bs_bs_banner
Botanical Journal of the Linnean Society, 2014, 176, 332–348. With 3 figures
Genetic diversity and ecological niche modelling of the
restricted Recordia reitzii (Verbenaceae) from southern
Brazilian Atlantic forest
VERÔNICA A. THODE1,2, GUSTAVO A. SILVA-ARIAS1, CAROLINE TURCHETTO1,
ANA LÚCIA A. SEGATTO1, GERALDO MÄDER1, SANDRO L. BONATTO3 and
LORETA B. DE FREITAS1,2*
1
Laboratory of Molecular Evolution, Department of Genetics, Universidade Federal do Rio Grande do
Sul, PoBox 15053, Porto Alegre, Brazil
2
Programa de Pós-Graduação em Botânica, Universidade Federal do Rio Grande do Sul, Bento
Gonçalves 9500, Porto Alegre, Brazil
3
Genomic and Molecular Biology Laboratory, Pontifícia Universidade Católica do Rio Grande do
Sul, Ipiranga 6681, 90610-001 Porto Alegre, RS, Brazil
Received 25 December 2013; revised 13 June 2014; accepted for publication 27 July 2014
Genetic diversity analyses, coupled with ecological niche modelling (ENM) of species with a restricted distribution,
may provide valuable information for understanding diversification patterns in endangered areas. We analyzed the
genetic diversity of Recordia reitzii, a tree restricted to the threatened and highly fragmented Brazilian Atlantic
forest, using three intergenic cpDNA spacers and ten microsatellite (SSR) loci. To assess the historical processes
that may have influenced the distribution of extant R. reitzii populations, the current potential distributions of
R. reitzii and Recordia boliviana, a closely related species, were modelled and projected onto the Last Glacial
Maximum (LGM) and Last Interglacial (LIG) periods. Niche divergence was quantified between these two. The
cpDNA and SSR data showed a north–south pattern of the diversity distribution and structured populations,
suggesting that gene flow is probably limited. According to our data, R. reitzii exhibits low genetic diversity, which
may be a result of a founder or distribution-reduction effect, narrow distribution or small population size. The
ecological niche models showed a wider palaeodistribution during the LIG and a retraction during the LGM for
both species. Tests of niche divergence and conservatism indicated that bioclimatic factors might have influenced
the diversification of these Recordia species. © 2014 The Linnean Society of London, Botanical Journal of the
Linnean Society, 2014, 176, 332–348.
ADDITIONAL KEYWORDS: cpDNA – microsatellite – niche conservatism – niche divergence.
INTRODUCTION
The Atlantic forest is widely known for the large
number of organisms and high endemism it harbours,
as well as for documented diversity losses (Mori, Boom
& Prance, 1981; Terborgh, 1992; Morellato & Haddad,
2000; Myers et al., 2000; Perret, Chautems &
Spichiger, 2006; Ribeiro et al., 2009; Stehmann et al.,
2009). This forest extends from the north to the south
of Brazil (3°S to 30°S) along its east coast, from sea
level up to an elevation of 2700 m (Metzger et al., 2009;
*Corresponding author. E-mail: [email protected]
332
Stehmann et al., 2009), and is characterized by strong
seasonality and high precipitation as a result of the
eastward winds from the tropical Atlantic (Fundação
Instituto Brasileiro de Geografia e Estatística, 1993).
The Atlantic forest currently covers only 7–11% of its
original area and is restricted to isolated fragments,
which leads to a reduction of biodiversity, especially
among species endemic to this kind of habitat (Ranta
et al., 1998; Ribeiro et al., 2009; Vieira et al., 2009).
Despite an alarming loss of diversity, many Atlantic
forest species are still unknown (Lewinsohn & Prado,
2005; Sobral & Stehmann, 2009) and additional
studies are necessary to investigate the evolutionary
© 2014 The Linnean Society of London, Botanical Journal of the Linnean Society, 2014, 176, 332–348
GENETIC DIVERSITY OF RECORDIA REITZII
processes that may have influenced the assemblages of
this ecoregion (Perret et al., 2006). Studying the distribution pattern of species that occur in the Atlantic
forest is of great importance to comprehend the factors
driving their diversity and the high number of endemic
species, in addition to understanding the consequences
of past, present and future environmental changes in
this region, particularly for its extant biodiversity
(Behling & Lichte, 1997; Behling & Negrelle, 2001;
Behling et al., 2002; Carnaval & Moritz, 2008).
Population genetic analyses have been conducted on
plants of the Atlantic forest and on the influence of
climatic changes and glacial/interglacial cycles on the
distribution, genetic diversity and population structure of its species (Reis et al., 2000; Conte et al., 2003;
Ribeiro et al., 2005; Alcantara, Semir & Solferini,
2006; Palma-Silva et al., 2009; Ramos, Lemos-Filho &
Lovato, 2009; Lorenz-Lemke et al., 2010; Pinheiro
et al., 2011; Collevatti et al., 2012a; Turchetto-Zolet
et al., 2013). However, little is known about the genetic
diversity of species with a restricted distribution and
the factors that might have influenced their present
isolation. Fine-scale studies are important to evaluate
larger events and to elucidate the diversification processes occurring in a biome as a whole (Silva et al.,
2012). Genetic diversity studies using plastid DNA
(cpDNA) (maternally inherited in most Angiosperm
species) and nuclear (biparentally inherited) markers
may provide a deeper understanding of ancient and
recent events that have influenced the current distribution of genetic variability (Ennos, 1994; Collevatti,
Grattapaglia & Hay, 2003; Palma-Silva et al., 2009).
Ecological niche modelling (ENM) has been used to
understand the relationships between the temporal
dynamics of species distributions and analyses of
genetic differentiation and species divergence
(Knowles, Carstens & Keat, 2007), as well as to study
niche evolution (Pearman et al., 2008). Studying the
differences and similarities of the niches of closely
related species may provide insight into the role of
their geographical distribution and ecological features
on diversification (Wiens & Graham, 2005; Warren,
Glor & Turelli, 2008; McCormack, Zellmer &
Knowles, 2010; Wellenreuther, Larson & Svensson,
2012; Wooten & Gibbs, 2012; Hung, Drovetski & Zink,
2013). Such analyses may shed light on important
evolutionary issues, such as speciation (Hua & Wiens,
2013), and serve as powerful tools for the conservation of this ecoregion and the protection of its species
(Bernardo-Silva et al., 2012). Niche conservatism is
manifested by the tendency of related species to
occupy environmental niches that are similar and the
inability of these organisms to survive under different
conditions (low suitability) (Harvey & Pagel, 1991;
Wiens & Graham, 2005). On the other hand, niche
divergence refers to evolution under new conditions
333
and to the expansion of niche breadth, increasing the
distribution of a clade (Wiens & Donoghue, 2004).
Tools based on geographical information system (GIS)
can be used to quantify niche conservatism and divergence between closely related taxa (Wiens &
Donoghue, 2004; Warren et al., 2008; McCormack
et al., 2010). Studying biotic interchanges can help in
understanding the historical assembly of biomes and
select biological corridors for future conservation
efforts (Antonelli & Sanmartín, 2011).
Recordia reitzii (Moldenke) Thode & O’Leary (Verbenaceae) is a tree restricted to the southern limit of
the Brazilian Atlantic rainforest (Salimena et al.,
2009; Thode et al., 2013a) (Fig. S1). Its geographical
distribution comprises a small area along the coast in
the Atlantic Ombrophilous Dense Forest, with an
extension of only ∼ 200 km, in the Brazilian states of
Rio Grande do Sul and Santa Catarina. It is found at
elevations of 10–550 m, not reaching the higher areas
of the southern Brazilian plateau (Serra Geral), which
can exhibit elevations up to ∼ 1000 m (Troncoso, 1974;
Reitz, Klein & Reis, 1978; Reitz, Klein & Reis, 1983;
Sobral et al., 2006; Behling & Pillar, 2007; Thode et al.,
2013a) (Fig. 1). Based on genetic and morphological
data, R. reitzii was inferred to be sister to Recordia
boliviana Moldenke (Thode et al., 2013a), a tree with a
narrow distribution in the semi-deciduous dry forests
of Bolivia. This species is found in the low mountains
west of Santa Cruz de la Sierra, between Samaipata
and Cochabamba (M. Nee, pers. comm.), occurring
from elevations of 50–1850 m. These two species share
several morphological traits and are the only members
of a genus with a large distributional gap (Thode et al.,
2013a). This disjunction has been previously described
as a contribution of the cool-seasonal Andean Piedemont forests to the cool-moist endemic portion of the
Tropical Atlantic forests, which nowadays are probably
disconnected by late expansions of aridity of the Chaco
Domain (Oliveira-Filho et al., 2013).
Using three plastid regions and ten microsatellite
(SSR) loci, we aimed to investigate the genetic diversity and structure of R. reitzii populations. In addition, niche modelling of current and past periods and
tests of niche divergence and conservatism for
R. reitzii and R. boliviana were performed to investigate the effects of past climate changes on the processes that may have played a role in the present
distribution of R. reitzii.
MATERIAL AND METHODS
POPULATION SAMPLING
A total of 135 individuals of R. reitzii, from 12 collection sites (hereafter referred to as populations), were
genotyped for 10 SSR loci, and three plastid regions
© 2014 The Linnean Society of London, Botanical Journal of the Linnean Society, 2014, 176, 332–348
334
V. A. THODE ET AL.
Figure 1. Map of Brazil indicating the area studied (A). Parts of Rio Grande do Sul (RS) and Santa Catarina (SC) States
showing the geographical distribution of cpDNA haplotypes (B) and a median-joining network (C) for the Recordia reitzii
populations sampled (H1–H5) and for the individuals of Recordia boliviana sampled (H6; Hrb). The size of the circles is
proportional to the sample size, and the colours represent the haplotypes. Darker areas in B represent higher altitudes.
H5 and H6 (Hrb) differ by eight mutations.
were sequenced in a representative subset of 81 individuals. The sampling sites were located throughout
the species distribution in southeastern Brazil (Fig. 1,
Table 1). With the exception of three populations
(OSO, PGR and ORL; Table 1), it was not possible to
collect a larger number of individuals owing to the
rarity of this taxon and anthropic habitat disturbances. Moreover, no individuals were found further
north than ORL (28°22′ S, 49°14′ W) or further south
than OSO (29°52′ S, 50°17′ W). Three individuals of
R. boliviana were sampled (Table S1) for the three
plastid regions and were used as the outgroup.
DNA
AMPLIFICATION, PLASTID
DNA
SEQUENCING
AND MICROSATELLITE GENOTYPING
Genomic DNA was extracted from field-collected
silica-dried leaves using the cetyltrimethylammonium
bromide (CTAB) protocol, according to Roy et al.
(1992). Three cpDNA intergenic spacers, trnH–psbA
(Sang, Crawford & Stuessy, 1997), trnS–trnG
(Hamilton, 1999) and rpl32–trnL (Shaw et al., 2007),
were amplified via the polymerase chain reaction
(PCR) in a Veriti thermocycler (Applied Biosystems,
Foster City, CA, USA). PCR was performed in a 25-μL
reaction containing 5 ng/μL of template genomic
DNA, 200 μM each deoxyribonucleotide triphosphate
(dNTP) (Invitrogen, Carlsbad, CA, USA), 0.2 μM each
primer, 2.0 mM MgCl2, 0.5 U of Taq Platinum DNA
polymerase (Invitrogen) and 1 × Taq Platinum reac-
tion buffer (Invitrogen), with an initial denaturation
step at 94 °C for 3 min, followed by 30 cycles of 94 °C
for 1 min, 57 °C for 1 min (50° for the rpl32-trnL) and
72 °C for 1 min, and a final extension cycle at 72 °C
for 8 min. The PCR products were purified according
to Dunn & Blattner (1987) and sequenced in a
MegaBACE™ 1000 automated sequencer (GE Healthcare Biosciences, Pittsburgh, PA, USA) using the
DYEnamicET Terminator Sequencing Premix Kit (GE
Healthcare). The sequences thus obtained were
assembled using Chromas 2.4 (Technelysium, Helensvale, Australia) and manually aligned and edited in
Mega5 (Tamura et al., 2011). Inversions and
insertion/deletion events (indels), consisting of more
than one base pair (bp), were coded as a single evolutionary event. The three cpDNA regions were concatenated for all analyses. Sequences generated in
this study were deposited in GenBank and accession
numbers are provided in Table S2.
PCR amplifications of the 10 SSR loci were performed using previously described protocols (Thode
et al., 2013b). The length of the amplified fragment of
DNA was measured on a MegaBACE™ 1000 automated sequencer using an ET-ROX 550 size ladder
(GE Healthcare). The data were analyzed using
Genetic Profiler 2.0 (GE Healthcare).
DATA ANALYSES: cpDNA
SEQUENCES
To characterize the genetic variability of the sampled
individuals, haplotype (h) and nucleotide (π)
© 2014 The Linnean Society of London, Botanical Journal of the Linnean Society, 2014, 176, 332–348
RS,
RS,
RS,
RS,
SC,
SC,
SC,
SC,
SC,
SC,
SC,
SC,
Osório
Três Forquilhas
Três Cachoeiras
Morro Azul
Praia Grande
Timbé do Sul
Siderópolis
Santo Antônio
Santo Antônio
Treviso
Lauro Muller
Orleans
S
S
S
S
S
S
S
S
S
S
S
S
29°
29°
29°
29°
29°
28°
28°
28°
28°
28°
28°
28°
52′
32′
27′
24′
10′
49′
43′
35′
34′
29′
23′
22′
54.8″,
55.8″,
38.6″,
18.5″,
58.4″,
12.3″,
10.1″,
17.1″,
03.1″,
34.6″,
16.5″,
16.7″,
W
W
W
W
W
W
W
W
W
W
W
W
50°
50°
49°
49°
49°
49°
49°
49°
49°
49°
49°
49°
17′
04′
56′
55′
59′
52′
22′
31′
32′
31′
30′
14′
15.0″
43.3″
10.0″
51.1″
43.0″
12.3″
12.6″
24.5″
36.4″
15.2″
13.6″
34.3″
ICN 166749
ICN 166751
ICN 166754
ICN 166750
ICN 166755
ICN 166757
ICN 166765
ICN 166759
ICN 166758
ICN 166766
ICN 175535
ICN 166763
Average
4.6
2.9
2.5
2.2
4.1
1.9
2.5
2.4
2.4
2.7
2.2
3.3
2.8
6
2
0
0
6
0
0
3
1
1
2
5
2.2
2.013
2.074
1.818
1.802
2.126
1.582
1.668
1.792
1.976
2.010
1.718
1.913
1.874
AR
0.317
0.321
0.347
0.405
0.390
0.467
0.300
0.294
0.456
0.241
0.306
0.354
0.350
Ho
0.477
0.589
0.423
0.557
0.536
0.484
0.451
0.435
0.557
0.548
0.505
0.500
0.424
He
0.339
0.489
0.195
0.295
0.278
0.040
0.350
0.350
0.203
0.589
0.436
0.297
0.322
FIS
A, average number of alleles per locus in the population; AR, allelic richness; FIS, fixation index; He, expected heterozygosity; Ho, observed heterozygosity; N,
sample size (N1 = cpDNA data, N2 = SSR data); PA, private alleles.
*ICN Herbarium, Department of Botany, Biosciences Institute, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil.
31
6
6
6
20
5
10
7
6
8
7
23
135
PA
18
4
2
4
4
4
11
4
5
10
5
10
81
Voucher*
OSO
TFO
TCA
MAZ
PGR
TSU
SID
S86
S84
TRE
LMU
ORL
Total
Coordinates
A
Collection site
N1
Population
N2
Microsatellite diversity
Local information
Table 1. Measures of genetic diversity for the ten microsatellite loci examined in 12 populations of Recordia reitzii
GENETIC DIVERSITY OF RECORDIA REITZII
© 2014 The Linnean Society of London, Botanical Journal of the Linnean Society, 2014, 176, 332–348
335
336
V. A. THODE ET AL.
diversities (Nei, 1987) and analyses of molecular variance (AMOVA, Excoffier, Smouse & Quattro, 1992)
were calculated with Arlequin 3.5.1.2 (Excoffier &
Lischer, 2010). AMOVA was performed using significance tests with 10 000 permutations. A medianjoining network (Bandelt, Forster & Röhl, 1999) of the
haplotypes was estimated in the program Network
4.1.0.9 (http://www.fluxus-engineering.com/) to determinate the evolutionary relationships between haplotypes. A Mantel test was conducted to test the
correlation between the genetic and geographical distances of the populations using the program Alleles in
Space 1.0 (AIS; Miller, 2005) with 1000 permutations.
DATA ANALYSES: SSR
LOCI
To characterize the SSR loci, genetic diversity indices
for each population (A, average number of alleles
per locus in the population; AR, allelic richness,
Ho, observed heterozygosity; He, expected heterozygosity) were calculated in Arlequin and Fstat 2.9.3.2
(Goudet, 2002). Significant departures from Hardy–
Weinberg equilibrium (HWE) and linkage disequilibrium with Bonferroni correction were tested using
Arlequin and Fstat, respectively. To identify evidence
of possible null alleles, we employed the software
Micro-Checker 2.2.3 (Van Oosterhout et al., 2004).
AMOVA were performed in Arlequin to verify the
degree of genetic differentiation among populations
and among individuals within populations. The pairwise population genetic differentiation was calculated
using Fst (which incorporates the infinite alleles
model) and Rst (designed for SSR data, incorporates
a stepwise mutation model) (Slatkin, 1995). Significance was tested using 10 000 permutations. The
Bayesian clustering method was implemented in
Structure 2.3.3 (Pritchard, Stephens & Donnelly,
2000) to investigate population genetic structure.
This approach allows an optimal number of genetic
clusters (K) to be determined based on individuals
and does not assume an a priori membership in a
population. The analyses were performed with 10
runs for each K (ranging between 1 and 20), with
burn-in of 250 000 Markov chain Monte Carlo
(MCMC) periods using 1 000 000 MCMC replicates,
allowing admixture, with no prior population information and correlated allele frequencies (Pritchard
et al., 2000; Falush, Stephens & Pritchard, 2003). To
infer the best K value, we employed the ΔK method
(Evanno, Regnaut & Goudet, 2005), which favours the
model with the greatest second-order rate of change
in lnPr (X|K). The individual ancestry coefficients for
the Structure runs were calculated using Clumpp
1.1.2 (Jakobsson & Rosenberg, 2007), with the
average pairwise similarity of individual assignments
across runs, applying the Greedy method using the G’
statistic. The Clumpp results were plotted using
Distruct 1.1 (Rosenberg, 2004).
POPULATION
DEMOGRAPHIC HISTORY ANALYSES
Tajima’s D neutrality tests (Tajima, 1989) and Fu’s Fs
(Fu, 1997) were carried out in Arlequin to assess the
population demographic history; significance was
determined based on 10 000 simulations. In addition,
changes in population size over time were estimated
with a Bayesian skyline plot (BSP; Drummond et al.,
2005) analysis performed in BEAST 1.6.1 (Drummond
& Rambaut, 2007). The priors for this analysis were a
relaxed molecular clock model with the substitution
rate previously estimated for chloroplast noncoding
regions (1.0 × 10−9; Wolfe, Li & Sharp, 1987) and an
HKY (Hasegawa, Kishino, and Yano) nucleotide substitution model. MCMC was performed for 100 000 000
steps, sampling every 10 000 steps. The computation of
the BSP and convergence checking were performed in
TRACER 1.5 (http://beast.bio.ed.ac.uk/Tracer).
In order to test possible past changes in population
size that could explain observed values of population
sizes and genetic diversity, we used the Approximate
Bayesian Computation (ABC) method (Beaumont,
Zhang & Balding, 2002) implemented in DIYABC2
(Cornuet et al., 2014), using the nuclear microsatellite
data. Two historical demographic scenarios were compared: (1) constant population size; and (2) recent
bottleneck. The scenario that best fitted the data and
parameters of interest was selected. For this analysis
we pooled all populations into a single sample. For
the constant population size scenario we defined prior
probability distribution of effective population size,
N1, with a uniform distribution bounded between
[5E2, 5E3]. For a recent bottleneck scenario we
defined prior probability distributions of effective
population sizes N1 and N2 with a uniform distribution bounded between [5E2, 5E3] and [5E3, 5E5],
respectively. For the time of population reduction we
set a prior probability distribution between [1E3,
1E5] generations. Several pre-analyses showed that
scenarios with older time for population reduction
were not distinguishable from constant population
scenario. Prior values for mutation model settings
were drawn for all loci. Prior values for mean mutation rates were then drawn from a uniform distribution bounded between [10E−4, 10E−3] and for an
individual locus mutation rate between [10E−5,
10E−2]. Prior values for coefficient P were set to zero
to define the Stepwise Mutation Model. Mutations
that inserted or deleted a single nucleotide from the
SSR sequence were not considered. For summary
statistics we included the mean number of alleles,
mean genic diversity, mean size variance and mean
Ganza–Willamson’s M. A reference table containing a
total of two million simulated datasets was produced.
© 2014 The Linnean Society of London, Botanical Journal of the Linnean Society, 2014, 176, 332–348
GENETIC DIVERSITY OF RECORDIA REITZII
After simulations, Euclidean distances between
observed and simulated datasets were drawn in order
to select the closest datasets for further analyses. To
assess the similarity between observed data and the
combination of priors of each scenario we used the
‘pre-evaluation scenario prior combination’. The scenario with the highest posterior probability and subsequent parameters was used to perform logistic
regression analyses. The posterior probability density
for each parameter, using 10% of the simulated datasets closest to the observed dataset, was estimated.
We assessed the confidence of the selected scenario by
calculating type I and type II errors. Finally, for the
best-fitting scenario, model checking, by simulating
10 000 pseudo observed datasets, was performed to
check the ability to produce simulated datasets
similar to the real dataset.
ECOLOGICAL
NICHE MODELLING
Georeferenced localities of R. reitzii and R. boliviana
(Table S3) were obtained from herbarium records, SpeciesLink (http://www.splink.org.br) and GBIF (http://
www.gbif.org). Only 21 records for R. reitzii (12
localities from the DNA analyses collected for this
study and nine obtained from databases) and 12
collection localities for R. boliviana (all from databases) were available, congruent with the restricted
distribution of both species. Only records with global
positioning system (GPS) coordinates and detailed
localities were used. All localities were plotted onto a
map of South America, encompassing southern Brazil
and Bolivia, using DIVA-GIS 7.1.7.2 (Hijmans et al.,
2004). The environmental data for both sets of locations were obtained from the 19 WorldClim bioclimatic
layers extracted through the programme DIVA-GIS
(http://www.worldclim.org/bioclim; Hijmans et al.,
2005), with a resolution of 30 arc seconds (∼ 1 km2).
The grid layers were cut to the areas that included both
Recordia species and were then exported to the ASCII
format with the software DIVA-GIS. To avoid highly
correlated variables in our dataset, Pearson correlation coefficients were calculated in ENMTools 1.3
(Warren et al., 2008) for all pairs of the 19 bioclimatic
layers. The pairs of variables showing an R > 0.75 were
identified and those presenting the lowest percentage
of importance to the model in a preliminary run were
discarded (Peterson, 2007; Nakazato, Warren & Moyle,
2010) (Table S4). Ten climatic variables were excluded
from the analyses as a result of high correlations and
nine were used in the ENMs (Table S5). The maximum
entropy algorithm in Maxent 3.3 (Phillips, Anderson &
Schapire, 2006) was employed to generate present
ENMs based on species presence data and the nine
bioclimatic predictors. Prediction models were run for
10 iterations using 20% random records to test each
337
run with the subsample option for each replicated run
type. The quality of the models was evaluated based on
the area under the curve (AUC) and True Skill Statistic (TSS; Allouche, Tsoar & Kadmon, 2006) scores. The
AUC scores range from 0.5 to 1, with values above 0.7
being acceptable (Pearce & Ferrier, 2000). TSS values
vary between −1 and 1: values of less than 0 indicate
that the model does not predict the known localities
better than random choice and a value of 1 indicates a
perfect discrimination power of the model between the
known localities and the geographical background. For
the TSS, the threshold value was the maximum of the
sum of the sensitivity and specificity of the test data. To
investigate potential past range shifts of R. reitzii and
R. boliviana, which potentially explain current patterns of geographical distribution, population size and
genetic diversity, we projected the models obtained to
two contrasting past climatic conditions; the Last
Glacial Maximum [(LGM) 21 000 years before present
(kBP)] and the Last Interglacial [(LIG) 120–140 kBP],
based on projections from the Model for Interdisciplinary Research on Climate (MIROC 3.2; Hasumi &
Emori, 2004), available at the WorldClim website
(http://www.worldclim.org/past).
TESTS
FOR NICHE CONSERVATISM AND DIVERGENCE
In addition to the allopatric distribution of these two
species of Recordia, we tested whether these taxa
present conservatism or divergence in their bioclimatic niche preferences. We implemented two
approaches to detect signals of conservatism or divergence between the Recordia species: one based on
similarity metrics between the species’ ENMs and
another based on the reduction of the raw bioclimatic
variables in a principal component analysis (PCA).
Considering that bioclimatic variables show low
spatial heterogeneity and therefore present spatial
autocorrelation (Soberón, 2007), the potential biases
of niche divergence measures associated with geographical distance (Godsoe, 2010) were assessed
using background tests developed for each approach.
This refinement dismisses possible confusion in the
inference of changes in environmental requirements
with changes in the environments available to the
distinct taxa (McCormack et al., 2010).
We employed two metrics to calculate the niche
similarity of R. reitzii and R. boliviana by comparing
the obtained estimates of habitat suitability from the
ENMs. Using ENMtools, we calculated the Schoener’s
D and Hellinger’s I metrics, which range from 0
(niche models show no overlap) to 1 (niche models are
identical, being equally suitable for both species)
(Schoener, 1968; van der Vaart, 1998; Warren et al.,
2008). Then, the identity test was executed to verify
whether the obtained niche-overlap scores exhibited
© 2014 The Linnean Society of London, Botanical Journal of the Linnean Society, 2014, 176, 332–348
338
V. A. THODE ET AL.
statistically significant values (Warren et al., 2008).
This test was performed with 100 replicates to generate a pseudoreplicated null distribution. In each
replicate, the test permutes the species’ localities,
produces new ENMs for each species and compares
them with the two metrics. The observed values of
niche overlap (Schoener’s D and Hellinger’s I) were
compared with the generated null distribution. The
null hypothesis of this test (the observed values of
niche overlap do not differ from random values) is
rejected when the estimated value for D and/or I is
significantly different from the null distribution.
The background test implemented in ENMTools
was used to determine whether the species’ niches
were more distinct than expected based on the environmental background differences between the two
taxa. Null models were taken into account to test
whether any possible observed niche differentiation
between the species was caused by differences
between the regions they occupied (geographical
background) (Warren et al., 2008). This test compares
the suitability scores of one species with the geographical background of the other, and vice versa,
based on 100 overlap values resulting from the comparison of the ENM of one species with ENMs created
from random points plotted within the geographical
range of the other. The number of background points
equals the sample size of the taxon from whose range
the random points were plotted. Two null distributions were generated (one for each taxon). Evidence of
niche divergence or conservatism requires two conditions: that niche characteristics are different/similar
between species and that these differences are
greater/smaller than a null distribution of the niche
similarity indices between one species and random
points within the other species’ range.
A PCA was implemented to explore differences, in
terms of environmental space, between the Recordia
species using the raw bioclimatic data from the nine
climatic variables employed in the ENMs (Table S5)
for the species’ occurrence points and for 1000 random
points plotted in a minimum convex polygon drawn
within the geographical ranges of the taxa using
DIVA-GIS. The values obtained for both species’
localities and background points for the nine variables were reduced through PCA of the correlation
matrix with Mystat (Student’s version of Systat;
Stepp & Leitner, 1989). The most important axes that
explained > 85% of the overall variance and presented
a clear biological interpretation based on the loading
scores of each variable were retained. We tested niche
divergence and conservatism against a null model of
background divergence on each of the axes by comparing the observed difference in mean niche values
for a given principal component with the difference in
mean background values. The significance was
assessed using 1000 Jackknife replicates of the mean
background values.
RESULTS
CPDNA DATA
The amplified sequences of the cpDNA intergenic
spacers trnH–psbA (354 bp), trnS–trnG (682 bp) and
rpl32–trnL (993 bp) were concatenated for all analyses, showing a total length of 2029 bp. The overall
alignment presented one inversion, of 17 bp, and
seven indels, which were coded as one evolutionary
step each. We also identified six single-base substitutions (four transitions and two transversions), resulting in five haplotypes for the 81 individuals from the
12 populations of R. reitzii. The overall haplotype (h)
and nucleotide (π) diversities were 0.441 [standard
deviation (SD) = 0.060] and 0.0015 (SD = 0.0009),
respectively. Haplotype H1 was the most frequent,
being present in all except for the northernmost population, ORL, which exhibited only haplotype H5. The
southernmost population, OSO (H1, H2, H3 and H4),
and population SID (H1 and H5) were the only populations to present more than one haplotype (Fig. 1).
AMOVA showed that the genetic variation caused by
differences among and within populations was 50%.
The Mantel test (r = 0.176, P < 0.001) indicated a
significant association between genetic and geographical distances.
The three sampled individuals of R. boliviana presented the same haplotype, H6 (Hrb), which differs
from H5 by eight mutations, fewer than the ten
mutations that differentiate H5 from the remaining
R. reitzii haplotypes (H1–H4) (Fig. 1).
SSR
DATA
The ten SSR loci analyzed were polymorphic across
the 12 populations (Table 1). Within the sampled
populations, the mean number of alleles (A) ranged
from 1.9 (TSU) to 4.6 (OSO) (mean = 2.8). The
number of private alleles (PA) per population ranged
from 0 (TCA, MAZ, TSU and SID) to 6 (OSO and
PGR). The allelic richness (AR) ranged from 1.6 (TSU)
to 2.1 (PGR) (mean = 1.9). The observed heterozygosity (Ho) varied between 0.241 (TRE) and 0.467
(TSU) (mean = 0.350), the expected heterozygosity
(He) ranged from 0.423 (TCA) to 0.589 (TFO)
(mean = 0.424), and the fixation index (FIS) ranged
from 0.040 (TSU) to 0.589 (TRE) (mean = 0.322). No
evidence of linkage disequilibrium or significant
departures from HWE was detected (Table 1). The
overall SSR data, verified using Micro-Checker, suggested no evidence of genotyping errors as a result of
high allelic dropout, but indicated that null alleles
might be present for Vx04, Vx06 and Vx13, in
© 2014 The Linnean Society of London, Botanical Journal of the Linnean Society, 2014, 176, 332–348
GENETIC DIVERSITY OF RECORDIA REITZII
339
Figure 2. Bayesian structure analysis of Recordia reitzii. Each vertical bar represents an individual apportioned into the
number of inferred clusters (K = 4), represented by different colours, and the black vertical lines delimit populations.
(Colour figure available online.)
addition to genotyping errors resulting from stuttering for Vx04 and Vx13. However, these potential
errors were double-checked with Genetic Profiler, and
all allele peaks were found to be clear and well
defined. The same analyses performed per population
suggested the presence of null alleles at five loci in
different populations: Vx01 (PGR and ORL), Vx04
(PGR and ORL), Vx06 (PGR, ORL and SID), Vx10
(SID), and Vx13 (PGR, ORL, SID and S86). AMOVA of
the SSR data revealed that most of the genetic variability exists within populations (81% based on the
stepwise mutation model and 76% based on the infinite alleles model). The Bayesian cluster analyses
identified the best group structure with K = 4 clusters
(Fig. 2).
POPULATION
DEMOGRAPHIC HISTORY ANALYSES
The results of the Tajima’s D and Fu’s Fs neutrality
tests were positive, but not significant (0.644,
P = 0.775; and 5.376, P = 0.957, respectively). The
BSPs (Fig. S2) suggested a more intense period of
population decline since ∼ 100 kBP.
The DIYABC analysis with the SSR data supported
a scenario of recent past population bottleneck. The
posterior probability of the logistic regression for the
bottleneck scenario was 1.0, with a type I error of
0.218 and a type II error of 0.066. Model check values
and marginal posterior probability densities for
parameters of the bottleneck scenario are provided in
Table S6.
ECOLOGICAL
The ENMs for R. reitzii and R. boliviana are displayed in Figure 3A–F. The ENMs under current climatic conditions predicted suitable localities for both
species that mostly matched their known distributions. Nevertheless, other suitable localities were also
indicated outside the present known ranges of the
species (Fig. 3A, D).
For the LGM, the potential geographical ranges
obtained for R. reitzii (Fig. 3B) and R. boliviana
(Fig. 3E) were somewhat wider than their present
ranges. However, the predictions for the LIG revealed
the widest ranges of suitable areas for both species
(Fig. 3C, F). The ENM for the LGM for R. boliviana
(Fig. 3E) indicated more suitable areas than the
current model (Fig. 3D), including part of the area
predicted for R. reitzii at the LGM along the Brazilian
coast (Fig. 3B).
The predicted distribution of R. reitzii at the LIG
suggested that the range of this species could have
been significantly wider than at present. This ENM
generated predictions, including several areas
forming a diagonal encompassing southern Brazil,
Paraguay and Bolivia, including areas within the
current distribution of R. boliviana (Fig. 3C). The predicted distribution of R. boliviana at the LIG (Fig. 3F)
indicated a similar diagonal pattern, although with
lower suitability values, and it did not reach the
suitable distribution of R. reitzii as it did under the
LGM prediction (Fig. 3E). In addition, the ENMs
identified suitability areas for R. boliviana along the
Pacific slope, surpassing the Andes, disjunct from the
present range of this species (Fig. 3E, F).
NICHE MODELLING
The replicated models for R. reitzii showed average
support values of AUC = 0.995 (SD = 0.004) and
TSS = 0.99 (SD = 0.010). For R. boliviana, these
values were AUC = 0.921 (SD = 0.090) and TSS = 0.74
(SD = 0.400).
Maxent analyses indicated that the variable ‘Annual
Temperature Range’ presented the greatest contribution to the R. boliviana model (42.8%), whereas the
‘Precipitation of the Driest Month’ made the greatest
contribution for R. reitzii (53.5%) (Table S5).
TESTS
OF NICHE CONSERVATISM AND DIVERGENCE
The niche identity test indicated that the niche
overlap indices obtained for R. reitzii and R. boliviana
were significantly different from the pseudoreplicated
datasets (Schoener’s D = 0.058; Hellinger’s I = 0.214;
P < 0.000) (Table 2).
The background tests did not indicate significant
evidence of either niche divergence or conservatism.
The observed niche overlap values, for Schoener’s D
© 2014 The Linnean Society of London, Botanical Journal of the Linnean Society, 2014, 176, 332–348
340
V. A. THODE ET AL.
Figure 3. Ecological niche models for Recordia reitzii (A–C) and Recordia boliviana (D–F) obtained with Maxent. Black
dots represent species localities on which the models were based. Scale bars on C and F indicate suitability scores
represented for R. reitzii and R. boliviana, respectively (the minimum suitability values shown are based on maximum
test sensitivity plus specificity threshold). In maps B and E, the coastline is shown for the LGM conditions and current
coastline is included as reference. The overview map (a) indicates the current extent of the Atlantic forest with the darker
area representing the Ombrophilous Dense Forest.
and Hellinger’s I, were not significantly different from
the null distributions (Table 2).
The tests of niche conservatism and divergence,
based on PCA of the raw bioclimatic data, showed
three axes that explained 85.8% of the total variation
(Table 3). For principal component 1 (PC1), accounting
for 47.9% of the total variation, variables that describe
temperature and precipitation seasonality presented
higher loadings, showing no signal of niche conserva-
tism or divergence and a high longitude/latitude correlation (Table 3). The other two principal components
(PC2 and PC3) indicated significant niche differences
between the Recordia species. PC2 accounted for
24.6% of the total variation and was mainly explained
by variables describing extreme temperature conditions, whereas PC3 accounted for 13.2% of the total
variation and was mainly explained by variables
describing temperature seasonality (Table 3).
© 2014 The Linnean Society of London, Botanical Journal of the Linnean Society, 2014, 176, 332–348
GENETIC DIVERSITY OF RECORDIA REITZII
341
Table 2. Niche overlap values for Helligers’ I and Schoener’s D compared with null distributions generated with the
identity test
Hellinger’s I
Schoener’s D
Niche identity test
Background test
R. reitzii vs. R. boliviana
R. reitzii vs. R. boliviana
background
R. boliviana vs. R. reitzii
background
Observed
values
Null
distribution
Observed
values
Null
distribution
Observed
values
Null
distribution
0.21
0.06
0.68–0.96
0.38–0.80
0.21
0.06
0.11–0.35
0.02–0.11
0.21
0.06
0.14–0.21
0.04–0.06
The similarity scores (I and D) are lower than the null hypotheses of niche equivalency, indicating that the bioclimatic
niches differ more than expected by chance.
Background test values for Helligers’ I and Schoener’s D are given.
The observed niche overlap values were not significantly different from the background null distributions, indicating no
significant evidence for either niche divergence or conservatism.
Table 3. Test of niche divergence vs. conservatism for Recordia reitzii and Recordia boliviana
Niche axes
Pairwise comparison
PC1
PC2
PC3
Observed differences
Null distribution
Explained (%)
Variable loading 1
Variable loading 2
1.91*
(1.88–1.92)
47.9
Precipitation of driest month
(−0.939)
Annual precipitation (−0.914)
Variable loading 3
Isothermality (0.896)
Variable loading 4
Temperature seasonality
(−0.829)
Weather seasonality and
range
0.96
−0.97
0.47*
(0.08–0.24)
24.6
Minimum temperature of
coldest month (0.837)
Mean temperature of driest
quarter (0.69)
Mean temperature of wettest
quarter (0.683)
–
0.77
(0.26–0.43)
13.2
Temperature annual range
(−0.923)
Temperature seasonality
(−0.382)
Mean temperature of wettest
quarter (−0.32)
–
Temperature extreme
conditions
0.02
0.02
Temperature seasonality
Interpretation
Longitude correlation
Latitude correlation
0.08
−0.02
Bold values indicate significant niche divergence between R. reitzii and R. boliviana.
*Niche values differ significantly between lineage pairs (Wilcoxon–Mann–Whitney test: P < 0.015).
DISCUSSION
GENETIC
VARIABILITY AND POPULATION STRUCTURE
OF
R.
REITZII
The cpDNA data showed that the R. reitzii populations presented lower haplotype and nucleotide diversities compared with other Neotropical trees (Ramos
et al., 2009; Novaes et al., 2010; Garcia et al., 2011;
Collevatti et al., 2012b), which might be related to the
sample size examined as well as the rarity and
restricted distribution of R. reitzii (covering an extension of only ∼ 200 km north–south along the Brazilian
southeastern coast). The low intrapopulation genetic
diversity observed, and the presence of fixed haplotypes in almost all populations, may be a result of
small population sizes or of a founder or distribution–
reduction effect. In fact, ABC and BSP analyses
support that recent population reduction seems the
most likely historical scenario that explains the
observed population sizes and genetic diversity
values. This finding matches with the ecological niche
models and their projections to LGM and LIG conditions that propose a wider distribution for R. reitzii in
the past (Fig. 3A–C).
© 2014 The Linnean Society of London, Botanical Journal of the Linnean Society, 2014, 176, 332–348
342
V. A. THODE ET AL.
The reduction of genetic variability observed in the
cpDNA, and the homozygote excess and positive FIS
values indicated for the SSR markers, may be a
consequence of the reduction of the species distribution since the LGM (Fig. 1 and Table 1). A Wahlund
effect (Wahlund, 1928) is unlikely, as the populations
did not show high substructuration in the STRUCTURE analysis. The analysis of nuclear markers did
not support the presence of exclusive haplotypes in
population OSO or, thus, its apparent isolation
(Figs 1, 2). The results from the Bayesian cluster
analyses shown in Figure 2 may indicate the presence
of some level of gene flow among populations. There is
no available information about the inheritance of
cpDNA in Recordia species or regarding their pollinator agents or seed-dispersion systems. The low
number of haplotypes may be related to a recent split,
resulting in a short divergence time, to restricted
gene flow and/or to the fact that R. reitzii may present
a low cpDNA substitution rate.
Mantel tests showed a significant association
between genetic and geographical distances, indicating a north–south pattern of diversity in the Atlantic
forest. A north–south pattern of diversity has been
previously observed in this ecoregion for other plant
groups, nevertheless with a broader distribution
along the Atlantic forest (Lorenz-Lemke et al., 2005;
Palma-Silva et al., 2009; Turchetto-Zolet et al., 2012).
Our cpDNA and SSR data revealed at least three
different genetic units (south, central and north) that
may be influenced by the distance between populations (Figs 1, 2). This result was confirmed by the
AMOVA results, showing that most of the variability
occurs within populations, demonstrating the existence of some level of gene flow. Additionally, each
population exhibits one unique haplotype in most
cases and presents one unique genetic component
predominant in the SSRs, which may be the result of
the population phenomena mentioned above, including distribution reduction.
Only eight mutations in the cpDNA markers separate the sampled individuals of the two Recordia
species, which is less than the number of mutations
that distinguish haplotypes H1 and H5 in R. reitzii
(Fig. 1), confirming the close relationship between
these species (Thode et al., 2013a).
Analyses of intraspecific genetic diversity are
important for identifying lineages that are geographically and genetically distinct for conservation purposes (Moritz, 1994; Moritz & Faith, 1998; Moritz
et al., 2000). The Atlantic forest and its species are
under threat from habitat destruction; therefore, the
conservation of areas that harbour narrowly distributed species is important to preserve genetic diversity
(Moritz et al., 2000). There are areas of protection
within the range of occurrence of R. reitzii, such as
the Serra Geral National Park, Serra Geral Biological
Reserve and Municipal Area of Environmental Protection of Caraá; however, more supervision of the
region and execution of the laws are necessary to
protect this threatened species.
SPECIES
DISTRIBUTION MODELLING
The AUC and TSS values indicated that good-quality
predictions were generated for both species. However,
for R. boliviana, lower and more variable values were
obtained (especially for TSS), which could be a result
of the reduced sample size. In addition, these TSS
values reflect a relatively high variability of the predictions produced across the different combinations of
localities included in each of the 10 replicates, suggesting that the models were influenced by the set of
localities used to train them. Similar results have
been found in other experimental studies using small
numbers of occurrence records (Pearson et al., 2007).
The ENMs under current climatic conditions also
predicted suitable localities outside the known
present ranges of the species (Fig. 3A, D). These areas
may display similar bioclimatic conditions to the
known species occurrence area and could be useful for
targeting field surveys to improve the discovery of
unknown populations (Pearson et al., 2007; de
Siqueira et al., 2009; Kamino et al., 2012).
The model of the LGM distribution of R. reitzii
showed that this species might have mainly occurred
in the same restricted region along the southern
Brazilian coast at that time, but in a greater number
of suitable areas (from its current distribution
towards the coast) owing to the presence of more
available land as a result of the lower sea levels at the
LGM (Fig. 3B).
The ENMs suggested that the current range of
Recordia species might represent a relict of a wider
population distribution during the LIG because of the
presence of more areas with suitable conditions into
which forest organisms could expand at that time
(warmer and more humid) (Behling et al., 2002;
Carnaval & Moritz, 2008; Lorenz-Lemke et al., 2010)
(Fig. 3C, F).
In the cooler and dryer LGM, the ancestral populations probably retracted and persisted in disjunct
areas, being locally restricted to a few locations with
suitable ecological conditions that allowed the survival of the two extant species (Fig. 3B, E). The
Atlantic forest is mainly separated from other Neotropical humid forests by a ‘dry diagonal’ of
seasonally dry biomes composed of open vegetation
and savannah-like areas, which include the Caatinga, Cerrado and Chaco (Vanzolini, 1963; Collevatti
et al., 2012a). Changes in species’ distributions and
persistence in refugia with adequate conditions occur
© 2014 The Linnean Society of London, Botanical Journal of the Linnean Society, 2014, 176, 332–348
GENETIC DIVERSITY OF RECORDIA REITZII
more frequently than complete extinctions (VegasVilarrubia et al., 2011; Turchetto-Zolet et al., 2013).
During past climatic conditions, plant lineages from
different vegetation formations (such as the Amazon
and dry forests) may have been connected with the
Atlantic forest by migration routes, such as forest
corridors (e.g. gallery forests along rivers) or a series
of patches (allowing migration through island
hopping) (Rizzini, 1963; Andrade-Lima, 1964;
Bigarella, Andrade-Lima & Riehs, 1975; Por, 1992;
Oliveira-Filho & Ratter, 1995; Costa, 2003).
Climate change and the evolutionary history of
species appear to be connected. Thus, the study of
past climate changes is extremely important for
obtaining information regarding future scenarios
and for biodiversity conservation (Rull, 2011). Climatic changes may affect the range of species as
they track suitable environments (Hewitt, 1996) and
may influence their distributions and rates of diversification. In the evolutionary past, corridors for the
migration of the organisms were available, allowing
movement of species during climate changes
(Donoghue, 2008). However, human habitat fragmentation might limit the dispersal of organisms at
present and under future climate changes and may
influence the assembly of regional species pools
(Donoghue, 2008).
TESTS
OF NICHE CONSERVATISM AND DIVERGENCE
The PCA-based niche-divergence test indicated that
variables related to extreme temperature conditions
(PC2) and temperature seasonality (PC3) might be
the factors showing the greatest influence over bioclimatic niche divergence between the Recordia species
(Table 3). This divergence could cause ecological specialization and possible changes in environmental tolerances between the species. The raw values of
variables that present highest loadings in PC2 for
both Recordia species show that R. reitzii presents
greater tolerance for colder periods. This may be
related to the area occupied by R. reitzii, which presents temperature values for extreme conditions that
are significantly different from and lower than those
for R. boliviana (bio, 6 Wilcoxon–Mann–Whitney test
P = 0.0144; bio8, t-test P = 0.0084; bio9, t-test
P = 0.0036). Additionally, the localities of R. reitzii
present significantly different and higher values of
temperature seasonality than do those of R. boliviana
(bio4, t-test P < 0.0001). These results indicate that
R. reitzii occupies areas with a higher amplitude of
temperature conditions, along the southern Atlantic
Brazilian coast, compared with the conditions of the
environment inhabited by R. boliviana in the low
mountains of the eastern Andes (see known localities
in Fig. 3A, D). These higher amplitude of conditions
343
could provide more environments (in temperature
range) to explore, which may have led to expansion of
the niche breadth of R. reitzii in relation to its sister
species.
Many theoretical and empirical contributions have
stressed the artificial increment of niche divergence
observed when comparing the bioclimatic niches
between allopatric species (Godsoe, 2010, 2012;
McCormack et al., 2010; Peterson, 2011). In addition,
niche conservatism has been accepted as a predominant feature in evolutionary biology (Peterson,
Soberón & Sánchez-Cordero, 1999; Prinzing et al.,
2001; Wiens & Graham, 2005). Taking into account
these two factors, our results regarding bioclimatic
niche divergence in Recordia should be carefully
interpreted. We may infer that the bioclimatic data
used in our analyses present a strong geographical
correlation and that most of the data on these variables indicate regional variation, rather than ecological niche characteristics. Therefore, the niche
divergence test based on ENM comparisons (Table 2)
did not reveal significant niche divergence between
the Recordia species (the differences between the
bioclimatic spaces occupied by the species did not
exceed the differences between regions). Fortunately,
the orthogonal transformation of the variables in the
PCA allowed all of this geographical correlation
among the variables and the overall data on regional
variation to be combined into a principal component
(PC1 is the only axis that showed a geographical
correlation; Table 3), after which the residual data
could be used to obtain ‘clean’ information (PC2
and PC3) to test the bioclimatic niche divergence of
Recordia.
Our results suggest that late Pleistocene fluctuations and bioclimatic factors might have played an
important role in the distribution of Recordia species
(Moritz et al., 2000), possibly through promoting
niche shifts among these species. Finally, to make
further inferences regarding the patterns of diversification observed in Recordia, dated phylogenies and
analyses of divergence time estimates are necessary
for the Verbenaceae.
The information generated in this study regarding
genetic variability, taking into account the genetically
distinct units (south, central and north) indicated by
the cpDNA and SSR data and the responses of
R. reitzii to climate change, which revealed a retraction of the populations of this species compared with
the LGM and LIG predictions, might be useful for
generating conservation strategies for this endangered species.
More in-depth biological studies, such as analyses
of pollination and dispersal systems, are required to
shed light on what factors influence the low genetic
diversity found in R. reitzii populations.
© 2014 The Linnean Society of London, Botanical Journal of the Linnean Society, 2014, 176, 332–348
344
V. A. THODE ET AL.
ACKNOWLEDGEMENTS
We thank Ana Luíza R. Cazé and Maikel ReckKortmann for help in the preliminary analyses, Alice
B. Rosa for assistance in the laboratory experiments
and Aline M.C. Ramos-Fregonezi, Jeferson N. Fregonezi, and Priscilla M. Zamberlan for assistance during
the field collections. We also thank Dr Alexandre
Antonelli and two anonymous reviewers for comments and suggestions that improved this manuscript. This project was supported by the Conselho
Nacional de Desenvolvimento Científico e Tecnológico
(CNPq) and Coordenação de Aperfeiçoamento de
Pessoal de Nível Superior (CAPES).
REFERENCES
Alcantara S, Semir J, Solferini VN. 2006. Low genetic
structure in an epiphytic Orchidaceae (Oncidium hookeri) in
the Atlantic Rainforest of South-eastern Brazil. Annals of
Botany 98: 1207–1213.
Allouche O, Tsoar A, Kadmon R. 2006. Assessing the
accuracy of species distribution models: prevalence, kappa
and the true skill statistic (TSS). Journal of Applied Ecology
43: 1223–1232.
Andrade-Lima D. 1964. Contribuição à dinâmica da flora do
Brasil. Arquivos do Instituto Ciências da Terra 2: 15–20.
Antonelli A, Sanmartín I. 2011. Why are there so many
plant species in the Neotropics? Taxon 60: 403–414.
Bandelt H-J, Forster P, Röhl A. 1999. Median-joining networks for inferring intraspecific phylogenies. Molecular
Biology and Evolution 16: 37–48.
Beaumont MA, Zhang WY, Balding DJ. 2002. Approximate
Bayesian computation in population genetics. Genetics 162:
2025–2035.
Behling H, Arz HW, Patzold J, Wefer G. 2002. Late Quaternary vegetational and climate dynamics in southeastern
Brazil, inferences from marine cores GeoB 3229-2 and GeoB
3202-1. Palaeogeography, Palaeoclimatology, Palaeoecology
179: 227–243.
Behling H, Lichte M. 1997. Evidence of dry and cold climatic
conditions at glacial times in tropical SE Brazil. Quaternary
Research 48: 348–358.
Behling H, Negrelle RRB. 2001. Tropical Rain Forest and
climate dynamics of the Atlantic lowland, Southern Brazil,
during the Late Quaternary. Quaternary Research 56: 383–
389.
Behling H, Pillar V. 2007. Late Quaternary vegetation,
biodiversity and fire dynamics on the southern Brazilian
highland and their implication for conservation and management of modern Araucaria forest and grassland ecosystems. Philosophical Transactions of the Royal Society B 362:
243–251.
Bernardo-Silva J, Martins-Ferreira C, Maneyro R, de
Freitas TRO. 2012. Identification of priority areas for
conservation of two endangered parapatric species of redbellied toads using ecological niche models and hotspot
analysis. Natureza & Conservação 10: 207–213.
Bigarella JJ, Andrade-Lima D, Riehs PJ. 1975.
Considerações a respeito das mudanças paleoambientaisna
distribuição de algumas espécies vegetais e animais no
Brasil. Anais da Academia Brasileira de Ciências 47: 411–
464.
Carnaval A, Moritz C. 2008. Historical climate modelling
predicts patterns of current biodiversity in the Brazilian
Atlantic forest. Journal of Biogeography 35: 1187–
1201.
Collevatti RG, de Castro TG, de Souza Lima J, de
Campos TM. 2012b. Phylogeography of Tibouchina
papyrus (Pohl) Toledo (Melastomataceae), an endangered
tree species from rocky savannas, suggests bidirectional
expansion due to climate cooling in the Pleistocene. Ecology
and Evolution 2: 1024–1035.
Collevatti RG, Grattapaglia D, Hay JD. 2003. Evidences
for multiple maternal lineages of Caryocar brasiliense populations in the Brazilian Cerrado based on the analysis of
chloroplast DNA sequences and microsatellite haplotype
variation. Molecular Ecology 12: 105–115.
Collevatti RG, Terrible LC, Lima-Ribeiro MS, Nabout
JC, Oliveira G, Rangel TF, Rabelo SG, Diniz-Filho JA.
2012a. A coupled phylogeographical and species distribution
modelling approach recovers the demographical history of a
Neotropical seasonally dry forest tree species. Molecular
Ecology 21: 5845–5863.
Conte R, Nodari RO, Vencovsky R, Reis MS. 2003.
Genetic Diversity and recruitment of the tropical palm
Euterpe edulis Mart., in a natural population from the
Brazilian Atlantic Forest. Heredity 91: 401–406.
Cornuet J-M, Pudlo P, Veyssier J, Dehne-Garcia A,
Gautier M, Leblois R, Marin J-M, Marin J-M. 2014.
DIYABC v2.0: a software to make Approximate Bayesian
Computation inferences about population history using
Single Nucleotide Polymorphism, DNA sequence and microsatellite data. Bioinformatics 30: 1187–1189.
Costa LP. 2003. The historical bridge between the Amazon
and the Atlantic Forest of Brazil: a study of molecular
phylogeography with small mammals. Journal of Biogeography 30: 71–86.
Donoghue MJ. 2008. A phylogenetic perspective on the distribution of plant diversity. Proceedings of the National
Academy of Sciences of the United States of America 105:
11549–11555.
Drummond AJ, Rambaut A. 2007. BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evolutionary
Biology 7: 214.
Drummond AJ, Rambaut A, Shapiro B, Pybus OG. 2005.
Bayesian coalescent inference of past population dynamics
from molecular sequences. Molecular Biology and Evolution
22: 1185–1192.
Dunn IS, Blattner FR. 1987. Sharons 36 to 40: Multienzyme, high capacity, recombination deficient replacement
vectors with polylinkers and polystuffers. Nucleic Acids
Research 15: 2677–2698.
Ennos RA. 1994. Estimating the relative rates of pollen and
seed migration among plant populations. Heredity 72: 250–
259.
© 2014 The Linnean Society of London, Botanical Journal of the Linnean Society, 2014, 176, 332–348
GENETIC DIVERSITY OF RECORDIA REITZII
Evanno G, Regnaut S, Goudet J. 2005. Detecting the
number of cluster of individuals using the software structure: a simulation study. Molecular Ecology 14: 2611–2620.
Excoffier L, Lischer HEL. 2010. Arlequin suite ver 3.5:
A new series of programs to perform population genetics
analyses under Linux and Windows. Molecular Ecology
Resources 10: 564–567.
Excoffier L, Smouse PE, Quattro JM. 1992. Analysis of
molecular variance inferred from metric distances among
DNA haplotypes: application to human mitochondrial
restriction data. Genetics 131: 479–491.
Falush D, Stephens M, Pritchard JK. 2003. Inference of
population structure using multilocus genotype data:
Linked loci and correlated allele frequencies. Genetics 164:
1567–1587.
Fu YX. 1997. Statistical tests of neutrality of mutations
against population growth, hitchhiking and background
selection. Genetics 147: 915–925.
Fundação Instituto Brasileiro de Geografia e Estatística. 1993. Recursos naturais e meio ambiente: uma visão do
Brasil. Rio de Janeiro: Fundação Instituto Brasileiro de
Geografia e Estatística.
Garcia MG, Silva RS, Carniello MA, Veldman JW, Rossi
AAB, de Oliveira LO. 2011. Molecular evidence of cryptic
speciation, historical range expansion, and recent intraspecific hybridization in the Neotropical seasonal forest tree
Cedrela fissilis (Meliaceae). Molecular Phylogenetics and
Evolution 61: 639–649.
Godsoe W. 2010. Regional variation exaggerates ecological
divergence in niche models. Systematic Biology 59: 298–306.
Godsoe W. 2012. Are comparisons of species distribution
models biased? Are they biologically meaningful? Ecography
35: 769–779.
Goudet J. 2002. FSTAT: a computer program to calculate
F-statistics (Version 2.9.3.2.). Journal of Heredity 86: 485–
486.
Hamilton MB. 1999. Four primer pairs for the amplification
of chloroplast intergenic regions with intraspecific variation.
Molecular Ecology 8: 521–523.
Harvey PH, Pagel M. 1991. The comparative method in
evolutionary biology. Oxford: Oxford University Press.
Hasumi H, Emori S. 2004. K-1 coupled model (MIROC)
description. K-1 Technical Report, Center for Climate
System Research, University of Tokyo, 34 pp.
Hewitt GM. 1996. Some genetic consequences of ice ages, and
their role in divergence and speciation. Biological Journal of
the Linnean Society 58: 247–276.
Hijmans R, Cameron S, Parra J, Jones P, Jarvis A. 2005.
Very high resolution interpolated climate surfaces for global
land areas. International Journal of Climatology 25: 1965–
1978.
Hijmans RJ, Guarino L, Bussink C, Mathur P, Cruz M,
Barrentes I, Rojas E. 2004. DIVA-GIS. v.7.1.7. A geographic information system for the analysis of species
distribution data. Available at: http://www.diva-gis.org
(accessed 1 February 2013).
Hua X, Wiens JJ. 2013. How does climate influence speciation? The American Naturalist 182: 1–12.
345
Hung C-M, Drovetski SV, Zink RM. 2013. Recent allopatric
divergence and niche evolution in a widespread Palearctic
bird, the common rosefinch (Carpodacus erythrinus).
Molecular Phylogenetics and Evolution 66: 103–111.
Jakobsson M, Rosenberg NA. 2007. Clumpp: a cluster
matching and permutation program for dealing with label
switching and multimodality in analysis of population structure. Bioinformatic 23: 1801–1806.
Kamino LHY, de Siqueira MF, Sánchez-Tapia A,
Stehmann JR. 2012. Reassessment of the extinction risk of
endemic species in the neotropics: How can modelling tools
help us? Natureza & Conservação 10: 191–198.
Knowles LL, Carstens BC, Keat ML. 2007. Coupling
genetic and ecological-niche models to examine how past
population distributions contribute to divergence. Current
Biology 17: 940–946.
Lewinsohn TM, Prado PI. 2005. How many species are
there in Brazil? Conservation Biology 19: 619–624.
Lorenz-Lemke AP, Muschner VC, Bonatto SL, Cervi AC,
Salzano FM, Freitas LB. 2005. Phylogeographic inferences concerning evolution of brazilian Passiflora acitinia
and P. elegans (Passifloraceae) based on ITS (nrDNA) variation. Annals of Botany 95: 799–806.
Lorenz-Lemke AP, Togni PD, Mäder G, Kriedt RA,
Stehmann JR, Salzano FM, Bonatto SL, Freitas LB.
2010. Diversification of plant species in a subtropical region
of eastern South American highlands: a phylogeographic
perspective on native Petunia (Solanaceae). Molecular
Ecology 19: 5240–5251.
McCormack JE, Zellmer AJ, Knowles LL. 2010. Does niche
divergence accompany allopatric divergence in Aphelocoma
jays as predicted under ecological speciation? Insights from
tests with niche models. Evolution 64: 231–1244.
Metzger JP, Martensen AC, Dixo M, Bernacci LC,
Ribeiro MC, Teixeira AMG, Pardini R. 2009. Time-lag in
biological responses to landscape changes in a highly
dynamic Atlantic forest region. Biological Conservation 142:
1166–1177.
Miller MP. 2005. Alleles in Space: computer software for the
joint analysis of interindividual spatial and genetic information. Journal of Heredity 96: 722–724.
Morellato LM, Haddad CFB. 2000. Introduction: The Brazilian Atlantic forest. Biotropica 32: 786–792.
Mori SA, Boom BM, Prance GT. 1981. Distribution patterns and conservation of eastern Brazilian coastal forest
tree species. Brittonia 33: 233–245.
Moritz C. 1994. Defining ‘Evolutionary Significant Units’ for
conservation. Trends in Ecology and Evolution 9: 373–375.
Moritz C, Faith DP. 1998. Comparative phylogeography and
the identification of genetically divergent areas for conservation. Molecular Ecology 7: 419–429.
Moritz C, Patton JL, Schneider CJ, Smith TB. 2000.
Diversification of rainforest faunas: an integrated molecular
approach. Annual Review of Ecology and Systematics 31:
533–563.
Myers N, Mittermeier RA, Mittermeier CG, da Fonseca
GAB, Kent J. 2000. Biodiversity hotspots for conservation
priorities. Nature 403: 853–858.
© 2014 The Linnean Society of London, Botanical Journal of the Linnean Society, 2014, 176, 332–348
346
V. A. THODE ET AL.
Nakazato T, Warren DL, Moyle LC. 2010. Ecological and
geographic modes of species divergence in wild tomatoes.
American Journal of Botany 97: 680–693.
Nei M. 1987. Molecular evolutionary genetics. New York:
Columbia University Press.
Novaes RML, Filho JP, Ribeiro RA, Lovato MB. 2010.
Phylogeography of Plathymenia reticulata (Leguminosae)
reveals patterns of recent range expansion towards northeastern Brazil and southern Cerrados in Eastern Tropical
South America. Molecular Ecology 19: 985–998.
Oliveira-Filho AT, Ratter JA. 1995. A study of the origin of
central Brazilian forests by the analysis of plant species
distribution patterns. Edinburgh Journal of Botany 52:
141–194.
Oliveira-Filho AT, Budke JC, Jarenkow JA, Eisenlohr
PV, Neves DRM. 2013. Delving into the variations in tree
species composition and richness across South American
subtropical Atlantic and Pampean forests. Journal of Plant
Ecology. doi: 10.1093/jpe/rtt058.
Palma-Silva C, Lexer C, Paggi GM, Barabará T, Bered F,
Bodanese-Zanettini MH. 2009. Range-wide patterns of
nuclear and chloroplast DNA diversity in Vriesea gigantea
(Bromeliaceae), a Neotropical forest species. Heredity 103:
503–512.
Pearce J, Ferrier S. 2000. Evaluating the predictive performance of habitat models developed using logistic regression.
Ecological Modelling 133: 225–245.
Pearman PB, Guisan A, Broennimann O, Randin CF.
2008. Niche dynamics in space and time. Trends in Ecology
& Evolution 23: 149–158.
Pearson RG, Raxworthy CJ, Nakamura M, Peterson AT.
2007. Predictingspeciesdistributionsfromsmallnumbers of
occurrence records: a test case usingcrypticgeckos in Madagascar. Journal of Biogeography 34: 102–117.
Perret M, Chautems A, Spichiger R. 2006. Dispersalvicariance analyses in the tribe Sinningieae (Gesneriaceae):
a clue to understanding biogeographical history of the Brazilian Atlantic forest. Annals of the Missouri Botanical
Garden 93: 340–358.
Peterson AT. 2007. Why not WhyWhere: the need for more
complex models of simpler environmental spaces. Ecological
Modelling 203: 527–530.
Peterson AT. 2011. Ecological niche conservatism: a timestructured review of evidence. Journal of Biogeography 38:
817–827.
Peterson AT, Soberón J, Sánchez-Cordero V. 1999. Conservatism of ecological niches in evolutionary time. Science
285: 1265–1267.
Phillips S, Anderson R, Schapire R. 2006. Maximum
entropy modeling of species geographic distributions. Ecological Modelling 190: 231–259.
Pinheiro F, Barros F, Palma-Silva C, Fay MF, Lexer C,
Cozzolino S. 2011. Phylogeography and genetic differentiation along the distributional range of the orchid Epidendrumfulgens: a Neotropical coastal species not restricted to
glacial refugia. Journal of Biogeography 38: 1923–1935.
Por FD. 1992. Sooretama: the Atlantic rain forest of Brazil.
The Hague: SPB Academic Publishing.
Prinzing A, Durka W, Klotz S, Brandl R. 2001. The niche
of higher plants: evidence for phylogenetic conservatism.
Proceedings of the Royal Society of London. Series B: Biological Sciences 268: 2383–2389.
Pritchard JK, Stephens M, Donnelly P. 2000. Inference of
population structure using multilocus genotype data. Genetics 155: 945–959.
Ramos ACS, Lemos-Filho JP, Lovato MB. 2009. Phylogeographical structure of the neotropical forest tree Hymenaeacourbaril (Leguminosae: Caesalpinioideae) and its
relationship with the vicariant Hymenaea stigonocarpa from
Cerrado. Journal of Heredity 100: 206–216.
Ranta P, Blom T, Niemelä J, Joensuu E, Siitonen M.
1998. The fragmented Atlantic rain forest of Brazil: size,
shape and distribution of forest fragments. Biodiversity and
Conservation 7: 385–403.
Reis MS, Fantini AC, Nodari RO, Reis A, Guerra MP,
Montovani A. 2000. Management and conservation of
natural populations in Atlantic Rain Forest: the case study of
palm heart (EuterpeedulisMartius). Biotropica 32: 894–902.
Reitz R, Klein RM, Reis A. 1978. Projeto Madeira de Santa
Catarina. Sellowia 30: 1–320.
Reitz R, Klein RM, Reis A. 1983. Projeto Madeira do Rio
Grande do Sul. Sellowia 34–35: 1–525.
Ribeiro MC, Metzger JP, Martensen AC, Ponzoni FJ,
Hirota MM. 2009. The Brazilian Atlantic Forest: how much
is left, and how is the remaining forest distributed? Implications for conservation. Biological Conservation 142: 1141–
1153.
Ribeiro RA, Ramos ACS, Lemos Filho JP, Lovato MB.
2005. Genetic variation in remnant populations of
Dalbergianigra(Papilionoideae), an endangered tree from
the Brazilian Atlantic Forest. Annals of Botany 95: 1171–
1177.
Rizzini CT. 1963. Nota prévia sobre a divisão fitogeográfica
do Brasil. Revista Brasileira de Geografia 25: 1–64.
Rosenberg NA. 2004. Distruct: a program for thegraphical
display of populationstructure. Molecular Ecology Notes 4:
137–138.
Roy A, Frascaria N, Mackay J, Bousquet J. 1992. Segregating random amplified polymorphic DNAs (RAPDS) in
Betulaalleghaniensis. Theoretical and Applied Genetics 85:
173–180.
Rull V. 2011. Neotropical biodiversity: timing and potential
drivers. Trends in Ecology & Evolution 26: 508–513.
Salimena FRG, Múlgura de Romero ME, O’ Leary N,
Peralta P, Thode VA. 2009. Verbenaceae. In: Stehmann
JR, Forzza RC, Salino A, Sobral M, Costa DP, Kamino LHY,
eds. Plantas da Floresta Atlântica. Rio de Janeiro: Jardim
Botânico do Rio de Janeiro, 492–494.
Sang T, Crawford DJ, Stuessy TF. 1997. Chloroplast DNA
phylogeny, reticulate evolution, and biogeography of
Paeonia (Paeoniaceae). American Journal of Botany 84:
1120–1136.
Schoener TW. 1968. Anolis lizards of Bimini: resource partitioning in acomplex fauna. Ecology 49: 704–726.
Shaw J, Lickey EB, Schilling EE, Small RL. 2007. Comparison of whole chloroplast genome sequences to choose
© 2014 The Linnean Society of London, Botanical Journal of the Linnean Society, 2014, 176, 332–348
GENETIC DIVERSITY OF RECORDIA REITZII
noncoding regions for phylogenetics studies in angiosperms:
the tortoise and the hare III. American Journal of Botany
94: 275–288.
Silva SM, Moraes-Barros N, Ribas CC, Ferrand N,
Morgante JS. 2012. Divide to conquer: a complex pattern
of biodiversity depicted by vertebrate components in the
Brazilian Atlantic Forest. Biological Journal of the Linnean
Society 107: 39–55.
de Siqueira MF, Durigan G, de Marco Júnior P,
Peterson AT. 2009. Something from nothing: using landscape similarity and ecological niche modeling to find rare
plant species. Journal for Nature Conservation 17: 25–32.
Slatkin M. 1995. A measure of population subdivision
based on microsatellite allele frequencies. Genetics 139:
457–462.
Soberón J. 2007. Grinnellian and Eltonian niches and geographic distributions of species. Ecology Letters 10: 1115–
1123.
Sobral M, Jarenkow JA, Brack P, Irgang B, Larocca J,
Rodrigues RS. 2006. Flora arbórea e arborescente do Rio
Grande do Sul, Brasil. São Carlos: RiMA/Novo Ambiente.
Sobral M, Stehmann JR. 2009. An analysis of new angiosperm species discoveries in Brazil (1990-2006). Taxon 58:
227–232.
Stehmann JR, Forzza RC, Salino A, Sobral M, Costa DP,
Kamino LHY. 2009. Plantas da Floresta Atlântica. Rio de
Janeiro: Jardim Botânico do Rio de Janeiro.
Stepp SL, Leitner DW. 1989. Software review: Mystat™:
A personal version of systat (statistical). Nonprofit Management Leadership 28: 33–34.
Tajima F. 1989. Statistical method for testing the neutral
mutation hypothesis by DNA polymorphism. Genetics 123:
585–595.
Tamura K, Peterson D, Peterson N, Stecher G, Nei M,
Kumar S. 2011. MEGA5: molecular evolutionary genetics
analysis using maximum likelihood, evolutionary distance,
and maximum parsimony methods. Molecular Biology and
Evolution 28: 2731–2739.
Terborgh J. 1992. Maintenance of the diversity in tropical
forests. Biotropic 24: 283–292.
Thode VA, Backes A, Mader G, Kriedt RA, Bonatto SL,
Freitas LB. 2013b. Development of microsatellites for Verbenoxylum reitzii (Verbenaceae), a tree endemic to the Brazilian Atlantic Rainforest. Applications in Plant Sciences 1:
1300005 (1–3).
Thode VA, O’Leary N, Olmstead RG, Freitas LB. 2013a.
Phylogenetic position of the monotypic genus Verbenoxylum
(Verbenaceae) and new combination under Recordia. Systematic Botany 38: 805–817.
Troncoso NS. 1974. Los Géneros de Verbenáceas de Sudamérica extratropical. Darwiniana 18: 295–412.
347
Turchetto-Zolet AC, Cruz F, Vendramin GG, Simon MF,
Salgueiro F, Margis-Pinheiro M, Margis R. 2012.
Large-scale phylogeography of the disjunct Neotropical tree
specie Schizolobium parahyba (Fabaceae-Caesalpinioideae).
Molecular Phylogenetics and Evolution 65: 174–182.
Turchetto-Zolet AC, Pinheiro F, Salgueiro F,
Palma-Silva F. 2013. Phylogeographical patterns shed
light on evolutionary process in South America. Molecular
Ecology 22: 1193–1213.
Van der Vaart AW. 1998. Asymptotic statistics. Cambridge:
Cambridge Univerity Press.
Van Oosterhout C, Hutchinson WF, Wills DPM, Shipley
P. 2004. Micro-Checker: software for identifying and correcting genotyping errors in microsatellite data. Molecular
Ecology Notes 4: 535–538.
Vanzolini PE. 1963. Problemas faunísticos do cerrado. In:
Ferri MG, ed. Simpósio sobre o Cerrado. São Paulo: University of São Paulo Press, 307–320.
Vegas-Vilarrubia T, Rull V, Montoya E, Safont E. 2011.
Quaternary palaeoecology and nature conservation:
a general review with examples from the neotropics. Quaternary Science Reviews 30: 2361–2388.
Vieira MV, Olifiers N, Delciellos AC, Antunes VZ,
Bernardo LR, Grelle CV, Cerqueira R. 2009. Land use
vs. fragment size and isolation as determinants of small
mammal composition and richness in Atlantic Forest remnants. Biological Conservation 142: 1191–1200.
Wahlund S. 1928. Zusammensetzung von population und
korrelationserscheinung vom standpunkt der vererbungslehre aus betrachtet. Hereditas 11: 65–106.
Warren DL, Glor RE, Turelli M. 2008. Environmental niche
equivalency versus conservatism: quantitative approaches
to niche evolution. Evolution 62: 2868–2883.
Wellenreuther M, Larson KW, Svensson EI. 2012. Climatic niche divergence or conservatism? Environmental
niches and range limits in ecologically similar damselflies.
Ecology 93: 1353–1366.
Wiens JJ, Donoghue MJ. 2004. Historical biogeography,
ecology and species richness. Trends in Ecology and Evolution 19: 639–644.
Wiens JJ, Graham CH. 2005. Niche conservatism: integrating revolution, ecology, and conservation biology. Annual
Review of Ecology and Systematics 36: 519–539.
Wolfe KH, Li W-H, Sharp PM. 1987. Rates of nucleotide
substitution vary greatly among plant mitochondrial, chloroplast and nuclear DNAs. Proceedings of the National
Academy of Sciences of the United States of America 84:
9054–9058.
Wooten JA, Gibbs HL. 2012. Niche divergence and lineage
diversification among closely related Sistrurus rattlesnakes.
Journal of Evolutionary Biology 25: 317–328.
SUPPORTING INFORMATION
Additional Supporting Information may be found in the online version of this article at the publisher’s web-site:
Figure S1. One specimen of Recordia reitzii in its natural habitat (Osório, Rio Grande do Sul, Brazil). Photo:
V. Thode.
© 2014 The Linnean Society of London, Botanical Journal of the Linnean Society, 2014, 176, 332–348
348
V. A. THODE ET AL.
Figure S2. Bayesian skyline plot showing the fluctuations in effective population size (Ne) over time (kBP,
thousands of years before present). The analysis was conducted with three cpDNA noncoding regions (trnHpsbA, trnS-trnG, and rpl32-trnL) sequences. Time estimates obtained using a relaxed molecular clock model
with the substitution rate previously estimated for chloroplast noncoding regions (1.0 × 10−9, Wolfe et al., 1987)
and a HKY (Hasegawa, Kishino, and Yano) nucleotide substitution model.
Table S1. Voucher information for the Recordia boliviana specimens sampled for cpDNA analyses.
Table S2. GenBank accession numbers for trnH-psbA, trnS-trnG, and rpl32-trnL sequences, respectively, used
in this study.
Table S3. Geographic coordinates of the Recordia reitzii (Rr) and R. boliviana (Rb) occurrence records used in
Ecological Niche Modeling.
Table S4. Correlation matrix for all pairs of the 19 climatic variables obtained from WorldClim (Hijmans et al.,
2005), calculated in ENMTools (Warren et al., 2008), with Pearson correlation coefficients.
Table S5. The nine climatic variables used for the Ecological Niche Modeling of Recordia reitzii and R.
boliviana.
Table S6. DIYABC analysis with SSR data. A = model check values and B = marginal posterior probability
densities for parameters of the bottleneck scenario.
© 2014 The Linnean Society of London, Botanical Journal of the Linnean Society, 2014, 176, 332–348