ArticlePDF Available

Late Pliocene population divergence and persistence despite Pleistocene climatic fluctuations in the Rio Doce snouted Treefrog (Ololygon carnevallii)

  • Unidad Ejecutora Lillo (CONICET - FML) Tucumán - Argentina


The causes of population genetic divergence within the Atlantic Forest of South America are diverse. For example, studies have pointed to the importance of regions of stable suitable habitat throughout the Pleistocene, large rivers acting as biogeographic barriers, as well as changes in elevation. Here, we generate a phylogeographic dataset for the Rio Doce Snouted Treefrog (Ololygon carnevallii), a species that is endemic to a narrow portion of the Atlantic Forest. Gene-tree analyses demonstrate that this species is composed of three distinct lineages that diverged from one another during the Pliocene. Ecological niche models projected to climate during the Pleistocene and mid-Holocene suggest that regions of suitable habitat would have shifted through time since the last glacial maximum, and regions of stable habitat were identified. Using generalized dissimilarity modeling, we find no association between genetic divergence and ecological niche models, riverine barriers, elevation, slope, or current climate suggesting that none of these variables have been responsible for lineage formation in the Rio Doce Snouted Treefrog. We suggest that additional phylogeographic studies of narrowly endemic species within the Atlantic Forest are needed to better understand the drivers of diversification and accumulation of biodiversity.
J Zool Syst Evol Res. 2021;00:1–11.
 1© 2021 Wiley‐VCH GmbH
Received: 8 June 2020 
Revised: 7 December 2020 
Accepted: 15 Decem ber 2020
DOI: 10.1111/jzs.12454
Late Pliocene population divergence and persistence despite
Pleistocene climatic fluctuations in the Rio Doce snouted
Treefrog (Ololygon carnevallii)
Edward A. Myers1| Henrique Folly2,3| Eric Ragalzi3| Renato Neves Feio2|
Diego José Santana3
1Depar tment of Vertebr ate Zoology,
Nationa l Museum of Natural History,
Smithsonian Institution, Washington, DC,
2Depar tamento de Biologia Animal, Museu
de Zoologia João Moojen, Univerisdade
Federal de Viços a, Viçosa, MG , Brazil
3Instituto de Biociências, Lab oratório de
Zoologia, Universidad e Federal de Mato
Grosso do Sul, Campo Gr ande, MS, Brazil
Edward A . Myers , Department of
Vertebr ate Zoology, National Museum of
Natural Histor y, Smithsonian Institution,
Washington, DC , USA.
The causes of population genetic divergence within the Atlantic Forest of South
America are diverse. For example, studies have pointed to the importance of regions
of stable suitable habitat throughout the Pleistocene, large rivers acting as biogeo-
graphic barriers, as well as changes in elevation. Here, we generate a phylogeographic
dataset for the Rio Doce Snouted Treefrog (Ololygon carnevallii), a species that is en-
demic to a narrow portion of the Atlantic Forest. Gene- tree analyses demonstrate
that this species is composed of three distinct lineages that diverged from one an-
other during the Pliocene. Ecological niche models projected to climate during the
Pleistocene and mid- Holocene suggest that regions of suitable habitat would have
shifted through time since the last glacial maximum, and regions of stable habitat were
identified. Using generalized dissimilarity modeling, we find no association between
genetic divergence and ecological niche models, riverine barriers, elevation, slope, or
current climate suggesting that none of these variables have been responsible for lin-
eage formation in the Rio Doce Snouted Treefrog. We suggest that additional phylo-
geographic studies of narrowly endemic species within the Atlantic Forest are needed
to better understand the drivers of diversification and accumulation of biodiversity.
Atlantic Forest, DNA barcode, phylogeography, Pliocene divergence
A divergência genética populacional de vários organismos na Mata Atlântica é resul-
tado de diversas causas. Por exemplo, estudos têm apontado para a importância de
regiões com habitat adequado estável ao longo do Pleistoceno, grandes rios atuando
como barreiras biogeográficas, bem como mudanças na topografia. Neste trabalho
nós geramos um conjunto de dados filogeográficos para a Perereca- de- Inverno- do-
Rio- Doce (Ololygon carnevallii), uma espécie que é endêmica de uma porção estreita
da Mata Atlântica. Análises de árvores de gene demonstram que esta espécie é com-
posta por três linhagens distintas que divergiram durante o Plioceno. Os modelos de
nicho ecológico projetados no Pleistoceno e Holoceno médio sugerem que as regiões
de habitat adequado teriam mudado ao longo do tempo desde o último máximo glacial,
   MYERS E t al.
Determining the factors that have promoted population genetic
divergence is important for understanding the process of specia-
tion and the accumulation of biodiversity. It is well documented
that genetic divergence accumulates between allopatric popula-
tions across a biogeographic barrier (Wallace, 1854). This can be
for example rivers that bisects the distribution of a species (e.g.,
the Amazon river; Hayes & Sewlal, 2004) or can be stretches of
unsuitable habitat (e.g., rainforest bisected by savannas; Lorenzen
et al., 2012). Changes in climate and associated changes in hab-
itat throughout the Pleistocene have also been responsible for
promoting allopatric divergence on a global scale (Hewitt, 2000).
Alternatively, genetic divergence can occur because of environ-
mental differences with local adaptation and reduced gene flow
between regions (Wang & Bradburd, 2014). This can occur across
ecotones or because of heterogeneity in climate across the dis-
tribution of a species (Cooke et al., 2012; Ribeiro et al., 2016).
Genetic differentiation across the distribution of a species can
also result from limited dispersal resulting in a pattern of isolation-
by- distance, where genetic distance is correlated with geographic
distances between populations (Wright, 1943). These factors can
work together to influence genetic divergence across the distribu-
tion of a species (Mitchell et al., 2015) and different factors may
be responsible for influencing divergence in codistributed species
(Myers et al., 2019).
The Atlantic Forest (AF) is a biodiversity hotspot of great con-
servation concern (Myers et al., 2000), and many studies have
sought to understand the high levels of endemic diversity. An
overview of these studies suggests that diversification within the
AF has been complex and that it is unlikely that any two codis-
tributed species will share the same evolutionary history (Carnaval
et al., 2014; Turchetto- Zolet et al., 2012). Refugia of stable forest
through time have been hypothesized to be important in driving
allopatric divergence in many taxa (Batalha- Filho & Miyaki, 2016;
Carnaval & Moritz, 2008). However, the timing of divergence in
some species predates the Pleistocene glacial periods suggesting
that climate cycles may not have been the sole factor driving diver-
sification (Paz et al., 2019; Thomé et al., 2010). It has also been sug-
gested that population expansion on to the Brazilian continental
shelf during periods of lower sea levels has been important in
structuring genetic diversity as opposed to the hypothesis that
mesic forests in the southern AF were replaced by drier vegeta-
tion (Leite et al., 2016). Numerous studies have also suggested that
biogeographic barriers, for example, large rivers within the AF,
have been responsible for driving population divergence and spe-
ciation (Amaro et al., 2012; Thomé et al., 2010), yet other studies
show little support for vicariance across riverine barriers (Colombi
et al., 2010). It is also possible that multiple factors are interact-
ing to structure genetic diversity in this biome (Sotelo- Muñoz
et al., 2020). For example, elevation and historical climate change
have influenced population genetic structure in codistributed birds
(Thom et al., 2020) and forest refugia and rivers interact in pro-
moting divergence in numerous taxa (Mascarenhas et al., 2019;
Menezes et al., 2016). Therefore, to understand diversification
within the AF multiple mechanisms need to be considered (Brunes
et al., 2015).
The AF can be subdivided into two distinct bioclimatic domains
with unique assemblages of plants and animals that are separated by
the Rio Doce (Carnaval et al., 2014). Regions north of the river are
warmer and the biotic communities are characterized by widespread
lowland species with affinities to eastern Amazonia (Batalha- Filho
et al., 2013). The southern portion of the AF is more seasonal in cli-
mate with more montane and subtropical taxa that are more shared
with Andean lineages (Batalha- Filho et al., 2013; Carnaval et al., 2014).
Phylogeographic patterns between these two forest blocks have also
been shown to differ where lineages have differential responses to
climatic cycling between these domains (Paz et al., 2019).
Th e Rio Do ce Sn out ed Tre efr og (Ololygon carnevallii Caramaschi
and Kisteumacher, 1989) is a small (<3.2 cm) frog that is nar-
rowly endemic within the Bahia refuge of central AF (Carnaval &
Moritz, 2008) at the turnover of the two distinct bioclimatic do-
mains and its distribution is bisected by the Doce River. Given
that this species is distributed across environmental turnover and
across a well- documented riverine barrier, it may be expected to
show population structure with these variables. Here, we generate
barcode sequence data sampled from across the known distribu-
tion of O. carnevallii to assess population genetic structure, date
the timing of divergence, and assess what factors have been re-
sponsible for population genetic divergence.
e então as regiões de habitat estável foram identificadas. Usando uma modelagem de
dissimilaridade generalizada não encontramos associação entre divergência genética
e modelos de nicho ecológico, rios como barreiras, elevação, declive ou clima atual,
sugerindo que nenhuma dessas variáveis foi responsável pela formação das linhagens
dessa espécie. Sugerimos que estudos filogeográficos adicionais de espécies estreita-
mente endêmicas dentro da Mata Atlântica são necessários para melhor compreender
as causas da diversificação e acúmulo de biodiversidade nessa ecorregião.
MYERS E t al.
2.1  |  Sample collection
We obtained a total of 30 tissue samples of Ololygon carnevallii from
12 localities to generate molecular data, by collecting individuals in
the field from August 2015 to December 2016 (Figure 1, Table 1).
Specimens were euthanized using 5% lidocaine chlorhydrate, fixed
with 10% formalin, and transferred for permanent storage to 70%
ethanol. Tissue samples (muscle or liver) were stored in 100% etha-
nol before specimen fixation in formalin (permits issued SISBIO
52251). We also verified 34 collection localities of this taxon by ex-
amining 429 specimens housed at several natur al histor y collections
for ecological niche modeling (see Table S1).
2.2  |  Laboratory procedures
Total DNA from the samples was extracted using the Blood & Tissue
DNA Mini Kit (Ludwig Biotec, Alvorada, Brazil) from a small piece
of muscle or liver tissue preserved in ethanol according to standard
DNA barcoding methods for anurans (Koroiva et al., 2020). A total
of 658 bp were amplified from the 5′ region of the mitochondrial
cytochrome c oxidase subunit 1 (COI) gene using the T3- AnF1 (5-
ADA RRT GTT G- 3′) primers (Lyra et al., 2017). PCR conditions for
amplification consisted of 1× buffer (Colorless GoTaq® Flexi Buffer;
Promega Corp., Madison, WI), 0.2 mM dNTP mix, 0.2 µM of each
primer, 2 mM MgCl2, 1U Taq polymerase (GoTaq® G2 hot st art poly-
merase, Promega Corp., Madison, WI), and 2 µl of template DNA, in
a tot al reaction volume of 25 µl. The PCR cycling program was run
as follows: initial denaturation step with 3 min at 95 °C, 35 cycles of
denaturation for 20 s at 95 °C, annealing for 20 s at 50 °C and ex ten-
sion for 1 min at 60 °C, and final extension for 5 min at 60 °C (see
Lyra et al., 2017). PCR products were purified with Ethanol/Sodium
Acetate and sequenced in both directions on an ABI 3130 Genetic
Analyzer (Applied Biosystems). These sequences were aligned in
Geneious v 9.0.5 (Biomatters Ltd.) using Muscle v3.11 (Edgar, 2004).
The number of variable sites was calculated using strataG (Archer
et al., 2017) in R, and the number of haplotypes was calculated with
haplotypes in R (Aktas, 2015).
FIGURE 1 The known geographic distribution of Ololygon carnevallii is shown in gray circles, all genetic samples used in this study are
illustrated by colored circles. All major rivers within southeastern Brazil are illustrated. Abbreviations are as follows, MG— Minas Gerais, BA—
Bahia, ES— Espírito Santo, R J— Rio de Janeiro. Inset is a picture of a live O. carnevallii (photo taken by P.S. Hote)
   MYERS E t al.
2.3  |  Phylogenetic analyses
We generated a Bayesian phylogeny of all of the O. carnevallii
sampled individuals to asses phylogeographic structure across its
dis tribu ti on usi ng BE AST v 2.6 (Bo uckaert et al., 2014). The most ap-
propriate substitution model for COI was inferred using jModelTest2
v 2.1.10 (Darriba et al., 2012). A COI sequence for Ololygon humilis
was downloaded from GenBank (accession number KU234705) and
used as an outgroup to root the gene tree (Faivovich et al., 2005).
We used a Yule speciation prior, implemented a strict clock rate of
2% per Myr (Crawford, 2003), and ran the analysis for 2 million gen-
erations sampling ever y 2,000 generations. Tracer v 1.7.1 (Rambaut
et al., 2018) was used to assess effective sample sizes of estimated
parameters and stationarit y, ensuring that all effective samples sizes
we re >200.
We implemented the Bayesian version of the general mixed
yule- coalescent model (bGMYC; Pons et al., 2006; Reid &
Carstens, 2012), to objectively define populations given the pos-
terior distribution of gene trees. We subsampled 100 trees from
the posterior distribution of the BEAST analysis, after a burn- in
of 10% using the R package evobiR (Blackmon & Adams, 2015).
bGM YC was run for 50,0 00 gen eratio ns, with a 40,0 00 generation
burn- in, and a thinning interval of 100. The threshold parameters
were set to t1 = 2 and t2 = 100. To generate a point estimate of
the number of distinct lineages from the bG MYC analysis, we used
a posterior probability cutoff of 0.05.
TABLE 1 Collecting localities, GenBank accession numbers for COI barcode sequence data, the corresponding lineage for each individual
sample sequenced, and the voucher accession information for all samples included in molecular analyses
Locality Latitude Longitude
GenBank accession
Braúnas −19.134111 −4 2. 7475 83 MT571617 2MZUF V16555
Caratinga/FMA −1 9.72 2 674 41.8 061 27 MT571611 3MZUF V17195
Caratinga/FMA −1 9.72 2 674 41.8 061 27 MT571612 3MZU FV17198
Caratinga/FMA −1 9.72 2 674 41.8 061 27 MT571613 3M ZUFV17196
Caratinga/PCH A. Branca −19. 61 0900 −41.8 03900 MT571629 3MAP- T425
Caratinga/UNEC −19.722786 −41. 80 610 4 MT571630 3UFMG- A17150
Caratinga/UNEC −19.722786 −41. 80 610 4 MT571631 3UFMG- A17151
Cataguases −21.215544 42 .7 56769 MT571618 3MZUFV16585
Cataguases −21.215544 42 .7 56769 MT571619 3MZUFV16586
Cataguases −21.215544 42 .7 56769 MT571620 3MZUFV16584
Cataguases −21.215544 42 .7 56769 MT5716 38 3C T2574
Dores de Guanhães −19. 0 6 0 838 −42 .924610 MT571634 1U F M G - A 1 7 2 7 0
Dores de Guanhães −19. 0 6 0 838 −42 .924610 MT571635 1U F M G - A 1 7 2 7 1
Leme do Prado −17. 0 8 33 0 0 −42.692500 MT571632 2UFMG- A1414 6
Leme do Prado −17. 0 8 33 0 0 −42.692500 MT571633 2UFMG - A14147
Marliéria −19.772821 −42.622016 MT571621 3MZUF V1716 3
Marliéria −19.772821 −42.622016 MT571622 3MZUF V17165
Marliéria −19.772821 −42.622016 MT571623 3MZU F V17151
Marliéria −19.772821 −42.622016 MT571624 3MZ UF V17171
Mesquita −19. 2 59 6 35 −42 .5541 23 M T571614 1MZUF V17160
Mesquita −19. 2 59 6 35 −42 .5541 23 M T571615 1MZUF V17159
Mesquita −19. 2 59 6 35 −42 .5541 23 MT571616 1MZU FV17169
Mesquita −19. 2 59 6 35 −42 .5541 23 M T5 760 18 3MZUF V17169
Muriaé −21 .01 3611 −42.446667 MT571637 3CT2571
Santo A. do Grama −20 . 28 762 5 42 .57417 8 MT571625 3MAP1312
Sapucaia de Guanhães −18.918369 −42 .61756 0 MT571626 2MAP1297
Sapucaia de Guanhães −18.918369 −42 .61756 0 MT571627 2MAP1298
Sapucaia de Guanhães −18.918369 −42 .61756 0 MT571628 2MAP1299
Sapucaia de Guanhães −18.918369 −42 .61756 0 MT576 017 2MAP1299
Virginópolis −18.822224 −42.705019 MT571636 1U F M G - A 1 7 2 7 9
Abbreviations: CT, Cytogenetics Lab Tissue Collection in the Beagle Lab, Universidade Federal de Viçosa; MAP, Mapinguari Lab from Universidade
Federal de Mato Grosso do Sul; MZUFV, Museu de Zoologia João Moojen, Universidade Federal de Viçosa; UFMG, Zoological Collec ation of the
Universidade Federal de Minas Gerais.
MYERS E t al.
2.4  |  Current and hindcast ecological niche models
To gen er a t e ecological niche mo de l s (EN M), we ge o r e f e r e n c e d all unique
collection localities for voucher specimens of O. carnevallii (a total of 34
localities). These voucher specimens were verified by examining speci-
mens at the following institutions: Museu de Zoologia João Moojen,
Universidade Federal de Viçosa (MZUF V), amphibian collection of
Universidade Federal de Minas Gerais (UFMGA), Museu de Ciências
Naturais, Pontifícia Universidade Católica de Minas Gerais (MCNAM),
and Coleção Zoológica da Universidade Federal de Mato Grosso do Sul
(ZUFMS- AMP) (Table S1). Locality records were spatially thinned so
that no localities were within 10 km of each other using the R package
spT hin (Aiello- Lammens et al., 2015). We downloaded the Worldclim
v1.4 (Hijmans et al., 20 05) 19 climate variables at 2.5 arcmin resolu-
tion. ENMtools (Warren et al., 2017) was used to test for correlations
among these variables and all correlated variables but one was removed
where Pearson’s correlations >0.7. This resulted in a total of six biocli-
matic variables: mean annual temperature, mean diurnal range, isother-
mality, temperature seasonality, annual precipitation, and precipitation
of driest month. To identify optimal model parameters to use in our
ENM, we tested all combinations of feature classes with regularization
multipliers between 0.5 and 4 at 0.5 intervals in ENMeval (Muscarella
et al., 2014). Ecological niche models were constructed using Biomod2
(Thuiller et al., 2013), using the best fit model parameters. We sampled
10,000 pseudoabsence points in a region surrounding the geographic
distribution of O. carnevallii (circ umscrib ed within: - 49, - 36 , - 24, - 12) an d
Maxent v3.4.1 (Phillips et al., 20 06) was used to construct ENMs using
the six uncorrelated bioclim variables. Maxent was run with 25 evalua-
tion run s, repl ic ate d for 5,000 iter at ions, and 25% of sample s we re used
as a training data set for model evaluation. We then produced an aver-
age of these ENMs runs and projected this model to both last glacial
maximum (LGM; 21 kya) and Mid- Holocene (6 kya) climate conditions
using the community climate system model general circulation model
(CCSM4; Gent et al., 2011) with the same six non- correlated bioclim
variables. Regions of habitat stability through time were identified by
stacking and averaging the current and two projected- paleo climate
ENMs. Regions highlighted in these stacked projects were inferred to
be regions of climate refugia through time for O. carnevallii. All ENMs
were normalized between 0 and 1 and exported as ASCII files.
2.5  |  Determinates of genetic structure
We sought to test what factors have been responsible for promoting
population genetic differentiation within O. carnevallii and focused on
the following variables: (1) geographic distance, (2) climate variables,
(3) the predicted ecological niche, (4) Pleistocene refugia during the
LGM, (5) mid- Holocene distribution, (6) habitat stability through time,
(7) rivers as biogeographic barriers, (8) rivers promoting gene flow,
(9) elevation, and (10) slope isolating populations. Each of these vari-
ables have been suggested to be responsible for population genetic
divergence and may be acting together to reduce gene flow between
populations (e.g., Myers et al., 2019; García- Rodríguez et al., 2020).
We downloaded shapefiles for rivers from the HydroSHEDS proj-
ect (Lehner et al., 2008) and elevation (Jarvis et al., 2008) to examine
how these features of the landscape have shaped genetic diversity.
Slope was derived from elevation using ras ter in R (Hijmans & van
Etten, 2012). These shapefiles were imported into R, converted to
ASCII format using raster (Hijmans & van Etten, 2012), and values
were normalized between 0 and 1. Here, higher values represent
higher costs to dispersal across the landscape. For example, when
rivers were tested as biogeographic barriers the rivers were set to
1, whereas when rives were considered to act as conductance they
had lower values compared with the rest of the landscape. Higher
elevations and steeper slop es were set to higher values . To assess th e
effect of geographic distances, we created an ASCII file with equal
values in each cell. We then used Circuitscape v0.1.0 (Anantharaman
et al., 2019; McRae et al., 2008) implemented in Julia (Bezanson
et al., 2017) to generate distance matrices for our sampled localities
given all of these potential factors (e.g., ENMs, rivers, elevation, and
slope). Current ENMs, LGM projected ENMs, stability ENMs, and riv-
ers were used as conductance surfaces in Circuitscape, while rivers,
elevation, and slop e were used as resistance surfaces. We used rivers
as both conductance and resistance because large rivers have been
proposed to be barriers to gene flow within the AF; however, given
that anurans have an aquatic larval stage, it is possible that rivers may
facilitate gene flow. Resistance and conductance distances were cal-
culated as pairwise in Circuitscape. The climate variables used were
th e same si x non- c orr ela ted va ria ble s used in the EN Ms an d her e wer e
used as the values extracted from each specimen collecting locality.
Using generalized dissimilarity modeling (GDM; Ferrier
et al., 2007), we tested how environmental variation between col-
lection localities and models of resistance or conductance matrices
have contributed to genetic differentiation. This approach is an ex-
tension of matrix regression and can accommodate non- linear rela-
tionships between variables (Ferrier et al., 20 07). We used hierfstat
(Goudet, 2005) in R to generate a matrix of Fst values between all
sampled geographic localities. This Fst matrix was then used as
the response variable, and the gdm R package (Manion et al., 2016)
was used to fit generalized dissimilarit y models. This allowed us to
test for correlations between genetic distances, the seven distance
matrices based on ENMs and geography, in addition to testing for
correlations with genetic distance and geographic distances (these
distances were estimated with the gdm func tion where ‘geo’ was set
to true) and current climate variables. We then used the gdm.varImp
function in the gdm R package to perform model and variable sig-
nificance testing and to estimate variable importance in our GDM.
3.1  |  DNA sequence data and phylogenetic analyses
COI sequ en ce data were genera te d for 30 O. carnevallii specimens
(Table 1) for a total of 551 base pairs when trimmed to the short-
est sequence, with 187 variable sites, and 23 unique haplotypes.
   MYERS E t al.
These sequence data were uploaded to GenBank (Table 1), and
the alignment is available in fasta format in the (Alignment S1).
The best fit model of sequence evolution inferred from jMod-
eltest2 was HKY+G. All effective sample size values from the
BEAST gene- tree analysis were >200, suggesting stationarity.
The timing of gene divergence within O. carnevallii began at ~3.5
mya (HPD: 2.6– 4.2 mya; Figure 2). bGMYC analyses suggested
that there are three distinct mtDNA lineages (Figure 2). These
lineages are united with strong posterior probabilit y; however,
the relationships between them is unresolved with low posterior
probability (Figure 2). Lineages 1 and 2 are distributed north of
the Rio Doce and largely in sympatry, while lineage 3 is south of
the Rio Doce with the exception of specimens from the type lo-
cality (Figure 1). Within lineage genetic variation is variable, with
10 haplotypes in lineage 3, 5 haplot ypes in lineage 1, and 4 hap-
lotypes in lineage 2.
3.2  |  Ecological niche models
After thinning specimen locality data, we retained a total of 37
localities for ENM. The best fit feature class was hinge with a
regularization multiplier of 1; however, the next best fit model
differed by a deltaAIC of only 0.22 which was LQHP with a regu-
la riz ati o n mul t ipl ier of 1 (the AI C score of th e next bes t model wa s
>8; Table S2). We therefore ran two separate ENMs using both of
these model settings. However, because the model projections
were similar, we only discuss the Hinge 1 model (see Figure S1
for LQHP projections). The current ENMs performed well with
AUC values of 0.91 and predicted the known geographic distri-
bution of O. carnevallii. However, the models did predict suitable
habitat to the southwest in the state of Minas Gerias and north-
ern Sao Paulo where this taxon has not been collected (Figure 3).
Both hindcast projections suggest shifts in suitable habitats for
O. carnevallii. For example, at 6 kya the potential suitable habitat
is much more expansive but with a shift in distribution to the in-
terior of Brazil. While at 21 kya, the projected suitable habitat is
shifted largely to the east of the current distribution (Figure 3).
The stability map generated with all three projections suggests
regions of suitable habitat in central and western Minas Gerais
(Figure 3).
3.3  |  Generalized dissimilarity modeling
The GDM with seven predictor variable matrices, current climate
conditions, and geographic distance explained 25.8% of the ge-
netic variation as measured by Fst. However, this model was not
significant with a p- value of 0.12. The most impor tant predictor
variables were geographic distance and elevation; however, after
FIGURE 2 Phylogeny and divergence times estimated in Beast and results from bGMYC analysis. Tip labels correspond to samples listed
in Table 1 and colored bars correspond to collecting localities in Figure 1. The colored bGMYC table corresponds to the posterior probability
that the corresponding sequences are conspecific and illustrates any uncertainty in inferred population limits
MYERS E t al.
permutation tests these were not significant (p- value > 0.14; for
relationships between FST and predictor variables see Figure S2).
The Rio Doce treefrog underwent diversification during the late
Pliocene (~3.5 mya) forming three distinct mtDNA lineages. These
lineages have remained distinct through multiple glacial cycles
throughout the Pleistocene, irrespective of suitable habitats.
Ecological niche models projected to the LGM and the mid- Holocene
demonstrate that the potential suitable habitat for this taxon would
have shifted through time, from east to west. However, given the
timing of lineage formation, this was not a primar y factor influenc-
ing divergence. While Pleistocene refugia have been important in
promoting diversification in many AF taxa (Carnaval et al., 2009;
De Mello Martins, 2011), many species have population divergence
that predates this period. In more widespread species, popula-
tion divergence into northern and southern lineages that diverged
in the Pliocene has been demonstrated (Grazziotin et al., 2006;
Thomé et al., 2010). Similarly, in these groups, hindcast ENMs have
demonstrated reduced availability of suitable habitat during the
Pleistocene but with limited congruence with population genetic
structure (Thomé et al., 2010; Tonini et al., 2013). It is possible that
FIGURE 3 Ecological niche models for Ololygon carnevallii. (a) ENM under current climate conditions, (b) ENM projected on to mid-
Holocene climate (6 kya), (c) ENM projected on to LGM climate conditions (21 kya), (d) Regions of climate suitability through time based on
all three projections. In all projections, the distribution of Ololygon carnevallii is shown as hollow circles
   MYERS E t al.
these earlier diverge nce ti me s ar e du e to Pliocene orogeny and hab i-
tat fragmentation with population divergence re- enforced by small
scale habitat fragmentation during late Pleistocene climatic oscilla-
tions (Grazziotin et al., 2006; Thomé et al., 2010).
The timing of lineage formation with O. carnevallii further sup-
ports discordant diversification mechanisms in the AF. Because
population and species divergence in the AF has been dated from
the Pliocene to the Pleistocene, the mechanisms generating the
species richness in this biome have been diverse. These seemingly
constant rates of lineage divergence and accumulation in the AF
biome are similar to patterns seen across the Neotropics, where the
timing of divergence has been continuous throughout the Tertiary
and Quaternary (Rull, 20 08). These results from the AF also further
support the notion that the processes that have given rise to the
biodiversity of the Neotropics are complex (Costa, 2003; Turchetto-
Zolet et al., 2013). These insights only point to the need for addi-
tional phylogeographic studies of AF endemics to better understand
the complex diversification mechanisms of this biome.
Most phylogeographic studies within the AF have focused on
the patterns and processes of diversification in widespread spe-
cies (Carnaval et al., 20 09; Paz et al., 2019; Thomé et al., 2010). The
Rio Doce treefrog is narrowly distributed approximately within the
Bahia refugium of the Atlantic Forest identified in Carnaval and
Moritz (2008). Because the distribution of O. carnevallii is restricted
to the Bahia refugium, it may be expected that there is no signature
of Pleistocene climate fluctuations driving population divergence.
Instead, we find deep divergence within this range restricted species
that dates to the Pliocene. Because phylogeographic patterns are
often species specific, associated with life- history traits and ecol-
ogy (Papadopoulou and Knowles, 2016), additional phylogeographic
studies of taxa with distributions restricted to these Pleistocene
forest refugia are needed. This would ultimately clarify if shared his-
torical processes versus species- specific responses have influenced
diversification within a region of climatic stability. Furthermore, such
st ud ies on na rro w en dem ics wil l she d li ght on wh at has be en resp on-
sible for generating endemicity and influencing patterns of genetic
variation at fine spatial scales. This may also highlight new regions
of conservation concern that have been overlooked by broad- scale
studies of widely distributed taxa.
The lineages we identify here began to diversify prior to the
onset of global climatic fluctuations of the Pleistocene. However,
irrespective of the timing of divergence among these clades, we ex-
plicitly tested for associations between patterns of genetic diver-
gence and environmental suitability through time (Paz et al., 2019).
ENMs do show reduced suitable habitat during the LGM and his-
torically stable regions. However, there is no association between
ENMs, whether current day, hindcast models or habitat stability
through time. Furthermore, there was no association with other
commonly cited variables that could be driving diversification (e.g.,
elevation, current climate variables, or rivers). While this suggests
that these variables have not been important in driving genetic di-
vergence within this taxon, it is also possible that there is not enough
signal in the barcode sequence data that was generated here.
The Rio Doce is often cited as an important barrier to gene
flow (Ribeiro et al., 2011; Thomé et al., 2010); however, this river
has not been important in structuring population genetic diver-
gence in O. carnevallii. This is demons trated in both the ge og raphi c
distribution of mtDNA lineages and results from generalized dis-
similarity models. Therefore, it is possible that earlier tectonic up-
lift of montane regions within the Atlantic Forest were important
for initial separation of lineages (Mello et al., 1999; Riccomini &
Assumpção, 1999). Furthermore, because lineages 1 and 2 seem
to co- occur nor th of the Rio Doce it is also possible that these
lineages have diverged in sympatry or parapatry as the result of
niche divergence. Future natural history studies focused on the
differences in calls, timing of breeding, or habitat use could help to
clarify divergence in niche between these two lineages. These lin-
eages may have also diverged in the climatic niche that is modeled
in ENMs. If this is the case, then treating all collecting localities
as a single unit in the ENM could lead to over- projection of the
model (e.g., Pearson et al., 2007). The future collection of spec-
imens, and sequence data to assign these specimens to lineages,
from new localities may allow for expanded tests of niche diver-
gence. However, given the narrow distribution of this taxon new
collecting localities may not lead to increased resolution in ENMs.
A potential caveat here is that the phylogeographic patterns found
in organelle versus nuclear genomes may not always be the same
across a potential barrier (Toews & Brelsford, 2012). This high-
li ght s the nee d to gene rat e a ddi t ion al ge nomi c data to di s crim i nat e
among opposing models of population divergence. Future studies
incorporating genomic data (e.g., RADseq or sequence capture;
Davey et al., 2011; Faircloth et al., 2012) with newer models of
past climate (Brown et al., 2018) will be particularly impor tant in
understanding these processes.
EAM was funded by the Peter Buck and Rathbone Bacon
Fellowship through the Smithsonian’s National Museum of Natural
History. HF and ER thank Coordenação de Aperfeiçoamento de
Pe sso al de Ní vel Su per i or— B ras il (C APES) Fin anc e Code 00 1. DJ S
thanks CNPq (Conselho Nacional de Desenvolvimento Científico
e Tecnológico) for his research fellowship (311492/2017- 7). We
thank Clodoald o L Assis and Pris cila Hote for their assistance with
fieldwork and Paulo C . A. Garcia for providing Ololygon carneval-
lii tissues. We thank P. Hote for the photograph of these species
in life.
The genetic data that support the findings of this study are openly
available in GenBank, accession numbers MT571611– MT576018.
An alignment of these sequences is also available in the supporting
Edward A. Myers
Diego José Santana
MYERS E t al.
Aiello- Lammens, M. E., Boria, R. A ., Radosavljevic, A., Vilela, B., &
Anderson, R. P. (2015). SpThin: An R Package for spatial thinning
of species occurrence records for use in ecological niche models.
Ecography, 38(5), 541– 545.
Aktas, C. (2015). Haplotypes: Haplotype Inference and Statistical Analysis
of Genetic Variation. R Package Version, 1.
Amaro, R. C., Rodrigues, M. T., Yonenaga- Yassuda, Y., & Carnaval, A. C.
(2012). Demographic processes in the montane Atlantic rainfor-
est: molecular and cytogenetic evidence from the endemic frog
Proceratophrys boiei. Molecular Phylogenetics and Evolution, 62(3),
880– 888.
Anantharaman, R., Hall, K ., Shah, V., & Edelman, A. (2019). Circuit scape
in Julia: High Performance Connec tivity Modelling to Support
Conservation Decisions. A rXiv Pre print A rXiv:1906 .035 42.
Archer, F. I., Adams, P. E., & Schneiders, B. B. (2017). Stratag: An r pack-
age for manipulating, summarizing and analysing population ge-
netic dat a. Molecular Ecolog y Resources, 17(1), 5– 11.
Batalha- Filho, H., Fjeldså, J., Fabre, P.- H., & Miyaki, C. Y. (2013).
Connections between the Atlantic and the Amazonian Forest
Avifaunas Represent Distinct Historical Events. Journal of
Ornithology, 154(1), 41– 50.
Batalha- Filho, H., & Miyaki, C. Y. (2016). L ate Pleistocene diver-
gence and postglacial expansion in the Brazilian Atlantic
Forest: Multilocus Phylogeography of Rhopias Gularis (Aves:
Passeriformes). Journal of Zoological Systematic s and Evolutionary
Research, 54(2), 137– 147.
Bezanson, J., Edelman, A., Karpinski, S., & Shah, V. B. (2017). Julia: A fresh
approach to numerical computing. SIAM Review, 59(1), 65– 98.
Blackmon, H., & Adams, R. H. (2015). EvobiR: Comparative and Population
Genetic Analyses. R package version 1.1. ht tps://CRAN.R- proje ge=evobiR
Bouckaert, R., Heled, J., Kühnert, D., Vaughan, T., Wu, C.- H., Xie, D.,
Suchard, M. A., Rambaut, A., & Drummond, A. J. (2014). BEA ST
2: A software platform for Bayesian evolutionary analysis. PLoS
Computational Biology, 10(4), e10 03537.
Brown, J. L ., Hill, D. J., Dolan, A . M., Carnaval, A. C., & Haywood, A . M.
(2018). PaleoClim, high spatial resolution paleoclimate surfaces for
global land areas. Scientific Data, 5(1), 1– 9.
Brunes, T. O., Maria, T. C ., Thomé, J. A., Haddad, C. F. B., & Sequeira,
F. (2015). Ancient divergence and recent population expansion
in a leaf frog endemic to the southern Brazilian Atlantic forest .
Organisms Diversity & Evolution, 15(4), 695– 710.
Carnaval, A. C., Hickerson, M. J., Haddad, C. F. B., Rodrigues, M. T., &
Moritz, C. (2009). Stability predic ts genetic diversity in the Brazilian
Atlantic forest hotspot . Science, 323(5915), 785– 789.
Carnaval, A. C., & Moritz, C. (2008). Historical climate modelling pre-
dicts patterns of current biodiversit y in the Brazilian Atlantic forest.
Journal of Biogeography, 35(7), 1187– 1201.
Carnaval, A. C., Walt ari, E., Rodrigues, M. T., Rosauer, D., VanDerWal, J.,
Damasceno, R., Prates, I., Strangas, M., Spanos, Z., Rivera, D., Pie,
M. R., Firkowski, C. R., Bornschein, M. R., Ribeiro, L. F., & Moritz, C.
(2014). Prediction of phylogeographic endemism in an environmen-
tally complex biome. Proceedings of the Royal Society B: Biological
Sciences, 281(1792), 20141461.
Colombi, V. H., Lopes, S. R ., & Fagundes, V. (2010). Testing the Rio Doce
as a riverine barrier in shaping the Atlantic rainforest population
divergence in the Rodent Akodon Cursor. Genetics and Molecular
Biology, 33(4), 785– 789.
Cooke, G. M., Chao, N. L., & Beheregaray, L. B. (2012). Divergent natural
selection with gene flow along major environmental gradients in
Amazonia: Insights from genome scans, population genetics and
phylogeography of the characin fish Triportheus albus. Molecular
Ecology, 21(10), 2410– 2427.
Costa, L . P. (20 03). The Historical Bridge between the Amazon and the
Atlantic Forest of Brazil: A study of molecular phylogeography with
small mammals. Journal of Biogeography, 30(1), 7186.
Crawford, A. J. (2003). Relative rates of nucleotide substitution in frogs.
Journal of Molecular Evolution, 57 (6), 636– 641.
Darriba, D., Taboada, G. L., Doallo, R., & Posada, D. (2012). JModelTest
2: More models, new heuristics and parallel computing. Nature
Methods, 9(8), 772.
Davey, J. W., Hohenlohe, P. A., Etter, P. D., Boone, J. Q., Catchen, J. M.,
& Blaxter, M. L. (2011). Genome- wide genetic marker discovery
and genotyping using next- generation sequencing. Nature Reviews
Genetics, 12(7), 499– 510.
De mello martins, F. (2011). Historical biogeography of the Brazilian
Atlantic Forest and the Carnaval- Moritz Model of Pleistocene
Refugia: What Do Phylogeographical Studies Tell Us? Biological
Journal of the Linnean Society, 104(3), 499– 509.
Edgar, R. C. (2004). MUSCLE: multiple sequence alignment with high
accuracy and high throughput. Nucleic Acids Research, 32(5),
1792– 1797.
Faircloth, B. C ., McCormack, J. E., Craw ford, N. G., Har vey, M. G .,
Brumf ie ld, R. T., & Glenn, T. C. (2012). Ultraco nser ve d element s an-
chor thousands of genetic markers spanning multiple evolutionary
timescales. Systematic Biology, 61(5), 717– 726.
Faivovich, J., Haddad, C. F. B., Garcia, P. C. A ., Frost, D. R., Campbell, J.
A., & Wheeler, W. C. (2005). Systematic review of the frog fam-
ily hylidae, with special reference to hylinae: phylogenetic analysis
and taxonomic revision. Bulletin of the American Museum of Natural
Histor y, 2005(294), 1– 240.
Ferrier, S., Manion, G., Elith, J., & Richardson, K. (2007). Using gener-
alized dissimilarity modelling to analyse and predict patterns of
beta diversity in regional biodiversity assessment. Diversity and
Distributions, 13(3), 252– 264.
García- Rodríguez, A., Guarnizo, C. E., Crawford, A. J., Garda, A. A., &
Costa, G. C. (2020). Idiosyncratic responses to drivers of genetic
differentiation in the complex landscapes of Isthmian Central
America. Heredity, 1– 15.
Gent, P. R., Danabasoglu, G., Donner, L. J., Holland, M . M., Hunke, E. C.,
Jayne, S. R., Lawrence, D. M., Neale, R. B., Rasch, P. J., & Vertenstein,
M. (2011). The community climate system model version 4. Journal
of Climate, 24 (19), 4973– 4991.
Goudet, J. (20 05). Hier fs tat , a Package for R to Compute an d test hierar-
chical F- statistics. Molecular Ecology Notes, 5(1), 184– 186.
Grazziotin, F. G., Monzel, M., Echeverrigaray, S., & Bonat to, S. L. (2006).
Phylogeography of the Bothrops Jararaca Complex (Serpentes:
Viperidae): Past Fragmentation and Island Colonization in the
Brazilian Atlantic Forest. Molecular Ecology, 15(13), 3969– 3982.
Hayes, F. E., & Sewlal, J.- A .- N . (20 04). The Amazon River as a dispersal
barrier to passerine birds: Effects of river width, habitat and ta xon-
omy. Journal of Biogeography, 31(11), 1809– 1818.
Hewitt, G. (2000). The genetic legacy of the quaternary ice ages. Nature,
405(6789), 907– 913.
Hi jma ns, R . J., C a m eron , S. E. , Pa r r a , J. L . , Jones, P. G ., & Jar v is, A . (20 0 5).
Very high resolution interpolated climate sur faces for global land
areas. International Journal of Climatology, 25(15), 1965– 1978.
Hijmans, R. J., & van Etten, J. (2012). Raster: Geographic Analysis and
Modeling with Raster Data. R Package Version 1:9– 92.
Jarvis, A., Reuter, H. I., Nelson, A., & Guevara, E. (2008). Hole- Filled SRTM
for the Globe Version 4. CGI AR- CSI SRTM 90mData base. Retrieved
from http://Srtm.Csi.Cgiar.Org
Koroiva, R., Rodrigues, L . R. R. & Santana, D. J. (2020). DNA barcoding
for identification of anuran species in the central region of South
America. Pee rJ, 8, e10189.
Lehner, B., Verdin, K., & Jarvis, A. (2008). New global hydrography de-
rived from spaceborne elevation data. Eos, Transactions American
Geophysical Union, 89(10), 93– 94.
   MYERS E t al.
Leite, Y. L. R., Cost a, L. P., Loss, A. C., Rocha, R. G., Batalha- Filho, H.,
Bastos, A. C., Quaresma, V. S., Fagundes, V., Paresque, R., Passamani,
M., & Pardini, R. (2016). Neotropical forest expansion during the
last glacial period challenges refuge hypothesis. Proceedings of the
National Academy of Sciences, 113(4), 1008– 1013
Lorenzen, E. D., Heller, R., & Siegismund, H. R. (2012). Comparative phy-
logeography of African Savannah Ungulates 1. Molecular Ecology,
21(15), 3656– 3670.
Lyra, M. L., Haddad, C. F. B., & de Azeredo- Espin, A. M. L. (2017). Meeting
the challenge of DNA barcoding neotropical amphibians: poly-
merase chain reaction optimization and new COI primers. Molecular
Ecology Resources, 17(5), 966– 980.
Mani on , G. , Li sk , M. , Fe rr ier, S. , Ni et o- Lugil de , D. , & Fitz pat ri ck, M. C. (2 016).
Gdm: Functions for Generalized Dissimilarity Modeling. R Package.
Mascarenhas, R., Miyaki, C . Y., Dobrovolski, R., & Bat alha- Filho,
H. (2019). Late pleistocene climate change shapes population
divergence of an Atlantic Forest Passerine: A model- based phy-
logeographic hypothesis test. Journal of Ornithology, 16 0 (3),
7 3 3 – 7 4 8 .
McRae, B. H., Dickson, B. G., Keitt, T. H., & Shah, V. B. (2008). Using
circuit the or y to mo del connec ti vity in ecolog y, evolution , and con-
servation. Ecology, 89(10), 2712– 2724.
Mello, C. L., Metelo, C. M. S., Suguio, K., & Kohler, H. C. (1999).
Quaternary sedimentation, neotectonics and the evolution of the
doce river middle valley lake system (Southeastern Brazil). Revista
do Instituto Geológico, 20( 1 - 2 ) , 2 9 – 3 6 .
Menezes, L., Canedo, C., Batalha- Filho, H., Garda, A. A., Gehara, M., &
Napoli, M. F. (2016). Multilocus Phylogeography of the Treefrog
Scinax eurydice (Anura, Hylidae) Reveals a Plio- Pleistocene
Diversification in the Atlantic Forest. PLoS One, 11 (6), e0154626.
Mitchell, M. W., Locatelli, S., Sesink Clee, P. R., Thomassen, H. A., &
Gonder, M. (2015). Environmental variation and rivers govern the
struc ture of chimpanzee genetic diversity in a biodiversity hotspot.
BMC Evolutionary Biology, 15(1), 1.
Muscarella, R., Galante, P. J., Soley- Guardia, M., Boria, R. A., Kass, J. M.,
Uriarte, M., & Anderson, R. P. (2014). ENM Eval: An R package for
conducting spatially independent evaluations and estimating opti-
mal model complexity for maxent ecological niche models. Methods
in Ecology and Evolution, 5(11), 1198– 1205.
Myers, E. A ., Xue, A. T., Gehara, M., Cox, C . L., Davis, A. R., Rabosky, J.-
L.- E., Martínez- mez, J. E., & Burbrink, F. T. (2019). Environmental
heterogeneity and not vicariant biogeographic barriers generate
community- wide population structure in desert- adapted snakes.
Molecular Ecology, 28(20), 4535– 4548.
Myers, N., Mittermeier, R. A., Mittermeier, C. G ., Da Fonseca, G . A. B., &
Kent, J. (2000). Biodiversit y hotspots for conservation priorities.
Nature, 403(6772), 853.
Papadopoulou, A., & Knowles, L. L. (2016). Toward a paradigm shift
in comparative phylogeography driven by trait based hypoth-
eses. Proceedings of the National Academy of Sciences, 113 (29),
8018– 8024.
Paz, A., Spanos, Z., Brown, J. L., Lyra, M., Haddad, C., Rodrigues, M., &
Carnaval, A. (2019). Phylogeography of Atlantic Forest Glassfrogs
(Vitreorana): When geography, climate dynamics and rivers mat ter.
Heredity, 122(5), 545– 557.
Pearson, R. G., Raxworthy, C. J., Nakamura, M., & Townsend Peterson,
A. (2007). Predic ting species distributions from small numbers of
occurrence records: a test case using cryptic geckos in Madagascar.
Journal of biogeography, 34(1), 102– 117.
Phillips, S. J., Anderson, R. P., & Schapire, R. E. (20 06). Maximum entropy
modeling of species geographic distributions. Ecological Modelling,
190 (3– 4), 231– 259.
Pons, J., Barraclough, T. G., Gomez- Zurita, J., Cardoso, A., Duran, D.
P., Hazell, S., Kamoun, S., Sumlin, W. D., & Vogler, A. P. (2006).
Seq uence- base d speci es deli mitat ion for the DNA taxonomy of un -
described insects. Systematic Biology, 55(4), 595– 609.
Rambaut, A., Drummond, A. J., Xie, D., Baele, G., & Suchard, M. A. (2018).
Posterior summarization in Bayesian Phylogenetics Using Tracer
1. 7. Systematic Biology, 67(5), 901.
Reid, N. M., & Carstens, B. C. (2012). Phylogenetic estimation error
can decrease the accurac y of species delimitation: A Bayesian im-
plementation of the general mixed yule- coalescent model. BMC
Evolutionary Biology, 12(1), 196.
Ribeiro, C., Priciane, J. P., Lemos- Filho, R. S., de Oliveira Buzatti, M. B. L.,
& Myriam, H. (2016). Species- Specific Phylogeographical Patterns
and Pleistocene East- West Divergence in Annona (Annonaceae)
in the Brazilian Cerrado. Botanical Journal of the Linnean Society,
181(1), 21– 36.
Ribeiro, R. A., Lemos- Filho, J. P., Ramos, A. C. S., & Lovato, M. B. (2011).
Phylogeography of the Endangered Rosewood Dalbergia Nigra
(Fabaceae): Insights into the evolutionary history and conservation
of the Brazilian Atlantic Forest. Heredity, 106(1), 46– 57.
Riccomini, C., & Assumpção, M. (1999). Quaternar y tectonics in Brazil.
Episodes, 22(3), 221– 225.
Rull, V. (2008). Speciation timing and neotropical biodiversity: the
tertiary- quaternary debate in the light of molecular phylogenetic
evidence. Molecular Ecology, 17(11), 2722– 2729.
Sotelo- Muñoz, M., Maldonado- Coelho, M., Svensson- Coelho, M., dos
Santos, S. S., & Miyaki, C. Y. (2020). Vicariance, dispersal, extinc-
tion and hybridization underlie the evolutionary history of Atlantic
Forest Fire- Eye Antbirds (Aves: Thamnophilidae). Molecular
Phylogenetics and Evolution, 14 8, 106820.
Thom, G., Smith, B. T., Gehara, M., Montesanti, J., Lima- Ribeiro, M. S.,
Piacentini, V. Q., Miy aki , C. Y., & do Amaral , F. R. (2020). Climatic dy-
namics and topography control genetic variation in Atlantic Forest
montane birds. Molecular Phylogenetics and Evolution, 14 8, 106812
Thomé, M. T. C., Zamudio, K. R ., Giovanelli, J. G . R., Haddad, C. F. B.,
Baldissera, F. A ., & Alexandrino, J. (2010). Phylogeography of en-
demic toads and post- Pliocene persistence of the Brazilian Atlantic
Forest. Molecular Phylogenetics and Evolution, 55(3), 1018– 1031.
Thuiller, W., Georges, D., & Engler, R. (2013). Biomod2: Ensemble Platform
for Species Distribution Modeling. R Package Version.
Toews, D. P. L ., & Brelsford, A. (2012). The biogeography of mitochon-
drial and nuclear discordance in animals. Molecular Ecology, 21(16),
3907– 3930.
Tonini, J. F. R., Costa, L. P., & Carnaval, A. C. (2013). Phylogeographic
struc ture is strong in the Atlantic forest; predic tive power of cor-
relative paleodistribution models, not always. Journal of Zoological
Systematic s and Evolutionary Research, 51(2), 114– 121.
Turchetto- Zolet, A. C., Cruz, F., Vendramin, G. G., Simon, M. F., Salgueiro,
F., Margis- Pinheiro, M., & Margis, R. (2012). Large- scale phyloge-
ography of the disjunc t neotropical tree species Schizolobium pa-
rahyba (Fabaceae- Caesalpinioideae). Molecular Phy logenetics and
Evolution, 65(1), 174– 182.
Turchetto- Zolet, A. C., Pinheiro, F., Salgueiro, F., & Palma- Silva, C. (2013).
Phylogeographical patterns shed light on evolutionary process in
South America. Molecular Ecology, 22(5), 1193– 1213.
Wallace, A . R. (1854). On the monkeys of the Amazon. Annals and
Magazine of Natural His tory, 14(84), 451– 454.
Wang, I. J., & Bradburd, G. S. (2014). Isolation by environment . Molecular
Ecology, 5649– 5662.
Warren, D., Dinnage, R., & Mat zke, N. (2017). ENMTools: Analysis of Niche
Evolution Using Niche and Distribution Models. R Pac kage (Ve rsion 0.2 ).
Wright, S. (1943). Isolation by distance. Genetics, 28(2), 114.
Additional supporting information may be found online in the
Supporting Information section.
Alignment S1. Seque nc e alignme nt of COI gene rate d and use d in this
stud y.
MYERS E t al.
Table S1. List of all museum numbers of Ololygon carnevallii speci-
mens examined and georeferenced for ENMs.
Table S2. Model parameter fit for ENMs obtained from ENMeval.
Figure S1. Cur rent EN M for O. carnevallii obtaine d using second best
fit model parameters (see Table S2).
Figure S2. Correlations between FST and potential explanatory vari-
ables from GDM.
How to cite this article: Myers EA, Folly H, Ragalzi E,
Feio RN, Santana DJ. Late Pliocene population divergence
and persistence despite Pleistocene climatic fluctuations in
the Rio Doce snouted Treefrog (Ololygon carnevallii). J Zool
Syst Evol Res. 2021;0 0:1–11. https: //
ResearchGate has not been able to resolve any citations for this publication.
Full-text available
Isthmian Central America (ICA) is one of the most biodiverse regions in the world, hosting an exceptionally high number of species per unit area. ICA was formed <25 million years ago and, consequently, its biotic assemblage is relatively young and derived from both colonization and in situ diversification. Despite intensive taxonomic work on the local fauna, the potential forces driving genetic divergences and ultimately speciation in ICA remain poorly studied. Here, we used a landscape genetics approach to test whether isolation by distance, topography, habitat suitability, or environment drive the genetic diversity of the regional frog assemblage. To this end, we combined data on landscape features and mitochondrial DNA sequence variation for nine codistributed amphibian species with disparate life histories. In five species, we found that at least one of the factors tested explained patterns of genetic divergence. However, rather than finding a general pattern, our results revealed idiosyncratic responses to historical and ecological processes, indicating that intrinsic life-history characteristics may determine the effect of different drivers of isolation on genetic divergence in ICA. Our work also suggests that the convergence of several factors promoting isolation among populations over a heterogeneous landscape might maximize genetic differentiation, despite short geographical distances. In conclusion, abiotic factors and geographical features have differentially affected the genetic diversity across the regional frog assemblage. Much more complex models (i.e., considering multiple drivers), beyond simple vicariance of Caribbean and Pacific lineages, are needed to better understand the evolutionary history of ICA’s diverse biotas.
Full-text available
The use of COI barcodes for specimen identification and species discovery has been a useful molecular approach for the study of Anura. Here, we establish a comprehensive amphibian barcode reference database in a central area of South America, in particular for specimens collected in Mato Grosso do Sul state (Brazil), and to evaluate the applicability of the COI gene for species-level identification. Both distance-and tree-based methods were applied for assessing species boundaries and the accuracy of specimen identification was evaluated. A total of 204 mitochondrial COI barcode sequences were evaluated from 22 genera and 59 species (19 newly barcoded species). Our results indicate that morphological and molecular identifications converge for most species, however, some species may present cryptic species due to high intraspecific variation, and there is a high efficiency of specimen identification. Thus, we show that COI sequencing can be used to identify anuran species present in this region.
Full-text available
Genetic structure can be influenced by local adaptation to environmental heterogeneity and biogeographic barriers, resulting in discrete population clusters. Geographic distance among populations, however, can result in continuous clines of genetic divergence that appear as structured populations. Here we evaluate the relevant importance of these three factors over a landscape characterized by environmental heterogeneity and the presence of a hypothesized biogeographic barrier in producing population genetic structure within 13 codistributed snake species using a genomic dataset. We demonstrate that geographic distance and environmental heterogeneity across western North America contribute to population genomic divergence. Surprisingly, landscape features long thought to contribute to biogeographic barriers play little role in divergence community wide. Our results suggest that isolation by environment is the most important contributor to genomic divergence. Furthermore, we show that models of population clustering that incorporate spatial information consistently outperform nonspatial models, demonstrating the importance of considering geographic distances in population clustering. We argue that environmental and geographic distances as drivers of community‐wide divergence should be explored before assuming the role of biogeographic barriers. This article is protected by copyright. All rights reserved.
Full-text available
Several evolutionary processes seem to have influenced the Atlantic Forest (AF) biogeographic history, as suggested by phylogeographic studies that have shown a multitude of patterns. Here, we use approximate Bayesian computation to test alternative historical hypotheses to investigate the phylogeographic pattern, historical demography, and palaeodistribution of the Grey-hooded Flycatcher Mionectes rufiventris, an endemic AF bird, distributed mainly in southern areas of the biome. Our goal was to integrate molecular and ecological data to test diversification hypotheses available for the AF. Our investigation revealed two mitochondrial phylogroups, geographically structured around the Doce River. Coalescence analyses revealed that these groups shared a common ancestor in the Late Pleistocene, between 200,000 and 300,000 years ago, and that divergence was probably associated with climatic fluctuations during this period. Demographic analyses suggested recent demographic expansion in both groups. Ecological niche modelling suggested larger ranges during the Last Glacial Maximum (LGM) than in the present, not in agreement with the genetic pattern recovered. We simulated alternative historical models to test these competing scenarios. Our findings support the existence of small populations during the LGM which expanded afterwards from putative refuges. Thus, these results suggest that the Pleistocene climate shaped patterns of diversification and demographic history of this species in accordance with the classical forest refuge hypothesis.
Full-text available
High-resolution, easily accessible paleoclimate data are essential for environmental, evolutionary, and ecological studies. The availability of bioclimatic layers derived from climatic simulations representing conditions of the Late Pleistocene and Holocene has revolutionized the study of species responses to Late Quaternary climate change. Yet, integrative studies of the impacts of climate change in the Early Pleistocene and Pliocene – periods in which recent speciation events are known to concentrate – have been hindered by the limited availability of downloadable, user-friendly climatic descriptors. Here we present PaleoClim, a free database of downscaled paleoclimate outputs at 2.5-minute resolution (~5 km at equator) that includes surface temperature and precipitation estimates from snapshot-style climate model simulations using HadCM3, a version of the UK Met Office Hadley Centre General Circulation Model. As of now, the database contains climatic data for three key time periods spanning from 3.3 to 0.787 million years ago: the Marine Isotope Stage 19 (MIS19) in the Pleistocene (~787 ka), the mid-Pliocene Warm Period (~3.264–3.025 Ma), and MIS M2 in the Late Pliocene (~3.3 Ma).
Full-text available
Bayesian inference of phylogeny using Markov chain Monte Carlo (MCMC) (Drummond et al., 2002; Mau et al., 1999; Rannala and Yang, 1996) flourishes as a popular approach to uncover the evolutionary relationships among taxa, such as genes, genomes, individuals or species. MCMC approaches generate samples of model parameter values - including the phylogenetic tree -drawn from their posterior distribution given molecular sequence data and a selection of evolutionary models. Visualising, tabulating and marginalising these samples is critical for approximating the posterior quantities of interest that one reports as the outcome of a Bayesian phylogenetic analysis. To facilitate this task, we have developed the Tracer (version 1.7) software package to process MCMC trace files containing parameter samples and to interactively explore the high-dimensional posterior distribution. Tracer works automatically with sample output from BEAST (Drummond et al., 2012), BEAST2 (Bouckaert et al., 2014), LAMARC (Kuhner, 2006), Migrate (Beerli, 2006), MrBayes (Ronquist et al., 2012), RevBayes (Höhna et al., 2016) and possibly other MCMC programs from other domains.
In order to gain insights into the biogeographic processes underlying biotic diversification in the Atlantic Forest (AF), we used a multi-locus approach to examine the evolutionary history of the White-shouldered Fire-eye (Pyriglena leucoptera) and the Fringe-backed Fire-eye (Pyriglena atra), two parapatric sister species endemic to the AF. We sequenced one mitochondrial, three Z chromosome-linked and three anonymous markers of 556 individuals from 66 localities. We recovered four lineages throughout the AF: P. atra and three populations within P. leucoptera. All populations diverged during the late Pleistocene and presented varying levels of admixture. One Z-linked locus showed the highest level of differentiation between the two species. On the other hand, a mitochondrial haplotype was shared extensively between them. Our data supported vicariance driving speciation along with extinction and dispersal as processes underlying intraspecific diversification. Furthermore, signatures of demographic expansion in most populations and areas of genetic admixture were recovered throughout the AF, suggesting that forest fragmentation was also important in differentiation. Genetic admixture areas are located between large rivers suggesting that AF rivers may diminish gene flow. Our results indicated a complex and dynamic biogeographic history of Pyriglena in the AF, with vicariance, extinction, dispersal and secondary contact followed by introgression likely influencing the current patterns of genetic distribution.
Recent advances in the field of landscape genetics provide ways to jointly analyze the role of present-day climate and landscape configuration in current biodiversity patterns. Expanding this framework into a phylogeographic study, we incorporate information on historical climatic shifts, tied to descriptions of the local topography and river configuration, to explore the processes that underlie genetic diversity patterns in the Atlantic Forest hotspot. We study two montane, stream-associated species of glassfrogs: Vitreorana eurygnatha and V. uranoscopa. By integrating species distribution modeling with geographic information systems and molecular data, we find that regional patterns of molecular diversity are jointly explained by geographic distance, historical (last 120 ky) climatic stability, and (in one species) river configuration. Mitochondrial DNA genealogies recover significant regional structure in both species, matching previous classifications of the northern and southern forests in the Atlantic Forest, and are consistent with patterns reported in other taxa. Yet, these spatial patterns of genetic diversity are only partially supported by nuclear data. Contrary to data from lowland taxa, historical climate projections suggest that these montane species were able to persist in the southern Atlantic Forest during glacial periods, particularly during the Last Glacial Maximum. These results support generally differential responses to climatic cycling by northern (lowland) and southern (montane) Atlantic Forest species, triggered by the joint impact of regional landscape configuration and climate change.