Perspectives in Plant Ecology, Evolution and Systematics 14 (2012) 205216

Contents lists available at SciVerse ScienceDirect

Perspectives in Plant Ecology, Evolution and Systematics

journal homepage: www.elsevier.de/ppees

Research article

The inuence of re on phylogenetic and functional structure of woody

savannas: Moving from species to individuals
Marcus V. Cianciaruso a, , Igor A. Silva b , Marco A. Batalha b , Kevin J. Gaston c , Owen L. Petchey d
a

Department of Ecology, Universidade Federal de Gois, CP 131, 74001-970 Goinia, GO, Brazil
Department of Botany, Universidade Federal de So Carlos, CP 676, 13565-905 So Carlos, SP, Brazil
c
Environment and Sustainability Institute, University of Exeter, Penryn, Cornwall TR10 9EZ, UK
d
Institute for Evolutionary Biology and Environmental Studies, University of Zurich, Winterthurerstrasse 190 8057 Zurich, Switzerland
b

a r t i c l e

i n f o

Article history:
Received 22 May 2010
Received in revised form 28 October 2011
Accepted 26 November 2011
Keywords:
Community assembly
Environmental ltering
Intraspecic variability
Plant traits

a b s t r a c t
Fire is a key determinant of tropical savanna structure and functioning. High re frequencies are expected
to assemble closely related species with a restricted range of functional trait values. Here we determined the effect of re on phylogenetic and functional diversity of woody species and individuals in
savanna communities under different re frequencies. We found phylogenetic signals for one third of
the functional traits studied. High numbers of res simultaneously led to phylogenetic overdispersion
and functional clustering when communities were represented by mean trait values with all traits that
putatively should be affected or respond to re. This nding is important, because it shows that the
relationship between ecological processes and the phylogenetic structure of communities is not straightforward. Thus, we cannot always assume that close relatives are more similar in their ecological features.
However, when considering a different set of traits representing different plant strategies (re resistance/avoidance, physiological traits and regeneration traits), the results were not always congruent.
When asking how communities are assembled in terms of individuals (not species) the outcome was
different from the species-based approach, suggesting that the realised trait values rather than mean
species trait values have an important role in driving community assembly. Thus, intraspecic trait
variability should be taken into account if we want fully to improve our mechanistic understanding of
assembly rules in plant communities.
2011 Elsevier GmbH. All rights reserved.

Introduction
A fundamental goal in ecology is to understand the process
by which local communities are assembled (Weiher and Keddy,
1999; Pavoine and Bonsall, 2011). In recent years, community
assembly rules have received increased attention from ecologists. A wide variety of processes may play important roles,
including mutualism, facilitation, dispersal limitation and random
colonisation-extinction events (see Leibold et al., 2004; Pausas and
Verd, 2010; Pavoine and Bonsall, 2011).
Two processes in particular have received much attention: environmental ltering and competitive interactions among species
(Webb et al., 2002; Kraft et al., 2007; Vamosi et al., 2009; Mouchet
et al., 2010; Pausas and Verd, 2010). These processes have
opposing (but not exclusive) effects on functional similarity and
phylogenetic relatedness of co-occurring species. Environmental

Corresponding author.
E-mail address: cianciaruso@gmail.com (M.V. Cianciaruso).
1433-8319/$ see front matter 2011 Elsevier GmbH. All rights reserved.
doi:10.1016/j.ppees.2011.11.004

lters select those species that can persist within a community

on the basis of their tolerance to the prevailing abiotic conditions.
The result is an assembly of species with similar characteristics
and niches (Fukami et al., 2005; Pausas and Verd, 2010). Consequently, a restricted range of species trait values is viewed as
evidence of environmental ltering (Weiher et al., 1998; Pausas
and Verd, 2008). Competitive interactions, by contrast, are predicted to result in co-occurring species with dissimilar traits, and
this can be interpreted as evidence for limiting similarity (niche
differentiation among species) (Weiher et al., 1998; Webb et al.,
2002). The predicted effects of these two processes on phylogenetic
relatedness depend on how functional traits evolved in species lineages. Phylogenetic signal in functional traits (i.e., traits are more
similar among closely related species; sensu Losos, 2008) and environmental ltering should result in species more closely related
than expected by chance (i.e., phylogenetic clustering). By contrast,
species living in communities driven by competition should be
less closely related (i.e., phylogenetic overdispersion; Webb et al.,
2002; Cavender-Bares et al., 2006; Kraft et al., 2007; Vamosi et al.,
2009; but see Mayeld and Levine, 2010). However, if functional

206

M.V. Cianciaruso et al. / Perspectives in Plant Ecology, Evolution and Systematics 14 (2012) 205216

traits are convergent in species lineages, environmental ltering

should assemble communities containing distantly related species,
whereas competition should remove any systematic associations
resulting in community assemblages indistinguishable from random (Webb et al., 2002). Note that the effects and relative strengths
of environmental ltering are likely to change with spatial scale
(Gmez et al., 2010; Thuiller et al., 2010), though this is not a focus
of our study.
Strongly parallel literatures have developed on the measurement of phylogenetic and functional diversity, and of the respective
clustering or over-dispersion of species. The concomitant use of
measures of phylogenetic diversity and functional diversity thus
seems likely to help in understanding the mechanisms of community assembly (Pavoine and Bonsall, 2011; Diniz Filho et al., 2011).
To date, however, there have been rather few such studies (but see
Pavoine et al., 2010; Meynard et al., in press; Sa et al., 2011). In an
experimental grassland, Cadotte et al. (2009) found more support
in using phylogenetic diversity than functional diversity to understand productivity patterns. However, a meta-analysis showed that
this is not always the case (Cadotte et al., 2008). Further, there is
evidence that species traits may be more important than species
relatedness to ecosystem processes (Daz and Cabido, 2001; Hooper
et al., 2002). If important traits were very labile with a low phylogenetic signal or when species exhibit considerable phenotypic
plasticity, phylogenetic distances will fail to resemble ecological
(functional) distance among species (Losos, 2008).
Although both phylogenetic and functional diversity have most
commonly been determined using species as the most nely
resolved taxonomic units, inclusion of intra-specic trait variation
is also important (Cianciaruso et al., 2009; Albert et al., 2010; de
Bello et al., 2011). Competition for resources, niche width expansion and, ultimately, natural selection occur at the individual level
(Pachepsky et al., 2007). Plant phenotypic diversity was found to
inuence ecosystem functioning in several studies (e.g., Madritch
and Hunter, 2002, 2003; and references in Schmitz et al., 2003), and
there is evidence for a role of phenotypic variability in species coexistence and responses to both biotic and abiotic lters (Jung et al.,
2010). Furthermore, environmental variation can cause important trait variation without species turnover (Fajardo and Piper,
2011). Intraspecic variability can generate trait-mediated indirect
interactions and thus inuence the strength of competition, facilitation, and species-environment relationships (Miner et al., 2005).
In ooded meadows Jung et al. (2010) found that plant species
were sorted not only according to their mean trait value but also
according to their ability to t their trait values to local abiotic and
biotic requirements. This implies that a priori unexpected species
(species whose mean trait values do not satisfy the abiotic and
biotic requirements) are still able to pass through environmental
lters (Jung et al., 2010). This said, empirical analyses of the role of
functional diversity at the individual level remain scarce.
Fire-prone communities are very useful to test community
assembly theories, because re may act as a strong environmental
lter for particular combinations of functional traits in communities with different re frequencies (Pausas and Verd, 2008; Silva
and Batalha, 2010). Savannas have experienced re for millions of
years and, as a consequence, savanna plants evolved re-tolerance
mechanisms, and sometimes even require re to maintain their
populations (Hoffmann, 1998). Among woody species, the main
adaptations are those that allow thermal isolation of living internal tissues, such as strong suberisation of the trunk and branches,
clonal reproduction, and the ability to sprout vigorously from
underground organs (Miranda et al., 2002; Pausas and Lavorel,
2003). Moreover, different processes related with recruitment,
owering, dispersal, and germination might be modied by re
(Coutinho, 1990). The role of re in assembling plant communities

has been widely investigated, especially in the Mediterranean basin

(see Verd and Pausas, 2007 for references), where it produces
functionally and phylogenetically clustered communities (Pausas
and Verd, 2008). However, there are few studies focusing on
community assembly in savannas and, as far as we know, none
including intraspecic trait variation.
Here, we determined the effect of re on the phylogenetic
and functional diversity of woody species in savanna communities under different re frequencies. We used null models to test
whether levels of community phylogenetic and functional structure were different from what we would expect by chance. In
summary, we asked the following questions: (i) Are the functional
traits more similar among closely related plant species? (ii) Do
phylogenetic and functional diversities differ between sites experiencing different re frequencies from that expected by the random
assembly of communities? (iii) Are there any differences between
species- and individual-based approaches to community assembly
based on functional attributes?

Methods
Study area and data
The Emas National Park (ENP) is located in the Brazilian
Central Plateau, southwestern Gois State (17 49 18 28 S and
52 39 53 10 W), and is one of the largest and most important savanna reserves in South America, covering ca. 133,000 ha.
Regional climate is tropical and humid, with a wet summer and
dry winter, classied as Aw following Kppen (1931). The dry season is from June to August and the wet season from September to
May. Annual rainfall and mean temperature lie around 1,745 mm
and 24.6 C, respectively. In the park, we nd a gradient from
open (68.1% of its area) to closed savannas (25.1%), as well as wet
grasslands (4.9%) and riparian and semideciduous forests (1.2%)
(Ramos-Neto and Pivello, 2000).
Historically, ENP was exploited by farmers for cattle ranching,
and dry season burnings were used to promote forage regrowth
every year. In 1984, the park was completely fenced, cattle were no
longer allowed inside, and a re exclusion policy was established
(Ramos-Neto and Pivello, 2000). As a consequence, uncontrolled
wildres occurred every 34 years, burning on average 80% of its
total area (Ramos-Neto and Pivello, 2000; Franca et al., 2007). Since
1994, when a catastrophic re burned almost 95% of ENPs area,
approximately 10 km2 of preventive rebreaks are burned annually in the dry season, and a re brigade is permanently stationed in
the park to prevent anthropogenic res during this period (Franca
et al., 2007). As a result, nowadays there are few occurrences of
anthropogenic burnings inside the park (almost all res are lightning res), and re frequency at a given point is around 67 years
on average.
We used a long time-series of satellite images to map re scars
within ENP from 1973 to 2009, creating a re map from which we
can determine the number of res in this period for any location
in the park. Using this information, we selected sites covered
with savanna vegetation that experience high (HiFi), intermediate
(MidFi) and low (LowFi) re frequencies. We randomly placed 64
plots (each a 25 m2 quadrat) across the study sites: 21 plots in
HiFi sites, 21 in MidFi sites, and 22 in LowFi sites. There was no
overlap in re frequency among each classication. On average,
the number of res between 1973 and 2009 (mean standard
deviation) was 16.00 1.12 in HiFi plots; 10.50 1.15 in MidFi
plots; and 7.00 1.03 in LowFi plots. In each plot, from September
2009 to January 2010, we sampled all woody individuals with
stem diameter equal to or higher than 3 cm at soil level. We

M.V. Cianciaruso et al. / Perspectives in Plant Ecology, Evolution and Systematics 14 (2012) 205216

207

Roupala montana
Davill a ell iptica
Eriotheca pubescens
Eriotheca gracilipes
Qualea parviflora
Myrcia lasiantha
Myrcia guianensis
Myrcia obovata
Myrcia camapuensis
Myrcia bella
Psidium laruotteanum
Eugenia punicifoli a
Eugenia bimarginata
Eugenia aurata
Miconia albicans
Lafoensia pacari
Sclerolobium aureum
Hymenaea stigonocarpa
Dimorphandra mollis
Mimosa amnisatri
Stryphnodendron adstringens
Anadenanthera peregrina
Albizia niopoides
Machaerium acutifolium
Acosmium dasycarpum
Rourea induta
Connarus suberosus
Plenckia populnea
Casearia sylvestris
Ouratea spectabil is
Ouratea acuminata
Byrsonima verbascifolia
Byrsonima coccolobifoli a
Erythroxylum tortuosum
Erythroxylum suberosum
Erythroxylum campestre
Kielmeyera coriacea
Caryocar brasiliense
Styrax ferrugineus
Pouteria torta
Pouteria ramiflora
Diospyros hispida
Piptocarpha rotundifolia
Eremanthus erythropappus
Schefflera malmei
Tabebuia ochracea
Tabebuia aurea
Pali courea rigida
Aspidosperma tomentosum

140

100

60

40

20

Fig. 1. The phylogenetic tree assembled for the cerrado species in all sampled plots (HighFi, MidFi, and LowFi) in Emas National Park, central Brazil. The relationship among
species was based on the current Phylomatic tree (tree R20080147, Webb and Donoghue, 2005). The age of marked nodes were based on minimum ages of nodes determined
for genus, families and higher orders from fossil data (Wikstrm et al., 2001). Other nodes were placed evenly between dated nodes and dated nodes. We improved tree
resolution by consulting recent phylogenies of Fabaceae (Simon et al., 2009), Malpighiales (Wurdack and Davis, 2009), and Myrtaceae (Costa, 2009). In these clades, we
evenly placed the genus nodes above family dated nodes (only including branches with more than 80% of support). The scale is in million years.

used Plantminer (Carvalho et al., 2010a) to search for families

(following the Angiosperm Phylogeny Group III, 2009), authors,
and synonyms concerning our species list. The soil in the study
areas is a typical Oxisol, poor in nutrients, well drained, and acidic.
Phylogenetic data
We constructed a phylogenetic tree for all sampled species
(Fig. 1) using the Phylomatic software, a phylogenetic database
and toolkit for the assembly of phylogenetic trees (Webb and
Donoghue, 2005). Phylogenetic distances among species from different families were estimated from the current Phylomatic tree
(R20080147). The backbone of the Phylomatic tree is the phylogenetic relationship among Angiosperm Phylogeny Group orders
(Stevens, 2001). Phylomatic tree is a supertree assembled by hand,
rather than by an automated supertree algorithm, and conicting branching patterns were resolved subjectively. It is, however,

intended to represent a pragmatic approximation of the true phylogeny of seed plants (Webb and Donoghue, 2005). We used a less
conservative Phylomatic tree which includes some branches with
bootstrap support less than 80% and more details of the decisions
involved in phylogenetic tree construction are given at the Phylomatic website (http://www.phylodiversity.net/phylomatic/). In
order to solve some of the politomies we improved tree resolution
by consulting recent phylogenies of some clades: Fabaceae (Simon
et al., 2009), Malpighiales (Wurdack and Davis, 2009), and Myrtaceae (Costa, 2009). In this case, only branches with more than
80% of support were included. Branch lengths were based on minimum ages of nodes determined for families and higher orders from
fossil data (Wikstrm et al., 2001). We placed undated nodes in the
tree evenly between dated nodes with the Branch Length Adjustment algorithm in Phylocom (Webb et al., 2008). This algorithm
took the phylogeny generated by Phylomatic, xed the root node
at 137 million years before present (i.e., the age of the eudicots

208

M.V. Cianciaruso et al. / Perspectives in Plant Ecology, Evolution and Systematics 14 (2012) 205216

Table 1
Traits used to measure functional diversity in savanna woody species at Emas National Park, Brazil.
Trait

Unit

Functional signicance

1. Plant height

2. Basal area
3. Bark thickness

m
mm

4. Specic leaf area

mm2 mg1

5. Leaf size
6. Leaf toughness
7. wood density
8. Leaf N
9. Leaf P
10. Leaf K
11. Resprouting due to top-kill
12. Pollination syndromes

mm2
mg g1
mg mm3
mg g1
mg g1
mg g1
Frequency
Categorical (bees, small insects, moths,
ies, birds, bats)
Categorical (anemochory, zoochory,
and autochory)

Associated with competitive vigour, whole plant fecundity, tolerance or avoidance

of disturbances and top-kill
Competitive vigour, survival ability after re
Protection of vital tissues against damage, thick barks can decrease mortality by
re or accelerate post-re recovery
Highly correlated with several physiological traits related to resource uptake and
use efciency and plant growth strategies
Ecological strategy, with respect to environmental nutrient stress and disturbances
Related to ammability, resistance to physical hazard, disturbed environments
Structural strength, resistance against physical damage

13. Dispersal mode

clade) and xed other nodes we had age estimative from Wikstrm
et al. (2001) (Fig. 1). It then sets all other branch lengths by placing
the nodes evenly between dated nodes, and between dated nodes
and terminals (Webb et al., 2008). This has the effect of minimising
variance in branch length, within the constraints of dated nodes. It
thus produces a pseudo-chronogram that can be useful for estimating phylogenetic distance (in units of time) between taxa mainly
for the analysis of phylogenetic community structure (Webb, 2000;
Webb et al., 2008; Vamosi et al., 2009).
We investigated whether the functional traits of the sampled
species tended to present a phylogenetic signal (i.e., trait similarity between species is related to phylogenetic relatedness, Losos,
2008) with a test based on the variance of phylogenetically independent contrasts (Blomberg et al., 2003). If related species are
similar to each other, the magnitude of independent contrasts will
generally be similar across the tree, resulting in a small variance
of contrast values (Blomberg et al., 2003). The observed contrast
variances are compared to the expectations under a null model of
randomly swapping trait values across the tips of the tree, with
1000 randomisations. For a detailed description of comparative
analyses using phylogenetically independent contrast, see Garland
et al. (1992). We excluded from this analysis the pollination and
dispersal traits, because they are categorical variables. We did this
analysis with the package picante (Kembel et al., 2010) for R (R
Development Core Team, 2010).
Functional traits
We used 13 plant traits (Table 1) that were relatively easy
to measure and represented functional characteristics related to
disturbances such as re and drought (Cornelissen et al., 2003;
Pausas and Paula, 2005). All traits, except pollination syndromes,
were measured or determined according to the protocol proposed by Cornelissen et al. (2003). We determined pollination
syndromes using data available for cerrado species (Gottsberger
and Silberbauer-Gottsberger, 2006). In this study we adopted the
species functional attribute approach proposed by Violle et al.
(2007). An attribute is the particular value taken by a trait at any
place and time. That is, within a species, trait values may exhibit
variation along environmental gradients or through time (Violle
et al., 2007). Here, functional information was plot-specic, and
we considered all the individuals sampled, allowing us to include
intraspecic differences among species (Cianciaruso et al., 2009).
Thus, we collected and analysed functional trait data for all individuals sampled in the plots. Because data on pollination and dispersal

Maximum photosynthetic rate, LNC:LPC ratio related to carbon cycling processes

Competitive vigour, persistence after environmental disturbances
Regeneration traits linked to (re)colonising ability and disturbances

modes was gathered from the literature, there was no intraspecic

variability for them.
Phylogenetic and functional structure of communities
To nd out whether re constrained the phylogenetic structure
of the savanna plant communities we calculated three continuously
distributed metrics of phylogenetic diversity: phylogenetic diversity index (PD; Faith, 1992), mean pairwise phylogenetic distance
(MPD), and mean nearest taxon distance (MNTD; Webb, 2000). PD
is the sum of branch lengths of the phylogenetic tree connecting
all species within a community. MPD is the mean pairwise phylogenetic distance between each species in the community, and it
is considered a tree-wide, or basal, measure of the phylogenetic
relatedness of co-occurring species (Webb, 2000). Lastly, MNTD
quanties the phylogenetic distance between each species and its
nearest neighbour on the phylogenetic tree with which it co-occurs
in the local assemblage. Thus, MNTD can be viewed as a terminal metric of the phylogenetic relatedness of co-occurring species
(Webb, 2000).
To assess the functional structure of communities we used the
same metrics above. This is straightforward since a functional dendrogram and phylogenetic tree can be represented by an identical
data structure, thus, any index applied to a phylogenetic tree can
also be applied to a functional dendrogram or to trait-based distance matrices (Pavoine and Bonsall, 2011). For example, one of
the functional measures we used (termed FD, Petchey and Gaston,
2002, 2006) has its roots on Faiths PD (Faith, 1992). Therefore,
FD is dened as the sum of branch lengths of the functional dendrogram necessary to connect all the species present in a local
community. Using the trait-based distance matrix (which is used
to calculated the functional dendrogram), MPD and MNTD can
be dened, respectively, as the mean pairwise functional distance
between each species in the assemblage (i.e., the mean functional
distance, MFD) and the functional distance between each species
and its nearest neighbour on the distance matrix with which it cooccurs in the local community (i.e., the mean nearest functional
distance, MNFD).
Constructing a functional dendrogram requires that functional
traits be transformed into functional distances, and that distances
be transformed into a dendrogram. Since we used both qualitative
and quantitative plant traits, we used a generalisation of Gower
distance (Pavoine et al., 2009) dedicated to the treatment of mixed
data and then UPGMA clustering. Because one of our aims was to
understand whether re selects for species (or individuals) more

M.V. Cianciaruso et al. / Perspectives in Plant Ecology, Evolution and Systematics 14 (2012) 205216

similar in their biological attributes than expected by chance, we

built species-based and individual-based dendrograms. For the
species-level analysis, for each trait and each species with more
than two individuals, we calculated species average trait values and
constructed a species-level functional dendrogram. For singleton
species, we used the measured value of each trait. We assembled
a trait matrix in which each species was labelled according to the
re treatment in which trait values were measured (HiFi, MidFi,
LowFi), since individuals from the same species but from different
sites may exhibit differences in trait values (see Cianciaruso et al.,
2009 for references). Thus, a species may occur in all three re frequencies, and exhibit different trait values in each. By doing this
we included the intraspecic variability of species because we had
functional trait information from species occurring in different re
treatments (i.e., among-sites spFD, Cianciaruso et al., 2009). To
include both the differences in individual traits among and withinsites (re treatments), we used data on all individuals sampled
in the 64 plots to construct an individual-based dendrogram. This
dendrogram was assembled using an individual level trait matrix.
Using this approach we captured the functional variation simultaneously at both the species and individual level (i.e., iFD approach,
see Cianciaruso et al., 2009).
Since choice of traits can inuence measures of functional diversity (Petchey et al., 2009), we constructed independent functional
dendrograms using different sets of traits thought to represent distinct ecological strategies (see Cornelissen et al., 2003; Pausas and
Paula, 2005 for references). The three trait sets were (i) re resistance/avoidance (plant height, basal area, resprouting after top-kill,
and bark thickness); (ii) physiological response to disturbances
(stem specic density, leaf size, SLA, leaf toughness, leaf N, P and
K contents); and (iii) regenerating traits (pollination and dispersal
mode). We calculated the same above-mentioned metrics for each
of these trait sets. Analyses were repeated separately for each functional dendrogram but for the individual-based dendrogram with
regenerative traits because there was no intraspecic variability
for this data. In this case spFD and iFD approaches should converge
(Cianciaruso et al., 2009).

209

Fig. 2. Standardised effect sizes of phylogenetic diversity (PD; circles) and mean
phylogenetic distance (MPD; squares) in woody savanna communities and their
95% condence intervals for each re frequency evaluated. Symbols in black indicate
values signicantly different from zero (random).

than zero indicate phylogenetic overdispersion (Webb et al., 2002;

Vamosi et al., 2009).
To test whether standard effects presented signicant deviations from a null expectation (mean = 0) we used a one sample
t-test. These metrics and the null model approach have been used
widely in explaining community assembly, searching for signals
of environmental lter or limiting similarity in plant assemblages
(e.g., Webb, 2000; Swenson et al., 2006; Kraft et al., 2007). In addition, FD (and by extension PD) was one of the best metrics to detect
assembly rules using null models under different simulation scenarios (Mouchet et al., 2010). We calculated all the metrics and did
all the analysis and randomisations using the functions ses.pd,
ses.mpd and ses.mntd of the picante package (Kembel et al.,
2010) for R (R Development Core Team, 2010).
Results

Data analysis
We used null models to determine whether coexisting species
were more or less similar than expected by chance in phylogenetic
and functional terms. We compared the observed values of each
metric found under each re treatment (HiFi, MidFi and LowFi)
with the corresponding mean value obtained from 1000 randomly
generated null communities using an independent swap algorithm
(Gotelli and Entsminger, 2001). By doing this we maintained the
observed species richness and abundance in the random communities. This null model combines good Type I error rates with the
power to detect niche-based assembly processes (Kembel, 2009).
Our species pool (or individual pool in the case of iFD) was dened
as all species (or individuals) found in all the 64 plots. For all metrics
we calculated the standardised effect sizes against null communities as
standardised effect =

(obsMetric rndMetric)
sd.rndMetric

where obsMetric is the observed value of the metric under analysis, rndMetric is the mean metric value of null communities, and
sd.rndMetric is the standard deviation of the 1000 random values
of the metric.
For MPD and MNTD these standardised effect size values are also
known as the nearest relative index (NRI) and the nearest taxon
index (NTI), respectively (Webb, 2000). Values of standard effects
greater than zero indicate phylogenetic clustering and values lower

In total we sampled 347 individuals, belonging to 49 species.

There were 15 species in the HiFi plots, 26 species in the MidFi and
38 species in the LowFi (Fig. 1; Appendix 1). All metrics we used to
describe the phylogenetic and functional structure of communities
led us to the same conclusions. So, for the sake of clarity hereafter
we just present the results for FD/PD and MPD/MFD.
Trait evolution and phylogenetic structure of communities
We found signicant phylogenetic signals for one third of the
functional traits analysed (Table 2). Strong phylogenetic signals
were observed in N, P, and K leaf content, and leaf toughness, such
that closely related species tended to have similar values for these
traits (Table 2). The other traits were randomly distributed across
the phylogeny (Table 2).
High re frequency (HiFi plots) were composed of species that
were more distantly related than expected by chance, that is, they
were phylogenetically overdispersed (Fig. 2). We did not observe
any signicant phylogenetic structure in the MidFi and LowFi plots
(Fig. 2).
Functional structure of communities: species level approach
When we used the 13 traits (Table 1) all metrics showed
that species co-occurring in HiFi plots were more similar in their
functional attributes than expected by chance (Fig. 3). However,

210

M.V. Cianciaruso et al. / Perspectives in Plant Ecology, Evolution and Systematics 14 (2012) 205216

Table 2
Tests for phylogenetic signal in functional traits of cerrado woody species, in Emas National Park, central Brazil. The observed variance and the mean of random variances
of the phylogenetic independent contrasts (PICs) are given. Signicant values at = 0.05 are presented in bold face.
Functional trait

Observed variance of PICs

Mean of random variances of PICs

Plant height
Basal area
Bark thickness
Specic leaf area
Leaf size
Leaf toughness
Stem specic density
Leaf N content
Leaf P content
Leaf K content
Resprouting due to top-kill

0.69
0.43
0.57
0.78
0.66
0.86
0.71
1.13
0.82
0.81
0.71

0.070
0.000
1.080
0.340
0.104
0.030
0.001
1.370
0.119
0.380
0.012

0.080
0.000
0.970
0.420
0.111
0.043
0.002
2.380
0.196
0.520
0.014

0.240
0.950
0.780
0.100
0.320
0.002
0.180
0.001
0.001
0.022
0.220

species were assembled at random with respect to their functional

attributes under MidFi and LowFi (Fig. 3). When using the set of
traits representing resistance or avoidance to re we found functional clustering for HiFi and LowFi plots whereas in MidFi plots the
values were not different from random (Fig. 3). When using traits
representing physiological response to disturbances we found that
HiFi and MidFi plots were composed of species that were more similar than expected by chance, whereas at LowFi we found a random
pattern (Fig. 3). Finally, when using the two regenerating traits (pollination and dispersal mode) we did not nd any deviations from
randomness (Fig. 3).
Functional structure of communities: individual level approach
Using all traits we found a consistent pattern of functional clustering in MidFi and LowFi plots and randomness for HiFi plots
(Fig. 4). However, when using trait subsets we found a different
picture: in LowFi, individuals were more similar in their re resistance/avoidance attributes than expected by chance (Fig. 4) but

were randomly assembled in HiFi and MidFi plots. For the physiological traits, individuals were more similar in the HiFi plots and
assembled at random in MidFi and LowFi plots (Fig. 4).

Discussion
Our main nding was that re frequency, the functional traits
considered, and intraspecic trait variability can each affect the
signature of community assembly in the savanna communities
studied. High number of res simultaneously led to phylogenetic
overdispersion and functional clustering patterns when we used
a species-level approach (i.e., species being represented by mean
trait values) with all traits that putatively should be affected or
respond to re. However, different sets of traits, representing
different strategies of plants sometimes led to different results. Furthermore, analyses of individuals (not species) produced different
conclusions, suggesting that including intraspecic variability may
produce shifts on how we understand community assembly.

Fig. 3. Species-based standardised effect sizes of functional diversity (FD; circles) and mean functional distance (MFD; squares) in woody savanna communities and their
95% condence intervals for each re frequency and trait set evaluated. Symbols in black indicate values signicantly different from zero (random).

M.V. Cianciaruso et al. / Perspectives in Plant Ecology, Evolution and Systematics 14 (2012) 205216

211

Fig. 4. Individual-based standardised effect sizes of functional diversity (FD; circles) and mean functional distances (MFD; squares) in woody savanna communities and
their 95% condence intervals for each re frequency and trait set evaluated. Symbols in black indicate values signicantly different from zero (random). Regenerative traits
were not analysed because there was no intra-specic variability for them.

Trait evolution and phylogenetic structure of communities

Wherever phylogenetic signals are found in plant traits, environmental lters are predicted to assemble phylogenetically
related species, causing phylogenetic clustering (Kraft et al., 2007;
Vamosi et al., 2009). In contrast to this prediction and to previous ndings showing phylogenetic clustering at local scales (Webb,
2000; Cavender-Bares et al., 2006; Swenson et al., 2006), including in other re-prone communities (e.g., Verd and Pausas, 2007
for Mediterranean vegetation), we found that co-occurring woody
species subjected to high numbers of res were less related than
expected by chance, whereas re frequencies had no apparent
effect on the phylogenetic structure of these communities. This
can be explained by the fact that we observed phylogenetic signals
in only one third of the functional traits studied. These traits (leaf
toughness, leaf N, P, and K concentrations), in spite of their phylogenetic signals, were not enough to produce phylogenetic clustering
pattern in our communities. However, they comprise about 58% of
the traits used to understand the physiological response to disturbances strategy. Indeed, for this set of traits, a phylogenetic signal
was detected (there was phylogenetic clustering of trait values).
Silva and Batalha (2010), when looking for phylogenetic signals in the community structure of six savannas in Brazil also
found that re did not promote phylogenetic clustering. They
postulated that the observed phylogenetic overdispersion was
mainly due to the persistence of long-lived resprouting species
from different plant lineages. Resprouting ability is an important
dimension of the persistence niche of plants that explains the survival of individuals subjected to environmental constraints such
as re or drought (Bond and Keeley, 2005). In general, most of
the plants of the Brazilian savannas are able to resprout after
recurrent res (Miranda et al., 2002; Gottsberger and SilberbauerGottsberger, 2006). For example, comparing to other re-prone

communities in South America, such as the southern temperate

grasslands in which around half of its species are resprouters
(Overbeck and Pfadenhauer, 2007), in the Brazilian savannas, more
than 90% of species can be classied as resprouters (Gottsberger
and Silberbauer-Gottsberger, 2006). Thus, the presence of a large
number of species able to resprout, belonging to different plant lineages, may prevent re from assembling closely related plants in
the savannas we studied. In the Mediterranean vegetation, alternatively, some specic adaptations to re (for example, obligate
seeders - plants that can only regenerate after re from seeds)
are concentrated in few lineages (such as Cistaceae and Fabaceae,
Verd and Pausas, 2007) and consequently frequent res promote phylogenetic clustering. Phylogenetic overdispersion may be
a common characteristic of plant communities in Brazilian savanna
in spite of re (Silva and Batalha, 2009a). Thus, frequent res did not
exclude entire clades from local assemblages; instead, re excludes
species relatively evenly across the phylogenetic tree.
This nding is conceptually important for studies investigating the role of re in these savannas because in fact disassembling
by removal of individuals, and consequently species, is probably much more important than the establishment (assembling) of
individuals that pursuit particular traits suited to one or another
re regime. In other re prone communities, such as Mediterranean heathlands, sclerophyllous shrublands, and temperate pine
forests, distinct strategies of colonisation and space occupation, for
example, resprouters versus obligate seeder species, explain community assembly after res (Pausas and Lavorel, 2003). However, in
these vegetation types, re recurrence is estimated to be between
25 and 40 years or more (DeBano et al., 1998). The re recurrence period is shorter in tropical savannas, and woody plants are
thought to face a strong bottleneck to seedling survival due to re,
drought, competition with grasses and herbivory (Miranda et al.,
2002; Gottsberger and Silberbauer-Gottsberger, 2006; Cianciaruso

212

M.V. Cianciaruso et al. / Perspectives in Plant Ecology, Evolution and Systematics 14 (2012) 205216

et al., 2010). Despite being an important aspect of plant population dynamics, the different establishment strategies after re are
relatively poorly understood for Neotropical savanna species.
The phylogenetic overdispersion of communities under high
levels of disturbance (e.g., high re frequencies) is consistent with
the environmental ltering hypothesis if species under frequent
re share similar traits due to convergence (Weiher et al., 1998;
Fukami et al., 2005; Pausas and Verd, 2008). Our ndings partially corroborate this view, since we found functional clustering
associated with high re frequency (Fig. 3) despite the observed
phylogenetic overdispersion (Fig. 2). Nevertheless, because the
phylogenetic structure of communities depends on the distribution of different traits among lineages (which can have different
evolutionary signals, as we found here) the phylogenetic framework discussed above can be improved by identifying specic traits
(or set of traits) that confer to different species groups the ability to survive under different environmental conditions (Mayeld
et al., 2006; Verd et al., 2009). Only after accounting for such information can we more thoroughly unveil the forces that structure
natural communities.

Functional structure of communities: species level approach

We found that HiFi plots were composed by species more similar
in their functional traits (functional clustering) than that of MidFi
and LowFi plots (Fig. 3). Similar ndings have been observed in
other re-prone communities around the world. In Mediterranean
woodlands, for instance, an increase in re frequency reduced the
functional diversity of plant communities due to the elimination
of species without effective mechanisms for post-re regeneration (Pausas and Verd, 2008). In Brazilian savannas, however,
most of the species sprout promptly after res (Gottsberger and
Silberbauer-Gottsberger, 2006), and the low functional diversity
is mainly due to the reduction in the range of trait values of the
species in HiFi plots. A reduction in the trait range was also observed
between neighbour woody species in an annually burned savanna
(Silva and Batalha, 2009b).
None of the set of traits (i.e., traits related to resistance, physiological response, and regeneration) seems to be favored by re
(Fig. 3). Frequent res decreased the range of traits associated
with physiological responses to disturbance. Thus, frequent res
predominantly assembled species with similar physiological functional traits: leaf traits and stem specic density. Leaf traits are
also important in the mediation of plant-herbivore interactions and
represent a fundamental dimension of the plant defenses against
herbivory. Herbivory may also act as an environmental lter selecting species with similar defenses against herbivory (Harley, 2003).
The fact that our results shows that re assembles more similar leaf
traits in an area than expected by chance encourages investigations
of the role of re in indirectly mediating an important top-down
process in these savannas.
Interestingly, traits related to resistance exhibited functional
clustering at both ends of the re frequency gradient: species were
more similar in their resistance traits under high and low re frequency. This could simply be explained by the fact that the plants
were taller and had greater basal area than expected by chance
under low re frequency, but were shorter and thinner under high
re frequency (Miranda et al., 2002). Knowing the position of the
species in trait space is an important issue, but up to now there
are few metrics which incorporate this aspect (but see Pavoine and
Bonsall, 2011).
Finally, regenerative traits were not affected by re. In Brazilian
savannas, where sprouting is a widespread strategy (Gottsberger
and Silberbauer-Gottsberger, 2006), it is perhaps unsurprising that

other regenerative traits, such as dispersal and pollination modes,

play a secondary role in the regeneration of the populations.
Functional structure of communities: individual level approach
When we considered the variation of traits among individuals,
the functional diversity was lower in areas with low re frequencies (Fig. 4). In other words, co-occurring individuals were more
similar in MidFi and in LowFi plots than those in HiFi plots. This is
the opposite pattern to that in the species based analyses, suggesting that accounting for intraspecic variation in the trait diversity
is important for fully understanding how communities are assembled. This functional clustering of individuals may be attributed
to the great number of co-occurring individuals of the same population in plots with lower re frequencies. Fires commonly kills
seedlings and saplings, decreasing the clustered pattern of spatial
distribution of the population (Greig-Smith, 1991). Thus, the exclusion of re may decrease the trait variability at small spatial scales.
Also, at low re frequencies individuals may have sufcient time
to recover from the last re and start to spread vegetatively via
clonal reproduction (San Jose et al., 1991; Hoffman, 2002). In this
case, many of the individuals within the same species we sampled
at MidFi and LowFi plots were in fact the same individuals which
could additionally explain the relatively low functional clustering
observed.
However, when we considered the set of traits representing
different plant strategies, the results were similar to those of the
species level approach. Frequent res also decreased the range of
traits associated with resistance and physiological responses to
disturbance.
Despite having observed that in some cases both species and,
especially, individuals were randomly assembled with respect to
their traits, overall we found that coexisting species were more
similar in their traits than expected by chance. While the environmental conditions where functional clustering was observed
varied, these ndings provide evidence against neutral theories of
biodiversity (Hubbell, 2001). These theories predict that the species
traits are unimportant for coexistence patterns. On the other hand,
we also found some situations in which species and individuals
were a random sample from the species (individuals) pool. This was
not always congruent between species-based and individual-based
approaches and depended on the traits under analysis. Indeed, a
critical point in understanding the functional structure of communities is the choice of functional traits with which organisms are
distinguished. Usually the recommendation is to use all the traits
that are thought to be important in the context under investigation (Petchey and Gaston, 2006). Another possibility would be to
use single-trait approaches (e.g., Lavorel et al., 2007) or a set of traits
related to more specic plant ecological strategies, as we did here.
This seems to be more informative because it allows us to understand better different aspects of the relationship between species
traits and the environment.
Finally, we showed that deciding how to summarise trait information (i.e., using species trait averages or individual values), is also
important. There is an increasing recognition of the importance
of including intraspecic variability in functional traits (Ackerly
and Cornwell, 2007; de Bello et al., 2011). Indeed, in some cases,
we found different results in species-based and individual-based
analyses. Recently, Jung et al. (2010) found evidence for the importance of intraspecic trait variability in detecting habitat ltering
and niche differentiation processes in meadow plant communities.
Here, by using an individual-level approach to functional diversity, we corroborated the view that realised trait values rather
than species mean trait values have an important role in driving
community assembly (Jung et al., 2010, Messier et al., 2010). This

M.V. Cianciaruso et al. / Perspectives in Plant Ecology, Evolution and Systematics 14 (2012) 205216

happens because species that are unexpected to occur in a given

assemblage according to its mean trait values (species-level trait
values) may in fact occur due to some plastic response (intraspecic
trait variability). Thus, approaches that incorporate intraspecic
trait variation could improve our capacity to understand community assembly processes and the evolution of functional traits in
plant communities (Pachepsky et al., 2001, 2007; Cianciaruso et al.,
2009; Albert et al., 2010).
This nding raises an old question about the basic biological
unit we should use in community ecology: species or individuals? Some authors suggest that intraspecic functional variability
can be considered as negligible, and others that individuals are the
most signicant ecological unit (e.g., Albert et al., 2010; de Bello
et al., 2011). We agree with Albert et al. (2010) when they state that
this decision cannot be generic; it of course depends on the levels
of intraspecic variation present in the system under study. This
will depend on both the kind of traits and also the kind of organisms. For example, plants are much more plastic than animals (see
Schlichting, 1986). Additionally, spatial and temporal scales, as well
as the feasibility of acquiring functional attributes at the individual
level are likely to inuence whether individual level trait values
are collected. At local scales such rened information seems more
important to consider, whereas at larger scales differences among
species could be more important.
Implications for re management strategies and concluding
remarks
Conservation-oriented studies frequently focus on species
diversity, and less attention is given to species traits or phylogenetic diversity, and very rarely to both (but see Mayeld et al., 2006;
Carvalho et al., 2010b; Devictor et al., 2010). Habitat modications
and management practices that change functional and phylogenetic diversity are likely to have large impacts on community
processes and evolutionary history. Perhaps, the most important
consequence of changes in species traits and relatedness have to
do with the largely unknown feedbacks of the altered environment
to further changes in biodiversity and ecosystem processes. Therefore, neglecting community patterns and aspects such as functional
diversity and evolutionary history of species, may render serious
long-term conservation efforts impossible and futile (Ernst et al.,
2006).
Here, we showed that re frequency has important consequences for the phylogenetic and functional structure of the

213

savannas studied. Whereas we observed functional clustering

under HiFi, which is in accordance with the observed pattern for
other re-prone vegetation, this was accompanied by phylogenetic overdispersion. This is important, because it shows that the
relationship between ecological processes and the phylogenetic
structure of communities is not always straightforward, that is, we
cannot always use PD as a surrogate for FD, with the assumption
that close relatives are always similar in their ecological features
(see also Losos, 2008). Also, we showed that including intraspecic
variability into the community assembly framework is relevant
and should improve our understanding of how community are
assembled. Nevertheless, even if there is evidence that species
intraspecic variability can regulate important ecosystem processes and that realised trait values rather than mean species
trait values have an important role in driving community assembly, as we showed here, most published studies continue to ignore
this scale of information.
Acknowledgements
We are grateful to Fapesp, for nancial support; to Fapesp and
Capes, for the scholarships granted to the rst author; to CNPq, for
the scholarship granted to the second author; to Helena Franca, for
providing us the updated re history map of Emas National Park; to
the park staff, for logistical assistance; to C.S. Goncalves, D.M. Silva,
M.V. Forzani, N.A. Escobar, N.B. Rossati, P. Loiola, P. Zava, T. Vecchi,
and V. Dantas, for valuable help in the eld; and to Jasper Slingsby,
Sara Blanchard, and two anonymous referees, which comments and
suggestions improved the manuscript. O.L.P. was partly funded by
the Royal Society and the University of Zurich.
Appendix 1.
Species sampled in Emas National Park, central Brazil
(17 49 18 28 S, 52 39 53 10 W) with the respective number of
individuals (n) and average trait values for each re frequency.
H = height (m), BA = basal area (m2 ), Brk = bark thickness (mm),
SLA = specic leaf area (mm2 mg1 ), LSz = leaf size (mm2 ), Tgh = leaf
toughness (N), Woo = wood density (mg mm3 ), N = leaf nitrogen content (mg g1 ), P = leaf phosphorus content (mg g1 ), K = leaf
potassium content (mg g1 ), top = resprouting due to top-kill (frequency), pol = polination mode (be = bees, si = small insects, = ies,
mo = moths, hb = humming birds, ba = bats), dis = dispersal mode
(ane = anemochory, autochory = auto, zoo = zoochory).

BA

Anadenanthera falcata
Aspidosperma tomentosum
Connarus suberosus
Diospyros hispida
Eremanthus erythropappus
Erythroxylum campestre
Erythroxylum suberosum
Hymenaea stigonocarpa
Mimosa amnis-atri
Myrcia camapuensis
Piptocarpha rotundifolia
Pouteria ramiora
Pouteria torta
Psidium laruotteanum
Stryphnodendron adstringens
Acosmium dasycarpum
Albizia hassleri
Anadenanthera falcata
Connarus suberosus
Davilla elliptica
Dimorphandra mollis
Eremanthus erythropappus
Eriotheca gracilipes
Erythroxylum tortuosum
Erythroxylum suberosum
Eugenia aurata
Eugenia bimarginata
Eugenia punicifolia
Kielmeyera coriacea
Machaerium acutifolium
Mimosa amnis-atri
Ouratea acuminata
Palicourea rigida
Piptocarpha rotundifolia
Pouteria ramiora
Pouteria torta
Psidium laruotteanum
Sclerolobium aureum
Stryphnodendron adstringens
Styrax ferrugineus
Tabebuia ochracea
Acosmium dasycarpum
Anadenanthera falcata
Byrsonima coccolobifolia
Byrsonima verbascifolia
Caryocar brasiliense
Casearia sylvestris
Connarus suberosus
Davilla elliptica
Didymopanax malmei
Dimorphandra mollis
Diospyros hispida
Eremanthus erythropappus
Eriotheca pubescens
Erythroxylum suberosum
Eugenia aurata
Hymenaea stigonocarpa
Kielmeyera coriacea

1
1
5
3
7
1
1
2
5
1
8
23
19
9
4
5
1
7
9
4
4
8
1
3
1
2
1
1
7
1
3
1
2
1
24
5
4
2
8
3
3
5
2
3
1
1
2
8
2
1
3
4
5
1
3
1
1
6

3.66
0.83
1.11
1.41
1.17
1.06
0.81
1.76
1.13
1.00
1.50
1.91
1.13
1.18
1.72
0.84
1.79
2.93
1.33
1.38
2.10
1.19
3.49
0.89
1.35
0.78
1.56
2.22
2.30
5.30
1.03
1.47
1.38
1.61
1.75
1.62
0.94
1.92
2.26
3.97
0.73
0.79
3.45
2.73
3.21
2.90
1.64
1.68
0.81
3.93
1.80
1.53
1.11
1.05
1.44
1.50
1.30
1.27

0.0298
0.0016
0.0024
0.0056
0.0041
0.0016
0.0018
0.0034
0.0026
0.0029
0.0040
0.0042
0.0025
0.0027
0.0037
0.0033
0.0008
0.0107
0.0037
0.0026
0.0096
0.0034
0.0121
0.0016
0.0030
0.0017
0.0046
0.0046
0.0097
0.0147
0.0011
0.0050
0.0038
0.0035
0.0032
0.0034
0.0015
0.0026
0.0042
0.0252
0.0022
0.0012
0.0053
0.0046
0.0076
0.0067
0.0015
0.0039
0.0009
0.0023
0.0046
0.0024
0.0038
0.0008
0.0021
0.0035
0.0008
0.0037

Brk

SLA

13.83
11.17
8.60
4.52
13.55
15.06
10.21
0.85
2.17
5.04
6.00
12.59
7.43
11.09
6.47
7.80
6.47
18.84
13.76
6.28
10.98
12.46
18.09
7.67
12.14
10.43
11.63
18.20
20.66
20.76
3.64
9.79
9.47
8.06
11.45
9.73
9.77
5.71
6.41
12.66
6.71
9.64
10.11
9.65
14.66
11.93
7.13
14.84
9.10
3.01
10.70
8.00
10.79
6.34
12.53
13.05
5.57
9.91

6.85
7.23
6.13
4.62
9.21
7.34
9.02
7.16
6.26
3.58
6.47
6.88
7.02
7.07
7.65
6.66
13.33
7.73
5.53
6.89
8.69
9.96
6.09
6.20
6.89
8.86
8.49
10.74
5.81
7.22
10.03
6.78
5.20
5.34
7.00
5.86
4.90
7.87
6.90
5.15
5.65
6.86
6.41
8.90
7.61
9.08
8.37
5.81
9.41
3.81
8.20
4.77
10.74
4.94
9.13
14.90
9.50
6.03

LSz
17068.00
5681.00
9898.32
17168.67
2785.29
2334.00
1845.00
11595.50
8771.94
8127.00
11859.19
9457.27
8519.41
4244.43
39570.11
10590.13
11511.00
15367.55
17068.94
3557.36
32941.93
3698.04
20150.00
5180.00
3364.60
2064.00
4026.00
1160.00
8889.82
20048.00
5246.48
4368.00
16986.50
5697.00
9036.90
9442.04
2513.35
31410.30
48283.71
2411.23
22489.47
11455.16
16023.89
6945.88
10873.00
51227.00
1037.76
14423.57
3876.89
12199.00
35013.12
18015.82
2070.18
59792.00
1966.40
2699.00
43291.00
9929.71

Tgh
0.18
0.88
1.75
1.28
0.86
1.39
0.98
1.36
0.15
1.92
1.63
1.19
0.78
1.68
0.93
1.48
0.74
0.27
2.19
1.17
0.11
0.66
2.30
1.22
1.24
0.97
1.54
0.91
1.97
2.37
0.17
1.42
2.31
2.12
0.78
0.71
1.68
0.97
0.73
1.81
2.09
2.20
0.61
0.75
0.53
0.55
0.76
1.93
1.23
1.08
0.14
1.15
0.55
2.89
1.13
1.34
1.77
2.37

Woo
0.63
0.41
0.42
0.29
0.47
0.62
0.48
0.58
0.55
0.85
0.48
0.42
0.46
0.62
0.48
0.47
0.55
0.59
0.41
0.56
0.44
0.44
0.35
0.49
0.51
0.41
0.45
0.54
0.29
0.56
0.61
0.57
0.23
0.52
0.44
0.48
0.61
0.41
0.45
0.49
0.38
0.39
0.54
0.45
0.46
0.40
0.42
0.43
0.56
0.40
0.46
0.38
0.52
0.26
0.52
0.39
0.41
0.27

P
25.90
19.98
19.21
14.14
18.47
15.90
18.52
15.59
22.12
14.62
17.69
15.01
18.45
11.79
21.42
18.65
32.10
25.25
14.48
16.46
33.73
17.93
19.21
15.95
17.88
19.57
22.67
28.69
14.42
31.77
27.00
11.47
18.80
15.36
15.18
16.23
12.64
22.61
20.42
12.96
13.17
22.15
31.78
16.03
15.83
21.59
15.77
14.70
20.01
20.09
38.39
12.78
22.04
19.01
21.57
36.47
20.89
17.19

1.21
1.13
1.03
1.12
1.54
1.02
1.05
0.98
1.24
0.87
1.05
1.03
1.21
0.77
0.97
1.04
1.61
1.36
0.85
1.22
1.19
1.31
0.98
0.96
1.18
1.20
1.21
2.06
0.99
1.05
1.45
0.70
0.82
0.86
1.04
1.14
0.67
1.29
1.01
0.76
0.88
1.00
1.46
1.11
1.25
1.63
1.16
0.75
1.13
1.01
1.64
0.98
1.79
1.39
1.31
2.25
1.65
1.28

K
5.36
7.65
5.15
9.78
14.68
3.83
8.16
3.44
4.54
4.85
6.44
5.02
5.03
7.91
4.34
4.23
7.14
4.63
5.16
6.50
3.64
12.02
10.97
3.06
4.65
6.50
7.14
10.20
4.74
4.59
6.63
3.57
6.12
5.10
5.14
5.00
5.87
3.58
3.19
3.32
3.83
5.82
8.16
8.25
11.22
6.63
6.63
4.94
11.23
5.61
6.29
10.71
16.73
9.18
9.01
14.03
9.69
6.89

top
0.00
1.00
0.20
0.67
0.00
1.00
0.00
0.50
0.00
0.00
0.25
0.09
0.53
0.33
0.25
0.20
0.00
0.14
0.22
0.00
0.00
0.38
0.00
0.00
0.00
0.00
1.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.04
0.60
0.25
1.00
0.13
0.00
0.00
0.00
0.50
0.00
0.00
0.00
0.50
0.25
0.00
0.00
0.67
0.50
0.40
0.00
0.33
0.00
1.00
0.33

pol

dis

si
mo
si
mo
si
si
si
ba
si
be
si
mo
mo
be
si
be
si
si
si
be
si
si
be
si
si
be
be
be
be
be
si
be
hb
si
mo
mo
be
si
si
be
be
be
si
be
be
ba

si
be
si
si
mo
si
be
si
be
ba
be

aut
ane
zoo
zoo
ane
zoo
zoo
zoo
ane
zoo
ane
zoo
zoo
zoo
zoo
ane
ane
aut
zoo
zoo
aut
ane
ane
zoo
zoo
zoo
zoo
zoo
ane
ane
ane
zoo
zoo
ane
zoo
zoo
zoo
aut
zoo
zoo
ane
ane
aut
zoo
zoo
zoo
zoo
zoo
zoo
zoo
aut
zoo
ane
ane
zoo
zoo
zoo
ane

M.V. Cianciaruso et al. / Perspectives in Plant Ecology, Evolution and Systematics 14 (2012) 205216

Species

HiFi
HiFi
HiFi
HiFi
HiFi
HiFi
HiFi
HiFi
HiFi
HiFi
HiFi
HiFi
HiFi
HiFi
HiFi
MidFi
MidFi
MidFi
MidFi
MidFi
MidFi
MidFi
MidFi
MidFi
MidFi
MidFi
MidFi
MidFi
MidFi
MidFi
MidFi
MidFi
MidFi
MidFi
MidFi
MidFi
MidFi
MidFi
MidFi
MidFi
MidFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi

214

Fire treatment

aut
zoo
ane
zoo
zoo
zoo
zoo
zoo
zoo
ane
ane
zoo
zoo
zoo
ane
aut
zoo
zoo
zoo
ane
ane
ba
be
si
be
be
be
be
be
be
si
si
mo
mo
be
be
mo
si
si
be
be
be

pol
top

1.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.14
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
11.48
3.06
7.27
2.68
8.16
4.08
2.55
4.32
9.34
6.89
14.54
8.58
6.60
7.93
9.95
3.06
2.81
4.34
3.06
4.98
7.10

K
P
Tgh

0.55
0.62
0.55
0.49
0.56
0.74
0.26
0.46
0.49
0.53
0.45
0.43
0.41
0.52
0.40
0.52
0.22
0.55
0.47
0.33
0.40
0.39
0.66
0.15
1.01
0.91
2.45
0.95
1.50
3.66
1.95
0.38
0.72
0.75
1.57
0.68
2.64
1.32
1.01
1.63
2.86
2.11
2718.00
5131.00
6173.45
661.94
4087.00
4691.00
936.00
3151.05
5070.79
9988.59
4079.00
10744.98
9464.43
3284.09
2888.00
9468.00
1997.00
46881.69
1683.00
46324.29
19954.49

LSz
SLA

13.23
5.09
13.00
7.68
9.14
3.87
6.38
6.40
5.02
5.22
15.75
8.37
7.10
4.94
12.13
4.69
6.20
7.03
5.30
3.10
5.82
6.37
5.42
3.22
12.03
6.61
7.91
15.23
11.87
12.12
11.02
11.21
11.31
6.89
10.58
9.19
3.96
9.52
9.40
11.57
13.39
12.09

Brk
BA

0.0064
0.0010
0.0030
0.0018
0.0035
0.0008
0.0046
0.0033
0.0069
0.0032
0.0115
0.0042
0.0043
0.0023
0.0076
0.0008
0.0008
0.0047
0.0176
0.0021
0.0028
2.20
1.85
0.95
1.02
1.13
0.71
1.54
1.53
2.62
1.87
5.37
2.24
1.39
1.08
5.15
1.93
0.75
2.43
4.20
1.50
1.30

Lafoensia pacari
Miconia albicans
Mimosa amnis-atri
Myrcia bella
Myrcia crassifolia
Myrcia guianensis
Myrcia lasiantha
Ouratea acuminata
Ouratea spectabilis
Piptocarpha rotundifolia
Plenckia populnea
Pouteria ramiora
Pouteria torta
Psidium laruotteanum
Qualea parviora
Roupala montana
Rourea induta
Stryphnodendron adstringens
Styrax ferrugineus
Tabebuia aurea
Tabebuia ochracea
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi
LowFi

n
Species
Fire treatment

1
1
4
2
1
1
1
23
5
2
1
22
8
10
1
1
1
3
1
2
6

Woo

26.21
11.73
27.99
13.10
17.85
8.22
11.97
13.79
13.11
13.66
33.31
18.55
17.89
15.10
17.98
8.57
13.80
23.64
12.67
14.13
23.55

1.79
0.49
1.41
0.74
1.28
0.49
0.77
0.76
1.02
1.02
2.17
1.33
1.12
1.06
1.57
0.51
0.34
1.04
0.90
1.17
1.42

dis

M.V. Cianciaruso et al. / Perspectives in Plant Ecology, Evolution and Systematics 14 (2012) 205216

215

References
Ackerly, D.D., Cornwell, W.K., 2007. A trait-based approach to community assembly: partitioning of species trait values into within- and among-community
components. Ecol. Lett. 10, 135145.
Albert, C.H., Thuiller, W., Yoccoz, N.G., Soudant, A., Boucher, F., Saccone, P., Lavorel,
S., 2010. Intraspecic functional variability: extent, structure and sources of
variation. J. Ecol. 98, 604613.
Angiosperm Phylogeny Group, 2009. An update of the Angiosperm Phylogeny Group
classication for the orders and families of owering plants: APG III. Bot. J. Linn.
Soc. 161, 105121.
Blomberg, S.P., Garland Jr., T., Ives, A.R., 2003. Testing for phylogenetic signal in
comparative data: behavioral traits are more labile. Evolution 57, 717745.
Bond, W.J., Keeley, J.E., 2005. Fire as a global herbivore: the ecology and evolution
of ammable ecosystems. Trends Ecol. Evol. 20, 387394.
Cadotte, M.W., Cardinale, B.J., Oakley, T.H., 2008. Evolutionary history and the effect
of biodiversity on plant productivity. PNAS 105, 1701217017.
Cadotte, M.W., Cavender-Bares, J., Tilman, D., Oakley, T.H., 2009. Using phylogenetic, functional and trait diversity to understand patterns of plant community
productivity. PLoS ONE 4, e5695.
Carvalho, G.H., Cianciaruso, M.V., Batalha, M.A., 2010a. Plantminer: a web tool for
checking and gathering plant species taxonomic information. Environ. Model.
Soft. 25, 815816, http://www.plantminer.com.
Carvalho, R.A., Cianciaruso, M.V., Trindade-Filho, J., Sagnori, M.D., Loyola, R.D.,
2010b. Drafting a blueprint for functional and phylogenetic diversity conservation in the Brazilian Cerrado. Natureza Conservacao 8, 171176.
Cavender-Bares, J., Keen, A., Miles, B., 2006. Phylogenetic structure of Floridian plant
communities depends on taxonomic and spatial scale. Ecology 87, S109S122.
Cianciaruso, M.V., Batalha, M.A., Gaston, K.J., Petchey, O.L., 2009. Including intraspecic variability in functional diversity. Ecology 90, 8189.
Cianciaruso, M.V., Silva, I.A., Batalha, M.A., 2010. Aboveground biomass of functional
groups in the ground layer of savannas under different re frequencies. Aust. J.
Bot. 58, 169174.
Cornelissen, J.H.C., Lavorel, S., Garnier, E., Daz, S., Buchmann, N., Gurvich, D.E., Reich,
P.B., Steege, H.ter., Morgan, H.D., van der Heijden, M.G.A., Pausas, J.G., Poorter,
H., 2003. A handbook of protocols for standardised and easy measurement of
plant functional traits worldwide. Aust. J. Bot. 51, 335380.
Costa, I.R., 2009. Estudos evolutivos em Myrtaceae: aspectos citotaxonmicos e logenticos em Myrteae, enfatizando Psidium e gneros relacionados. PhD Thesis.
Universidade de Campinas, Campinas.
Coutinho, L.M., 1990. Fire in the ecology of the Brazilian cerrado. In: Goldammer,
J.G. (Ed.), Fire in the Tropical Biota. Springer, pp. 81103.
DeBano, L.F., Neary, D.G., Ffolliott, P.F., 1998. Fires Effects on Ecosystems. Jon Wiley
& Sons, New York.
de Bello, F., Lavorel, S., Albert, C.H., Thuiller, W., Grigulis, K., Dolezal, J., Janecek,
Leps, J., 2011. Quantifying the relevance of intraspecic trait variability for
S.,
functional diversity. Meth. Ecol. Evol. 2, 163174.
Devictor, V., Mouillot, D., Meynard, C., Jiguet, F., Thuiller, W., Mouquet, N., 2010.
Spatial mismatch and congruence between taxonomic, phylogenetic and functional diversity: the need for integrative conservation strategies in a changing
world. Ecol. Lett. 13, 10301040.
Daz, S., Cabido, M., 2001. Vve la diffrence: plant functional diversity matters to
ecosystem process. Trends Ecol. Evol. 16, 646655.
Diniz Filho, J.A.F., Cianciaruso, M.V., Rangel, T.F., Bini, L.M., 2011. Eigenvector estimation of phylogenetic and functional diversity. Funct. Ecol. 25, 735744.
Ernst, R., Linsenmair, K.E., Rdel, M.O., 2006. Diversity erosion beyond the species
level: dramatic loss of functional diversity after selective logging in two tropical
amphibian communities. Biol. Cons. 133, 143155.
Faith, D.P., 1992. Conservation evaluation and phylogenetic diversity. Biol. Cons. 61,
110.
Fajardo, A., Piper, F.I., 2011. Intraspecic trait variation and covariation in a
widespread tree species (Nothofagus pumilio) in southern Chile. New Phytol.
189, 259271.
Franca, H., Ramos-Neto, M.B., Setzer, A., 2007. O fogo no Parque Nacional das Emas.
Instituto do Meio Ambiente e dos Recursos Naturais Renovveis.
Fukami, T., Bezemer, T.M., Mortimer, S.R., van der Putten, W.H., 2005. Species divergence and trait convergence in experimental plant community. Ecol. Lett. 8,
12831290.
Garland Jr., T., Harvey, P.H., Ives, A.R., 1992. Procedures for the analysis of comparative data using phylogenetically independent contrasts. Syst. Biol. 41, 1832.
Gmez, J.P., Bravo, G.A., Brumeld, R.T., Tello, J.G., Cadena, C.D., 2010. A phylogenetic approach to disentangling the role of competition and hbitat ltering in
community assembly of Neotropical forest birds. J. Anim. Ecol. 79, 11811192.
Gotelli, N.J., Entsminger, G.L., 2001. Swap and ll algorithms in null model analysis:
rethinking the Knights Tour. Oecologia 129, 281291.
Gottsberger, G., Silberbauer-Gottsberger, I., 2006. Life in the Cerrado: A South American Tropical Seasonal Vegetation. Vol. 2. Pollination and Seed Dispersion, Reta.
Greig-Smith, P., 1991. Pattern in a derived savanna in Nigeria. J. Trop. Ecol. 7, 91502.
Harley, C.D.G., 2003. Abiotic stress and herbivory interact to set range limits across
a two-dimensional stress gradient. Ecology 84, 14771488.
Hoffmann, W., 1998. Post-burn reproduction of woody plants in a neotropical
savanna: the relative importance of sexual and vegetative reproduction. J. App.
Ecol. 35, 422433.

216

M.V. Cianciaruso et al. / Perspectives in Plant Ecology, Evolution and Systematics 14 (2012) 205216

Hoffman, W., 2002. Post-burn reproduction of woody plants in a neotropical

savanna: the relative importance of sexual and vegetative reproduction. J. App.
Ecol. 35, 422433.
Hooper, D.H., Naeem, S., Inchausti, P., 2002. Species diversity, functional diversity
and ecosystem functioning. In: Loreau, M. (Ed.), Biodiversity and Ecosystem
Functioning: Syntheses and Perspectives. Oxford University Press, Oxford, UK,
pp. 195208.
Hubbell, S., 2001. The Unied Neutral Theory of Biodiversity and Biogeography.
Princeton University.
Jung, V., Violle, C., Mondy, C., Hoffmann, L., Muller, S., 2010. Intraspecic variability
and trait-based community assembly. J. Ecol. 98, 11341140.
Kembel, S.W., 2009. Disentangling niche and neutral inuences on community
assembly: assessing the performance of community phylogenetic structure
tests. Ecol. Lett. 12, 949960.
Kembel, S.W., Cowan, P.D., Helmus, M.R., Cornwell, W.K., Morlon, H., Ackerly, D.D.,
Blomberg, S.P., Webb, C.O., 2010. Picante: R tools for integrating phylogenies
and ecology. Bioinformatics 26, 14631464.
Kppen, W., 1931. Grundriss der Klimakunde, Berlin.
Kraft, N.J.B., Cornwell, W.K., Webb, C.O., Ackerly, D.D., 2007. Trait evolution, community assembly, and the phylogenetic structure of ecological communities. Am.
Nat. 170, 271283.
Lavorel, S., Daz, S., Cornelissen, J.H.C., Garnier, E., Harrison, S.P., McIntyre, S., Pausas,
J.G., Prez-Harguindeguy, N., Roumet, C., Urcelay, C., 2007. Plant functional
types: are we getting any closer to the holy grail? In: Canadell, J.G., Pitelka, L.F.,
Pataki, D. (Eds.), Terrestrial Ecosystems in a Changing World. The IGBP Series.
Springer-Verlag, Berlin/Heidelberg.
Leibold, M.A., Holyoak, M., Mouquet, N., Amarasekare, P., Chase, J.M., Hoopes, M.F.,
Holt, R.D., Shurin, J.B., Law, R., Tilman, D., Loreau, M., Gonzalez, A., 2004. The
metacommunity concept: a framework for multi-scale community ecology.
Ecol. Lett. 7, 601613.
Losos, J.B., 2008. Phylogenetic niche conservatism, phylogenetic signal and the relationship between phylogenetic relatedness and ecological similarity among
species. Ecol. Lett. 11, 9951007.
Madritch, M.D., Hunter, M.D., 2002. Phenotypic diversity inuences ecosystem functioning in an oak sandhills community. Ecology 83, 20842090.
Madritch, M.D., Hunter, M.D., 2003. Intraspecic litter diversity and nitrogen deposition affect nutrient dynamics and soil respiration. Oecologia 136, 124128.
Mayeld, M.M., Ackerly, D., Daily, G.C., 2006. The diversity and conservation of
plant reproductive and dispersal functional traits in human-dominated tropical
landscapes. J. Ecol. 94, 522536.
Mayeld, M.M., Levine, J.M., 2010. Opposing effects of competitive exclusion on the
phylogenetic structure of communities. Ecol. Lett. 13, 10851093.
Messier, J., McGill, B.J., Lechowicz, M.J., 2010. How do traits vary across ecological
scales? A case for trait-based ecology. Ecol. Lett. 13, 838848.
Meynard, C., Devictor, V., Mouillot, D., Thuiller, W., Jiguet, F., Mouquet, N. Beyond
taxonomic diversity patterns: how do alpha, beta and gamma components of
bird functional and phylogenetic diversity respond to environmental gradients
across France? Global Ecol. Biog. in press.
Miner, B.G., Sultan, S.E., Morgan, S.G., Padilla, D.K., Relyea, R.A., 2005. Ecological
consequences of phenotypic plasticity. Trends Ecol. Evol. 20, 685692.
Miranda, H.S., Bustamante, M.C., Miranda, A.C., 2002. The re factor. In: Oliveira,
P.S., Marquis, J.R. (Eds.), The Cerrados of Brazil: Ecology and Natural History of
a Neotropical Savanna. Columbia University, New York, pp. 5168.
Mouchet, M., Villger, S., Mason, M., Mouillot, D., 2010. Functional diversity measures: an overview of their redundancy and their ability to discriminate
community assembly rules. Funct. Ecol. 24, 867876.
Overbeck, G.E., Pfadenhauer, J., 2007. Adaptive strategies in burned subtropical
grassland in Southern Brazil. Flora 202, 2749.
Pachepsky, E., Bown, J.L., Eberst, A., Bausenwein, U., Millard, P., Squire, G.R., Squire,
G., Crawford, J.W., 2007. Consequences of intraspecic variation for the structure
and function of ecological communities. Part 2: Linking diversity and function.
Ecol. Modell. 207, 277285.
Pausas, J.G., Lavorel, S., 2003. A hierarchical deductive approach for functional types
in disturbed ecosystems. J. Veg. Sci. 14, 409416.
Pausas, J.G., Paula, S., 2005. Plant Functional Traits Database for Euro-Mediterranean
Ecosystems. Eurelab Deliverable D-04-06, http://eurelab.org.
Pausas, J.G., Verd, M., 2008. Fire reduces morphospace occupation in plant communities. Ecology 89, 21812186.
Pausas, J.G., Verd, M., 2010. The jungle of methods for evaluating phenotypic and
phylogenetic structure of communities. BioScience 60, 614625.
Pavoine, S., Vallet, J., Dufour, A.B., Gachet, S., Daniel, H., 2009. On the challenge of
treating various types of variables: application for improving the measurement
of diversity. Oikos 118, 391402.

Pavoine, S., Vela, E., Gachet, S., De Belair, G., Bonsall, M.B., 2010. Linking patterns
in phylogeny, traits, abiotic variables and space: a novel approach to linking
environmental ltering and plant community assembly. J. Ecol. 99, 165175.
Pavoine, S., Bonsall, M., 2011. Measuring biodiversity to explain community assembly: a unied approach. Biol. Rev., doi:10.1111/j.1469-185X.2010.00171.x.
Petchey, O.L., Gaston, K.J., 2002. Functional diversity (FD), species richness, and
community composition. Ecol. Lett. 5, 402411.
Petchey, O.L., Gaston, K.J., 2006. Functional diversity: back to basics and looking
forward. Ecol. Lett. 9, 741758.
Petchey, O.L., OGorman, E., Flynn, D.F.B., 2009. A functional guide to functional
diversity measures. In: Naeem, S., Bunker, D.E., Hector, A., Loreau, M., Perrings,
C. (Eds.), Biodiversity and Human Impacts. Oxford University Press, Oxford.
R Development Core Team, 2010. R: A Language and Environment for Statistical
Computing. R Foundation for Statistical Computing (http://www.r-project.org).
Ramos-Neto, M.B., Pivello, V.R., 2000. Lightning res in a Brazilian savanna National
Park: rethinking management strategies. Environ. Manag. 26, 675684.
Sa, K., Cianciaruso, M.V., Loyola, R.D., Brito, D., Armour-Marshall, K., Diniz-Filho,
J.A.F., 2011. Understanding global patterns of mammalian functional and phylogenetic diversity. Philos. Trans. R. Soc. Lond. B 366, 25362544.
San Jose, J.J., Farinas, M.R., Rosales, J., 1991. Spatial patterns of trees and structuring
factors in a Trachypogon savanna of the Orinoco llanos. Biotropica 23, 114123.
Schlichting, C.D., 1986. The evolution of phenotypic plasticity in plants. Ann. Rev.
Ecol. Syst. 17, 667693.
Schmitz, O.J., Adler, F.R., Agrawal, A.A., 2003. Linking individual-scale trait plasticity
to community dynamics. Ecology 84, 10811082.
Silva, I.A., Batalha, M.A., 2009a. Phylogenetic overdispersion of plant species in
southern Brazilian savannas. Braz. J. Biol. 69, 843849.
Silva, I.A., Batalha, M.A., 2009b. Co-occurrence of tree species at ne spatial scale in
a woodland cerrado, southeastern Brazil. Plant Ecol. 200, 277286.
Silva, I.A., Batalha, M.A., 2010. Phylogenetic structure of Brazilian savannas under
different re regimes. J. Veg. Sci. 21, 10031013.
Simon, M.F., Grether, R., Queiroz, L.P., Skema, C., Pennington, R.T., Hughes, C.E., 2009.
Recent assembly of the Cerrado, a neotropical plant diversity hotspot, by in situ
evolution of adaptations to re. PNAS 106, 2035920364.
Stevens, P.F., 2001. Angiosperm Phylogeny Website. Version 9, June 2008,
http://www.mobot.org/MOBOT/research/APweb/.
Swenson, N.G., Enquist, B.J., Pither, J., Thompson, J., Zimmerman, J.K., 2006. The problem and promise of scale dependency in community phylogenetics. Ecology 87,
24182424.
Thuiller, W., Gallien, L., Boulangeat, I., de Bello, F., Munkemuller, T., Roquet, C.,
Lavergne, S., 2010. Resolving Darwins naturalization conundrum: a quest for
evidence. Div. Distr. 16, 461475.
Verd, M., Pausas, J.G., 2007. Fire drives phylogenetic clustering in Mediterranean
Basin woody plant communities. J. Ecol. 95, 13161323.
Verd, M., Rey, P.J., Alcntara, J.M., Siles, G., Valiente-Banuet, A., 2009. Phylogenetic
signatures of facilitation and competition in successional communities. J. Ecol.
97, 11711180.
Vamosi, S.M., Heard, S.B., Vamosi, J.C., Webb, C.O., 2009. Emerging patterns in
the comparative analysis of phylogenetic community structure. Mol. Ecol. 18,
572592.
Violle, C., Navas, M.L., Vile, D., Kazakou, E., Fortunel, C., Hummel, I., Garnier, E., 2007.
Let the concept of trait be functional! Oikos 116, 882892.
Webb, C.O., 2000. Exploring the phylogenetic structure of ecological communities:
an example for rain forest trees. Am. Nat. 156, 145155.
Webb, C.O., Ackerly, D.D., McPeek, M.A., Donoghue, M.J., 2002. Phylogenies and
community ecology. Ann. Rev. Ecol. Syst. 33, 475505.
Webb, C.O., Donoghue, M.J., 2005. Phylomatic: tree assembly for applied phylogenetics. Mol. Ecol. Not. 5, 181183.
Webb, C.O., Ackerly, D.D., Kembel, S.W., 2008. Phylocom: software for the analysis
of phylogenetic community structure and trait evolution. Bioinformatics 24,
20982100.
Weiher, E., Clarke, P.G.D., Keddy, P.A., 1998. Community assembly rules, morphological dispersion, and the coexistence of plant species. Oikos 81, 309322.
Weiher, E., Keddy, P.A., 1999. Ecological Assembly Rules Perspectives, advances,
retreats. Cambridge University Press.
Wikstrm, N., Savolainen, V., Chase, M.W., 2001. Evolution of the angiosperms:
calibrating the family tree. Proc. R. Soc. Lond. Biol. Sci. 268, 22112220.
Wurdack, K.J., Davis, C.C., 2009. Malpighiales phylogenetics: gaining ground on one
of the most recalcitrant clades in the angiosperm tree of life. Am. J. Bot. 96,
15511570.