533–2.069)
1.000
1.000
1.000
1.000 (0.583–2.316)
1.000
1.000 (0.422–1.700)
1.000
1.000 (0.425–1.712)
1.000
Small islands
Large island other
1.527 (0.956–2.451)
1.000
1.583
1.804 (1.130–2.895)
1.000
2.655 (1.654–4.283)
1.858 (1.154–3.010)
1.717 (1.180–3.076)
1.000
1.241
1.000
1.232
1.469
Large island TC & TG
0.412 (0.245–0.734)
0.337 (0.198–0.610)
0.494 (0.184–0.563)
0.486 (0.289–0.867)
0.309 (0.183–0.556)
1.000
1.000
0.904
1.000
1.241
1.000
1.232
0.460
Rates of diversification in female snout vent length. Details follow Table 1.
M4
M5
M9
M1
M3
M2
M11
M7
All equal
M8
M10
M6
Model average
Model
Table 2.
1.000
1.000
1.583
1.267 (0.782–2.139)
0.791 (0.489–1.336)
1.905 (1.177–3.217)
1.000
0.904
1.000
1.241
1.057 (0.653–1.784)
1.265 (0.634–1.734)
1.164
Mainland
0.000
0.483
1.116
1.854
1.948
2.974
5.695
7.802
10.037
11.835
12.147
14.008
deltaAICc
0.290
0.228
0.166
0.115
0.110
0.066
0.017
0.006
0.002
0.001
0.001
0.000
wtAIC
150.862
149.529
150.304
151.041
149.888
149.375
146.923
146.961
143.673
143.853
143.697
143.858
Maximum
likelihood
∗∗∗
(7)
(6)
∗∗
(7)
∗∗
(8)
∗∗
(7)
∗∗
(7)
∗
(6)
∗
(7)
(5)
NS (6)
NS (6)
NS (7)
∗∗∗
P (k)
G AV I N H . T H O M A S E T A L .
EVOLUTION AUGUST 2009
Figure 2.
and (C) sexual size dimorphism among Anolis lizards. For the four
ecomorph/island categories, the maximum-likelihood value of the
Rates of diversification of (A) male SVL, (B) female SVL,
relative rate estimates along with approximate 95% confidence
intervals are shown for the full four-rate model. The four categories are labeled as: Small island (small-island species), Large
TCTG (large-island trunk-ground and trunk-crown species), Large
other (large-island species that are not trunk-ground and trunkcrown anoles), and Mainland (mainland species).
parameter estimates were also consistent when we first transformed the phylogeny according to the maximum-likelihood estimate of kappa, even though we found evidence for long-branch
stasis (κ = 0.666; comparison with κ = 1: χ2 = 5.081, P =
0.024). The intermediate rates among large-island trunk-crown
and trunk-ground species are not present when we used the
(7)
(6)
∗∗
(8)
∗∗
(7)
∗∗
(7)
∗∗
(7)
∗
(6)
∗
(6)
NS (7)
(5)
NS (7)
NS (6)
∗∗∗
∗∗∗
256.094
254.072
256.178
254.767
254.573
254.468
250.364
250.234
250.375
248.190
250.249
248.438
0.406
0.160
0.146
0.108
0.089
0.080
0.004
0.003
0.001
0.001
0.001
0.001
0.000
1.855
2.049
2.653
3.042
3.250
9.270
9.531
11.437
11.459
11.689
13.122
0.293
0.354
0.272 (0.164–0.471)
0.275 (0.432–1.246)
0.403 (0.244–0.697)
0.388
1.000
0.565 (0.341–0.982)
1.000
1.000
0.574 (0.346–0.996)
1.000
0.321
0.293
0.354
0.307 (0.194–0.492)
0.383
0.455 (0.288–0.730)
0.311 (0.506–1.283)
0.605 (0.382–0.970)
1.000
0.595 (0.376–0.954)
1.000
1.000
1.000
0.337
1.000 (1.831–7.047)
1.000 (1.519–5.829)
1.000 (0.537–2.067)
1.000 (1.400–5.382)
1.000
1.000 (1.386–5.318)
1.000
1.000
1.000
1.000
1.000
1.000
1.000
M9
M8
M1
M6
M2
M7
M11
M10
M4
All equal
M3
M5
Model average
0.505 (1.032–2.981)
0.354
0.505 (0.303–0.873)
0.383
1.000
0.388
1.000
1.000
0.958 (0.576–1.651)
1.000
1.049 (0.632–1.801)
1.204 (0.726–2.068)
0.508
Mainland
Large island other
Large island TC & TG
Small islands
Model
Table 3.
Rates of diversification in sexual size dimorphism in snout vent length. Details follow Table 1.
deltaAICc
wtAIC
Maximum
likelihood
P (k)
S I Z E D I V E R S I F I C AT I O N I N A N O L E S
reduced dataset: instead this group has a similar low rate to
the large-island other ecomorph species and mainland species
(see Supplementary Appendix S7). Taken together, these results
strongly suggest exceptionally high rates of diversification in SSD
among small-island species.
Discussion
Dispersal to novel, previously unoccupied, habitats can result in
changes to both the strength and direction of selection pressures
(Simpson 1944; Barton 1996; Blondel 2000; Herrel et al. 2008;
Price 2008). Phenotypic change may be driven by differences in,
for example, climate, community structure, and predation risk experienced by colonizing species (Blondel 2000; Blumstein 2002).
Typically, studies of ecologically driven variation in rates of morphological evolution have considered islands as novel environments (Millien 2006; Harmon et al. 2008). Our results show that
lineages of Anolis lizards that disperse to novel mainland environments have similar rates of body size diversification to lineages
that dispersed to small-island (i.e., novel island) environments.
However, whether rates of trait diversification among mainland and small-island lineages differ from those of the (largeisland) source pool depends on the definition of the source pool.
Compared to all large-island taxa, rates of body size diversification on small islands or the mainland are not high: they are
indistinguishable from the adaptive radiation of anoles on the
Greater Antilles. Yet if the source pool is restricted to include
only those lineages that appear to be ecologically predisposed to
being successful dispersers and colonizers, that is the trunk-crown
and trunk-ground ecomorphs (Poe et al. 2007), then rates of body
size diversification are elevated among small-island lineages. Furthermore, rates of morphological diversification in sexual size dimorphism are high among small-island lineages, but not among
mainland lineages, regardless of how the source pool is defined.
We also note that large-island species of the other four ecomorphs
have a higher rate of diversification in body size than large-island
species of the trunk-crown and trunk-ground ecomorphs. This
may imply that rather than high rates among small-island anoles,
there is a low rate among large-island trunk-crown and trunkground species. Although we suggest that it is more parsimonious
to infer high rates among small-island species, we also discuss
the alternatives below.
A restricted definition of the source pool is valid and important in interpreting our results. Small-island and mainland taxa
are similar to the trunk-crown or trunk-ground ecomorphs (Losos
and De Queiroz 1997) because they are descended from them
not because small-island or mainland lineages have converged
toward these two ecomorphs (our Results and Poe et al. 2007).
Why, then, are diversification rates higher in both small-island and
mainland lineages than in the restricted source pool lineages? One
EVOLUTION AUGUST 2009
2025
G AV I N H . T H O M A S E T A L .
explanation is that rates of diversification are low among trunkcrown and trunk-ground species because there is something unusual about these two particular ecomorphs. Male–male competition is known to be particularly strong with sexual selection
favoring large male size in trunk-crown and trunk-ground anoles
(Butler et al. 2000; Butler and Losos 2002; Butler et al. 2007).
This may limit the extent to which ecological factors can influence
diversification in body size. Indeed, a substantial proportion of
morphospace occupancy in trunk-crown and trunk-ground anoles
is determined by sexual dimorphism rather than by interspecific
variation (Butler et al. 2007). Consequently, sexual selection may
be a stronger constraint on body size divergence in trunk-crown
and trunk-ground anoles than in the other ecomorphs. The difference in rates between our two groups of large-island ecomorphs
may also be partly artefactual. Species’ ecomorphs are designated
on the basis of ecology, habitat use, and behavior but they also
form distinct clusters in morphological space such that there are
greater morphological difference between ecomorphs than within
them (Losos et al. 1998; Beuttell and Losos 1999). When we
compare the species that belong to the colonizing ecomorphs to
the rest of the large island species, we are comparing two ecomorphs with four ecomorphs and unique species that do not fit to
any particular ecomorph. Hence, a lower rate among the group of
trunk-crown and trunk-ground species is not surprising because it
captures little of the overall between ecomorph variation.
Although either of these explanations may explain the difference between the two ecomorph groupings of large island species,
they cannot explain why small-island species have a much higher
rate of body size diversification than the trunk-crown and trunkground ecomorphs. There is no evidence to suggest that species
of any of the four other ecomorphs have successfully colonized
small islands. We therefore suggest that the difference in rates that
we identified between small-island species and trunk-crown and
trunk-ground species is most likely due to net increases in rate
among small-island species. One nonadaptive explanation that
may apply over short periods of time is that founder effects (e.g.,
Mayr 1954; Carson and Templeton 1984) or random genetic drift
acting on standing genetic variation (Kimura 1968) has resulted
in elevated rates of trait diversification. However, this is unlikely
to explain our results given that most field studies indicate that
phenotypic differences between populations are generally best explained by selection rather than by purely neutral processes (e.g.,
Merilä and Crnokrak 2001; Clegg et al. 2002; Leinonen et al.
2008).
Two nonmutually exclusive ecological explanations may be
important. First, the variation in selection pressure (particularly
the direction of selection) encountered by lineages colonizing
new environments may result in each colonizing species having
distinct optima in each new environment (Price 2008). If this
is the case, then variation in optima from one species or novel
2026
EVOLUTION AUGUST 2009
environment to the next will result in elevated rates of phenotypic diversification across species. However, variation in selection pressure is expected to be greatest among island settings, and
will increase as island area decreases, due to greater variation in
community composition (Price et al. 2009). The elevated rates
among mainland taxa (compared to the restricted source pool)
that we observed are therefore not expected in this model. One
possible reason is that while species identity within communities
may be variable in island settings and may link to variation in
selection pressure, selection and trait optima may be influenced
by other factors such as the greater complexity and variety of possible species interactions in the more species-rich mainland communities. The second, and most frequently invoked mechanism
is that ecological opportunity is high on islands largely because
some communities have few or no competitors, and this allows
rapid trait diversification. If so, rates among small-island lineages
may be high because they are not competing with smaller (twig
ecomorphs) and larger (crown-giant ecomorph) competitors that
may inhibit the size evolution of trunk-crown and trunk-ground
species on species-rich large islands. In contrast, the mainland recolonizers may come into contact and compete with members of
the species-rich Dactyloa sister clade (Nicholson et al. 2005). If
there is variation in the direction of selection on different islands,
as in the Price model (Price 2008), then ecological opportunity
would elevate rates among small-island but not mainland anoles.
A recent study by Pinto et al. (2008), however, suggests that
anoles that have recolonized the mainland may not compete with
Dactyloa anoles. They argue that Caribbean anoles and their descendents that recolonized the mainland have better-developed
toe-pads than the Dactyloa species (Macrini et al. 2003; Velasco
and Herrel 2007). The toe-pad may therefore be a key innovation
or exaptation (Simpson 1944; Gould and Vrba 1982) that has increased ecological opportunity for the mainland colonizers. Our
results are clearly consistent with this explanation and suggest
a role for variation in the direction of selection as suggested by
Price in combination with ecological opportunity both on small
islands and the mainland. At present, there is insufficient phylogenetic data to test Pinto et al.’s (2008) hypothesis that the large
clade of mainland recolonizers represents an adaptive radiation.
Our results for rates of diversification in body size are consistent with a reduction in the number of competitors and a role
for unusual toe-pad evolution (Macrini et al. 2003; Velasco and
Herrel 2007; Pinto et al. 2008). However, it has also been suggested that evolution in mainland anoles is regulated by predators
whereas evolution in Caribbean anoles is regulated by intraspecific interactions (Andrews 1979; Pinto et al. 2008). Intraspecific interactions may be particularly important on small islands,
where anole population densities are often exceptionally high
(Lister 1976; Schoener and Schoener 1980; Wright 1981; Buckley
and Jetz 2007). This may explain the relatively high rates of
S I Z E D I V E R S I F I C AT I O N I N A N O L E S
diversification in sexual size dimorphism on small islands in two
ways, both related to the niche variation hypothesis (Van Valen
1965). First, high population density should, all else being equal,
increase intraspecific competition and may promote resource partitioning between the sexes if the resource base is sufficiently large
(Fitch 1981; Dayan and Simberloff 1998). Second, some islands
have a narrow resource base such that males and females cannot diverge from one another, causing sexual size dimorphism to
diminish (Lack 1947; Meiri et al. 2005). Increased trait diversification rates could therefore arise simply because species on some
small islands become more dimorphic whereas species on other
small islands become less dimorphic so the range of dimorphism
across all small-island species increases. An alternative, although
not mutually exclusive explanation, is that sexual selection that
favors larger males in the battle for breeding territories may be
intensified at high densities (Grant 1968; Stamps et al. 1997).
The importance of sexual selection relative to ecological explanations is likely to depend on the colonizing lineages. Species of the
trunk-crown and trunk-ground ecomorphs typically display more
pronounced male–male competition than other ecomorphs (Butler
et al. 2000; Butler and Losos 2002; Butler et al. 2007) and at high
densities competition may be stronger. This implies that the high
rate of diversification in SSD on small islands is at least partly
due a combination of nonrandom colonization and increased sexual selection. This is further supported because species on small
islands show, on average, more extreme male-biased dimorphism
than large-island trunk-crown and trunk-ground species.
THE MULTIPLE RATES MODEL
The multiple rate method that we introduce here is a simple extension of Thomas et al’s (2006) method for comparing rates of
phenotypic diversification. It differs from the noncensored approach of O’Meara et al. (2006) by allowing each different partition of the phylogeny to have a different phylogenetic mean. We
suggest that, contrary to Revell (2008), many hypotheses that infer different rates imply different evolutionary regimes and hence
different means. This is not a trivial distinction because assuming
a common mean can have serious consequences for the inferred
differences in rates. Our simulations (Supplementary Appendix
S5) show that models that assume a single mean (e.g., the noncensored test in O’Meara et al. 2006), but not our multiple-means
model, can infer differences in rate even if only the means differ. Where there is no difference in means, or if that difference
is small, then the common mean and multiple mean models are
similar. When should each model be used? The common mean
approach is appropriate if there is no mean difference between
groups and the interest is in a net overall difference in rate between groups, or when means differ and the interest is in any form
of rate shift. In contrast, the multiple means model is appropriate
if the interest is in a net overall difference in rate between groups
regardless of whether they differ in mean. In practice, it will often
be informative to use both to explore whether observed differences
between groups can be explained by differences in rates, means,
or both. The common mean model is a special (nested) case of
our multiple means model and they can be readily compared using maximum likelihood or AIC. In general, both the common
mean and multiple mean models should be regarded as tests for
differences in the net rate of phenotypic diversification but they
may not reflect the true rate of evolution if, for example, there
is parallel directional selection across lineages (where evolution
can be fast, but diversification slow).
Our model also differs from the most frequently used implementation of the Ornstein–Uhlenbeck (OU) model in which
different groups are allowed to have different optima but only a
single rate (Butler and King 2004). OU models are designed primarily to test for evidence of stabilizing selection and each group
has a parameter that reflects the strength of selection (sometime
referred to the “rubber band” parameter) and a single drift parameter across all groups. In principle, it is possible to fit OU
models with multiple drift parameters and explore nested models
in which both optima and rate can vary (O’Meara et al. 2006).
Our model is similar to this variant of the OU model except that
we effectively set the strength of selection parameter to zero.
CONCLUSIONS
We have shown that in the light of nonrandom island colonization
both properties of novel environments and ecological properties
characteristic of small islands influence morphological diversification rates in anoles. Ecological opportunity may be high on
small-islands as a result of a reduction in the number of competing species, most obviously the lack of the twig, trunk, crowngiant, and grass-bush ecomorphs. In contrast, on the mainland a
unique toe-pad may allow colonizing species to minimize competition with Dactyloa and hence they also have enhanced ecological opportunity (Pinto et al. 2008). Ecological opportunity
promotes morphological variation and if the direction of selection encountered by different colonizing species also varies then
traits may diverge rapidly (Price 2008; Price et al. 2009). Although
novel environments promote body size diversification in lineages
relative to their ancestral stock, the evolutionary trajectories of
males and females appear to differ depending on the properties
of those novel environments. Where species richness is low and
abundance is high (on small islands), the sexes diverge from one
another. Where species richness and potential interspecific competition is high, and abundance is presumably lower (on the mainland) body sizes diverge but the sexes evolve in parallel with one
another.
ACKNOWLEDGMENTS
We are indebted to J. Losos for invaluable advice; N. Cooper, T. Ezard,
R. Freckleton, S. Fritz, R. Grenyer, J. Hortal, W. Jetz, O. Jones, R. Lande,
EVOLUTION AUGUST 2009
2027
G AV I N H . T H O M A S E T A L .
B. Langerhans, J. Losos, L. Mahler, L. McInnes, D. Orme, A. Pigot, T.
Price, A. Purvis, and L. Revell for comments on the manuscript and/or
insightful discussion; L. Harmon and B. O’Meara for thorough and helpful
reviews; J. Losos and R. Powell for data; R. McDiarmid & J. Rosado for
help measuring anoles in museum collections; L. Butcher and B. Sanger
from the Michael Way Library for their invaluable help in data collecting;
and NERC for funding.
LITERATURE CITED
Andrews, R. M. 1979. Evolution of life histories: a comparison of Anolis
lizards from matched island and mainland habitats. Breviora 454:1–51.
Barton, N. H. 1996. Natural selection and random genetic drift as causes of
evolution on islands. Philos. Trans. R. Soc. B 351:785–795.
Bellemain, E., and R. E. Ricklefs. 2008. Are islands the end of the colonization
road? Trends Ecol. Evol. 23:461–468.
Beuttell, K., and J. B. Losos. 1999. Ecological morphology of Caribbean
anoles. Herpetol. Monogr. 13:1–28.
Blondel, J. 2000. Evolution and ecology of birds on islands: trends and
prospects. Vie et Milieu 50:205–220.
Blumstein, D. T. 2002. Moving to suburbia: ontogenetic and evolutionary
consequences of life on predator-free islands. J. Biogeogr. 29:685–692.
Buckley, L. B., and W. Jetz. 2007. Insularity and the determinants of lizard
population density. Ecol. Letts. 10:481–489.
Burnham, K. P., and D. R. Anderson. 2002. Model selection and multimodel
inference: a practical information-theoretic approach. Springer-Verlag,
New York.
Butler, M. A., and A. A. King. 2004. Phylogenetic comparative analysis: a
modeling approach for adaptive evolution. Am. Nat. 164:683–695.
Butler, M. A., and J. B. Losos. 2002. Multivariate sexual dimorphism, sexual selection, and adaptation in Greater Antillean Anolis lizards. Ecol.
Monogr. 72:541–559.
Butler, M. A., T. W. Schoener, and J. B. Losos. 2000. The relationship between
sexual size dimorphism and habitat use in Greater Antillean Anolis
lizards. Evolution 54:259–272.
Butler, M. A., S. A. Sawyer, and J. B. Losos. 2007. Sexual dimorphism and
adaptive radiation in Anolis lizards. Nature 447:202–205.
Campbell, T. S., and A. C. Echternacht. 2003. Introduced species as moving
targets: changes in body sizes of introduced lizards following experimental introductions and historical invasions. Biol. Invas. 5:193–212.
Carson, H. L., and A. R. Templeton. 1984. Genetic revolutions in relation
to speciation phenomena: the founding of new populations. Annu. Rev.
Ecol. Syst. 15:97–131.
Case, T. J. 1975. Species numbers, density compensation, and colonizing
ability of lizards on island in the Gulf of California. Ecology 56:3–18.
———. 1978. A general explanation for insular body size trends in terrestrial
vertebrates. Ecology 59:1–18.
Clegg, S. M., S. M. Degnan, C. Moritz, A. Estoup, J. Kikkiawa, and I. P.
F. Owens. 2002. Microevolution in island forms: the roles of drift and
directional selection in morphological divergence of a passerine bird.
Evolution 56:2090–2099.
Dayan, T., and D. Simberloff. 1998. Size patterns among competitors: ecological character displacement and character release in mammals, with
special reference to island populations. Mammal Rev. 28:99–124.
Diaz, L. M., N. Navarro, and O. H. Garrido. 1998. Nueva especie de Chamaeleolis (Sauria: Iguanidae) de las Meseta de Cabo Cruz, Granma, Cuba.
Avicennia 8/9:27:34.
Edwards, A. W. F. 1972. Likelihood. Cambridge Univ. Press, Cambridge.
Eldredge, N., and S. J. Gould. 1972. Punctuated equilibria: an alternative
to phyletic gradualism. Pp. 82–115 in T. J. M. Schopf, ed. Models in
paleobiology. Freeman, Cooper & Co, San Francisco.
2028
EVOLUTION AUGUST 2009
Filardi, C. E., and R. G. Moyle. 2005. Single origin of a pan-Pacific bird group
and upstream colonization of Australasia. Nature 438:216–219.
Fitch, H. S. 1981. Sexual size differences in reptiles. Univ. Kansas Mus. Nat.
Hist. Miscellaneous Publications 70:1–72.
Freckleton, R. P., P. H. Harvey, and M. Pagel. 2002. Phylogenetic analysis and
comparative data: a test and review of the evidence. Am. Nat. 125:1–
15.
Garland, T. 1992. Rate tests for phenotypic evolution using phylogenetically
independent contrasts. Am. Nat. 140:509–519.
Garrido, O. H., and S. B. Hedges. 2001. A new anole from the northern slope of
the Sierra Maestra in eastern Cuba (Squamata: Iguanidae). J. Herpetol.
35:378–383.
Glor, R. E., J. Losos, and A. Larson. 2005. Out of Cuba: overwater dispersal
and speciation among lizards in the Anolis carolinensis subgroup. Mol.
Ecol. 14:2419–2432.
Gould, S. J., and E. S. Vrba. 1982. Exaptation—a missing term in the science
of form. Paleobiology 8:4–15.
Grant, P. R. 1968. Bill size, body size, and the ecological adaptations of bird
species to competitive situations on islands. Syst. Biol. 17:319–333.
———. 1972. Convergent and divergent character displacement. Biol. J. Linn.
Soc. 4:39–68.
Hansen, T. F., and E. P. Martins. 1996. Translating between microevolutionary process and macroevolutionary patterns: the correlation structure of
interspecific data. Evolution 50:1404–1417.
Harmon, L. J., J. Melville, A. Larson, and J. B. Losos. 2008. The role of geography and ecological opportunity in the diversification of day geckos
(Phelsuma). Syst. Biol. 57:562–573.
Hedges, S. B. 2008. At the lower size limit in snakes: two new species of
threadsnakes (Squamata: Leptotyphlopidae: Leptotyphlops) from the
Lesser Antilles. Zootaxa 1841:1–30.
Herrel, A., K. Huyghe, B. Vanhooydonck, T. Backeljau, K. Breugelmans,
I. Grbac, R. Van Damme, and D. J. Irschick. 2008. Rapid large-scale
evolutionary divergence in morphology and performance associated with
exploitation of a different dietary resource. Proc. Natl. Acad. Sci. USA
105:4792–4795.
Hooijer, D. A. 1967. Indo-Australian insular elephants. Genetica 38:143–162.
Irschick, D. I., L. J. Vitt, P. A. Zani, and J. B. Losos. 1997. A comparison
of evolutionary radiations in mainland and Caribbean Anolis lizards.
Ecology 78:2191–2203.
Keogh, J. S., I. A. W. Scott, and C. Hayes. 2005. Rapid and repeated origin
of insular gigantism and dwarfism in Australian tiger snakes. Evolution
59:226–233.
Kimura, M. 1968. Evolutionary rate at the molecular level. Nature 217:624–
626.
Lack, D. L. 1947. Darwin’s finches. Cambridge Univ. Press, Cambridge.
Leinonen, T., R. B. O’Hara, J. M. Cano, and J. Merila. 2008. Comparative studies of quantitative trait and neutral marker divergence: a meta-analysis.
J. Evol. Biol. 21:1–17.
Lister, B. C. 1976. The nature of niche expansion in West Indian Anolis
lizards I: ecological consequences of reduced competition. Evolution
30:659–676.
Lister, A. M. 1989. Rapid dwarfing of red deer on Jersey in the last interglacial.
Nature 342:539–542.
Losos, J. 1994. Integrative approaches to evolutionary ecology: Anolis lizards
as model systems. Annu. Rev. Ecol. Syst. 25:467–493.
Losos, J., and K. De Queiroz. 1997. Evolutionary consequences of ecological
release in Caribbean Anolis lizards. Biol. J. Linn. Soc. 61:459–483.
Losos, J. B., T. R. Jackman, A. Larson, K. De Queiroz, and
L. Rodriguez-Schettino. 1998. Contingency and determinism in
replicated adaptive radiations of island lizards. Science 279:2115–
2118.
S I Z E D I V E R S I F I C AT I O N I N A N O L E S
Losos, J. B., M. Leal, R. E. Glor, K. De Queiroz, P. E. Hertz, L. RodriguezSchettino, L. A. Chamizo, T. R. Jackman, and A. Larson. 2003. Niche lability in the evolution of a Caribbean lizard community. Nature 424:542–
545.
Losos, J., R. E. Glor, J. J. Kolbe, and K. E. Nicholson. 2006. Adaptation,
speciation, and convergence: a hierarchical analysis of adaptive radiation
in Caribbean Anolis lizards. Ann. Mo. Bot. Gardens 93:24–33.
Lovette, I. J., E. Bermingham, and R. E. Ricklefs. 2001. Clade-specific morphological diversification and adaptive radiation in Hawaiian songbirds.
Proc. R. Soc. Lond. B 269:37–42.
MacArthur, R., J. M. Diamond, and J. R. Karr. 1972. Density compensation
in island faunas. Ecology 53:330–342.
Macrini, T. E., D. I. Irschick, and J. B. Losos. 2003. Ecomorphological differences in toepad characteristics between mainland and island anoles.
J. Herpetol. 37:52–58.
Maddison, W. P., and D. R. Maddison. 2006. StochChar: a package of Mesquite
modules for stochastic models of character evolution. Version 1.1.
———. 2007. Mesquite: a modular system for evolutionary analysis. Version
2.0. http://mesquiteproject.org.
Mayr, E. 1954. Changes of genetic environment and evolution. Pp. 157–180
in E. B. Ford, ed. Evolution as a process. Allen & Unwin, London.
McPeek, M. A. 1995. Testing hypotheses about evolutionary change on single
branches of phylogeny using evolutionary contrasts. Am. Nat. 145:686–
703.
Meiri, S. 2008. Evolution and ecology of lizard body sizes. Glob. Ecol. Biogeogr. 17:724–734.
Meiri, S., T. Dayan, and D. Simberloff. 2005. Variability and sexual size dimorphism in carnivores: testing the niche variation hypothesis. Ecology
86:1432–1440.
Merilä, J., and P. Crnokrak. 2001. Comparison of genetic differentiation at
marker loci and quantitative traits. J. Evol. Biol. 14:892–903.
Millien, V. 2006. Morphological evolution is accelerated among island mammals. PLoS Biol. 4:e321.
Mooers, A. O., S. M. Vamosi, and D. Schluter. 1999. Using phylogenies to
test macroevolutionary hypotheses of trait evolution in cranes (Gruidae).
Am. Nat. 154:249–259.
Nicholson, K. E., R. E. Glor, J. J. Kolbe, A. Larson, S. Blair Hedges, and
J. Losos. 2005. Mainland colonization by island lizards. J. Biogeogr.
32:929–938.
O’Meara, B. C., C. Ané, M. J. Sanderson, and P. C. Wainwright. 2006. Testing for different rates of continuous trait evolution using likelihood.
Evolution 60:922–933.
Pagel, M. 1997. Inferring evolutionary processes from phylogenies. Zool.
Script. 26:331–348.
Pinto, G., D. L. Mahler, L. J. Harmon, and J. B. Losos. 2008. Testing the
island effect in adaptive radiation: rates and patterns of morphological
diversification in Caribbean and mainland Anolis lizards. Proc. R. Soc.
Lond. B 275:2749–2757.
Poe, S., J. R. Goheen, and E. P. Hulebak. 2007. Convergent exaptation and
adaptation in solitary island lizards. Proc. R. Soc. Lond. B 274:2231–
2237.
Price, T. 2008. Speciation in birds. Roberts & Company Publishers, Greenwood Village, CO.
Price, T. D., A. B. Phillimore, M. Awodey, and R. Hudson. 2009. Ecological
and geographical influences on the allopatric phase of island specia-
tion. In P. R. Grant and B. R. Grant, eds. From field observations to
mechanisms. A program in evolutionary biology. Princeton Univ. Press,
Princeton, NJ. In press.
Raxworthy, C. J., M. R. J. Forstner, and R. A. Nussbaum. 2002. Chameleon
radiation by oceanic dispersal. Nature 415:784–787.
Revell, L. J. 2008. On the analysis of evolutionary change along single
branches in a phylogeny. Am. Nat. 172:140–147.
Roth, V. L. 1992. Inferences from allometry and fossils: dwarfing of elephants
on islands. Pp. 259–288 in D. Futuyama and J. Antonovics, eds. Oxford
Surveys in Evolutionary Biology, vol. 8. Oxford Univ. Press, Oxford.
Russell, I. C. 1877. The giant birds of New Zealand. Am. Nat. 11:11–
21.
Schoener, T. W. 1969. Size patterns in West Indian Anolis lizards: I. Size and
species diversity. Syst. Zool. 18:386–391.
Schoener, T. W., and A. Schoener. 1980. Densities, sex ratios, and population
structure in four species of Bahamian Anolis lizards. J. Anim. Ecol.
49:19–53.
Selander, R. K. 1966. Sexual dimorphism and differential niche utilization in
birds. Condor 68:113–151.
Simpson, G. G. 1944. Tempo and Mode in Evolution. Columbia Univ. Press,
New York.
Smith, R. J. 1999. Statistics of sexual size dimorphism. J. Hum. Evol. 36:423–
459.
Sondaar, P. Y. 1977. Insularity and its effect on mammal evolution. Pp. 671–
707 in M. K. Hecht, P. C. Goody, and B. M. Hecht, eds. Major patterns
in vertebrate evolution. Plenum, New York.
Stamps, J. A., and R. M. Andrews. 1992. Estimating asymptotic size using
the largest individuals per sample. Oecologia 92:503–512.
Stamps, J. A., J. B. Losos, and R. M. Andrews. 1997. A comparative study
of population density and sexual size dimorphism in lizards. Am. Nat.
149:64–90.
Thomas, G. H., R. P. Freckleton, and T. Székely. 2006. Comparative
analyses of the influence of developmental mode on phenotypic diversification rates in shorebirds. Proc. R. Soc. Lond. B 273:1619–
1624.
Trewick, S. A. 1997. Flightlessness and phylogeny amongst endemic rails
(Aves: Rallidae) of the New Zealand region. Philos. Trans. R. Soc. B
352:429–446.
Van Valen, L. M. 1965. Morphological variation and the width of the ecological
niche. Am. Nat. 99:377–390.
Velasco, J. A., and A. Herrel. 2007. Ecomorphology of Anolis lizards of
the Chocó region in Colombia and comparisons with Greater Antillean
ecomorphs. Biol. J. Linn. Soc. 92:29–39.
Williams, E. E. 1969. The ecology of colonization as seen in the zoogeography
of anoline lizards. Q. Rev. Biol. 44:345–389.
———. 1972. The origin of faunas. Evolution of lizard congeners in a complex
island fauna: a trial analysis. Evol. Biol. 6:47–89.
———. 1983. Ecomorphs, faunas, island size, and diverse end points in island
radiations of Anolis. Pp. 326–370 in R. B. Huey, E. R. Pianka, and T.
W. Schoener, eds. Lizard ecology: studies of a model organism. Harvard
Univ. Press, Cambridge.
Wright, S. J. 1981. Extinction-mediated competition: the Anolis lizards and
insectivorous birds of the West Indies. Am. Nat. 117:181–192.
Associate Editor: G. Hunt
EVOLUTION AUGUST 2009
2029
G AV I N H . T H O M A S E T A L .
Supporting Information
The following supporting information is available for this article:
Appendix S1. Data and Sample size (see separate tab-delimited text file).
Appendix S2. References for data sources.
Appendix S3. Species that have been subject to morphological analyses of ecomorph.
Appendix S4. Maximum likelihood reconstruction of Anolis ecomorphs.
Appendix S5. Common mean simulations.
Appendix S6. Phylogeny with ecomorph/geographic setting as node labels (see separate nexus file).
Appendix S7. Results of common mean, reduced dataset, and kappa-transformed analyses.
Appendix S8. Source code for rates analyses in R (see separate R file).
Appendix S9. Example of rates analyses (see separate R file).
Appendix S10. Multiple-rates simulations.
Supporting Information may be found in the online version of this article.
(This link will take you to the article abstract).
Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting information supplied by the
authors. Any queries (other than missing material) should be directed to the corresponding author for the article.
Additional results and discussion can be found in a document at http://www.repository.naturalis.nl/record/289893.
2030
EVOLUTION AUGUST 2009
Supplementary Appendices for: Thomas, Meiri and Phillimore Body size diversification in Anolis: novel environments and island
effects.
Contents
Supplementary Appendix S1 – data and sample size (see separate tab-delimited text
file)
Supplementary Appendix S2 – references for data sources (this file)
Supplementary Appendix S3 – species that have been subject to morphological
analyses of ecomorph (this file)
Supplementary Appendix S4 – maximum likelihood reconstruction of Anolis
ecomorphs (this file)
Supplementary Appendix S5 – common mean simulations (this file)
Supplementary Appendix S6 – phylogeny with ecomorph/geographic setting as node
labels (see separate nexus file)
Supplementary Appendix S7 – supplementary results (this file)
Supplementary Appendix S8 – source code for rates analyses in R (see separate R
file)
Supplementary Appendix S9 – example of rates analyses (see separate R file)
Supplementary Appendix S10 – multiple-rates simulations (this file)
Supplementary Appendix S1 – data and sample size
Species
A acutus
Anolis acutus
Island
type
SmallIsland
ecomorph
NA
geo
ecomorph
0
Female
SVL
49
Male
SVL
min n
females
67
A ahli
Anolis ahli
LargeIsland
TrunkGround
1
52.4
61.7
A alayoni
A alfaroi
Anolis alayoni
Anolis alfaroi
LargeIsland
LargeIsland
Twig
GrassBush
2
2
38.9
33
46.8
36
A aliniger
Anolis aliniger
LargeIsland
TrunkCrown
1
57
60
A allisoni
Anolis allisoni
LargeIsland
TrunkCrown
1
75
A allogus
A altae
Anolis allogus
Anolis altae
LargeIsland
Mainland
TrunkGround
NA
1
3
A altitudi
Anolis altitudinalis
LargeIsland
TrunkCrown
A alumina
Anolis alumina
LargeIsland
GrassBush
min n
males
37
76
3
n females
37 (Lazell 1972)
n males
refs
notes
76 (Lazell 1972)
Clobert et al. 1998, Dunham and
Miles 1985, Stamps et al. 1997,
Perry and Garland 2002,
Roughgarden 1995, Cox et al. 2003,
Schwartz and Henderson 1991,
Andrews and Rand 1974, Lazell
1972, Stamps and Andrews 1992,
Dunham et al. 1988,
NA
Fitch 1981, Schettino 1999,
Schwartz and Henderson 1991
NA
Schettino 1999, Uetz 2006
Schettino 1999
NA
NA
8
3 (Schettino 1999)
16
13 (Schettino 1999)
NA
8 (Schettino 1999)
16 (Schettino
1999)
NA
4
10
4 (Butler et al. 2000)
10 (Butler et al.
2000)
100
88
170
88 (Schoener 1970),
16 (Schettino 1999)
170 (Schoener
1970), 27
(Schettino 1999)
49
52
62.8
52
367
1
777
NA
367 (Schoener
1970), 10 (Schettino
1999)
1 (Taylor 1956)
777 (Schoener
1970), 10
(Schettino 1999)
NA
1
51
52
NA
NA
NA
NA
2
37
40
NA
NA
13
NA
NA
A
alutaceu
Anolis alutaceus
LargeIsland
GrassBush
2
37
37.5
334
295
A angustic
Anolis angusticeps
LargeIsland
Twig
2
47
53
60
41
NA
NA
334 (Schoener
1970), 8 (Schettino
1999)
295 (Schoener
1970), 9
(Schettino 1999)
60 (Butler et al.
2000)
40 (Schoener
1970), 41 (Butler
et al. 2000)
Schwartz and Henderson 1991,
Williams 1983, Butler et al. 2000
Schoener 1970, Fitch 1981,
Schettino 1999, Schwartz and
Henderson 1991, Herrel et al. 2004,
Butler et al. 2000, McCranie et al.
2005
Schoener 1970, Fitch 1981,
Schettino 1999, Schwartz and
Henderson 1991, Butler et al. 2000,
Rogner 1997
Savage 2002, Taylor 1956
Schettino 1999, Schwartz and
Henderson 1991
Fitch 1981, Schwartz and
Henderson 1991, Williams 1983
Schoener 1970, Fitch 1981,
Schettino 1999, Schwartz and
Henderson 1991, Herrel et al. 2004,
Butler et al. 2000
Schoener 1970, Stamps et al. 1997,
Fitch 1981, Perry and Garland 2002,
Cox et al. 2003, Schettino 1999,
Schwartz and Henderson 1991,
Herrel et al. 2004, Butler et al. 2000,
NA
NA
NA
NA
Sometimes
considered a
subspecies of
isolepis but data
from Garrido and
Hedges suggest
that alt is the
larger of the two.
No data can be
confidently
assigned to
isolepis rather
than altitudinalis
so we omit
isolepis
NA
NA
NA
Species
Island
type
ecomorph
geo
ecomorph
Female
SVL
Male
SVL
min n
females
min n
males
n females
n males
refs
notes
Uetz 2006
A
annecten
A
aquaticu
A
argenteo
A argillac
A armouri
Anolis annectens
Anolis aquaticus
Anolis argenteolus
Anolis argillaceus
Anolis armouri
Mainland
Mainland
LargeIsland
LargeIsland
LargeIsland
NA
NA
Unique
Unique
TrunkGround
3
3
2
2
1
63.3
62
51
44.8
56
77.6
9
71
10
59.8
12
62
46.2
67
10
23
NA
A auratus
Anolis auratus
Mainland
NA
3
57
51
A
bahoruco
A
baleatus
A
baracoae
Anolis
bahorucoensis
LargeIsland
GrassBush
2
44
51
NA
Anolis baleatus
LargeIsland
CrownGiant
2
148
180
NA
Anolis baracoae
LargeIsland
CrownGiant
2
155
172
A
barahona
Anolis barahonae
LargeIsland
CrownGiant
2
148
160
A
barbatus
Anolis barbatus
LargeIsland
Unique
2
157
170
A barbouri
Anolis barbouri
LargeIsland
Unique
2
55
44
25
62 (Schettino 1999)
54
23 (Schoener 1970),
19 (Schettino 1999)
NA
21
NA
45
NA
NA
10
NA
10
NA
2
NA
2
NA
A bartschi
Anolis bartschi
LargeIsland
Unique
2
63.6
80
31
25
A
bimacula
Anolis bimaculatus
SmallIsland
NA
0
70
123
77
113
A biporcat
Anolis biporcatus
Mainland
NA
3
105
105
19
9 (Barros et al. 2007)
9 (Fitch 1976, Fitch
1981), 1 (Taylor
1956), 10 (Fitch et al.
1976)
24
1 (Williams 1974),
10 (Barros et al.
2007)
10 (Fitch 1976,
Fitch 1981), 2
(Taylor 1956), 12
(Fitch et al. 1976)
Williams 1974, Barros et al. 2007
NA
Fitch 1976, Fitch 1981, Cox et al.
2003, Savage 2002, Fitch and Hillis
1984, Taylor 1956, Fitch et al. 1976
NA
52 (Schettino
1999)
54 (Schoener
1970), 18
(Schettino 1999)
Schettino 1999, Schwartz and
Henderson 1991, Rogner 1997
Schoener 1970, Fitch 1981,
Schettino 1999, Schwartz and
Henderson 1991
Schwartz and Henderson 1991,
Herrel et al. 2004, Williams 1983
19 (Fitch 1981), 8
(Fitch 1976), 15 (Vitt
and de Carvalho
1995), 10+11
(Hoogmoed 1973)
NA
45 (Fitch 1981),
31 (Fitch 1976),
21 (Vitt and de
Carvalho 1995),
7+14 (Hoogmoed
1973), 1
(Boulenger 1896)
NA
NA
NA
10 (Schettino 1999)
NA
10 (Schettino
1999)
NA
NA
2 (Leal and Losos
2000)
2 (Leal and Losos
2000)
NA
NA
31 (Schettino 1999)
25 (Schettino
1999)
77 (Lazell 1972)
113 (Lazell 1972)
Stamps et al. 1997, Fitch 1976, Fitch
1981, Cox et al. 2003, Avila-Pires
1995, Vitt and de Carvalho 1995,
Andrews and Rand 1974,
Hoogmoed 1973, Herrel et al. 2004,
Boulenger 1896
Fitch 1981, Schwartz and
Henderson 1991, Rogner 1997,
Herrel et al. 2004, Williams 1983
Fitch 1981, Schwartz and
Henderson 1991, Williams 1983
Schettino 1999, Schwartz and
Henderson 1991
Fitch 1981, Schwartz and
Henderson 1991, Herrel et al. 2004,
Williams 1983
Schettino 1999, Schwartz and
Henderson 1991, Herrel et al. 2004,
Leal and Losos 2000
Schwartz and Henderson 1991,
Herrel et al. 2004, Losos Pers.
Comm
Schettino 1999, Schwartz and
Henderson 1991, Rogner 1997,
Losos Pers. Comm
Stamps et al. 1997, Fitch 1981,
Roughgarden 1995, Schwartz and
Henderson 1991, Andrews and
Rand 1974, Herrel et al. 2004, Lazell
1972, Stamps and Andrews 1992,
Powell et al. 2005
19 (Fitch 1981), 2
(Guyer and Donnelly
2005)
24 (Fitch 1981), 1
(Ruthven 1916), 2
(Guyer and
Donnelly 2005)
Fitch 1976, Fitch 1981, Campbell
1999, Andrews and Rand 1974,
Herrel et al. 2004, Ruthven 1916,
Guyer and Donnelly 2005
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Species
A bitectus
Anolis bitectus
Island
type
Mainland
ecomorph
NA
geo
ecomorph
3
Female
SVL
54.7
Male
SVL
min n
females
55.8
min n
males
2
2
n females
n males
refs
notes
2 (Meiri,
unpublished)
2 (Meiri,
unpublished)
Meiri, unpublished
NA
9 (Fitch 1981)
9 (Fitch 1981)
Fitch 1981, Schettino 1999,
Schwartz and Henderson 1991
15 (Schwartz 1980)
27 (Schwartz
1980)
A bremeri
Anolis bremeri
LargeIsland
TrunkGround
1
52.3
72
9
9
A breslini
Anolis breslini
LargeIsland
TrunkGround
1
45
60
15
27
A breviros
A
brunnneu
Anolis brevirostris
LargeIsland
Trunk
2
45
51
175
451
175 (Schoener 1970)
451 (Schoener
1970)
Anolis brunneus
SmallIsland
TrunkCrown
0
70
76
A capito
Anolis capito
Mainland
NA
3
97
91
15
NA
13 (Fitch 1981), 4
(Taylor 1956), 12
(Guyer and Donnelly
2005)
NA
13 (Fitch 1981), 3
(Taylor 1956), 15
(Guyer and
Donnelly 2005)
A caroline
Anolis carolinensis
Mainland
NA
3
57.5
NA
NA
13
71
12
47
14 (Fitch 1981, Fitch
et al. 1976), 8 (Guyer
and Donnelly 2005)
14 (Fitch 1976),
47 (Gerber and
Echternacht
2000), 1
(Boulenger 1885)
6 (Fitch 1981,
Fitch et al. 1976),
6 (Guyer and
Donnelly 2005)
12 (Fitch 1976), 1
(Boulenger 1885)
Schwartz 1980
Schoener 1970, Fitch 1981,
Schwartz and Henderson 1991,
Williams 1983, Butler et al. 2000
Schwartz and Henderson 1991
Stamps et al. 1997, Fitch 1976, Fitch
1981, Savage 2002, Herrel et al.
2004, Taylor 1956, Guyer and
Donnelly 2005
Tinkle et al. 1970, Dunham and
Miles 1985, Stamps et al. 1997,
Fitch 1970, 1981, Perry and Garland
2002, Cox et al. 2003, Andrews and
Rand 1974, Gerber and Echternacht
2000, Boulenger 1885, Dunham et
al. 1988
Fitch 1976, Fitch 1981, Fitch and
Hillis 1984, Herrel et al. 2004, Guyer
and Donnelly 2005, Fitch et al. 1976
Schwartz and Henderson 1991,
Williams 1983
Schettino 1999, Schwartz and
Henderson 1991
Schoener 1970, Fitch 1981,
Schettino 1999, Schwartz and
Henderson 1991
Schoener 1970, Fitch 1981,
Schwartz and Henderson 1991,
Williams 1983, Butler et al. 2000
Schoener 1970, Fitch 1981,
Schwartz and Henderson 1991,
Williams 1983
Schettino 1999, Schwartz and
Henderson 1991
Schoener 1970, Fitch 1981,
Schwartz and Henderson 1991,
Herrel et al. 2004, Williams 1983,
Butler et al. 2000
A
carpente
Anolis carpenteri
Mainland
NA
3
45
41
A caudalis
Anolis caudalis
LargeIsland
Trunk
2
44
48
NA
NA
NA
NA
A centrali
Anolis centralis
LargeIsland
Unique
2
46
47.2
NA
NA
NA
NA
A
chamaele
Anolis
chamaeleonides
79 (Schoener), 33
(Fitch 1981)
42 (Schoener), 12
(Fitch 1981)
A chlorocy
Anolis
chlorocyanus
LargeIsland
TrunkCrown
1
54.8
80
197
396
197 (Schoener 1970)
396 (Schoener
1970)
A christop
Anolis christophei
LargeIsland
Unique
2
45
49
33
27
33 (Schoener)
27 (Schoener)
A clivicol
Anolis clivicola
LargeIsland
GrassBush
2
45
49.4
10
9
10 (Schettino 1999)
9 (Schettino 1999)
A coelesti
A
confusus
Anolis coelestinus
LargeIsland
TrunkCrown
1
60.3
84
572
1242
Anolis confusus
LargeIsland
TrunkGround
1
48
53
3
12
3 (Schettino 1999)
1242 (Schoener
1970)
12 (Schettino
1999)
Anolis conspersus
LargeIsland
TrunkCrown
1
47
76
35
NA
35 (Gerber and
Echternacht 2000)
Schettino 1999
Cox et al. 2003, Gerber and
Echternacht 2000, Licht and Gorman
1970
6 (Butler et al. 2000)
22 (Butler et al.
2000)
Fitch 1981, Roughgarden 1995,
Schwartz and Henderson 1991,
A
conspers
A cooki
Anolis cooki
LargeIsland
LargeIsland
Unique
TrunkGround
2
1
172
59
14
177
70
79
NA
6
6
42
22
572 (Schoener 1970)
NA
Sometimes
considered a
subspecies of
whitemani but
distinguished by
Schwartz (1980)
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Species
Island
type
ecomorph
geo
ecomorph
Female
SVL
Male
SVL
min n
females
min n
males
n females
n males
refs
notes
Williams 1983, Butler et al. 2000
A crassulu
Anolis crassulus
Mainland
NA
3
56
53
7
8
A cristate
A
cupeyale
Anolis cristatellus
Anolis
cupeyalensis
LargeIsland
TrunkGround
1
76
75
498
876
LargeIsland
GrassBush
2
31
33
9
10
A cupreus
Anolis cupreus
Mainland
NA
3
51
57
461
657
A cuvieri
A
cyanople
A cybotes
A darlingt
A
desechen
A distichu
A dolichoc
A equestri
Anolis cuvieri
Anolis
cyanopleurus
Anolis cybotes
LargeIsland
CrownGiant
2
135
137
LargeIsland
GrassBush
2
36.2
43
LargeIsland
TrunkGround
1
Anolis darlingtoni
LargeIsland
Twig
2
Anolis desechensis
SmallIsland
TrunkGround
0
Anolis distichus
Anolis
dolichocephalus
LargeIsland
Trunk
LargeIsland
LargeIsland
66
NA
16
NA
81
27
NA
148
74
NA
45
57
NA
2
48
58
GrassBush
2
52
51
CrownGiant
2
170
190
245
3
NA
616
NA
1022
NA
385
468
7 (Fitch 1976, Fitch
1981), 6 (McCranie
et al. 1992)
498 (Schoener
1970), 19 (Butler and
Losos 2002)
8 (Fitch 1976,
Fitch 1981), 3
(McCranie et al.
1992)
9 (Schettino 1999)
876 (Schoener
1970), 20 (Butler
and Losos 2002)
10 (Schettino
1999)
461 (Fitch 1981),
428 (Fitch 1976)
657 (Fitch 1981),
640 (Fitch 1976)
16 (Schoener 1970),
3 (Butler and Losos
2002)
27 (Schoener
1970), 6 (Butler
and Losos 2002)
NA
NA
148 (Schoener
1970), 133 (Fitch),
27 (Fobes et al.
1992)
245 (Schoener
1970), 230 (Fitch
1981), 18 (Fobes
et al. 1992)
NA
3 (Thomas and
Hedges 1991)
NA
NA
616 (Schoener 1970)
1022 (Schoener
1970)
NA
NA
385 (Schoener 1970)
468 (Schoener
1970)
A ernestwi
Anolis equestris
Anolis
ernestwilliamsi
SmallIsland
TrunkGround
0
60
82
NA
NA
NA
NA
A etheridg
A
Anolis etheridgei
Anolis
LargeIsland
LargeIsland
Unique
Unique
2
2
43
61
43
72
NA
NA
NA
NA
NA
NA
NA
NA
Fitch 1976, Fitch 1981, Cox et al.
2003, Kohler et al. 2006, McCranie
et al. 1992
Schoener 1970, Fitch 1981, Perry
and Garland 2002, Butler and Losos
2002, Roughgarden 1995, Savage
2002, Schwartz and Henderson
1991, Herrel et al. 2004, Smith
1934b, Williams 1983, Butler et al.
2000
Schettino 1999, Schwartz and
Henderson 1991
Clobert et al. 1998, Stamps et al.
1997, Fitch 1973a, 1976, 1981,
Perry and Garland 2002, Cox et al.
2003, Savage 2002, Dunham et al.
1988, Fitch and Hillis 1984
Schoener 1970, Butler and Losos
2002, Schwartz and Henderson
1991, Andrews and Rand 1974,
Herrel et al. 2004, Butler et al. 2000,
Perry and garland 2002,
Roughgarden 1995, Williams 1983
Schettino 1999, Schwartz and
Henderson 1991
Schoener 1970, Fitch 1981,
Schwartz and Henderson 1991, Licht
and Gorman 1970, Fobes et al.
1992, Herrel et al. 2004, Williams
1983, Butler et al. 2000
Thomas and Hedges 1991,
Schwartz and Henderson 1991,
Williams 1983
Schwartz and Henderson 1991
Schoener 1970, Stamps et al. 1997,
Fitch 1981, Perry and Garland 2002,
Cox et al. 2003, Schwartz and
Henderson 1991, Rogner 1997,
Herrel et al. 2004, Williams 1983,
Stamps and Andrews 1992, Butler et
al. 2000
Fitch 1981, Schwartz and
Henderson 1991, Williams 1983
Schoener 1970, Fitch 1981,
Schettino 1999, Schwartz and
Henderson 1991, Dalrymple 1980,
Herrel et al. 2004, Butler et al. 2000
Roughgarden 1995, Schwartz and
Henderson 1991
Schwartz and Henderson 1991,
Williams 1983
Schwartz and Henderson 1991,
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Species
Island
type
ecomorph
geo
ecomorph
Female
SVL
Male
SVL
min n
females
eugenegr
eugenegrahami
A
evermann
Anolis evermanni
LargeIsland
TrunkCrown
1
52.4
78
A ferreus
Anolis ferreus
SmallIsland
NA
0
65
119
NA
A fowleri
Anolis fowleri
LargeIsland
Unique
2
75
77
NA
A
fuscoaur
Anolis fuscoauratus
Mainland
NA
3
52
min n
males
n males
refs
notes
19 (Butler and Losos
2002)
17 (Butler and
Losos 2002)
NA
NA
NA
NA
NA
NA
47
22 (Fitch 1981), 10
(Duellman and
Mendelson 1995), 11
(Hoogmoed 1973),
15+12 (Duellman
2005), 1 (Lotzkat
2007), 108 (Vitt et al.
2003)
NA
21 (Fitch 1981), 6
(Duellman and
Mendelson 1995),
11 (Hoogmoed
1973), 7
(Duellman 2005),
1 (Lotzkat 2007),
47 (Vitt et al.
2003)
Williams 1983
Losos 1990, Fitch 1981, Butler and
Losos 2002, Roughgarden 1995,
Schwartz and Henderson 1991,
Herrel et al. 2004, Williams 1983,
Butler et al. 2000
Roughgarden 1995, Schwartz and
Henderson 1991, Herrel et al. 2004,
Lazell 1972, Lazell 1964
Schwartz and Henderson 1991,
Williams 1983
68 (Schoener 1970),
12 (Butler and Losos
2002), 7 (Herrel et al.
2004)
3 (Schettino 1999)
86 (Schoener
1970), 10 (Butler
and Losos 2002),
6 (Herrel et al.
2004)
5 (Schettino 1999)
61 (Lazell 1972)
111 (Lazell 1972)
19
49
17
108
A garmani
A garridoi
Anolis garmani
Anolis garridoi
LargeIsland
LargeIsland
CrownGiant
Twig
2
2
95
36.8
132
41.8
68
3
86
5
A gingivin
Anolis gingivinus
SmallIsland
NA
0
55
72
61
111
n females
A grahami
Anolis grahami
LargeIsland
TrunkCrown
1
64
75
21
23
21 (Butler and Losos
2002), 1 (Boulenger
1885), 17 (Herrel et
al. 2004)
A guafe
A
guamuhay
A
guazuma
Anolis guafe
LargeIsland
TrunkGround
1
40
48.8
13
24
13 (Schettino 1999)
18 (Butler and
Losos 2002), 1
(Boulenger 1885),
23 (Herrel et al.
2004)
24 (Schettino
1999)
1 (Schettino 1999)
NA
3 (Schettino 1999)
7 (Schettino 1999)
19 (Butler and Losos
2002)
18 (Butler and
Losos 2002)
NA
NA
128 (Schoener 1970)
203 (Schoener
1970?)
A
gundlach
A
haetianu
A
henderso
Anolis guamuhaya
LargeIsland
Unique
2
162
Anolis guazuma
LargeIsland
Twig
2
41
48.5
3
7
Anolis gundlachi
LargeIsland
TrunkGround
1
52
75
19
18
Anolis haetianus
LargeIsland
TrunkGround
1
60
75
Anolis hendersoni
LargeIsland
GrassBush
2
42.6
NA
1
49.3
NA
NA
NA
128
203
Fitch 1976, Fitch 1981, Duellman
and Mendelson 1995, Avila-Pires
1995, Duellman 1978, Andrews and
Rand 1974, Dixon and Soini 1986,
Hoogmoed 1973, Duellman 2005,
Herrel et al. 2004, Lotzkat 2007, Vitt
et al. 2003, Lotzkat 2007
Schoener 1970, Stamps et al. 1997,
Fitch 1981, Trivers 1976, Butler and
Losos 2002, Schwartz and
Henderson 1991, Rogner 1997,
Butler et al. 2000, Losos 1990,
Herrel et al. 2004, Williams 1983
Schettino 1999
Roughgarden 1995, Schwartz and
Henderson 1991, Herrel et al. 2004,
Lazell 1972, Stamps and Andrews
1992, Powell et al. 2005
Fitch 1981, Butler and Losos 2002,
Cox et al. 2003, Schwartz and
Henderson 1991, Andrews and
Rand 1974, Boulenger 1885, Licht
and Gorman 1970, Rogner 1997,
Herrel et al. 2004, Williams 1983,
Butler et al. 2000
NA
NA
NA
NA
NA
NA
NA
Schettino 1999
NA
Schettino 1999
Schettino 1999, Schwartz and
Henderson 1991
Losos 1990, Fitch 1981, Butler and
Losos 2002, Roughgarden 1995,
Schwartz and Henderson 1991,
Andrews and Rand 1974, Rogner
1997, Herrel et al. 2004, Williams
1983, Stamps and Andrews 1992,
Butler et al. 2000
NA
Schwartz and Henderson 1991
Schoener 1970, Fitch 1981,
Schwartz and Henderson 1991,
Williams 1983, Butler et al. 2000
NA
NA
NA
NA
Species
A
homolech
A humilis
Anolis homolechis
Anolis humilis
Island
type
LargeIsland
Mainland
ecomorph
TrunkGround
geo
ecomorph
1
Female
SVL
55.8
Male
SVL
70
NA
3
44
45
min n
females
NA
min n
males
NA
222
216
n females
n males
NA
NA
106 (Fitch), 1 (Taylor
1956), 222 (Guyer
and Donnelly 2005)
155 (Fitch 1981),
29 (Fitch and Hillis
1984), 1 (Taylor
1956), 216 (Guyer
and Donnelly
2005)
A imias
Anolis imias
LargeIsland
TrunkGround
1
46.5
67.4
2
5
2 (Schettino 1999)
A inexpect
Anolis inexpectatus
LargeIsland
GrassBush
2
35
37
15
16
15 (Schettino 1999)
5 (Schettino 1999)
16 (Schettino
1999)
19 (Butler et al.
2000)
20 (Butler et al.
2000)
A insolitu
Anolis insolitus
LargeIsland
Twig
2
44
47
19
20
A
intermed
Anolis intermedius
Mainland
NA
3
54
54
98
241
A isthmicu
Anolis isthmicus
Mainland
NA
3
58
63
9
25
A jubar
Anolis jubar
LargeIsland
TrunkGround
1
53.2
62
A krugi
Anolis krugi
LargeIsland
GrassBush
2
39.3
55
19
A laeviven
Anolis laeviventris
Mainland
NA
3
65
61
A leachi
Anolis leachii
SmallIsland
NA
0
70
123
A
lemurinu
A limifron
Anolis lemurinus
Anolis limifrons
Mainland
Mainland
NA
NA
3
3
79
51
79
51
98 (Fitch), 3 (Taylor
1956)
9 (Fitch 1981), 3
(Fitch 1978)
241 (Fitch 1981)
25 (Fitch 1981), 6
(Fitch and Hillis
1984), 8 (Fitch
1978)
NA
NA
18
19 (Butler and Losos
2002)
13
24
13 (Fitch 1976)
18 (Butler and
Losos 2002)
24 (Fitch 1976), 1
(Barbour
1932a&b)
41
47
41 (Lazell 1972)
13
16 (Fitch 1976), 15
(Guyer and Donnelly
2005)
47 (Lazell 1972)
13 (Fitch 1976), 3
(Fitch and Hillis
1984), 10 (Guyer
and Donnelly
2005)
284 (Fitch 1981), 2
(Taylor 1956), 46
(Guyer and Donnelly
2005), 21 (Kohler
and Sunyer 2008),
33 (Fitch et al. 1976)
400 (Fitch 1981),
65 (Fitch and Hillis
1984), 1 (Taylor
1956), 102 (Guyer
and Donnelly
2005), 50 (Kohler
and Sunyer 2008),
42 (Fitch et al.
1976)
NA
NA
16
284
400
refs
Schoener 1970, Fitch 1981,
Schettino 1999, Schwartz and
Henderson 1991, Butler et al. 2000
Stamps et al. 1997, Fitch 1973a,
1973b, 1976, 1981, Savage 2002,
Andrews 1979, Andrews and Rand
1974, Herrel et al. 2004, Taylor
1956, Stamps and Andrews 1992,
Vitt et al. 2002, Guyer and Donnelly
2005, Dunham et al. 1988, Kohler et
al. 2006, Fitch and Hillis 1984
Schettino 1999, Schwartz and
Henderson 1991
Schettino 1999, Schwartz and
Henderson 1991
Schwartz and Henderson 1991,
Williams 1983, Butler et al. 2000
Clobert et al. 1998, Fitch 1973a,
1976, 1981, Savage 2002, Taylor
1956, Fitch and Hillis 1984
Fitch 1981, Fitch and Hillis 1984,
Fitch 1978
Schettino 1999, Schwartz and
Henderson 1991
Losos 1990, Fitch 1981, Butler and
Losos 2002, Roughgarden 1995,
Schwartz and Henderson 1991,
Herrel et al. 2004, Williams 1983,
Butler et al. 2000
Fitch 1976, Barbour 1932a&b
Roughgarden 1995, Lazell 1972,
Kolbe et al. 2008
Fitch 1976, Savage 2002, Andrews
and Rand 1974, Fitch and Hillis
1984, Herrel et al. 2004, Klutsch et
al. 2007, McCranie et al. 2005,
Guyer and Donnelly 2005
Tinkle et al. 1970, Clobert et al.
1998, Fitch 1973a, 1973b, 1976,
1981, Perry and Garland 2002,
Andrews 1976, Cox et al. 2003,
Savage 2002, Andrews 1979,
Andrews and Rand 1974, Herrel et
al. 2004, Taylor 1956, Stamps and
Andrews 1992, Guyer and Donnelly
2005, Dunham et al. 1988, Kohler
and Sunyer 2008, Fitch et al. 1976,
Stamps et al. 1997, Smith 1981
notes
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Species
Island
type
ecomorph
geo
ecomorph
Female
SVL
Male
SVL
min n
females
min n
males
n females
n males
refs
21 (Butler and
Losos 2002), 58
(Herrel et al.
2004)
Losos 1990, Stamps et al. 1997,
Fitch 1981, Perry and Garland 2002,
Butler and Losos 2002, Cox et al.
2003, Schwartz and Henderson
1991, Andrews and Rand 1974,
Licht and Gorman 1970, Herrel et al.
2004, Williams 1983, Butler et al.
2000, Losos Pers. Comm
A lineatop
Anolis lineatopus
LargeIsland
TrunkGround
1
51
73
45
58
24 (Butler and Losos
2002), 45 (Herrel et
al. 2004)
A lionotus
Anolis lionotus
Mainland
NA
3
70
76
44
59
44 (Campbell 1973)
59 (Campbell
1973)
A lividus
A longicep
Anolis lividus
Anolis longiceps
SmallIsland
SmallIsland
NA
TrunkCrown
0
0
55
76
70
83
64
NA
44 (Lazell 1972)
NA
64 (Lazell 1972)
NA
A longitib
A loveridg
Anolis longitibialis
Anolis loveridgei
LargeIsland
Mainland
TrunkGround
NA
1
3
59
72
117.9
NA
NA
NA
2 (McCranie 1992)
A loysiana
Anolis loysiana
LargeIsland
Trunk
2
204 (Schoener
1970), 16 (Schettino
1999), 1 (Boulenger
1885)
NA
NA
57 (Schoener
1970), 11
(Schettino 1999)
298 (Schoener
1970), 17
(Schettino 1999),
1 (Boulenger
1885)
NA
38.2
44
NA
NA
2
47.2
40
57
40 (Schoener), 10
(Schettino 1999)
Campbell 1973
Fitch 1981, Roughgarden 1995,
Schwartz and Henderson 1991,
Andrews and Rand 1974, Herrel et
al. 2004, Lazell 1972
Schwartz and Henderson 1991
Schwartz and Henderson 1991,
Williams 1983
McCranie 1992
Schoener 1970, Fitch 1981,
Schettino 1999, Schwartz and
Henderson 1991, Butler et al. 2000
notes
NA
A. lionotus
measurements
from Panama (as
suggested by
Savage), all other
"lionotus"
measurements
are taken to be
from oxylophus
NA
NA
NA
NA
NA
A lucius
Anolis lucius
LargeIsland
Unique
2
60
70
A luteogul
A macilent
A
marcanoi
Anolis luteogularis
Anolis macilentus
LargeIsland
LargeIsland
CrownGiant
GrassBush
2
2
176
36
191
41
NA
NA
NA
NA
NA
NA
NA
NA
Anolis marcanoi
LargeIsland
TrunkGround
1
49
65
NA
NA
NA
NA
A
marmorat
Anolis marmoratus
SmallIsland
NA
0
57
82
NA
NA
NA
NA
A marron
A
maynardi
Anolis marron
LargeIsland
Trunk
2
42
50
NA
NA
NA
NA
Schoener 1970, Fitch 1981,
Schettino 1999, Schwartz and
Henderson 1991, Boulenger 1885,
Rogner 1997
Schettino 1999, Schwartz and
Henderson 1991
Schettino 1999
Schwartz and Henderson 1991,
Williams 1983
Roughgarden 1995, Schwartz and
Henderson 1991, Rogner 1997,
Lazell 1972, Lazell 1964
Schwartz and Henderson 1991,
Williams 1983
Anolis maynardi
SmallIsland
TrunkCrown
0
48
76
NA
NA
NA
NA
Schwartz and Henderson 1991
NA
A
meridion
Anolis meridionalis
Mainland
NA
3
46.5
48
9 (Vitt and Caldwell
1993)
3 (Vitt and
Caldwell 1993)
Vitt and Caldwell 1993
NA
204
298
9
3
A mestrei
Anolis mestrei
LargeIsland
TrunkGround
1
48.5
56.5
10
8
A
monensis
Anolis monensis
SmallIsland
TrunkGround
0
46
59
34
24
A monticol
Anolis monticola
LargeIsland
Unique
2
42
56
NA
NA
10 (Schettino 1999)
8 (Schettino 1999)
34 (Losos Pers.
Comm.)
24 (Losos Pers.
Comm.)
NA
NA
Fitch 1981, Schettino 1999,
Schwartz and Henderson 1991
Schwartz and Henderson 1991,
Herrel et al. 2004, Losos Pers.
Comm, Powell pers.com.
Fitch 1981, Schwartz and
Henderson 1991, Williams 1983,
Losos Pers. Comm
NA
NA
NA
NA
NA
NA
NA
NA
NA
Species
A
nebuloid
Anolis nebuloides
Island
type
Mainland
ecomorph
NA
geo
ecomorph
3
Female
SVL
Male
SVL
44
min n
females
55.5
min n
males
4
7
A nitens
Anolis nitens
Mainland
NA
3
85
83
A noblei
Anolis noblei
LargeIsland
CrownGiant
2
172.1
190
A nubilus
Anolis nubilis
SmallIsland
NA
0
52
81
6
19
A occultus
Anolis occultus
LargeIsland
Twig
2
39.2
40
10
4
A oculatus
Anolis oculatus
SmallIsland
NA
0
65
96
A olssoni
Anolis olssoni
LargeIsland
GrassBush
2
44
50
A onca
Anolis onca
Mainland
NA
3
67
75
A opalinus
Anolis opalinus
LargeIsland
TrunkCrown
1
46
56
A
ophiolep
A
oporinus
Anolis ophiolepis
LargeIsland
GrassBush
2
39.5
Anolis oporinus
LargeIsland
TrunkCrown
1
46.7
39.8
NA
36
NA
63
NA
NA
NA
97
114
NA
1
118
232
20
1
10
NA
n females
4 (Smith 1934)
1 (Duellman and
Mendelson 1995),
(Lotzkat 2007), 36
(Vitt et al. 2008), 1
(dos Santos et al.
2007)
refs
notes
2 (Fitch and Hillis
1984), 3 (Fitch
1978), 7 (Smith
1934)
Fitch and Hillis 1984, Smith 1934
NA
2 (Duellman and
Mendelson 1995),
(Lotzkat 2007), 63
(Vitt et al. 2008)
NA
NA
6 (Lazell 1972)
19 (Lazell 1972)
10 (Butler and Losos
2002)
4 (Butler and
Losos 2002)
NA
NA
NA
114 (Schoener
1970)
1 (Boulenger
1885)
118 (Schoener), 84
(Fitch), 21 (Butler
and Losos 2002)
232 (Schoener),
18 (Butler and
Losos 2002)
97 (Schoener 1970)
20 (Schettino 1999)
1 (Garrido and
Hedges 2001)
Mainland
NA
3
52
57
8
8
A oscellos
Anolis ortonii
Anolis
ocelloscapularis
8 (Fitch 1976), 5
(Hoogmoed 1973),
1+2 (Duellman 2005)
Mainland
NA
3
46.5
42
5
1
5 (Kohler et al. 2001)
A
oxylophu
A
Anolis oxylophus
Anolis pachypus
Mainland
Mainland
NA
NA
3
3
72
50
85
50
24
24
19
32
24 (Fitch), 14 (Guyer
and Donnelly 2005)
24 (Fitch 1981)
A ortonii
n males
10 (Schettino
1999)
NA
8 (Fitch 1976), 4
(Hoogmoed
1973), 2
(Duellman 2005)
1 (Kohler et al.
2001)
19 (Fitch 1981), 3
(Fitch and Hillis
1984), 1
(Boulenger 1894),
13 (Guyer and
Donnelly 2005)
32 (Fitch 1981)
Duellman and Mendelson 1995, Vitt
and Zani 1998, Herrel et al. 2004,
Lotzkat 2007, Vitt et al. 2002, Vitt et
al. 2008, dos Santos et al. 2007
Schettino 1999, Schwartz and
Henderson 1991
Fitch 1981, Roughgarden 1995,
Schwartz and Henderson 1991,
Lazell 1972
Losos 1990, Fitch 1981, Butler and
Losos 2002, Roughgarden 1995,
Schwartz and Henderson 1991,
Williams 1983, Butler et al. 2000
Stamps et al. 1997, Fitch 1981,
Andrews 1976, Roughgarden 1995,
Andrews 1979, Schwartz and
Henderson 1991, Andrews and
Rand 1974, Rogner 1997, Herrel et
al. 2004, Lazell 1972, Dunham et al.
1988
Schoener 1970, Fitch 1981,
Schwartz and Henderson 1991,
Herrel et al. 2004, Williams 1983,
Butler et al. 2000
Andrews and Rand 1974, Boulenger
1885
Schoener 1970, Losos 1990, Fitch
1981, Butler and Losos 2002, Cox et
al. 2003, Schwartz and Henderson
1991, Andrews and Rand 1974,
Rogner 1997, Herrel et al. 2004,
Williams 1983, Butler et al. 2000
Schettino 1999, Schwartz and
Henderson 1991, Herrel et al. 2004,
Butler et al. 2000
NA
NA
NA
NA
NA
NA
NA
NA
NA
Garrido and Hedges 2001
Cox et al. 2003, Fitch 1976, AvilaPires 1995, Duellman 1978, Dixon
and Soini 1986, Hoogmoed 1973,
Duellman 2005, Herrel et al. 2004
NA
NA
Kohler et al. 2001
NA
Fitch 1981, Fitch and Hillis 1984,
Boulenger 1894, Guyer and
Donnelly 2005
Fitch 1976, Fitch 1981, Savage 2002
see notes on
lionotus
NA
Island
type
ecomorph
Anolis pandoensis
Mainland
NA
Anolis paternus
LargeIsland
Twig
Species
pachypus
A
pandoens
A
paternus
geo
ecomorph
Female
SVL
Male
SVL
3
60
53
2
47.8
50
min n
females
min n
males
NA
NA
14
A placidus
Anolis placidus
LargeIsland
Twig
2
46
45.3
A poecilop
Anolis poecilopus
Mainland
NA
3
68
74
9
8
A pogus
Anolis pogus
SmallIsland
NA
0
42
50
25
A polylepi
Anolis polylepis
Mainland
NA
3
53
57
48
A
polyrhac
Anolis polyrhachis
Mainland
NA
3
50
A
poncensi
Anolis poncensis
LargeIsland
GrassBush
2
40
48
6
6
A
porcatus
Anolis porcatus
LargeIsland
TrunkCrown
1
61.4
74.3
300
688
A porcus
Anolis porcus
LargeIsland
Unique
2
172
NA
16
NA
NA
3
162
NA
n females
n males
refs
notes
NA
NA
16 (Schettino
1999)
Savage 2002
Schettino 1999, Schwartz and
Henderson 1991
Hedges and Thomas 1989,
Schwartz and Henderson 1991
Stamps et al. 1997, Fitch 1981,
Williams 1984b
Roughgarden 1995, Powell et al.
2005
Stamps et al. 1997, Fitch 1976, Fitch
1981, Perry and Garland 2002,
Savage 2002, Andrews 1979,
Andrews and Rand 1974, Fitch and
Hillis 1984, Stamps and Andrews
1992
NA
14 (Schettino 1999)
NA
NA
9 (Fitch 1981)
8 (Fitch 1981)
40
25 (Lazell 1972)
40 (Lazell 1972)
40
48 (Fitch 1981)
1 (Smith 1968), 3
(Campbell et al.
1989, McCranie et
al. 1993)
NA
6 (Butler and Losos
2002)
6 (Butler and
Losos 2002)
300 (Schoener 1970)
688 (Schoener
1970)
NA
NA
40 (Fitch 1981), 7
(Fitch and Hillis
1984)
NA
NA
131 (Schoener), 20
(Butler and Losos
2002)
251 (Schoener
1970), 19 (Butler
and Losos 2002)
A pulchell
Anolis pulchellus
LargeIsland
GrassBush
2
37.5
51
131
251
A pumilis
A
purpurgu
A
quadrioc
Anolis pumilus
Anolis
purpurgularis
Anolis
quadriocellifer
LargeIsland
Unique
2
39.2
34.2
11
7
Mainland
NA
3
58.1
59.3
4
9
11 (Schettino 1999)
4 (McCranie et al.
1993)
LargeIsland
TrunkGround
1
48.5
55
10
10
10 (Schettino 1999)
A
quercoru
Anolis quercorum
Mainland
NA
3
41
46
16
26
16 (Fitch 1981), 6
(Fitch 1978)
7 (Schettino 1999)
9 (McCranie et al.
1993)
10 (Schettino
1999)
26 (Fitch 1981),
16 (Fitch and Hillis
1984), 12 (Fitch
1978)
A recondit
Anolis reconditus
LargeIsland
Unique
2
84
100
NA
NA
A rejectus
Anolis rejectus
LargeIsland
GrassBush
2
36.7
37
7
6
7 (Schettino 1999)
6 (Schettino 1999)
A ricordii
Anolis ricordi
LargeIsland
CrownGiant
2
151
160
40
88
40 (Butler et al.
2000)
A rubribar
Anolis rubribarbus
LargeIsland
TrunkGround
1
47.5
65.9
4
10
4 (Schettino 1999)
88 (Butler et al.
2000)
10 (Schettino
1999)
NA
NA
Smith 1968, Campbell et al. 1989,
McCranie et al. 1993
Losos 1990, Fitch 1981, Butler and
Losos 2002, Roughgarden 1995,
Schwartz and Henderson 1991,
Williams 1983, Butler et al. 2000
Schoener 1970, Fitch 1981,
Schettino 1999, Schwartz and
Henderson 1991, Rogner 1997,
Herrel et al. 2004, Butler et al. 2000
Schettino 1999, Schwartz and
Henderson 1991, Herrel et al. 2004
Schoener 1970, Losos 1990, Fitch
1981, Butler and Losos 2002,
Roughgarden 1995, Schwartz and
Henderson 1991, Herrel et al. 2004,
Williams 1983, Butler et al. 2000
Schettino 1999, Schwartz and
Henderson 1991
Cox et al. 2003, McCranie et al.
1993
Schettino 1999, Schwartz and
Henderson 1991
Fitch 1981, Fitch and Hillis 1984,
Fitch 1978
Schwartz and Henderson 1991,
Herrel et al. 2004, Williams 1983,
Losos Pers. Comm
Schettino 1999, Schwartz and
Henderson 1991
Fitch 1981, Schwartz and
Henderson 1991, Williams 1983,
Butler et al. 2000
Schettino 1999
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Species
A
sabanus
Anolis sabanus
Island
type
SmallIsland
ecomorph
NA
geo
ecomorph
0
Female
SVL
51
Male
SVL
min n
females
69
min n
males
26
57
n females
n males
26 (Lazell 1972)
57 (Lazell 1972)
146 (Schoener
1970), 25 (Butler and
Losos 2002), 451
(Campbell and
Echternacht 2003)
426 (Schoener
1970), 21 (Butler
and Losos 2002),
82 (Gerber and
Echternacht
2000), 461
(Campbell and
Echternacht 2003)
Fitch 1981, Roughgarden 1995,
Schwartz and Henderson 1991,
Lazell 1972, Powell et al. 2005
Schoener 1970, Losos 1990,
Stamps et al. 1997, Fitch 1981,
Perry and Garland 2002, Butler and
Losos 2002, Campbell 1999,
Schettino 1999, Smith 1946,
Schwartz and Henderson 1991,
Andrews and Rand 1974, Gerber
and Echternacht 2000, Licht and
Gorman 1970, Rogner 1997, Herrel
et al. 2004, Maisano 2002, Campbell
and Echternacht 2003, Williams
1983, Stamps and Andrews 1992,
Butler et al. 2000, McCranie et al.
2005, Cox et al. 2003
42 (Lazell 1972)
77 (Lazell 1972)
Lazell 1972
1 (Cope 1895)
1 (Cope 1895)
A sagrei
A
schwartz
Anolis sagrei
LargeIsland
TrunkGround
1
57
73
451
461
Anolis schwartzi
SmallIsland
NA
0
43
49
42
77
A scriptus
Anolis scriptus
SmallIsland
NA
0
64
76
1
1
A semiline
Anolis semilineatus
LargeIsland
GrassBush
2
43
47
43
70
43 (Schoener 1970)
47
34 (Fitch 1981), 12
(Alvarez del Toro
and Smith 1956)
70 (Schoener
1970)
47 (Fitch 1981),
34 (Fitch and Hillis
1984), 5 (Alvarez
del Toro and
Smith 1956)
NA
NA
48 (Butler et al.
2000)
28 (Butler et al.
2000)
NA
A sericeus
Anolis sericeus
Mainland
NA
3
47
52
A sheplani
Anolis sheplani
LargeIsland
Twig
2
40
41
A shrevei
Anolis shrevei
LargeIsland
TrunkGround
1
50
60
A singular
Anolis singularis
LargeIsland
TrunkCrown
1
45
52
A
smallwoo
Anolis smallwoodi
LargeIsland
CrownGiant
2
165
190
A
smaragdi
Anolis smaragdinus
SmallIsland
TrunkCrown
0
51
64
A
sminthus
Anolis sminthus
Mainland
NA
3
58
51
A strahmi
A stratulu
Anolis strahmi
Anolis stratulus
LargeIsland
LargeIsland
TrunkGround
TrunkCrown
1
1
64
46
79
50
34
NA
NA
48
NA
28
NA
NA
NA
NA
NA
NA
7
NA
NA
10
NA
48
129
refs
7 (Fitch 1981), 6
(McCranie et al.
1992)
Schwartz and Henderson 1991,
Herrel et al. 2004, Cope 1895
Schoener 1970, Fitch 1981,
Schwartz and Henderson 1991,
Andrews and Rand 1974, Williams
1983, Butler et al. 2000
Fitch 1973a, 1976, 1981, Savage
2002, Fitch and Hillis 1984, Fitch
1978, Alvarez del Toro and Smith
1956, McCranie et al. 2005, Kohler
et al. 2006
Hedges and Thomas 1989,
Schwartz and Henderson 1991,
Williams 1983
notes
NA
NA
NA
NA
NA
NA
NA
NA
Schwartz and Henderson 1991,
Williams 1983, Butler et al. 2000
Schwartz and Henderson 1991,
Williams 1983
NA
NA
Schettino 1999, Schwartz and
Henderson 1991, Herrel et al. 2004
NA
Stamps et al. 1997, Schwartz and
Henderson 1991, Herrel et al. 2004
NA
Fitch 1976, Fitch 1981, Fitch and
Hillis 1984, Kohler et al. 2006,
McCranie et al. 1992
NA
NA
10 (Fitch 1981),
2+7 (Fitch and
Hillis 1984), 9
(McCranie et al.
1992)
NA
NA
48 (Schoener 1970),
26 (Butler and Losos
2002)
129 (Schoener
1970), 11 (Butler
and Losos 2002)
Schwartz and Henderson 1991,
Herrel et al. 2004, Williams 1983
Schoener 1970, Losos 1990, Fitch
1981, Butler and Losos 2002,
Roughgarden 1995, Avila-Pires
1995, Schwartz and Henderson
1991, Herrel et al. 2004, Williams
1983
NA
NA
NA
Species
Island
type
ecomorph
geo
ecomorph
Female
SVL
Male
SVL
min n
females
min n
males
n females
n males
refs
notes
Fitch 1976, Fitch 1981, Duellman
and Mendelson 1995, Avila-Pires
1995, Duellman 1978, Andrews and
Rand 1974, Dixon and Soini 1986,
Herrel et al. 2004, Vitt et al. 2002,
Amaral 1933
NA
A tropidon
Anolis tropidonotus
Mainland
NA
3
55
56
34
16
34 (Fitch)
129 (Fitch 1981),
2 (Duellman and
Mendelson 1995),
1 (Amaral 1933),
50 (Vitt et al.
2002)
24 (Fitch 1976,
Fitch 1981), 1
(Ruthven 1916), 2
(Barbour
1932a&b)
16 (Fitch 1981),
10 (Fitch and Hillis
1984), 1 (Alvarez
del Toro and
Smith 1956)
A uniformi
Anolis uniformis
Mainland
NA
3
40.5
40.3
13
29
13 (Fitch 1981)
29 (Fitch 1981)
A
trachyde
Anolis trachyderma
Mainland
NA
3
58
61
101
129
101 (Fitch 1981), 28
(Vitt et al. 2002)
A tropidog
Anolis
tropidogaster
Mainland
NA
3
54
63
15
24
15 (Fitch 1976, Fitch
1981)
98
71 (Lazell 1972)
98 (Lazell 1972)
43
10
55
4
NA
43 (Schwartz 1990),
18 (Butler et al.
2000)
10 (Fitch 1981)
NA
55 (Schwartz
1990), 19 (Butler
et al. 2000)
4 (Fitch 1981)
Fitch 1976, Fitch 1981, Cox et al.
2003, Ruthven 1916, Barbour
1932a&b
Stamps et al. 1997, Fitch 1976, Fitch
1981, Fitch and Hillis 1984, Rogner
1997, Alvarez del Toro and Smith
1956, Kohler et al. 2006, D'Cruze
and Stafford 2006
Fitch 1976, Fitch 1981, Kohler et al.
2006
Stamps et al. 1997, Losos 1990,
Fitch 1981, Schoener 1970, Perry
and Garland 2002, Butler and Losos
2002, Schwartz and Henderson
1991, Herrel et al. 2004, Williams
1983, Butler et al. 2000, Andrews
and Rand 1974, Losos Pers. Comm
Schettino 1999, Schwartz and
Henderson 1991
Schettino 1999, Schwartz and
Henderson 1991, Herrel et al. 2004,
Losos Pers. Comm
Stamps et al. 1997, Fitch 1981,
Roughgarden 1995, Schwartz and
Henderson 1991, Lazell 1972,
Stamps and Andrews 1992, Powell
et al. 2005, Kolbe et al. 2008
Schwartz and Henderson 1991,
Williams 1983
Schwartz and Henderson 1991,
Herrel et al. 2004, Schwartz 1980,
Williams 1983, Butler et al. 2000
Fitch 1976, Fitch 1981, Savage 2002
8
6
8 (Kohler and
McCranie 2001)
6 (Kohler and
McCranie 2001)
Kohler and McCranie 2001
A valencie
Anolis valencienni
LargeIsland
Twig
2
74
86
29
40
29 (Schoener), 15
(Butler and Losos
2002), 21 (Herrel et
al. 2004)
A vanidicu
Anolis vanidicus
LargeIsland
GrassBush
2
37
39
9
10
9 (Schettino 1999)
40 (Schoener), 29
(Butler and Losos
2002), 25 (Herrel
et al. 2004)
10 (Schettino
1999)
12 (Schettino 1999)
21 (Schettino
1999)
A vermicul
Anolis vermiculatus
LargeIsland
Unique
2
84.9
124.5
A wattsi
Anolis wattsi
SmallIsland
NA
0
49
58
A websteri
Anolis websteri
LargeIsland
Trunk
2
47
51
A
whiteman
A woodi
Anolis whitemani
Anolis woodi
LargeIsland
Mainland
TrunkGround
NA
1
3
54
86
67
95
A zeus
Anolis zeus
Mainland
NA
3
44
40
12
21
144
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Supplementary Appendix S2
Data sources cited in Supplementary Appendix S1
Alvarez Del Toro, M., and H. M. Smith. 1956. Notulae herpetologicae Chiapasiae. I.
Herpetologica 12:3-17.
Amaral, A. D. 1933. Estudos sobre Lacertilios neotropicos. I. novos generos e
especies de largartos do Brasil. Memorias do Instituto de Butantan 7.
Andrews, R. M. 1976. Growth rate in island and mainland anoline lizards. Copeia
1976:477-482.
Andrews, R. M. 1979. Evolution of life histories: a comparison of Anolis lizards from
matched island and mainland habitats. Breviora 454:1-51.
Andrews, R. M., and A. S. Rand. 1974. Reproductive effort in anoline lizards.
Ecology 55:1317-1327.
Avila-Pires, T. C. S. 1995. Lizards of Brazilian Amazonia. Backhuys Publishers,
Leiden.
Barbour, T. 1932a. New Anoles. Proceedings of the New England Zoological Club
12:97-102.
Barbour, T. 1932b. On a new Anolis from Western Mexico. Copeia 1932:11-12.
Barros, T. R., L. F. Esqueda, A. Mijares-Urrutia, E. La Marca, and K. E. Nicholson.
2007. The anoline “lost link” rediscovered: variation and distribution of Anolis
annectens Williams 1974 (Squamata Polychrotidae). Tropical Zoology 20:4153.
Boulenger, G. A. 1885. Catalogue of the Lizards in the British Museum (Nat. Hist.)
II. Iguanidae, Xenosauridae, Zonuridae, Anguidae, Anniellidae,
Helodermatidae, Varanidae, Xantusiidae, Teiidae, Amphiesbaenidae. Trustees
of the British Museum, London.
Boulenger, G. A. 1894. Second report on additions to the lizard collection in the
Natural History Museum. Proceedings of the Zoological Society of London
1894:722-736.
Boulenger, G. A. 1896. Descriptions of new reptiles and batrachians from Colombia.
Annals and Magazine of Natural History 17:16-21.
Butler, M. A., and J. B. Losos. 2002. Multivariate sexual dimorphism, sexual
selection, and adaptation in Greater Antillean Anolis lizards. Ecological
Monographs 72:541-559.
Butler, M. A., T. W. Schoener, and J. B. Losos. 2000. The relationship between
sexual size dimorphism and habitat use in Greater Antillean Anolis lizards.
Evolution 54:259-272.
Campbell, H. W. 1973. Ecological observations on Anolis lionotus and Anolis
poecilopus (Reptilia, Sauria) in Panama. American Museum Novitates 2516:129.
Campbell, J. A. 1999. Amphibians and Reptiles of Northern Guatemala, the Yucatan,
and Belize. University of Oklahoma Press, Norman.
Campbell, J. A., D. M. Hillis, and W. W. Lamar 1989. A new lizard of the genus
Norops (Sauria: Iguanidae) from the cloud forest of Hidalgo, Mexico.
Herpetologica 45:232-242.
Campbell, T. S., and A. C. Echternacht. 2003. Introduced species as moving targets:
changes in body sizes of introduced lizards following experimental
introductions and historical invasions. Biological Invasions 5:193-212.
Clobert, J., T. Garland, and R. Barbault. 1998. The evolution of demographic tactics
in lizards: a test of some hypotheses concerning life history evolution. Journal
of Evolutionary Biology 11:329-364.
Cope, E. D. 1895. The Batrachia and Reptilia of the University of Pennsylvania West
Indian expedition of 1890 and 1891. Proceedings of the Academy of Natural
Sciences of Philadelphia 46:429-442.
Cox, R. M., S. L. Skelly, and H. B. John-Alder. 2003. A comparative test of adaptive
hypotheses for sexual size dimorphism in lizards. Evolution 57:1653-1669.
D'Cruze, N. C., and P. J. Stafford. 2006. Resource partitioning of sympatric Norops
(Beta Anolis) in a subtropical mainland community. Herpetological Journal
16:273-280.
Dalrymple, G. H. 1980. Comments on the density and. diet of a giant anole, Anolis
equestris. Journal of Herpetology 14:412-415
Dixon, J. R., and P. Soini. 1986. The reptiles of the upper Amazon basin, Iquitos
Region, Peru. 2nd edition. Milwaukee Public Museum., Milwaukee.
dos Santos, A. R., J. G. da Frota, and F. R. V. Ribeiro. 2007. Reptilia, Squamata,
Polychrotidae, Anolis nitens tandai: Distribution extension, new state record,
and geographic distribution map. Check List 3:9-10.
Duellman, W. E. 1978. The biology of an equatorial herpetofauna in Amazonian
Ecuador. University of Kansas Museum of Natural History Miscellaneous
publications 65:1-352.
Duellman, W. E. 2005. Cusco Amazónico: The lives of amphibians and reptiles in an
Amazonian Rainforest. Cornell University Press, Ithaca.
Duellman, W. E., and J. R. Mendelson. 1995. Amphibians and reptiles from Northern
Departamento Loreto, Peru: Taxonomy and Biogeography. University of
Kansas Science Bulletin 55:329-376.
Dunham, A. E., and D. B. Miles. 1985. Patterns of covariation in life history traits of
squamate reptiles: the effects of size and phylogeny reconsidered. American
Naturalist 126:231-257.
Dunham, A. E., P. J. Morin, and H. M. Wilbur. 1988. Methods for the study of reptile
populations. Pp. 331-373 in C. Gans, and R. B. Huey, eds. Biology of the
Reptilia. Vol. 16. Ecology B. Defense and life history. Liss, New York.
Fitch, H. S. 1970. Reproductive cycles of lizards and snakes. University of Kansas
Museum of Natural History Miscellaneous publications 52:1-247.
Fitch, H. S. 1973a. A field study of Costa Rican lizards. University of Kansas Science
Bulletin 50:39-126.
Fitch, H. S. 1973b. Population structure and survivorship in some Costa Rican lizards.
Occasional Papers of the Museum of Natural History, University of Kansas
18:1-41.
Fitch, H. S. 1976. Sexual size differences in the mainland anoles. Occasional Papers
of the Museum of Natural History, University of Kansas 50:1-21.
Fitch, H. S. 1978. Two new anoles (Reptilia: Iguanidae) from Oaxaca with comments
on other Mexican species. Milwaukee Public Museum Contributions in
Biology and Geology 20:1-15.
Fitch, H. S. 1981. Sexual size differences in reptiles. University of Kansas Museum
Natural History Miscellaneous Publications 70:1-72.
Fitch, H. S., A. F. Echelle, and A. A. Echelle. 1976. Field observations on rare or little
known mainland anoles. University of Kansas Science Bulletin 51:91-128.
Fitch, H. S., and D. M. Hillis. 1984. The Anolis dewlap: interspecific variability and
morphological associations with habitat. Copeia 1984:315-323.
Fobes, T. M., R. Powell, J. S. Parmerlee, A. Lathrop, and D. D. Smith. 1992. Natural
history of Anolis cybotes (Sauria: Polychrotidae) from an altered habitat in
Barahona, Dominican Republic. Caribbean Journal of Science 28:200-207.
Garrido, O. H., and S. B. Hedges. 2001. A new anole from the northern slope of the
Sierra Maestra in eastern Cuba (Squamata: Iguanidae). Journal of Herpetology
35:378-383.
Gerber, G. P., and A. C. Echternacht. 2000. Evidence for asymmetrical intraguild
predation between native and introduced Anolis lizards. Oecologia 124:599607.
Guyer, C., and M. A. Donnelly. 2005. Amphibians and reptiles of La Selva, Costa
Rica, and the Caribbean slope: A comprehensive guide. University of Chicago
Press, Chicago.
Hedges, S. B., and R. Thomas. 1989. A new species of Anolis (Sauria: Iguanidae)
from the Sierra de Neiba, Hispaniola Herpetologica 45:330-336.
Herrel, A., B. Vanhooydonck, R. Joachim, and D. J. Irschick. 2004. Frugivory in
polychrotid lizards: effects of body size. Oecologia 140:160-168.
Hoogmoed, M. S. 1973. Notes on the herpetofauna of Surinam IV.: the lizards and
amphisbaenians of Surinam. W. Junk Publishers, Den Haag.
Klutsch, C. F. C., B. Misof, W.-R. Grosse, and R. F. A. Moritz. 2007. Genetic and
morphometric differentiation among island populations of two Norops lizards
(Reptilia: Sauria: Polychrotidae) on independently colonized islands of the
Islas de Bahia (Honduras). Journal of Biogeography 34:1124-1135.
Kohler, G. 2001. Type material and use of the name Anolis bourgeaei Bocourt
(Sauria: Polychrotidae). Copeia 2001.
Kohler, G., and J. R. McCranie. 2001. Two new species of anoles from northern
Honduras (Squamata: Polychrotidae). Senckenbergiana biologica 81:235-245.
Kohler, G., J. R. McCranie, and L. D. Wilson. 2001. A new species of anole from
western Honduras (Squamata: Polychrotidae). Herpetologica 57:247-255.
Kohler, G., and J. Sunyer. 2008. Two new species of anoles formerly referred to as
Anolis limifrons (Squamata, Polychrotidae). Herpetologica 64:92-108.
Kohler, G., M. Vesely, and E. Greenbaum. 2006. The Amphibians and Reptiles of El
Salvador. Krieger Publishing Company, Malabar.
Kolbe, J. J., P. L. Colbert, and B. E. Smith. 2008. Niche relationships and
interspecific interactions in Antiguan lizard communities. Copeia 2008:261272.
Lazell, J. D. 1964. The anoles (Sauria: Iguanidae) of the Guadeloupeen Archipelago.
Bulletin of the Museum of Comparative Zoology 131:359-401.
Lazell, J. D. 1972. The anoles (Sauria: Iguanidae) of the Lesser Antilles. Bulletin of
the Museum of Comparative Zoology 143:1-115.
Leal, M., and J. B. Losos. 2000. Natural history of the Cuban lizard Chamaeleolis
barbatus. Journal of Herpetology 34:318-322.
Licht, P., and G. C. Gorman. 1970. Reproductive and fat cycles in Caribbean Anolis
lizards. University of California Publications in Zoology 35:1-52.
Losos, J. B. 1990. Ecomorphology, performance ability and scaling of West Indian
Anolis lizards: an evolutionary analysis. Ecological Monographs 60:369-388.
Lotzkat, S. 2007. Taxonomie und zoogeographie der herpetofauna des NirguaMassivs, Venezuela. Johann Wolfgang Goethe-Universität, Frankfurt am
Main.
Maisano, J. A. 2002. Terminal fusions of skeletal elements as indicators of maturity in
squamates. Journal of Vertebrate Paleontology 22:268-275.
McCranie, J. R. 1992. Rediscovery of the Honduran Giant Anole—Norops loveridgei
(Sauria: Polychrotidae). Caribbean Journal of Science 28:233-234.
McCranie, J. R., L. D. Wilson, and G. Kohler. 2005. Amphibians and reptiles of the
Bay Islands and Cayos Cochinos, Honduras. Bibliomania.
McCranie, J. R., L. D. Wilson, and K. L. Williams. 1992. A new species of anole of
the Norops crassulus group (Sauria: Polychridae) from northwestern
Honduras. Caribbean Journal of Science 28:208-215.
McCranie, J. R., L. D. Wilson, and K. L. Williams. 1993. Another new species of
lizard of the Norops schiedei group (Sauria: Polychrotidae) from northern
Honduras. Journal of Herpetology 27:393-399.
Perry, G., and T. J. Garland. 2002. Lizard home ranges revisited: effects of sex, body
size, diet, habitat, and phylogeny. Ecology 83:1870-1885.
Powell, R., H. R. W, and J. S. Parmerlee. 2005. Powell, R., Henderson R. W. and
Parmerlee, J. S. 2005. The Reptiles and Amphibians of the Dutch Caribbean
St Eustatius, Saba, and St Maarten. International Union for the Conservation
of Nature.
Rogner, M. 1997. Lizards. Volume 1. Krieger Publishing Company, Malabar, FL.
Roughgarden, J. 1995. Anolis Lizards of the Caribbean: Ecology, Evolution, and Plate
Tectonics. Oxford University Press.
Ruthven, A. G. 1916. Three new species of Anolis from the Santa Marta Mountains,
Colombia. Occasional Papers of the Museum of Zoology, University of
Michigan 32:1-8.
Savage, J. M. 2002. The Amphibians and Reptiles of Costa Rica. The University of
Chicago Press, Chicago.
Schettino, L. R. 1999. The iguanid lizards of Cuba. University Press of Florida,
Miami.
Schoener, T. W. 1970. Size patterns in West Indian Anolis Lizards: II. Correlations
with sizes of particular sympatric species - displacement and convergence.
The American Naturalist 104:155-174.
Schwartz, A. 1980. Variation in Hispaniolan Anolis whitemani Williams. Journal of
Herpetology 14:399-406.
Schwartz, A., and R. W. Henderson. 1991. Amphibians and Reptiles of the West
Indies. University of Florida Press, Gainesville.
Smith, H. M. 1934. Notes on some Mexican lizards of the genus Anolis with the
description of a new species, A. megapholidotus. Transactions of the Kansas
Academy of Science 36:318-320.
Smith, H. M. 1946. Handbook of lizards. Lizards of the United States and Canada.
Cornell University Press, Ithaca.
Smith, H. M. 1968. Another new lizard from Mexico of the schiedi group of Anolis.
Southwestern Naturalist 13:368-370.
Smith, R. J. 1981. Interpretation of correlations in intraspecific and interspecific
allometry. Growth 45.
Stamps, J. A., and R. M. Andrews. 1992. Estimating asymptotic size using the largest
individuals per sample. Oecologia 92:503-512.
Stamps, J. A., J. B. Losos, and R. M. Andrews. 1997. A comparative study of
population density and sexual size dimorphism in lizards. American Naturalist
149:64-90.
Taylor, E. H. 1956. A review of the lizards of Costa Rica. University of Kansas
Science Bulletin 38:3-322.
Thomas, R., and S. B. Hedges. 1991. Rediscovery and description of the Hispaniolan
lizard Anolis darlingtoni (Sauria, Iguanidae) Caribbean Journal of Science
27:90-93.
Tinkle, D. W., H. M. Wilbur, and S. G. Tilley. 1970. Evolutionary strategies in lizard
reproduction. Evolution 24:55-74.
Trivers, R. L. 1976. Sexual selection and resource accruing abilities in Anolis
garmani. Evolution 30:253-269.
Uetz, P. 2006. The reptile database CD-ROM edition, October 2006, Heidelberg,
Germany.
Vitt, L. J., T. C. S. Avila-Pires, Z. P. A, and M. C. Esposito. 2002. Life in shade: the
ecology of Anolis trachyderma (Squamata: Polychrotidae) in Amazonian
Ecuador and Brazil, with comparisons to ecologically similar anoles. Copeia
2002:275-286.
Vitt, L. J., T. C. S. Avila-Pires, P. A. Zani, S. S. Sartorius, and M. C. Esposito. 2003.
Life above ground: ecology of Anolis fuscoauratus in the Amazon rain forest,
and comparisons with its nearest relatives. Canadian Journal of Zoology
81:142-156.
Vitt, L. J., and J. P. Caldwell. 1993. Ecological observations on cerrado lizards in
Rondonia, Brazil. Journal of Herpetology 27:46-52.
Vitt, L. J., and C. M. de Carvalho. 1995. Niche partitioning in a tropical wet season,
lizards in the lavrado area of northern Brazil. Copeia 1995:305-329.
Vitt, L. J., D. B. Shepard, G. H. C. Vieira, J. P. Caldwell, G. R. Colli, and D. O.
Mesquita. 2008. Ecology of Anolis nitens brasiliensis in Cerrado Woodlands
of Cantao. Copeia 2008:144-153.
Vitt, L. J., and P. A. Zani. 1998. Ecological relationships among sympatric lizards in a
transitional forest in the northern Amazon of Brazil Journal of Tropical
Ecology 14:63-86.
Williams, E. E. 1974. A case history in retrograde evolution: the onca lineage in
anoline lizards. I. Anolis annectens new species, intermediate between the
genera Anolis and Tropidodactylus. Breviora 421:1-21.
Williams, E. E. 1983. Ecomorphs, faunas, island size, and diverse end points in island
radiations of Anolis. in R. B. Huey, E. R. Pianka, and T. W. Schoener, eds.
Lizard Ecology: Studies of a Model Organism. Harvard University Press,
Cambridge.
Williams, E. E. 1984. New or problematic Anolis from Colombia. 3. Two new
semiaquatic anoles from Antioquia and Choco, Colombia. Breviora 478:1-22.
Supplementary Appendix S3
Species that have been subject to morphometric analyses of ecomorphs. List of
species included in Losos et al. (1998) courtesy of JB Losos.
Species only in Losos et al.
(1998)
Species only in Beuttell and
Losos (1999)
Species in both papers
A. ahli
A. acutus
A. aliniger
A. allisoni
A. aenus
A. angusticeps
A. allogus
A. bimaculatus
A. bahorucoensis
A. alutaceus
A. carolinensis
A. brevirostris
A. barahonae
A. christophei
A. chlorocyanus
A. bremeri
A. conspersus
A. coelestinus
A. guazuma
A. cooki
A. cristatellus
A. homolechis
A. darlingtoni
A. cuvieri
A. loysiana
A. etheridgei
A. cybotes
A. luteogularis
A. extremus
A. distichus
A. mestrei
A. ferreus
A. equestris
A. ophiolepis
A. fowleri
A. evermanni
A. paternus
A. gingivinus
A. garmani
A. singularis
A. griseus
A. grahami
A. vanidicus
A. leachi
A. gundlachi
A. lividus
A. insolitus
A. longiceps
A. krugi
A. luciae
A. lineatopus
A. marmoratus
A. longitibialis
A. maynardi
A. marcanoi
A. monensis
A. occultus
A. pogus
A. olssoni
A. reconditus
A. opalinus
A. richardi
A. poncensis
A. roquet
A. porcatus
A. sabanus
A. pulchellus
A. schwartzi
A. sagrei
A. sheplani
A. semilineatus
A. trinitatis
A. strahmi
A. wattsi
A. stratulus
A. valencienni
Literature cited
Beuttell, K., and J. B. Losos. 1999. Ecological morphology of Caribbean anoles.
Herpetological Monographs 13:1-28.
Losos, J. B., T. R. Jackman, A. Larson, K. De Queiroz, and L. Rodriguez-Schettino.
1998. Contingency and determinism in replicated adaptive radiations of island
lizards. Science 279:2115-2118.
Supplementary Appendix S4
Maximum likelihood reconstruction of Anolis ecomorphs. The ancestral nodes of the
seven separate colonisations of small islands are marked as 1-7 and reinvasions of the
mainland from the Greater Antilles are marked as 8 & 9. The insert shows the
proportional likelihoods of ancestral ecomorphs for the ancestral nodes of the small
island lineages. Only ecomorphs with a proportional likelihood >0.05 are shown.
Supplementary Appendix S5
Effects of assuming a common mean
Thomas et al’s (2006) method for comparing rates of phenotypic diversification
differs from the non-censored approach of O’Meara et al. (2006) by allowing each
different partition of the phylogeny to have a different phylogenetic mean. Revell
(2008) questioned why branches that are united only by a different rate regime should
have different means. We suggest that many hypotheses that infer different rate
regimes imply different evolutionary regimes such that a difference in mean is also a
likely outcome. This may seem trivial when the interest is in rates of trait evolution
rather than their means, however, assuming a common mean can have serious
consequences for the inferred differences in rates.
As a simple example, we simulated the effects of different means on rates for a
phylogeny divided into two groups. For convenience, we used Nicholson et al’s.
(2005) Anolis phylogeny and divided branches into large island versus small island
and mainland branches. We then simulated trait evolution in which the means of the
two groups differed by varying amounts. Because the importance of differences in
mean of any trait depends on the scale on which the trait is measured, we set the
difference in means according to the number of expected standard deviations.
Ricklefs (2006) recently showed that the expected variance for a trait across species
under a Brownian model can be estimated by taking the mean of the off-diagonal
elements of the variance-covariance matrix representation of the phylogeny. We used
this measure to estimate the expected variance for the anole phylogeny and took the
square root to calculate the expected standard deviation. We then used the function
rmvnorm in the R library mvtnorm (Genz et al. 2008) to simulate trait evolution in
which there was no difference in rates between groups but the mean of the first group
was set at 0 and the mean of the second group was either 0, 0.5, 1, 2, 3, 4, 5, or 6
standard deviations larger. For each degree of difference in trait means we ran 1000
simulations. We then fit one set of models that allow different means and one set of
models that force a common mean to each simulated trait. We examined type I error
using a likelihood ratio test to compare the maximum likelihood model against the
null equal rates model. In addition we recorded the parameter estimate in each case.
The simulations (Fig 1.) show that where the difference in means is small (two or
fewer standard deviations difference), neither the common mean nor the multi-mean
model inferred a rate shift thus, type I errors are similar (and acceptably low) for the
multi-mean and common mean tests across this range. Also, the parameter estimates
are close to one for both the multi-mean and common mean tests when differences in
means are small. However, as the magnitude of differences in means increases, the
rate at which the common mean model infers rates shifts increases and the type I error
rate is unacceptably high. In contrast, the multi-mean model is unaffected by the
magnitude of the difference in simulated means. Similarly, under the common mean
model, the rate-parameter estimates become increasingly different from one as the
differences in mean become larger whereas the parameter estimates in the multi-mean
model remains unaffected.
Figure 1. Type I error rates (A) and parameter estimates (B) for a Brownian trait
where the mean differs in different parts of the tree.
References
Genz, A., F. Bretz, T. Hothorn, T. Miwa, X. Mi, F. Leisch, and F. Scheipl. 2008.
mvtnorm: Multivariate Normal and t Distributions. R package version 0.9-2.
Nicholson, K. E., R. E. Glor, J. J. Kolbe, A. Larson, S. Blair Hedges, and J. Losos.
2005. Mainland colonization by island lizards. Journal of Biogeography
32:929-938.
O'Meara, B. C., C. Ané, M. J. Sanderson, and P. C. Wainwright. 2006. Testing for
different rates of continuous trait evolution using likelihood. Evolution
60:922-933.
Revell, L. J. 2008. On the analysis of evolutionary change along single branches in a
phylogeny. The American Naturalist 172:140-147.
Ricklefs, R. E. 2006. Time, species, and the generation of trait variation in clades.
Systematic Biology 55:151-159.
Thomas, G. H., R. P. Freckleton, and T. Székely. 2006. Comparative analyses of the
influence of developmental mode on phenotypic diversification rates in
shorebirds. Proceedings of the Royal Society of London Series B 273:16191624.
Supplementary Appendix S6 – phylogeny with ecomorph/geographic setting
as node
labels
#NEXUS
[Phylogeny from: Nicholson, K. E., R. E. Glor, J. J. Kolbe, A. Larson, S. Blair
Hedges, and J. Losos. 2005. Mainland colonization by island lizards. Journal of
Biogeography 32:929-938.]
[Downloaded from: http://biosgi.wustl.edu/~lososlab/anolis_mbg_2005/]
begin taxa;
dimensions ntax=165;
taxlabels
A_tropidog
A_equestri
A_placidus
A_baleatus
A_tropidon
A_guamuhay
A_biporcat
A_poecilop
A_garridoi
A_distichu
A_leachi
A_clivicol
A_humilis
A_meridion
A_polyrhac
A_websteri
A_ferreus
A_crassulu
A_breviros
A_vanidicu
A_polylepi
A_capito
A_marron
A_zeus
A_sminthus
A_gundlach
A_dolichoc
A_sagrei
A_uniformi
A_imias
A_aquaticu
A_conspers
A_cristate
A_evermann
A_altae
A_altitudi
A_lionotus
A_krugi
A_smaragdi
A_etheridg
A_olssoni
A_oporinus
A_pachypus
A_ahli
A_singular
A_monticol
A_armouri
A_homolech
A_henderso
A_rubribar
A_annecten
A_onca
A_barbatus
A_guazuma
A_laeviven
A_oxylophu
A_bartschi
A_baracoae
A_caroline
A_pulchell
A_alutaceu
A_cupeyale
A_noblei
A_marcanoi
A_bahoruco
A_vermicul
A_trachyde
A_sheplani
A_limifron
A_intermed
A_scriptus
A_allisoni
A_porcus
A_paternus
A_argillac
A_lividus
A_mestrei
A_ernestwi
A_eugenegr
A_acutus
A_ortonii
A_ricordii
A_oscellos
A_barahona
A_stratulu
A_carpente
A_bremeri
A_christop
A_confusus
A_nubilus
A_valencie
A_macilent
A_auratus
A_cybotes
A_haetianu
A_oculatus
A_wattsi
A_luteogul
A_opalinus
A_insolitu
A_nebuloid
A_pandoens
A_darlingt
A_cooki
A_isthmicu
A_breslini
A_cuvieri
A_lemurinu
A_bimacula
A_longicep
A_gingivin
A_poncensi
A_nitens
A_recondit
A_sericeus
A_coelesti
A_pumilis
A_lineatop
A_barbouri
A_inexpect
A_maynardi
A_quadrioc
A_monensis
A_alfaroi
A_alayoni
A_longitib
A_guafe
A_whiteman
A_strahmi
A_fuscoaur
A_semiline
A_aliniger
A_ophiolep
A_grahami
A_purpurgu
A_bitectus
A_pogus
A_rejectus
A_brunnneu
A_sabanus
A_jubar
A_occultus
A_quercoru
A_garmani
A_fowleri
A_marmorat
A_allogus
A_lucius
A_centrali
A_schwartz
A_angustic
A_chlorocy
A_desechen
A_shrevei
A_smallwoo
A_porcatus
A_loysiana
A_cyanople
A_chamaele
A_loveridg
A_cupreus
A_argenteo
A_caudalis
A_woodi
A_alumina
;
end;
begin trees;
tree [&r] con_50_majrule =
((A_occultus:85.167845,((A_coelesti:58.072167,(A_chlorocy:34.907719,(A_aliniger:
19.402478,A_singular:19.402478)1:15.505241)1:23.164448)1:20.913282,(A_darlingt
:69.543072,(A_monticol:62.340379,(A_bahoruco:38.836483,(A_dolichoc:16.979814,
A_henderso:16.979814)2:21.856669)2:23.503897)2:7.202693)2:9.442377,(A_equestr
i:9.671404,(A_luteogul:6.984718,(A_baracoae:5.382205,(A_noblei:1.687602,A_smal
lwoo:1.687602)2:3.694603)2:1.602512)2:2.686686)2:69.314045,(A_bartschi:46.4094
93,A_vermicul:46.409493)2:32.575956)2:6.182396)2:6.948849,((A_marcanoi:51.462
757,((A_longitib:23.484257,A_strahmi:23.484257)1:15.788668,(A_breslini:28.50547
8,(A_whiteman:26.241261,((A_armouri:14.25956,A_shrevei:14.25956)1:7.05933,(A
_cybotes:18.962348,A_haetianu:18.962348)1:2.356542)1:4.922372)1:2.264216)1:10.
767448)1:12.189831)1:33.4638,((((A_alutaceu:11.263887,A_inexpect:11.263887)2:3
1.07247,(A_vanidicu:35.407633,((A_alfaroi:20.416245,A_macilent:20.416245)2:12.6
31686,(A_clivicol:22.768465,(A_rejectus:18.037547,(A_cupeyale:4.36141,A_cyanop
le:4.36141)2:13.676137)2:4.730918)2:10.279465)2:2.359702)2:6.928724)2:28.20928
5,(((A_alayoni:32.14993,(A_angustic:18.946972,A_paternus:18.946972)2:13.202958
)2:23.388897,(A_sheplani:14.811765,A_placidus:14.811766)2:40.727062)2:6.384934
,(((A_garridoi:0.56475,A_guazuma:0.56475)2:50.360642,(A_loysiana:37.741894,(A_
pumilis:13.443447,(A_centrali:11.499654,A_argillac:11.499654)2:1.943794)2:24.298
446)2:13.183498)2:3.834091,((A_oporinus:16.311783,A_altitudi:16.311783)1:32.446
959,((A_caroline:17.248509,A_porcatus:17.248509)1:19.355642,((A_allisoni:17.911
995,A_smaragdi:17.911995)1:8.571143,(A_brunnneu:23.745879,(A_longicep:21.104
349,A_maynardi:21.104349)0:2.64153)0:2.73726)1:10.121013)1:12.154591)1:6.0007
41)2:7.164278)2:8.62188)2:11.171156,((((A_pogus:32.676182,(A_wattsi:13.405478,
A_schwartz:13.405478)0:19.270704)0:19.767807,(A_leachi:32.457343,((A_bimacula
:18.211987,A_gingivin:18.211987)0:10.233462,(A_oculatus:20.276694,(A_ferreus:1
2.366709,(A_lividus:11.075687,(A_nubilus:10.080339,(A_marmorat:8.118164,A_sa
banus:8.118164)0:1.962175)0:0.995348)0:1.291022)0:7.909985)0:8.168755)0:4.0118
95)0:19.986645)0:18.411217,((A_distichu:33.675375,(A_websteri:23.447331,(A_bre
viros:19.558374,(A_caudalis:12.032684,A_marron:12.032684)2:7.52569)2:3.888957)
2:10.228044)2:31.339455,((A_acutus:26.069079,(A_evermann:19.614502,A_stratulu
:19.614502)1:6.454577)1:26.316453,((A_krugi:32.023689,A_pulchell:32.023689)2:1
4.184029,((A_gundlach:32.451484,A_poncensi:32.451484)1:10.336754,((A_monensi
s:21.844737,A_cooki:21.844737)1:13.916855,(A_scriptus:33.145342,(A_cristate:16.
38016,(A_desechen:12.628035,A_ernestwi:12.628035)0:3.752126)1:16.765181)1:2.6
1625)1:7.026646)1:3.419481)1:6.177814)1:12.629297)1:5.840375)1:8.230574,(((A_i
mias:43.381302,(A_rubribar:18.080283,(A_ahli:8.713256,A_allogus:8.713256)1:9.36
7027)1:25.301019)1:15.347195,((A_guafe:21.949519,(A_jubar:15.987089,(A_confus
us:11.48692,A_homolech:11.48692)1:4.500168)1:5.96243)1:17.735279,(A_mestrei:2
6.741691,(A_ophiolep:20.072286,(A_sagrei:17.326956,(A_bremeri:7.271714,A_qua
drioc:7.271714)1:10.055243)1:2.74533)1:6.669404)1:12.943108)1:19.043699)1:8.64
9016,(((A_lineatop:34.754433,A_recondit:34.754433)1:15.079204,(A_valencie:45.23
1942,((A_conspers:18.939305,A_grahami:18.939305)1:6.73591,(A_garmani:21.1613
17,A_opalinus:21.161317)1:4.513898)1:19.556727)1:4.601694)1:9.332218,(((A_ann
ecten:18.835076,A_onca:18.835076)3:27.647264,(A_nitens:40.539112,A_meridion:4
0.539112)3:5.943227,A_auratus:46.48234)3:10.384131,((A_loveridg:34.175879,A_p
urpurgu:34.175879)3:18.819437,(((A_nebuloid:23.211312,A_quercoru:23.211312)3:
25.787487,(A_bitectus:45.257381,(A_biporcat:33.884313,(A_woodi:16.409412,A_aq
uaticu:16.409412)3:17.474901)3:11.373068)3:3.741419,(A_polyrhac:41.352073,A_u
niformi:41.352073)3:7.646727,A_crassulu:48.998799)3:1.649719,(A_sminthus:44.80
7,((A_isthmicu:12.759526,A_sericeus:12.759526)3:26.709163,((A_ortonii:32.369411
,(A_intermed:16.489409,A_laeviven:16.489409)3:15.880002)3:3.224796,(((A_cupre
us:21.478969,A_polylepi:21.478969)3:7.878584,(A_altae:16.810402,(A_fuscoaur:13.
932759,A_pandoens:13.932759)3:2.877643)3:12.547151)3:3.305345,(((A_capito:21.
464239,A_tropidon:21.464239)3:7.810471,(A_humilis:24.59371,A_pachypus:24.593
71)3:4.681)3:1.575314,(A_oscellos:23.055949,(A_carpente:20.564469,((A_lemurinu:
17.467386,(A_limifron:7.430975,A_zeus:7.430975)3:10.036412)3:1.6924,((A_lionot
us:12.203853,A_oxylophu:12.203853)3:2.589853,(A_tropidog:11.857214,(A_trachyd
e:9.018625,A_poecilop:9.018625)3:2.838589)3:2.936491)3:4.366082)3:1.404681)3:2
.49148)3:7.794075)3:1.812874)3:2.931309)3:3.874481)3:5.338311)3:5.841519)3:2.3
46797)3:3.871155)3:2.299385)1:8.211659)1:11.708266)1:2.631018)2:3.209759,((A_
argenteo:54.710554,A_lucius:54.710554)2:23.263736,((A_barbatus:11.057573,(A_po
rcus:6.959118,(A_chamaele:5.917888,A_guamuhay:5.917888)2:1.04123)2:4.098455)
2:54.992843,(A_cuvieri:54.869185,(A_christop:41.31477,(A_eugenegr:31.445243,(A
_ricordii:10.104409,(A_baleatus:4.155331,A_barahona:4.155331)2:5.949078)2:21.34
0834)2:9.869528)2:13.554415)2:11.181231)2:11.923874)2:6.952267,(A_barbouri:77.
08065,((A_etheridg:50.203579,(A_fowleri:32.636205,A_insolitu:32.636205)2:17.567
374)2:21.743955,(A_olssoni:56.831733,(A_alumina:35.053007,A_semiline:35.05300
7)2:21.778726)2:15.115801)2:5.133117)2:7.845906)2:7.190137)2;
end;
Supplementary Appendix S7
The following tables contain full results for rates tests on male SVL (Tables S7-1 - S7-3), female SVL (Tables S7-4 - S76), and sexual size dimorphism in SVL (Tables S7-7 - S7-9). Each table contains 12 models (see main text for details)
with maximum likelihood estimates for each parameter. Confidence limits are provided for each parameter except where
the parameter was fixed in the model. Models are ranked by AICc and both delta AICc and Akaike weights are
provided. The Akaike weights were used to calculate model averaged parameter estimates. In addition, the maximum
likelihood of each model is provided, along with results of likelihood ratio tests against a model in which all rate
parameters are forced to be equal. All other details are as per Table 1 in the main text.
For each of male SVL, female SVL, and SSD, three sets of models are included. The first forces a single mean across
each of the four categories (small islands, large island trunk crown and trunk ground ecomorphs, large island other
ecomorphs, and mainland) with the full data set. The second allows different means but uses a reduced data set in which
only taxa with size estimates based on 20 or more individuals are included. The third model set uses the full data set and
allows different means, but the phylogeny was transformed according to the maximum likelihood estimate of kappa
prior to analyses (see main text).
Male SVL
Table S7-1. Male SVL assuming a single mean
Model
Small islands Large island TC & TG Large island other Mainland deltaAIC wtAIC Max. likelihood
M5 theta
1.000
0.415
1.000
1.000
0.000
0.376
120.373
M5 LCI
0.245
M5 UCI
0.743
M3 theta
1.000
0.383
1.000
0.807
1.504
0.177
120.672
M3 LCI
0.227
0.498
M3 UCI
0.684
1.364
M4 theta
1.000
0.451
1.212
1.000
1.554
0.173
120.647
M4 LCI
0.268
0.756
M4 UCI
0.802
1.953
M9 theta
1.000
0.418
1.011
1.011
2.102
0.131
120.373
M9 LCI
0.535
0.244
M9 UCI
2.025
0.741
M1 theta
1.000
0.413
1.108
0.872
3.556
0.064
120.711
M1 LCI
0.541
0.245
0.691
0.538
M1 UCI
2.047
0.734
1.786
1.474
M11 theta
1.000
1.000
1.449
1.000
5.975
0.019
117.385
M11 LCI
0.898
M11 UCI
2.354
All equal
1.000
1.000
1.000
1.000
6.203
0.017
116.233
M7 theta
1.000
0.651
1.055
0.651
6.508
0.015
118.170
M7 LCI
0.828
1.004
M7 UCI
3.155
2.627
M2 theta
1.000
1.000
1.723
1.404
6.722
0.013
118.063
M2 LCI
1.069
0.866
M2 UCI
2.793
2.375
M8 theta
1.000
0.814
0.814
0.814
7.897
0.007
116.424
M8 LCI
0.660
M8 UCI
2.530
M10 theta
1.000
1.000
1.000
1.035
8.262
0.006
116.242
M10 LCI
0.638
M10 UCI
1.751
M6 theta
1.000
0.798
0.798
0.858
9.923
0.003
116.463
M6 LCI
0.674
0.662
M6 UCI
2.580
1.819
Model average
1.000
0.455
1.062
0.958
P (k)
** (3)
* (4)
* (4)
* (4)
* (5)
NS (3)
(2)
NS (4)
NS (4)
NS (3)
NS (3)
NS (4)
Table S7-2. Male SVL with reduced data set (allowing multiple means)
Model
Small islands Large island TC & TG Large island other Mainland deltaAIC wtAIC Max. likelihood
M5 theta
1.000
0.384
1.000
1.000
0.000
0.312
84.397
M5 LCI
0.208
M5 UCI
0.766
M3 theta
1.000
0.342
1.000
0.639
0.768
0.212
85.148
M3 LCI
0.187
0.344
M3 UCI
0.675
1.297
M9 theta
1.000
0.319
0.771
0.771
1.772
0.128
84.646
M9 LCI
0.670
0.223
M9 UCI
2.795
0.831
M4 theta
1.000
0.407
1.130
1.000
2.117
0.108
84.474
M4 LCI
0.221
0.644
M4 UCI
0.806
1.989
M7 theta
1.000
0.433
0.841
0.433
2.911
0.073
84.077
M7 LCI
1.189
1.101
M7 UCI
4.994
3.440
M1 theta
1.000
0.313
0.870
0.580
2.957
0.071
85.211
M1 LCI
0.516
0.170
0.495
0.312
M1 UCI
2.154
0.619
1.530
1.178
All equal
1.000
1.000
1.000
1.000
4.816
0.028
80.874
M8 theta
1.000
0.612
0.612
0.612
5.042
0.025
81.876
M8 LCI
0.837
M8 UCI
3.550
M11 theta
1.000
1.000
1.402
1.000
5.694
0.018
81.549
M11 LCI
0.793
M11 UCI
2.495
M10 theta
1.000
1.000
1.000
0.834
6.766
0.011
81.013
M10 LCI
0.448
M10 UCI
1.699
M6 theta
1.000
0.621
0.621
0.575
7.263
0.008
81.901
M6 LCI
0.824
0.497
M6 UCI
3.494
1.883
M2 theta
1.000
1.000
1.398
0.991
7.965
0.006
81.550
M2 LCI
0.791
0.533
M2 UCI
2.487
2.017
Model average
1.000
0.414
0.961
0.808
P (k)
** (6)
* (7)
* (7)
* (7)
* (7)
* (7)
(5)
NS (6)
NS (6)
NS (6)
NS (7)
NS (7)
Table S7-3. Male SVL after kappa transformation (allowing multiple means)
Model
Small islands Large island TC & TG Large island other Mainland deltaAIC wtAIC Max. likelihood
M5 theta
1.000
0.411
1.000
1.000
0.000
0.390
121.754
M5 LCI
0.242
M5 UCI
0.738
M4 theta
1.000
0.446
1.213
1.000
1.636
0.172
122.028
M4 LCI
0.265
0.754
M4 UCI
0.795
1.958
M3 theta
1.000
0.381
1.000
0.818
1.659
0.170
122.017
M3 LCI
0.226
0.505
M3 UCI
0.681
1.382
M9 theta
1.000
0.421
1.031
1.031
2.178
0.131
121.757
M9 LCI
0.525
0.241
M9 UCI
1.985
0.733
M1 theta
1.000
0.415
1.128
0.896
3.767
0.059
122.070
M1 LCI
0.541
0.247
0.701
0.553
M1 UCI
2.046
0.740
1.821
1.512
M11 theta
1.000
1.000
1.441
1.000
6.139
0.018
118.684
M11 LCI
0.889
M11 UCI
2.349
All equal
1.000
1.000
1.000
1.000
6.181
0.018
117.584
M2 theta
1.000
1.000
1.741
1.437
6.779
0.013
119.457
M2 LCI
1.076
0.887
M2 UCI
2.831
2.428
M7 theta
1.000
0.665
1.065
0.665
6.914
0.012
119.389
M7 LCI
0.810
0.990
M7 UCI
3.093
2.608
M8 theta
1.000
0.827
0.827
0.827
8.013
0.007
117.747
M8 LCI
0.649
M8 UCI
2.494
M10 theta
1.000
1.000
1.000
1.054
8.296
0.006
117.606
M10 LCI
0.650
M10 UCI
1.782
M6 theta
1.000
0.807
0.807
0.882
10.082 0.003
117.805
M6 LCI
0.666
0.674
M6 UCI
2.556
1.847
Model average
1.000
0.453
1.065
0.967
P (k)
** (6)
* (7)
* (7)
* (7)
* (8)
NS (6)
(5)
NS (7)
NS (7)
NS (6)
NS (6)
NS (7)
Female SVL
Table S7-4. Female SVL assuming a single mean
Model
Small islands Large island TC & TG Large island other Mainland deltaAIC wtAIC Max. likelihood
M4 theta
1.000
0.424
1.460
1.000
0.000
0.265
147.964
M4 LCI
0.247
0.916
M4 UCI
0.770
2.340
M5 theta
1.000
0.351
1.000
1.000
0.036
0.260
146.895
M5 LCI
0.203
M5 UCI
0.646
M9 theta
1.000
0.482
1.434
1.434
1.201
0.145
147.364
M9 LCI
0.363
0.194
M9 UCI
1.477
0.618
M3 theta
1.000
0.320
1.000
0.785
1.377
0.133
147.276
M3 LCI
0.186
0.485
M3 UCI
0.584
1.326
M1 theta
1.000
0.471
1.616
1.156
2.000
0.098
148.029
M1 LCI
0.520
0.274
1.014
0.714
M1 UCI
2.119
0.858
2.590
1.952
M2 theta
1.000
1.000
2.414
1.772
3.035
0.058
146.447
M2 LCI
1.509
1.094
M2 UCI
3.883
2.992
M11 theta
1.000
1.000
1.778
1.000
4.823
0.024
144.502
M11 LCI
1.109
M11 UCI
2.868
M7 theta
1.000
0.887
1.620
0.887
6.822
0.009
144.553
M7 LCI
0.586
1.140
M7 UCI
2.402
2.947
All equal
1.000
1.000
1.000
1.000
8.455
0.004
141.648
M8 theta
1.000
1.221
1.221
1.221
10.231 0.002
141.798
M8 LCI
0.425
M8 UCI
1.746
M10 theta
1.000
1.000
1.000
1.047
10.498 0.001
141.664
M10 LCI
0.647
M10 UCI
1.768
M6 theta
1.000
1.214
1.214
1.238
12.328 0.001
141.800
M6 LCI
0.428
0.630
M6 UCI
1.756
1.722
Model average
1.000
0.460
1.352
1.094
P (k)
** (4)
** (3)
** (4)
** (4)
** (5)
** (4)
* (3)
NS (4)
(2)
NS (3)
NS (3)
NS (4)
Table S7-5. Female SVL with reduced data set (allowing multiple means)
Model
Small islands Large island TC & TG Large island other Mainland deltaAIC wtAIC Max. likelihood
M7 theta
1.000
0.504
1.933
0.504
0.000
0.291
91.171
M7 LCI
0.995
0.445
2.112
0.574
M7 UCI
4.377
1.855
6.918
2.749
M11 theta
1.000
1.000
2.878
1.000
0.584
0.218
89.726
M11 LCI
1.571
M11 UCI
5.228
M4 theta
1.000
0.571
2.452
1.000
0.968
0.180
90.687
M4 LCI
0.296
1.353
M4 UCI
1.216
4.409
M1 theta
1.000
0.452
1.949
0.575
2.082
0.103
91.306
M1 LCI
0.502
0.235
1.076
0.286
M1 UCI
2.201
0.958
3.503
1.320
M3 theta
1.000
0.280
1.000
0.346
2.210
0.097
90.066
M3 LCI
0.144
0.172
M3 UCI
0.609
0.794
M2 theta
1.000
1.000
2.731
0.838
2.717
0.075
89.812
M2 LCI
1.489
0.416
M2 UCI
4.967
1.927
M5 theta
1.000
0.340
1.000
1.000
5.569
0.018
87.234
M5 LCI
0.170
M5 UCI
0.769
M9 theta
1.000
0.473
1.562
1.562
6.660
0.010
87.841
M9 LCI
0.321
0.154
M9 UCI
1.411
0.672
M10 theta
1.000
1.000
1.000
0.464
8.638
0.004
85.700
M10 LCI
0.230
M10 UCI
1.070
All equal
1.000
1.000
1.000
1.000
9.677
0.002
84.052
M6 theta
1.000
1.309
1.309
0.581
10.485 0.002
85.928
M6 LCI
0.378
0.220
M6 UCI
1.715
1.024
M8 theta
1.000
1.180
1.180
1.180
11.751 0.001
84.143
M8 LCI
0.420
M8 UCI
1.899
Model average
1.000
0.636
2.175
0.740
P (k)
*** (7)
*** (6)
** (7)
** (8)
** (7)
** (7)
* (6)
* (7)
NS (6)
(5)
NS (7)
NS (6)
Table S7-6. Female SVL after kappa transformation (allowing multiple means)
Model
Small islands Large island TC & TG Large island other Mainland deltaAIC wtAIC Max. likelihood
M4 theta
1.000
0.412
1.521
1.000
0.000
0.288
149.509
M4 LCI
0.245
0.954
M4 UCI
0.732
2.440
M5 theta
1.000
0.337
1.000
1.000
0.440
0.231
148.197
M5 LCI
0.198
M5 UCI
0.609
M9 theta
1.000
0.493
1.579
1.579
1.088
0.167
148.965
M9 LCI
0.338
0.184
M9 UCI
1.311
0.562
M1 theta
1.000
0.486
1.796
1.266
1.855
0.114
149.688
M1 LCI
0.534
0.289
1.126
0.782
M1 UCI
2.069
0.864
2.881
2.139
M3 theta
1.000
0.310
1.000
0.793
1.921
0.110
148.549
M3 LCI
0.183
0.490
M3 UCI
0.556
1.340
M2 theta
1.000
1.000
2.649
1.907
2.997
0.064
148.010
M2 LCI
1.652
1.178
M2 UCI
4.269
3.220
M11 theta
1.000
1.000
1.855
1.000
5.737
0.016
145.548
M11 LCI
1.154
M11 UCI
3.002
M7 theta
1.000
0.902
1.712
0.902
7.841
0.006
145.589
M7 LCI
0.585
1.180
M7 UCI
2.319
3.069
All equal
1.000
1.000
1.000
1.000
10.076 0.002
142.301
M8 theta
1.000
1.239
1.239
1.239
11.880 0.001
142.477
M8 LCI
0.423
M8 UCI
1.702
M10 theta
1.000
1.000
1.000
1.058
12.183 0.001
142.325
M10 LCI
0.654
M10 UCI
1.787
M6 theta
1.000
1.228
1.228
1.264
14.052 0.000
142.483
M6 LCI
0.427
0.636
M6 UCI
1.715
1.738
Model average
1.000
0.458
1.462
1.162
P (k)
*** (7)
*** (6)
** (7)
** (8)
** (7)
** (7)
* (6)
* (7)
(5)
NS (6)
NS (6)
NS (7)
Sexual size dimorphism in SVL
Table S7-7. SSD SVL assuming a single mean
Model
Small islands Large island TC & TG Large island other Mainland deltaAIC wtAIC Max. likelihood
M9 theta
1.000
0.533
0.274
0.274
0.000
0.474
251.129
M9 LCI
1.933
1.145
M9 UCI
7.626
3.399
M1 theta
1.000
0.533
0.288
0.252
1.931
0.181
251.230
M1 LCI
0.530
0.314
0.182
0.152
M1 UCI
2.091
0.931
0.464
0.438
M2 theta
1.000
1.000
0.416
0.365
2.368
0.145
249.946
M2 LCI
0.263
0.221
M2 UCI
0.670
0.630
M8 theta
1.000
0.342
0.342
0.342
3.637
0.077
248.259
M8 LCI
1.549
M8 UCI
6.085
M6 theta
1.000
0.378
0.378
0.256
3.898
0.068
249.181
M6 LCI
1.403
0.406
M6 UCI
5.525
1.184
M7 theta
1.000
0.384
0.292
0.384
4.567
0.048
248.846
M7 LCI
1.381
0.477
M7 UCI
5.431
1.226
M11 theta
1.000
1.000
0.572
1.000
10.433 0.003
244.861
M11 LCI
0.359
M11 UCI
0.924
M10 theta
1.000
1.000
1.000
0.531
10.691 0.002
244.732
M10 LCI
0.319
M10 UCI
0.926
M4 theta
1.000
1.056
0.584
1.000
12.505 0.001
244.877
M4 LCI
0.619
0.367
M4 UCI
1.856
0.943
M3 theta
1.000
1.158
1.000
0.555
12.535 0.001
244.862
M3 LCI
0.684
0.335
M3 UCI
2.017
0.966
All equal
1.000
1.000
1.000
1.000
13.525 0.001
242.276
M5 theta
1.000
1.342
1.000
1.000
14.451 0.000
242.852
M5 LCI
0.789
M5 UCI
2.349
Model average
1.000
0.572
0.314
0.297
P (k)
*** (4)
*** (5)
*** (4)
*** (3)
** (4)
** (4)
* (3)
* (3)
NS (4)
NS (4)
(2)
NS (3)
Table S7-8. SSD SVL with reduced data set (allowing multiple means)
Model
Small islands Large island TC & TG Large island other Mainland deltaAIC wtAIC Max. likelihood
M8 theta
1.000
0.236
0.236
0.236
0.000
0.350
157.580
M8 LCI
2.127
M8 UCI
9.301
M6 theta
1.000
0.254
0.254
0.140
0.652
0.252
158.413
M6 LCI
1.977
0.259
M6 UCI
8.650
1.397
M9 theta
1.000
0.302
0.210
0.210
1.303
0.182
158.088
M9 LCI
2.395
0.742
M9 UCI
10.465
2.956
M7 theta
1.000
0.240
0.234
0.240
2.312
0.110
157.583
M7 LCI
2.096
0.538
M7 UCI
9.163
1.772
M1 theta
1.000
0.297
0.232
0.142
2.602
0.095
158.623
M1 LCI
0.503
0.154
0.128
0.066
M1 UCI
2.200
0.608
0.422
0.359
M2 theta
1.000
1.000
0.384
0.247
7.580
0.008
154.949
M2 LCI
0.216
0.115
M2 UCI
0.697
0.630
M4 theta
1.000
0.482
0.365
1.000
11.597 0.001
152.941
M4 LCI
0.248
0.202
M4 UCI
0.990
0.664
M11 theta
1.000
1.000
0.465
1.000
12.557 0.001
151.302
M11 LCI
0.260
M11 UCI
0.845
M10 theta
1.000
1.000
1.000
0.353
14.139 0.000
150.511
M10 LCI
0.165
M10 UCI
0.897
M3 theta
1.000
0.642
1.000
0.311
14.942 0.000
151.268
M3 LCI
0.338
0.146
M3 UCI
1.288
0.790
All equal
1.000
1.000
1.000
1.000
16.535 0.000
148.179
M5 theta
1.000
0.750
1.000
1.000
18.127 0.000
148.517
M5 LCI
0.393
M5 UCI
1.512
Model average
1.000
0.266
0.237
0.200
P (k)
*** (6)
*** (7)
*** (7)
*** (7)
*** (8)
** (7)
** (7)
* (6)
* (6)
* (7)
(5)
NS (6)
Table S7-9. SSD SVL after kappa transformation (allowing multiple means)
Model
Small islands Large island TC & TG Large island other Mainland deltaAIC wtAIC Max. likelihood
M9 theta
1.000
0.515
0.280
0.280
0.000
0.470
257.721
M9 LCI
1.917
1.098
M9 UCI
7.363
3.176
M1 theta
1.000
0.515
0.285
0.273
2.196
0.157
257.731
M1 LCI
0.537
0.308
0.180
0.165
M1 UCI
2.064
0.890
0.459
0.473
M8 theta
1.000
0.349
0.349
0.349
2.829
0.114
255.212
M8 LCI
1.543
M8 UCI
5.900
M2 theta
1.000
1.000
0.418
0.402
3.034
0.103
256.204
M2 LCI
0.265
0.243
M2 UCI
0.673
0.695
M7 theta
1.000
0.394
0.290
0.394
3.546
0.080
255.948
M7 LCI
1.366
0.463
M7 UCI
5.225
1.186
M6 theta
1.000
0.375
0.375
0.278
3.882
0.067
255.780
M6 LCI
1.433
0.446
M6 UCI
5.489
1.287
M11 theta
1.000
1.000
0.558
1.000
9.317
0.004
251.969
M11 LCI
0.352
M11 UCI
0.899
M10 theta
1.000
1.000
1.000
0.579
11.261 0.002
250.997
M10 LCI
0.349
M10 UCI
1.007
M4 theta
1.000
0.972
0.552
1.000
11.496 0.001
251.973
M4 LCI
0.583
0.348
M4 UCI
1.676
0.889
All equal
1.000
1.000
1.000
1.000
12.855 0.001
249.120
M3 theta
1.000
1.093
1.000
0.596
13.348 0.001
251.047
M3 LCI
0.658
0.359
M3 UCI
1.879
1.035
M5 theta
1.000
1.247
1.000
1.000
14.314 0.000
249.470
M5 LCI
0.750
M5 UCI
2.146
Model average
1.000
0.531
0.314
0.314
P (k)
*** (7)
*** (8)
*** (6)
*** (7)
** (7)
** (7)
* (6)
. (6)
. (7)
NS (7)
NS (6)
Supplementary Appendix S8 – source code for rates analyses in R
library(mvtnorm)
library(MASS)
library(ape)
### Sets up the variance-covariance matrices.
### Input is a phylogeny in ape format with node labels corresponding to group
(e.g. "small island", "mainland" etc).
### Requires the discrete states assigned for each branch in the same order as
edge lengths in the phy format. States should be numeric but can be any value.
### The function returns a list of matrices. The matrices are ordered in the
rank order of the states of the discrete trait.
multiThetaMat <- function(phy, discrete, dat, clades="ancestral") {
switch(clades,
"ancestral" = {
# Get the tip labels in alphabetical order
leaves <- c(1:length(phy$tip.label))
a <- data.frame(leaves, row.names=phy$tip.label)
anms <- rownames(a)
asnms <- sort(anms, index.return = TRUE)
a <- a[asnms$x, ]
# Get the discrete data in alphatical order
b <- data.frame(discrete, row.names=rownames(dat))
bnms <- rownames(b)
bsnms <- sort(bnms, index.return = TRUE)
b <- b[bsnms$x, ]
if(sum(ifelse(asnms$x == bsnms$x, 0, 1)) > 0){
stop("Taxon names in phylogeny do not match those in data
frame")
}
leaves.order <- data.frame(tip.index = asnms$ix, data.index =
bsnms$ix, anc.state = b, row.names = asnms$x)
leaves.anc.state <- as.numeric(leaves.order$anc.state)
# Makes data frame of node numbers that correspond to edge index
with ancestral state for internal branches
first.node.number <- length(phy$tip.label)+2
branches <- c(first.node.number:max(phy$edge))
branches.anc.state <as.numeric(phy$node.label[c(2:length(phy$node.label))])
branch.recon <- data.frame(node = branches, anc.state =
branches.anc.state)
# Put together node and tip indices and ancestral states
node.ancestor <- data.frame(edge.index = c(branch.recon$node,
leaves.order$tip.index), anc.state = c(branches.anc.state, leaves.anc.state))
edge.ind <- sort(node.ancestor$edge.index, index.return = TRUE)
node.ancestor <- node.ancestor[edge.ind$ix, ]
# Put tree edge in node order and then put it all together and order
according to the tree edge index
tree.edge <- sort(phy$edge[,2], index.return = TRUE)
tree.edge.dat <- data.frame(node.ancestor, tree.edge = tree.edge$x,
tree.edge.index = tree.edge$ix)
tree.edge.dat <- tree.edge.dat[order(tree.edge.dat$tree.edge.index),]
ancestor <- as.numeric(as.factor(tree.edge.dat$anc.state))
nlevels <- unique(ancestor)
ThetaMat <- vector(mode="list", length = length(nlevels))
for(i in nlevels) {
x.anc <- ifelse(ancestor == i, 1, 0)
state.edge <- x.anc * phy$edge.length
state.phy <- list(edge=phy$edge, edge.length=state.edge,
Nnode=phy$Nnode, tip.label=phy$tip.label)
class(state.phy) <- "phylo"
state.matrix <- vcv.phylo(state.phy)
ThetaMat[[i]] <- state.matrix
}
},
"monophyletic" = {
phy.mat <- vcv.phylo(phy)
nms <- rownames(phy.mat)
if(length(nms) == 0) stop("Need to supply row names for the
Variance matrix")
snms <- sort(nms, index.return = TRUE)
phy.mat <- phy.mat[snms$ix, snms$ix]
disc <- data.frame(discrete, row.names=rownames(dat))
disc.nms <- rownames(disc)
if(length(disc.nms) == 0) stop("Need to supply row names for
the data")
disc.snms <- sort(disc.nms, index.return = TRUE)
disc <- disc[disc.snms$ix, ]
if(sum(ifelse(snms$x == disc.snms$x, 0, 1)) > 0){
stop("Taxon names in phylogeny do not match those in data
frame")
}
ancestor <- as.numeric(as.factor(disc))
nlevels <- unique(ancestor)
ThetaMat <- vector(mode="list", length = length(nlevels))
for(i in nlevels) {
x.anc <- ifelse(discrete == i, 1, 0)
state.matrix <- x.anc * phy.mat
ThetaMat[[i]] <- state.matrix
}
}
)
new.discrete <- vector(mode="integer", length=1)
new.discrete <- data.frame(new.discrete)
ThetaMat <- list(ThetaMat=ThetaMat, states=new.discrete)
class(ThetaMat) <- "theta.mat"
return(ThetaMat)
}
### Internal function to set up design matrix.
make.anc <- function(y, discrete.trait, data=NULL, common.mean=FALSE) {
if(is.factor(discrete.trait) == "FALSE"){
stop("The discrete trait must be a factor")
}
m <- model.frame(y ~ discrete.trait, data)
x <- model.matrix(y ~ discrete.trait, m)
return(x)
}
### Loads the traits
load.traits <- function(y, phy.mat, discrete.trait, data) {
dat <- data.frame(discrete.trait, y, row.names = rownames(data))
nms <- rownames(dat)
snms <- sort(nms, index.return = TRUE)
dat <- dat[snms$x, ]
y <- dat$y
if(length(phy.mat$states[,1])==1){
x <- dat$discrete.trait
} else {
new.discrete <- phy.mat$states
nms <- rownames(new.discrete)
snms <- sort(nms, index.return = TRUE)
new.discrete <- new.discrete[snms$x, ]
x <- new.discrete
}
idx <- which( is.na(dat$y) == FALSE)
Vmat <- vector(mode="list", length = length(phy.mat))
phy.mat <- phy.mat$ThetaMat
for(i in 1:length(phy.mat)) {
Vmatrix <- phy.mat[[i]]
nms <- rownames(Vmatrix)
snms <- sort(nms, index.return = TRUE)
Vmatrix <- Vmatrix[snms$ix, snms$ix]
Vmatrix <- Vmatrix[idx, idx]
Vmat[[i]] <- Vmatrix
}
x <- as.matrix(x)
y <- y[idx]
x <- x[idx, ]
traits <- list(y = y, x = x, Vmat = Vmat)
class(traits) <- "theta.phyo"
return(traits)
}
# Puts together a single variance covariance matrix from the
# individual parts and theta. Requires a "theta.phylo" object and the number of
states
# of the discrete trait
form.V <- function(traits, theta="all.one") {
V <- traits$Vmat
if(theta[1] == "all.one")
c(rep(1,length(traits$Vmat)))
} else { theta <- theta }
{ theta <-
nV <- length(theta)
if(length(theta) != length(traits$Vmat)){
stop("The number of theta's specified does not correspond to the
number of levels in the explantory variable")
}
v1 <- V[[1]]
thetaMats <- vector(mode="list", length = nV)
retMat <- matrix(0, nrow = dim(v1)[1], ncol = dim(v1)[2])
for(i in 1:nV) {
thetaMats[[i]] <- theta[i] * V[[i]]
retMat <- retMat + thetaMats[[i]]
}
retMat <- retMat
return(retMat)
}
# Estimates the mean of a given trait (accounting for phylogeny).
est.mean <- function(traits, theta="all.one", common.mean=FALSE) {
if(theta[1] == "all.one")
c(rep(1,length(traits$Vmat)))
} else { theta <- theta }
{ theta <-
if(length(theta) != length(traits$Vmat)){
stop("The number of theta's specified does not correspond to the
number of levels in the explantory variable")
}
y <- traits$y
x <- as.factor(traits$x)
V <- form.V(traits, theta)
x <- make.anc(y, x)
if(common.mean==FALSE) {x <- x} else { x <- rep(1, length(x[,1]))}
iV <- solve(V)
xVix <- crossprod(x, iV %*% x)
xViy <- crossprod(x, iV %*% y)
mu <- solve(xVix) %*% xViy
return(mu)
}
# Estimates the variance of a given trait (accounting for phylogeny)
est.var <- function(traits, theta="all.one", common.mean=FALSE) {
if(theta[1] == "all.one")
c(rep(1,length(traits$Vmat)))
} else { theta <- theta }
{ theta <-
if(length(theta) != length(traits$Vmat)){
stop("The number of theta's specified does not correspond to the
number of levels in the explantory variable")
}
y <- traits$y
x <- as.factor(traits$x)
if (common.mean==FALSE) {k <- nlevels(x)} else {k <- 1}
V <- form.V(traits, theta)
x <- make.anc(y, x)
if(common.mean==FALSE) {
x <- x} else { x <- rep(1, length(x[,1]))}
mu <- est.mean(traits, theta, common.mean=common.mean)
iV <- solve(V)
e <- y - x %*% mu
s2 <- crossprod(e, iV %*% e)
n <- length(y)
phylo.var <- ( s2 / (n - k) )
return(phylo.var)
}
# Full ML estimation for given x and V
mv.lik <- function(traits, theta="all.one", common.mean=FALSE) {
if(theta[1] == "all.one") { theta <c(rep(1,length(traits$Vmat)))
} else { theta <- theta }
logDetFun <- function(mat) {
svdMat <- La.svd(mat)
d <- svdMat$d
n <- length(d)
logDet <- sum(log(d))
return(logDet)
}
y <- traits$y
x <- as.factor(traits$x)
V <- form.V(traits, theta)
x <- make.anc(y, x)
logDetV <- logDetFun(V)
mu <- est.mean(traits, theta, common.mean)
s2 <- est.var(traits, theta, common.mean)
n <- length(x[,1])
ll <- -n / 2.0 * log( 2 * pi) - n / 2.0 * log(s2) - logDetV /
2.0 - (n - 1)/2.0
max.lik.theta <- ( list(ll = ll, mu = mu, s2 = s2) )
return(max.lik.theta)
}
# Constructor function for estimating max likelihood of models with parameters
fixed to a particular value
make.mv.lik <- function(traits, fixed, common.mean=FALSE) {
op <- fixed
function(theta){
op[!fixed] <- theta
logDetFun <- function(mat) {
svdMat <- La.svd(mat)
d <- svdMat$d
n <- length(d)
logDet <- sum(log(d))
return(logDet)
}
y <- traits$y
x <- as.factor(traits$x)
if (common.mean==FALSE) {k <- nlevels(x)} else {k <- 1}
V <- traits$Vmat
v1 <- V[[1]]
nV <- length(op)
thetaMats <- vector(mode="list", length = nV)
vmat <- matrix(0, nrow = dim(v1)[1], ncol = dim(v1)[2])
for(i in 1:nV) {
thetaMats[[i]] <- op[i] * V[[i]]
vmat <- vmat + thetaMats[[i]]
}
x <- make.anc(y, x)
if(common.mean==FALSE) {
x <- x} else { x <- rep(1, length(x[,1]))}
logDetV <- logDetFun(vmat)
iV <- solve(vmat)
xVix <- crossprod(x, iV %*% x)
xViy <- crossprod(x, iV %*% y)
mu <- solve(xVix) %*% xViy
e <- y - x %*% mu
s2 <- crossprod(e, iV %*% e)
n <- length(y)
phylo.var <- ( s2 / (n - k) )
n <- length(y)
ll <- -n / 2.0 * log( 2 * pi) - n / 2.0 * log(phylo.var) logDetV / 2.0 - (n - 1)/2.0
ypred <- x%*%mu
max.lik.theta <- ( list(ll = ll, mu = mu, phylo.var =
phylo.var) )
return(-1 * max.lik.theta$ll)
}}
# Estimates the ML
optim.var.param <- function(traits, theta="all.one", fixed = "est.thetas",
thetaMIN = 0.001, thetaMAX = 50, common.mean=FALSE) {
if(theta[1] == "all.one")
c(rep(1,length(which(fixed==FALSE))))
} else { theta <- theta }
{ theta <-
if(fixed[1] == "est.thetas") { op <c(rep(FALSE,length(traits$Vmat) - 1), TRUE)
} else { op <- fixed }
mvl <- make.mv.lik(traits, op, common.mean=common.mean)
vo <- try(optim(theta, mvl, method = "L-BFGS-B", lower =
thetaMIN, upper = thetaMAX))
MLTheta <- vo$par
fixed[which(fixed==FALSE)] <- MLTheta
MLTheta <- fixed
ML <- -vo$value
convergence <- vo$convergence
n <- length(traits$y)
if(length(op)!=length(which(op==FALSE))) {
if(common.mean==TRUE) {k <- 2 +
length(which(op==FALSE))
} else { k <(length(which(op==FALSE)) +1 + length(op)) }
} else {
if(common.mean==TRUE) {k <- 1 +
length(which(op==FALSE))
} else { k <(length(which(op==FALSE)) + length(op)) }
}
aic <- -2 * ML + 2 * k
aicc <- -2 * ML + 2 * k + ((2*k*(k+1))/(n-k-1))
max.lik.theta <- list(MLTheta = MLTheta, Max.lik = ML, aic = aic, aicc =
aicc, convergence=convergence, n.parameters = k)
return(max.lik.theta)
}
# Estimate confidence intervals for one of thetas while fixing others at given
value
theta.CI <- function(traits, MLtheta, fixed, thetaMIN = 0.001, thetaMAX=50,
common.mean=FALSE) {
MLtheta <- as.numeric(format(MLtheta))
ML <- mv.lik(traits, MLtheta, common.mean=common.mean)$ll
fixed.thetas <- data.frame(MLtheta, fixed)
use.theta.ind <- which(fixed.thetas$fixed==FALSE)
use.theta <- fixed.thetas$MLtheta[use.theta.ind]
var.fun <- function(vary.theta) {
fixed.thetas$MLtheta[use.theta.ind] <- vary.theta
test.theta <- fixed.thetas$MLtheta
ll <- mv.lik(traits, test.theta,
common.mean=common.mean)$ll
return( ll - ML + 1.92)
}
if(var.fun(thetaMIN) < 0) {
Lci <- uniroot(var.fun, interval = c(thetaMIN,
use.theta))$root
}
if(var.fun(thetaMAX) < 0) {
Uci <- uniroot(var.fun, interval = c(use.theta,
thetaMAX))$root
}
return(c(Lci=Lci, Uci=Uci))
}
# Estimate confidence intervals for all thetas while fixing others at given
value in turn
all.theta.CI <- function(traits, MLtheta, fixed, thetaMIN = 0.001, thetaMAX=50,
common.mean=FALSE) {
n.theta <- length(MLtheta)
all.CIs <- matrix(nrow=n.theta, ncol=2, dimnames=list(c(1:n.theta),
c("Lci", "Uci")))
for(i in 1:n.theta) {
fix.now <- c(rep(TRUE, n.theta))
fix.now[i] <- FALSE
if(fixed[i] == FALSE) {
CI <- theta.CI(traits, MLtheta, fix.now,
common.mean=common.mean)
all.CIs[i,1] <- CI[1]
all.CIs[i,2] <- CI[2]
} else { all.CIs[i,1] <- NA
all.CIs[i,2] <- NA }
}
return(all.CIs)
}
#Estimates the ML and CIs, with option of a report
ML.fun <- function(traits, theta="all.one", fixed = "est.thetas", pretty = TRUE,
thetaMIN = 0.001, thetaMAX = 50, common.mean=FALSE) {
if(theta[1] == "all.one")
c(rep(1,length(which(fixed==FALSE))))
} else { theta <- theta }
{ theta <-
if(fixed[1] == "est.thetas") { op <c(rep(FALSE,length(traits$Vmat) - 1), TRUE)
} else { op <- fixed }
ovp <- optim.var.param(traits, theta, fixed, thetaMIN,
thetaMAX, common.mean=common.mean)
max.lik.theta <- ovp$MLTheta
max.lik <- ovp$Max.lik
lik1 <- mv.lik(traits, rep(1,length(traits$Vmat)),
common.mean=common.mean)$ll
D <- 2 * (max.lik - lik1)
if (length(op) == length(which(op==FALSE)))
{ k2 <-
length(op) - 1
} else { k2 <- length(which(op==FALSE)) }
if(length(op)!=length(which(op==FALSE))) {
if(common.mean==TRUE) {k <- 2 +
length(which(op==FALSE))
} else { k <(length(which(op==FALSE)) +1 + length(op)) }
} else {
if(common.mean==TRUE) {k <- 1 +
length(which(op==FALSE))
} else { k <(length(which(op==FALSE)) + length(op)) }
}
pval <- 1- pchisq(D, k2)
n <- length(traits$y)
CIs.theta <- all.theta.CI(traits, max.lik.theta,
fixed=rep("FALSE", length(traits$Vmat)), common.mean=common.mean)
aic <- -2 * max.lik + 2 * k
aicc <- -2 * max.lik + 2 * k + ((2*k*(k+1))/(n-k-1))
if(common.mean==TRUE) {
aic.theta1 <- -2 * lik1 + 2 * 2
aicc.theta1 <- -2 * lik1 + 2 * 2 +
((2*2*(2+1))/(n-2-1))
} else {
aic.theta1 <- -2 * lik1 + 2 * (1 +
length(op))
aicc.theta1 <- -2 * lik1 + 2 * (1 +
length(op)) + ((2*(1 + length(op))*((1 + length(op))+1))/(n-(1 + length(op))-1))
}
if(pretty == TRUE) {
cat("____________________________\n")
cat("Maximum likelihood estimation, rates
model:\n\n")
cat("ML estimate of theta: ", max.lik.theta, "\n")
cat("Lower confidence intervals for theta:",
CIs.theta[,1], "\n")
cat("Upper confidence intervals for theta:",
CIs.theta[,2], "\n")
cat("Number of parameters: ", k, "\n")
cat("Maximised log likelihood: ", max.lik, "\n")
cat("Log likelihood (theta 1): ", lik1, "\n")
cat("LR statistic (test vs theta = 1):", D)
cat(" P = ", pval )
cat(" df = ", k2, " \n")
cat(" AIC = ", aic, " \n")
cat(" AICc = ", aicc, " \n")
cat(" Theta 1 AIC = ", aic.theta1, " \n")
cat(" Theta 1 AICc = ", aicc.theta1, " \n")
cat("____________________________\n")
}
if(pretty == FALSE) {
max.lik.theta.list <- list(MLTheta =
max.lik.theta, LCI = CIs.theta[,1], UCI = CIs.theta[,2], nParam = k, Max.lik =
max.lik, Lik1 = lik1, LR = D, P = pval, df = k2, AIC = aic, AICc=aicc,
AIC.theta1=aic.theta1, AICc.theta1=aicc.theta1)
return(max.lik.theta.list) }
}
# Generate some random data
dummy.data <- function(traits, theta="all.one", group.means="all.equal") {
if(theta[1] == "all.one") { theta <- c(rep(1,length(traits$Vmat)))
} else { theta <- theta }
V <- form.V(traits, theta=theta)
expect.sd <- sqrt(mean(V[upper.tri(V)]))
if (group.means[1]=="all.equal") {ydum <- as.matrix(t(rmvnorm(1, sigma =
(V) ))) }
else {
x.means <- unique(traits$x)
n.means <- length(x.means)
samp.means <- rep(NA, length(traits$x))
ydum <- vector(mode="list", length=length(group.means))
for (i in 1:n.means) {
samp.means[which(traits$x == (i-1))] <rep(0+(expect.sd*group.means[i]), length(which(traits$x == (i-1))))
}
ydum <- as.matrix(t(rmvnorm(1, mean=samp.means, sigma =
}
return(ydum)
}
(V) )))
Supplementary Appendix S9 – example of rates analyses
The example below is a simple analysis of male snout vent length
source("Thomas_et_al_Anoles_supplement_S8.R")
First read in phylogeny and data. Note that the rownames of the data frame and
the tip labels of the phylogeny should be the same
but they need not be in the same order (ordering is done internally in the
load.traits function below
Read in phylogeny. Phylogeny should be in nexus format with node labels
corresponding to inferred ancestral states for each branch.
tree <- read.nexus("Thomas_et_al_Anoles_supplement_S6.nex")
Read in data to be analysed
main.dat <- read.table("Thomas_et_al_Anoles_supplement_S1.dat", sep="\t",
header=TRUE)
Subset data frame down to relevant columns and use first column as row names
(corresponding to tip labels in phylogeny).
main.dat <- data.frame(geo_ecomorph = main.dat$geo_ecomorph, Male_SVL =
main.dat$Male_SVL, row.names=main.dat[,1])
Now need to convert the phylogeny into multiple variance-covariance matrices
and set up the data for analysis
Set up matrices. Requires a tree, specification of the data to be used to
define matrices and the data frame
tree.mat <- multiThetaMat(tree, main.dat$geo_ecomorph, main.dat)
Now put all the data together. This produces a list of vcv matrices, response
variable (male svl) and the groups to be compared
male.full <- load.traits(log10(main.dat$Male_SVL), tree.mat,
main.dat$geo_ecomorph, main.dat)
Now some analyses
First, a model with a different rate in each of the four groups. The 'fixed'
command is used to determine whether a particular rate is
to be constrained or not. Use '1' to fix a group and 'FALSE' to show that the
parameter is not fixed and should be estimated. The
values should be entered in the same order as the ranking of the groups. That
is, group 0 (small islands) takes position one in the
fixed vector, group 1 (large island trunk crown and trunk ground) takes
position 2 and so on.
The default is to allow each group to take a different mean.
ML.fun(male.full, fixed=c(1,FALSE,FALSE, FALSE), pretty=TRUE)
A different model, force small islands, large island other, and mainland to be
1 and only estimate large island trunk crown and trunk ground
ML.fun(male.full, fixed=c(1,FALSE,1, 1), pretty=TRUE)
Run the same two models, but this time assuming a common mean across all four
groups
ML.fun(male.full, fixed=c(1,FALSE,FALSE, FALSE), pretty=TRUE, common.mean=TRUE)
ML.fun(male.full, fixed=c(1,FALSE,1, 1), pretty=TRUE, common.mean=TRUE)
The output from the last model should look like this:
____________________________
Maximum likelihood estimation, rates model:
ML estimate of theta: 1 0.4145383 1 1
Lower confidence intervals for theta: 0.5357232 0.2446358 0.7154475 0.4946122
Upper confidence intervals for theta: 2.026233 0.7430008 1.95334 1.451399
Number of parameters: 3
Maximised log likelihood: 120.3727
Log likelihood (theta 1): 116.2331
LR statistic (test vs theta = 1): 8.279331 P = 0.00400989 df = 1
AIC = -234.7455
AICc = -234.5936
Theta 1 AIC = -228.4662
Theta 1 AICc = -228.3907
____________________________
The estimates of relative rate (theta) are returned in the same order as in the
input, as are approximate confidence intervals.
Note that CIs are estimated based on the ML model - that is, CIs are calculated
for each parameter in turn while the other
parameters are held at their ML values for that model.
The number of parameters is the number of means estimated (one in this case)
and the number of different rates allowed (two)
The ML for the model is reported along with the log likelihood for the model
where rates are assumed to be all equal.
The likelihood ratio statistic is a comparison of the ML of the model against
the equal rates model.
AIC and AICc are returned for the ML model and the equal rate model.
This is a function for generating dummy data. Traits can be simulated for a
given phylogeny and group definition.
Both rates and means of the groups can be simulated. Means for different groups
are generated in relation to the
number of expected standard deviations.
dummy.data(male.full, theta=c(1,1,1,1), group.means=c(0,0,0,0))
dummy.data(male.full, theta=c(1,2,2,1), group.means=c(0,0,0,0))
dummy.data(male.full, theta=c(1,1,1,1), group.means=c(0,1,2,3))
Supplementary Appendix S10
Parameter estimates (with 2.5 and 97.5 percentiles in parentheses) and type I error
rates from 10,000 simulations per model.
Small island
1
1
1
1
1
0.973
(0.426-1.926)
0.960
(0.427-1.859)
0.972
(0.421-1.903)
1
ML estimate !
Large island
Large island
trunk-ground
other
trunk-crown
1.006
1.031
(0.457-2.418)
(0.505-2.380)
1
1.014
(0.592-1.717)
0.981
1
(0.541-1.715)
0.972
0.998
(0.532-1.745)
(0.591-1.658)
0.971
1
(0.553-1.612)
1
1
Mainland
Type I error
1.034
(0.479-2.468)
1.023
(0.555-1.810)
0.999
(0.567-1.712)
1
0.056
1
0.051
0.053
0.057
0.053
0.052
1
1
1.009
(0.580-1.655)
1
0.977
(0.554-1.653)
1
1
1
0.056
1
0.051
1
1
0.974
(0.428-1.958)
1
1.000
(0.633-1.575)
1.002
(0.614-1.621)
1.002
(0.586-1.627)
1
1
0.053
0.054
0.048