https://orcid.org/0000-0003-4426-3286
Figures
Abstract
Rhynchonycteris is a monotypic genus of Embalonurid bats, whose geographic distribution extends from southern Mexico to tropical regions of the South American continent, including Trinidad and Tobago. Although species that have a wide geographic distribution are frequently revealed to be polytypic, to date, no study has evaluated the taxonomic status of populations of Rhynchonycteris naso. Thus, the aim of this study is to address the patterns of phylogeographic structure and taxonomic subdivision of R. naso using molecular phylogenetics, morphometric data and ecological niche modeling. Phylogenetic results recovered using the genes COI, Cytb, Chd1, Dby, and Usp9x, supported the monophyly of the genus Rhynchonycteris, in addition, a deep phylogeographic structure was revealed by the mitochondrial gene COI for the populations of Belize and Panama in comparison to those of South America. The PCA, and the linear morphometry indicated an apparent differentiation between the cis-Andean and trans-Andean populations. Furthermore, according to the skull morphology, at least two morphotypes were identified. Ecological niche modeling projections in the present have shown that the Andean cordillera acts as a climatic barrier between these two populations, with the depression of Yaracuy (Northwest Venezuela) being the only putative climatically suitable path that could communicate these two populations. On the other hand, projections for the last glacial maximum showed a drastic decrease in climatically suitable areas for the species, suggesting that cycles of lower temperatures played an important role in the separation of these populations.
Citation: Biganzoli-Rangel AJ, Leon-Alvarado OD, Robe LJ, Meza MA, Gutiérrez EE, Paladini A (2023) Systematics and phylogeography of bats of the genus Rhynchonycteris (Chiroptera: Emballonuridae): Integrating molecular phylogenetics, ecological niche modeling and morphometric data. PLoS ONE 18(5): e0285271. https://doi.org/10.1371/journal.pone.0285271
Editor: Aristeidis Parmakelis, National and Kapodistrian University of Athens, GREECE
Received: August 17, 2022; Accepted: April 18, 2023; Published: May 4, 2023
Copyright: © 2023 Biganzoli-Rangel et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting information files.
Funding: We are thankful to the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior — Brasil (CAPES) — For the Master degree fellowship of A.J.B-R and O.D.L-A. To the Conselho Nacional de Desenvolvimento Cientifico e Tecnológico — Brasil (CNPq) — in which A.P. is a productivity fellow (process number: 309191/2019-0), and L.J.R is a pqCNPq: research fellow (process number: 314206/2021-3). Also, The IDEA WILD organization provided funding through valuable equipment to A.J.B-R (code: BIGABRAZ1119) and O.D.L-A (code: LEONBRAZ1119). Finally, to the The American Society of Mammalogists (ASM) for the Latin American Fellowship to O.D.L-A. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Bats of the Emballonuridae family are currently recognized by eight genera and 22 species in the New World, and five genera and 30 species in the Old World [1–3]. These strictly insectivorous bats are found in tropical and subtropical areas, where they provide important ecosystem services, acting in the control of insect populations [4–6]. Its wide geographic distribution in the Neotropics extends from southern Mexico, central America and central South America, and into the Old World across Africa, Madagascar, South Asia, part of Australia and the Pacific Islands from eastern Samoa [3].
Within the New World Emballonurids, Rhynchonycteris W. Peters, 1867 [7] is a monotypic genus of bats presenting a neotropical distribution [8]. These bats generally inhabit mangroves and lowland forests, close to streams, rivers and lakes [9], where they occur from 0 to 900 meters above sea level (m.a.s.l.) [8,10]. Their distribution extends from southern Mexico, through Central America, eastern Peru, northern Bolivia and central Brazil, including Trinidad and Tobago [1,8,10–13]. A striking feature of the wide geographic distribution of Rhynchonycteris naso (Wied-Neuwied, 1820) [14] is the division of these populations by the Andes into two large groups: the trans-Andean, occurring north of the Andean slopes, and the cis-Andean, distributed in the southern part of the Andean slopes [15].
Several studies have investigated the phylogenetic relationships within the Emballonuridae family. These studies evaluated skull morphology [16], protein electrophoresis [17], hyoid bone morphology [18], morphology and behavior [19] and molecular analysis with nuclear and mitochondrial genes [20]. Nevertheless, none of these studies tested the monophyly of Rhynchonycteris naso with a comprehensive sampling of terminal taxa or evaluated differences among its populations taking into account its biogeography, ecology and phylogeography.
Currently, due to the ubiquity of cryptic diversification, the discovery of new species is rarely related to morphological variation. In fact, new species are usually delimitated by integrating multiple approaches, as behavior, ecology and molecular markers [21], with subtle or unnoticed morphological differences. This is why, for example, in other genera of bats, genetic tools have helped to uncover cryptic diversity and to clarify interspecific relationships [22–24]. Other approaches as Ecological Niche Models (ENMs) also frequently contribute to reduce the Linnean and Wallacean impediments [25,26], related to shortages in the knowledge about the number and the distribution of most species, respectively.
Due to the lack of studies assessing the monophyly and the taxonomic status of different populations of Rhynchonycteris naso, the aim of this study is to address the patterns of distribution of genetic and morphological variation of this species. Furthermore, we evaluated geographic and non-geographic morphometric variation. Finally, we assessed whether ENM projections performed for the species in various past climate scenarios are congruent with population divergence patterns.
Materials and methods
Taxonomic sampling
To test the monophyly of R. naso, we selected DNA sequences available in GenBank from all recognized neotropical species of the Emballonuridae family [1,2]. The ingroup was composed by 20 species (Balantiopteryx infusca, Balantiopteryx io, Balantiopteryx plicata, Centronycteris centralis, Centronycteris maximiliani, Cormura brevirostris, Cyttarops alecto, Diclidurus albus, Diclidurus ingens, Diclidurus isabellus, Diclidurus scutatus, Peropteryx kappleri, Peropteryx leucoptera, Peropteryx macrotis, Peropteryx pallidoptera, Peropteryx trinitatis, Saccopteryx bilineata, Saccopteryx canescens, Saccopteryx gymnura and Saccopteryx leptura) in addition to R. naso. The only species of neotropical Emballonuridae that could not be included in the analysis due to the absence of any sequence is Saccopteryx antioquensis, endemic to the northern Andes of Colombia [27].
Furthermore, we included sequences of six Old World embalonurids as an outgroup (Emballonura beccarii, Emballonura monticola, Emballonura raffrayana, Mosia nigrecens, Taphozous melanopogon and Taphozous longimanus). These species encompass the group most closely related to New World embalonurids [20,28].
Genetic data
For the phylogenetic analysis we used mitochondrial and nuclear (autosomal and sexual) gene sequences available in GenBank. These sequences included fragments of Cytochrome C Oxidase I (COI), Cytochrome b (Cytb), Chromodomain-Helicase-DNA-Binding Protein 1 (Chd1), DEAD-Box Helicase 3 Y-Linked (Dby), and Ubiquitin Specific Peptidase 9 X-Linked (Usp9x) (see [20]) (for accession numbers, see S1 Table).
Phylogenetic analysis
We aligned the sequences of each gene using the MUSCLE algorithm (K-mer distance) [29] implemented in MEGA X [30], with default options. Individual gene trees were reconstructed by maximum likelihood (ML) in iqTree v.2 [31], using ModelFinder [32] to select the best substitution model. Clade support was measured through 1000 ultrafast bootstrap replicates [33]. A partitioned Bayesian inference analysis was performed in MrBayes v3.2.2 [34] and implemented in the Cipres Science Gateway v3.3. The best nucleotide evolution model was selected for each gene using jModelTest2 [35,36], using the Akaike Information Criterion (AIC) [37]. The Bayesian analysis consisted of a cold chain and three heated chains, using a sampling approach by the Markov Chain Monte Carlo (MCMC) algorithm, with 50,000,000 generations, sampling every 5,000 generations. The initial 25% of the samples were discarded as burn-in. Samples from each run were evaluated in Tracer [38] and a cutoff point of 200 for the effective sample size (ESS) was used to determine if the Markov chain reached stationarity and converged. Finally, the majority-rule tree was rooted in Taphozous longimanus.
Phylogeographic structure
To assess the phylogeographic structure of R. naso, we used COI sequences available for the species in GenBank. The haplotypes and their frequencies were determined using the software DnaSP v6.0 [39]. Pairwise ΦST values among populations were estimated in Arlequin v3.5 [40], with significance inferred with 10,000 permutations. A Mantel test was then performed to assess whether spatial distance (Euclidean geographic distance) is correlated with genetic distance, in an isolation by distance pattern.
Data were also analyzed using the software SAMOVA v1.0 [41], which defines the best group of populations (K) that are geographically homogeneous but genetically differentiated. To this task, K values ranging from 2 to 10 were evaluated with 100 initial conditions. The K value that produced the highest Fct was determined, and individuals were assigned to groups following this classification. The general structure of the population was then inferred with an Analysis of Molecular Variance (AMOVA). Finally, median-joining haplotype networks [42] for the main haplogroups were generated using the software Network, v10 (http://www.fluxus-engineering.com/).
Genetic distance
Pairwise intraspecific and interspecific genetic distances of the COI gene were calculated in the MEGA X software [30] using two approaches: Kimura 2 parameters substitution model and p-distance.
Morphometric analysis
Museum acronyms for specimens examined are: AMNH, American Museum of Natural History, New York, USA; IAVH, Instituto Alexander Von Humboldt, Bogotá, Colombia; ICN, Instituto de Ciencias Naturales Universidad Nacional de Colombia, Bogotá, Colombia; UIS, Universidad Industrial de Santander, Bucaramanga, Colombia; MUSM, Museo Universidad Nacional Mayor de San Marcos, Lima, Peru; MZUSP, Museu de Zoologia da Universidade de São Paulo, São Paulo, Brazil; UFES, Museu da Universidade Federal de Espírito Santo, Espírito Santo, Brazil; QCAZ, Museo de Zoología de la Pontificia Universidad Católica del Ecuador, Quito, Ecuador.
A digital caliper with an accuracy of 0.01 mm was used to take one external (forearm) and 10 craniodental measurements of 121 specimens, identified according to Hood and Garner [12]. The measurements were previously used in other studies of the Emballonuridae family [8,43]. The craniodental, and external measurements used in this study were:
- Greatest length of skull (GLS): Distance from the posteriormost point on the occiput to the anteriormost point on the premaxilla (excluding the incisors).
- Condilobasal length (CBL): Distance between a line connecting the posteriormost margins of the occipital condyles and the anteriormost point on the premaxilla.
- Height upper canine (UP CANIN): Greatest length from point immediately dorsal to cingulum to end of tooth.
- Breadth braincase (BR BCASE): Breadth just dorsal to posterior juncture of zygomatic process.
- Breadth at mastoids (MASTOID): Greatest breadth at mastoid processes.
- Maxillary toothrow (MTR): Length from anterior alveolar border of canine to posterior alveolar border of M3.
- Width at upper molars (M3 M3): Width between alveolar borders of upper third molars.
- Width at upper canines (C1 C1): Width between alveolar borders of upper canines.
- Dentary length (DENT LEN): From midpoint of mandibular condyle to anteriormost point of dentary.
- Maxillary toothrow (MTR): Length from anterior alveolar border of canine to posterior alveolar border of M3.
- Forearm length (FA): Distance from the elbow (tip of the olecranon process) to the wrist (including the carpals).
Some structures could not be measured due to skull deterioration. Therefore, some data were inferred with the missMDA package [44] applied in R v3.6.1 [45].
All measurements were transformed into log base 10. Morphometric differences between populations of R. naso and between sexes were evaluated using a Principal Component Analysis (PCA) based on the variance-covariance matrix. To show the relationships between groups in the morphospace, the values of the principal components (PC) were graphically represented. Furthermore, statistical differences between sexes and between the cis-Andean and trans-Andean populations were evaluated with a Multivariate Permutation Analysis of Variance (PERMANOVA). The methods were applied in R v3.6.1 [45].
Finally, the skulls were photographed with a Nikon D3500 with an AF-S 40mm 1:2.8G Nikkor Micro lens. The skulls were photographed in ventral, dorsal and lateral views to assess possible differences in morphology.
Ecological niche modeling
To assess whether the climate suitability model projections to past climate scenarios are congruent with the inferred divergence patterns, we build the ecological niche models (ENM) using the maximum entropy algorithm implemented in Maxent. v4.0 [46]. A total of 3904 occurrences of R. naso were obtained from Global Biodiversity Information Facility—GBIF (http://www.gbif.org). Analyzes were based exclusively on data from voucher specimens deposited in natural history museums. Occurrences without location-specific information were removed, as well as occurrences without geographic uncertainty or with uncertainty > 5000 meters. To reduce the sampling bias that may exist and the spatial autocorrelation of the data [47], we used the spThin package [48] in R v3.6.1, so that samples have a minimum distance of 10 kilometers between each other (see [49,50]). Thus, we got 227 filtered occurrences, divided into 170 calibration or training points and 57 test points.
To create the models and effectively infer areas of climate suitability for the target organism, the environmental data was composed of 19 bioclimatic layers (spatial resolution = 2.5’, which corresponds to the uncertainty inherent in the occurrence data) downloaded from WorldClim v1.4 [51]. Four bioclimatic variables (BIO7, BIO8, BIO18, BIO19) were excluded from the analysis because these variables include known spatial artifacts [52]. Four sets of environmental variables were chosen, based on jackknife tests applied in Maxent. Variables that contributed the least were sequentially removed. The sets used for the analyzes consisted of 15, 8, 6, and 4 variables.
The models were calibrated in the known range of distribution for the species (from southern Mexico to central South America), which is considered a working hypothesis of M for the species [53]. To acquire an adequate representation of the environments available for these species, 20,000 random background points were included in the delimited study area.
Models were created with 10 bootstrap replicates with a variety of different combinations of feature classes (Linear (L), Quadratic (Q), Product (P), Threshold (T), and Hinge (H)), and 18 regularization multiplier values (0.1–1 with intervals of 0.1; 1–6 with intervals of 1; 8, and 10). Model performance was evaluated considering statistical significance (partial ROC (Receiver Operating Curve); [54]), predictive power (omission rates, E = 5%, [55]), and complexity level (AICc; [56]), in this order.
After calibrating and selecting the models, model projections were performed to present and past scenarios. Model projections involves transferring the niche models to a different space than the one used for model calibration; a Mobility-Oriented Parity analysis (MOP) was performed to quantify the similarity between the calibration and transfer regions [57].
Past environmental data were characterized for the Last Interglacial (approx. 120,000–140,000 years ago), whose conditions were obtained from the PaleoClim [58,59], and for the Last Glacial Maximum (approx. 22,000 years ago), and the Mid Holocene (approx. 6,000 years ago), whose conditions were obtained from WorldClim [51]. The Last Glacial Maximum and the Mid Holocene periods were evaluated under three different climate models (GCMs; MPI-ESM-P, CCSM4 and MIROC-ESM).
To identify and graphically represent changes in suitability, and changes in suitable areas in past scenarios, the approach of Cobos et al., [60] was used. All analyzes were performed in R v3.6.1 with the kuenm package [61].
In addition to the models generated considering the total distribution of the species (hereafter called the “Total model”), we generated models for the “cis-Andean” (85 occurrences to calibrate and 32 occurrences to test the models), and “trans-Andean” populations (85 occurrences to calibrate and 25 occurrences to test the models). The methodology remained the same as described above.
Results
Phylogenetic analysis
The multi-locus dataset encompassed 3521 nucleotides, characterized for a total set of 249/27 ingroup specimens/species. The selected nucleotide substitution models were: HKY + I for Chd1 (423 nucleotides in length), GTR + I + Γ for COI (658 nucleotides), HKY + I + Γ for Cytb (1140 nucleotides), GTR + I for Dby (729 nucleotides), and HKY + I for Usp9x (571 nucleotides). The Bayesian analysis recovered Rhynchonycteris naso as a monophyletic group with a high Bayesian posterior probability (100%), a result that was also supported by individual gene trees (S3 Fig). The target species was positioned as a sister group of the Centronycteris genera, but the clade formed by these two groups had a low posterior probability (72%) (Fig 1), and was only supported by the Dby gene tree. Our analysis also suggested a geographic structure for the populations of R. naso, with populations from South America (cis-Andean) being recovered as monophyletic in regard to those from Belize and Panama (trans-Andean), with high posterior probability (99%) (Fig 1), and high bootstrap value (100%) in the COI gene tree (S3 Fig). Nevertheless, in the concatenated analysis, the trans-Andean specimens were recovered as a grade, possibly because of shortages related to missing data. In fact, most individuals from Belize and individuals from Panama presenting only COI sequences were recovered as a clade, whereas “Rhynchonycteris naso 11” from Belize was represented only by Chd1 sequences and appeared as a sister group of the cis-Andean specimens with low posterior probability (76%).
Bayesian posterior probabilities (> 50%) are indicated at each node with color dots. Branch lengths are proportional to the scale, given in substitutions per nucleotide.
Phylogographic structure
Out of a total of 106 COI sequences recovered for R. naso, 104 were evaluated for the presence of phylogeographic structure, and two sequences were excluded due to the high number of missing nucleotides. These resulted in 48 different haplotypes (S2 Table), none of which was shared between trans-Andean and cis-Andean populations. In fact, these groups encompassed different haplogroups in the network, differing by at least 47 substitutions (Fig 2). Furthermore, the populations of Belize (Hap14) and Panama (Hap10, Hap11, Hap12 and Hap13) differed by at least 15 substitutions.
Each circle represents a different haplotype, whose size is proportional to frequency. Colors subdivide haplotypes according to their sampling points as shown by the legend. Red small circles represent median vectors, and dashes in the lines connecting different haplotypes represent the number of substitutions between them.
All the 10 groups evaluated with SAMOVA showed significant differences between them (P < 0.001) (Table 1). The highest value of ΦCT was reached with K = 3 groups (ΦCT = 0.90809) and presented a subdivision among the population of Belize (harboring haplotype Hap14), the population of Panama (harboring haplotypes Hap10, Hap11, Hap12 and Hap13), and a third group including all other populations (see S2 Table for nomenclature). Thus, SAMOVA analysis support the hypothesis obtained with the phylogeny results (Fig 1), suggesting trans-Andean and cis-Andean populations of R. naso are deeply differentiated.
The general genetic structure among populations of R. naso (grouped according to the SAMOVA test) was highly significant (ΦST = 0.91, P < 0.001), while the pairwise ΦST values among the three major clusters varied from 0.83 to 0.92 (Fig 3, S3 Table). The cis-Andean populations were the most divergent, first compared to those of Belize (Hap14) (ΦST = 0.92, P < 0.001), then compared to those of Panama (Hap10, Hap11, Hap12 and Hap13) (ΦST = 0.91, P < 0.001) (Fig 3, S3 Table). In general, for the populations analyzed, a moderate correlation between genetic distance and geographic distance was detected, and was significant (r = 0.55, P = 0.00).
The color scale refers to the gradient of ΦST values, from no difference (ΦST = 0) to complete difference (ΦST = 1).
The COI distances recovered with the Kimura two-parameter (K2P) model among different species of Emballonuridae varied from 0.071 to 0.300. The mean distance within the R. naso species was 0.030, while the divergence between the cis-Andean and trans-Andean lineages was 0.110. On the other hand, the divergences recovered with p-distance varied from 0.066 to 0.244 among different species of the Emballonuridae family, whereas the distance between cis-Andean and trans-Andean lineages of R. naso was 0.100 (S4–S6 Tables).
Morphometric analysis
When comparing sexes, the PCA did not show any difference in the morphospace (Fig 4). On the other hand, there were apparent differences between cis-Andean and trans-Andean populations in the morphospace, specially regarding PCA1, which explained 32% of the total variation (Fig 5).
Some specimens did not have information about the sex, and were identified as “unknown”.
Regarding skull morphology, we were able to identify two general morphotypes without a particular geographic pattern (Fig 6). Morphotype I (Fig 6A–6F and 6J–6L) is characterized by the union of the upper incisors by the palate region (absence of diastema); in ventral view, the maxillary bone protrudes laterally to the molars. Morphotype II (Fig 6G–6I) presents a diastema between the upper incisors in the palate region; in ventral view, the maxillary bone does not protrude laterally to the molars; in lateral view, it seems that the anterior part of the skull (maxilla and nasal region) is narrower when compared with morphotype I. Also, the skull of morphotype II has a more pronounced curvature in the transition between the frontal bone and the maxilla when compared with morphotype I.
(A-C) sex not determined, from Costa Rica (trans-Andean). (D-F) male from Colombia (trans-Andean). (G-I) male from Ecuador (trans-Andean). (J-L) male from Ecuador (cis-Andean). Photo authors (J-L): Andrea Caicedo y Cristian Poveda.
Although the PCA did not show differences between sexes in the morphospace, the PERMANOVA analysis indicated that there are significant differences between them (P < 0.001). The same happened when cis-Andean and trans-Andean populations were compared. In general, trans-Andean populations are smaller than cis-Andean populations (Table 2). For skull length, the mean of cis-Andean populations were higher (males, 11.6 mm, n = 50; females, 11.6 mm, n = 40) than those from trans-Andean populations (males, 11.4 mm, n = 11; females, 11.2 mm, n = 9). The same happened with the forearm length, with the mean of cis-Andean populations higher (males, 37.9 mm, n = 48; females, 38.4 mm, n = 37) than those of trans-Andean (males, 35.0 mm, n = 10; females, 37.3 mm, n = 9). Therefore, it can be assumed that a clear differentiation in skull and wing size distinguishes cis-Andean from trans-Andean populations of R. naso.
Descriptive statistics are presented as mean (minimum–maximum) standard deviation [sample size].
Ecological niche modeling
The best ENM configurations obtained for R. naso as a whole or for each of its clusters are presented in Table 3. The projection of the total model for current climatic conditions (Fig 7 and S1 Fig) shows a high suitability in coastal regions from south of Mexico to Northeast Brazil, the Amazon region, the Llanos region in Colombia and Venezuela and a patch covering the central-North of Bolivia and the southeast of Peru. Furthermore, ENM analysis suggests that the Andean Cordillera and the Sistema Coriano in Venezuela (mountain system, transition between the Cordillera de la Costa and the Andes Mountains) could act as a climatic barrier (suitability = 0) between cis-Andean and trans-Andean populations. Also, according to this model, part of the Brazilian Caatinga and Cerrado biomes would act as a climatic barrier, isolating populations of R. naso in the forests of the east coast of the country.
The cis-Andean and trans-Andean models showed a similar result compared to the total model. For the cis-Andean model (Fig 8A and S1 Fig), the Andean Cordillera seems to be acting as a climatic barrier, but the Sistema Coriano and northeast of Caatinga biome do not seem to be a problem. The trans-Andean model (Fig 8B and S1 Fig) is similar to the cis-Andean one, but the particularity of this one is that it presents the Yaracuy depression (a geographic feature located in northwest Venezuela) as a putative connection between the two populations.Regarding the importance presented by each variable to the models, precipitation of driest quarter (40.6%) and annual precipitation (13.8%) were the most influential to the cis-Andean population. On the other hand, seasonal temperature (20.8%) and precipitation of the wettest month (17.6%) were the most important in the trans-Andean population.
In past climate model projections, the mid-Holocene map (Fig 9) showed that cis-Andean populations would have an increase in climatically suitable areas in the Andes region of Venezuela. It is important to note that the increase in this area would allow the Táchira depression (a geographic feature in the east of the Andes, separating the Tamá massif to the west and the Sierra de Mérida to the east) in Venezuela to act as a new passage to cis-Andean populations across the Cordillera. Also, there would be a reduction in some areas of southern and eastern Brazil. On the other hand, trans-Andean populations would have an increase in climatically suitable areas in the Andes region in Venezuela, but this increase does not suggest the Táchira depression region as a climatically suitable region for the passage of trans-Andean populations through the mountain range. One of the general circulation models (GCM) shows the loss of climatically suitable areas in the Sistema Coriano in the northwest of Venezuela, which would leave the trans-Andean populations isolated from the cis-Andean populations in the middle Holocene.
At the last glacial maximum (Fig 10), climatically suitable areas for cis-Andean populations are reduced, especially in the Sistema Coriano, Los Llanos from Venezuela and Colombia, and the Andes region from Venezuela to Peru. In addition to this reduction of areas, the populations of the east coast of Brazil (Rio Grande do Norte, to Rio do Janeiro) would be separated from those encountered in the rest of Brazil by the loss of suitable areas in the northeast region of the country (Maranhão, to Ceará, and part of Rio Grande do Norte). Trans-Andean populations would have a considerable reduction in climatically suitable areas, limiting their populations to the coasts of the Caribbean Sea in Nicaragua, in southern Costa Rica, on the coasts of the Pacific Ocean in Panama, and in northwest Colombia.
In the last Interglacial (Fig 11) the models showed a distribution of climatically suitable areas that is similar to those found in the present projections for both populations.
Discussion
Integrative approaches are being commonly helping to overcome the taxonomic impediment, leading to an acceleration on the delimitation and on the discovery of new species [62–64]. This study adds to this scenario in supporting the presence of cryptic or unrecognized diversity within the morphotope previously assigned to Rhynchonycteris naso. In fact, molecular analysis suggested a deep subdivision of this species in two different clusters, that seem to yield incipient morphological and ecological differences. ENM strategies further reinforced this outcome, by presenting a climatic barrier that extends along the Andean Cordillera and the Sistema Coriano in Venezuela that could have resulted in the divergence between cis-Andean and trans-Andean (Belize and Panama) populations of R. naso.
The genetic divergence found here between trans and cis-Andean populations for the COI gene was 11%, which is above the mean of 7.8% of genetic difference found by Clare et al. [65] for the same marker when they compared 47 genera of bats. Clare et al. [65] and Clare et al. [66] also analyzed the mean COI intraspecific divergence of 87 species of bats belonging to seven families in Guyana, and 163 species of bats belonging to nine Neotropical families, respectively, and found mean intraspecific values of 0.6% and 1.38%, which are quite lower than the 3% found here for R. naso. On the other hand, Martins et al. [67] found a genetic diversity range of 6–11% for Desmodus rotundus when they analyzed the mitochondrial Cytb gene. The authors state that these values are among the highest ones that have already been described for Neotropical bats. Although Martins et al. [67] used a mitochondrial gene different from that used in our work, which is likely to have a distinct evolution rate, Tobe et al. [68] affirm that the Cytb gene is more variable at intra and interspecific levels when compared to the COI gene. This reinforces the idea that the genetic distance of 11% found in the populations of R. naso suggest the presence of cryptic species. This is a very common occurrence in bats, and is usually associated with the low number of taxonomic works, which are needed to clarify the status of species [65,69,70].
Another source of evidence supporting subdivision of R. naso in two clusters is provided by the Network analysis, where each of them encompasses a distinct haplogroup, differing by at least 47 substitutions. This is also quite above the results previously presented for different species by other authors. Baird et al. [71], for example, studied the relationships between two species of Vespertilionidae bats, Aeoreste cinereus and Aeorestes semotus, of North America and Hawaii, and found a difference of at least 13 substitutions using the COI gene, and genetic divergences of 4.2% using the Cytb gene. On the other hand, Demos et al. [72] studied the genetic diversity of particular clades of Nycteridae bats using the Cytb gene, and detected that populations of Nycteris thebaica clade 1 and Nycteris thebaica clade 3 differ by approximately 40 substitutions, and 5% of genetic divergence. Thus, the differences found between the two clusters (cis-Andean and trans-Andean) of R. naso provide significant support to their diversification and isolation, suggesting the species could be subdivided in at least two species.
Although PCA analyzes (Fig 4) showed that males and females of R. naso are homogeneous within morphospace, it recovered a distinct pattern when cis-Andean and trans-Andean populations are compared. Besides, trans-Andean populations revealed smaller than cis-Andean populations. The visual analysis of the morphology of the skull reveals the presence of two distinct morphotypes without an apparent geographic pattern. This is not surprising, since morphological stasis can be favored by selection during cladogenesis, leading to the emergence of new species without apparent morphological change [73]. Alternatively, the faint morphological difference found in this study can be due to a recent speciation process and the absence of the disruptive selection in morphological characters [23]. Afterall, similar ecological niches can direct the morphological homogeneity [23].
In general, the projections resulting from ecological niche modeling in the present and during the Middle Holocene (Figs 8 and 9) showed low suitability along the entire Andes cordillera, to the exception of regions of high suitability in the Yaracuy depression in Venezuela. These regions may have acted as barriers to gene flow between the trans-Andean and cis-Andean populations, promoting a vicariance between populations occurring in both sides of the Andes. Similar results were obtained by Gutiérrez-Pinto et al. [74] using birds of the Parulidae family. According to Plumpton and Jones [8] and Hood and Gardner [10], R. naso has a specific distribution in lowland tropical areas, rarely reaching a maximum of 900 meters of elevation. According to this information, the Andes cordillera would be one of the most efficient barriers to the dispersion of this species.
Furthermore, projections to the last glacial maximum (Fig 10) showed a significant decrease in climatic suitability regions (compared to those in the present), with restricted areas presenting suitable climatic conditions putatively acting as refuges for populations of R. naso. These patterns may have influenced the genetic diversification, the speciation and the distribution area of the populations studied here [75]. Thus, glacial and interglacial cycles may have presented important effects on the isolation of populations in warmer and humid areas in Central America (e.g., Panama and Colombia) and South America (e.g., North and West of the Amazon), where individuals could survive during more harsh conditions. Although in the last interglacial the increase in temperatures would have benefitted the occupation of new areas, the Andes cordillera could still have acted as a climatic barrier keeping the two populations apart. Even though, the Sistema Coriano, the Yaracuy depression, and possibly the Tachira depression seem to constitute climatically suitable passages, enabling a putative gene flow between populations that could not be detected here.
Conclusion
The speciation process is not always accompanied by drastic morphological changes. This commonly makes traditional taxonomy practices to underestimate the real number of biological species currently existing. Nevertheless, especially in the face of the taxonomic impediment and the biodiversity crisis, we emphasize here the importance of integrative approaches to the identification of cryptic species. This task is of utmost importance for the formulation of conservation plans and strategies [73]. In this work we present several evidence demonstrating that there are two clusters within the general morphotype previously assigned to Rhynchonycteris naso, and that these probably represent two distinct species. Such recognition is important to understand the real distribution of each taxa, to evaluate their threatens and to promote further biological studies. Only in the face of this new status it will be possible to protect important providers of ecosystem services, as are the insectivorous bats.
Supporting information
S1 Fig. Summary of ENM calibration and projection areas and extrapolation risks for Rhynchonycteris naso considered as a whole or for each of its clusters in current conditions.
https://doi.org/10.1371/journal.pone.0285271.s001
(TIFF)
S2 Fig. Majority-rule tree inferred from the combined data set of Embalonurid bats through a partitioned Bayesian analysis.
The numbers following the species names correspond to the unique identifiers. Bayesian posterior probabilities are indicated at each node. Branch lengths are proportional to the scale, given in substitutions per nucleotide.
https://doi.org/10.1371/journal.pone.0285271.s002
(PDF)
S3 Fig. Maximum likelihood individual gene trees, and mitochondrial data combined data set, of Embalonurid bats.
Bootstrap probabilities are indicated at each node.
https://doi.org/10.1371/journal.pone.0285271.s003
(PDF)
S1 Table. List of specimens used for morphological and molecular evaluations.
Species with unique identifier number, locality, associated voucher specimen, and GenBank accession number.
https://doi.org/10.1371/journal.pone.0285271.s004
(PDF)
S2 Table. List of Rhynchonycteris naso haplotypes used in this work.
The numbers following the species names correspond to the unique identifiers.
https://doi.org/10.1371/journal.pone.0285271.s005
(PDF)
S3 Table. Pairwise values of genetic differentiation (ΦST) obtained with COI sequences among the three clusters of populations of Rhynchonycteris naso suggested by SAMOVA analysis.
Values below the diagonal are distances. Values above the diagonal are the p values.
https://doi.org/10.1371/journal.pone.0285271.s006
(PDF)
S4 Table. Pairwise values of interspecific genetic distance obtained with COI sequences among different species of Embalonurids bats, using the Kimura 2-parameters substitution model.
The numbers highlighted in gray refer to the level of sequence divergence within groups.
https://doi.org/10.1371/journal.pone.0285271.s007
(PDF)
S5 Table. Pairwise values of interspecific genetic distance obtained with COI sequences among different species of Embalonurids bats, using p-distance.
The numbers highlighted in gray refer to the level of sequence divergence within groups.
https://doi.org/10.1371/journal.pone.0285271.s008
(PDF)
S6 Table. Mean values of intraspecific genetic distance obtained with COI sequences for different species of Embalonurids, using kimura 2-parameters substitution model.
https://doi.org/10.1371/journal.pone.0285271.s009
(PDF)
Acknowledgments
We are very grateful to the curators and staff of institutions that allowed access to material, and for the work they do in these important research spaces. Especially to Alejandra Camacho (QCAZ), Catalina Cárdenas (ICN), Elizabeth Escobar (MUSM), Hugo López-Arévalo (ICN), Julian Lozano-Flores (IavH), Juliana Gualda de Barros (MZUSP), Karina Rebelo (ZUEC), Mercedes Molina (MUSM), Monique Pereira (UFES), Nancy Simmons (AMNH), Santiago Burneo (QCAZ), Victor Pacheco (MUSM), and Victor Hugo Serrano (UIS).
References
- 1.
Simmons NB. Mammal species of the world: a taxonomic and geographic reference. Order Chiroptera. 3rd edition. Baltimore: Johns Hopkins University Press; 2005.
- 2. Lim BK, Engstrom MD, Reid FA, Simmons NB, Voss RS, Fleck DW. A new species of Peropteryx (Chiroptera: Emballonuridae) from western Amazonia with comments on phylogenetic relationships within the genus. American Museum Novitates. 2010; 3686: 1–20.
- 3.
Vaughan TA, Ryan JM, Czaplewski NJ. Mammalogy. 6th edition. Massachusetts: Jones & Bartlett Publishers; 2013.
- 4.
Ghanem SJ, Voigt CC. Increasing awareness of ecosystem services provided by bats. In: Brockmann HJ, Roper TJ, Naguib M, Mitani JC, editors. Advances in the Study of Behavior Amsterdam: Academic Press; 2013. pp. 279–302.
- 5. Russo D, Bosso L, Ancillotto L. Novel perspectives on bat insectivory highlight the value of this ecosystem service in farmland: Research frontiers and management implications. Agriculture, Ecosystems & Environment. 2018; 266: 31–38.
- 6. Kunz TH, Braun de Torrez E, Bauer D, Lobova T, Fleming TH. Ecosystem services provided by bats. Annals of the New York academy of sciences. 2011; 1223: 1–38. pmid:21449963
- 7.
Peters W. Ober die zu den Gattungen Mimon und Saccopterix gehorigen Flederthiere. Monatsber. Preuss. Akad. Wiss. Berlin. 1867; 470–478.
- 8. Plumpton D, Jones J. Rhynchonycteris naso. Mammalian Species. 1992; 413: 1–5.
- 9.
Emmons LH, Feer F. Neotropical rainforest mammals: a field guide. 2nd edition. Chicago: The Chicago University Press; 1997.
- 10.
Hood CS, Gardner AL. Mammals of South America. Volume I: Marsupials, Xenarthrans, Shrews, and Bats. Family Emballonuridae. Chicago: The University of Chicago Press; 2008.
- 11. Goodwin GG. Mammals of Costa Rica. Bulletin of the American Museum of Natural History. 1946; 87: 271–473.
- 12. Goodwin GG. Mammals of Honduras. Bulletin of the American Museum of Natural History. 1942; 79: 107–195.
- 13.
Lim BK, Miller B. Rhynchonycteris naso. The IUCN Red List of Threatened Species 2016: e.T19714A22010818. 2016. http://dx.doi.org/10.2305/IUCN.UK.2016-2.RLTS.T19714A22010818.en. Downloaded on 17 April 2019.
- 14.
Wied-Neuwied APM. Reise nach Brasilien in den Jahren 1815, bis, 1817 (Travels in Brazil in the years 1815, 1816, 1817). London: Printed for H. Colburn & Co; 1820.
- 15. Weksler M. Phylogenetic relationships of oryzomine rodents (Muroidea: Sigmodontinae): separate and combined analyses of morphological and molecular data. Bulletin of the american Museum of natural History. 2006; 296: 1–149.
- 16.
Barghoorn SF. New material of Vespertiliavus Schlosser (Mammalia, Chiroptera) and suggested relationships of emballonurid bats based on cranial morphology. American Museum novitates. 1977; 2618.
- 17. Robbins LW, Sarich VM. Evolutionary relationships in the family Emballonuridae (Chiroptera). Journal of Mammalogy. 1988; 69: 1–13.
- 18. Griffiths TA, Smith AL. Systematics of emballonuroid bats (Chiroptera, Emballonuridae and Rhinopomatidae), based on hyoid morphology. Bulletin of the American Museum of Natural History. 1991; 206: 62–83.
- 19.
Dunlop JM. The Evolution of Behaviour and Ecology in Emballonuridae (Chiroptera). Ph.D. thesis, York University, Toronto. 1998. https://www.collectionscanada.gc.ca/obj/s4/f2/dsk2/ftp02/NQ27289.pdf.
- 20. Lim BK, Engstrom MD, Bickham JW, Patton JC. Molecular phylogeny of New World sheath-tailed bats (Emballonuridae: Diclidurini) based on loci from the four genetic transmission systems in mammals. Biological Journal of the Linnean Society. 2008; 93: 189–209.
- 21. Clare EL, Adams AM, Maya-Simões AZ, Eger JL, Hebert PD, Fenton MB. Diversification and reproductive isolation: cryptic species in the only New World high-duty cycle bat, Pteronotus parnellii. BMC evolutionary biology. 2013; 13: 1–18.
- 22. Barratt EM, Deaville R, Burlan TM, Bruford MW, Jones G, Racey PA, et al. DNA answers the call of pipistrelle bat species. Nature. 1997; 387: 138–139. pmid:9144281
- 23. Mayer F, Helversen OV. Cryptic diversity in European bats. Proceedings of the Royal Society of London. Series B: Biological Sciences. 2001; 268: 1825–1832.
- 24. Molinari J, Bustos XE, Burneo SF, Camacho MA, Moreno SA, Fermin G. A new polytypic species of yellow-shouldered bats, genus Sturnira (Mammalia: Chiroptera: Phyllostomidae), from the Andean and coastal mountain systems of Venezuela and Colombia. Zootaxa. 2017; 4243: 75–96.
- 25. Flanders J, Wei L, Rossiter SJ, Zhang S. Identifying the effects of the Pleistocene on the greater horseshoe bat, Rhinolophus ferrumequinum, in East Asia using ecological niche modelling and phylogenetic analyses. Journal of Biogeography. 2011; 38(3): 439–452.
- 26. Mota-Vargas C, Rojas-Soto OR. Taxonomy and ecological niche modeling: Implications for the conservation of wood partridges (genus Dendrortyx). Journal for Nature Conservation. 2016; 29: 1–13.
- 27. Muñoz J, Cuartas CA. Saccopteryx antioquensis n. sp. (Chiroptera: Emballonuridae) del noroeste de Colombia. Actualidades Biologicas. 2001; 23: 53–61.
- 28. Teeling EC, Springer MS, Madsen O, Bates P, O’brien SJ, Murphy WJ. A molecular phylogeny for bats illuminates biogeography and the fossil record. Science. 2005; 307: 580–584. pmid:15681385
- 29. Edgar RC. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic acids research. 2004; 32: 1792–1797. pmid:15034147
- 30. Kumar S, Stecher G, Li M, Knyaz C, Tamura K. MEGA X: molecular evolutionary genetics analysis across computing platforms. Molecular biology and evolution. 2018; 35: 1547–1549. pmid:29722887
- 31. Minh BQ, Schmidt HA, Chernomor O, Schrempf D, Woodhams MD, Von Haeseler A, et al. IQ-TREE 2: new models and efficient methods for phylogenetic inference in the genomic era. Molecular biology and evolution. 2020; 37(5): 1530–1534. pmid:32011700
- 32. Kalyaanamoorthy S, Minh BQ, Wong TK, Von Haeseler A, Jermiin LS. ModelFinder: fast model selection for accurate phylogenetic estimates. Nature methods. 2017; 14(6): 587–589. pmid:28481363
- 33. Hoang DT, Chernomor O, Von Haeseler A, Minh BQ, Vinh LS. UFBoot2: improving the ultrafast bootstrap approximation. Molecular biology and evolution. 2018; 35(2): 518–522. pmid:29077904
- 34. Ronquist F, Teslenko M, Van Der Mark P, Ayres DL, Darling A, Höhna S, et al. MrBayes 3.2: efficient Bayesian phylogenetic inference and model choice across a large model space. Systematic biology. 2012; 61: 539–542. pmid:22357727
- 35. Guindon S, Gascuel O. A simple, fast and accurate method to estimate large phylogenies by maximum-likelihood. Systematic Biology. 2003; 52: 696–704.
- 36. Darriba D, Taboada GL, Doallo R, Posada D. jModelTest 2: more models, new heuristics and parallel computing. Nature Methods. 2012; 9(8): 772. pmid:22847109
- 37. Akaike H, Petrov PN, Csaki F. Information theory and an extension of the maximum likelihood principle. Second international symposium on information theory Budapest. Akad Kiado. 1973; 267–281.
- 38. Rambaut A, Drummond AJ, Xie D, Baele G, Suchard MA. Posterior summarization in Bayesian phylogenetics using Tracer 1.7. Systematic biology. 2018; 67: 901. pmid:29718447
- 39. Rozas J, Ferrer-Mata A, Sánchez-Delbarrio JC, Guirao-Rico S, Li-Brado P, Ramos-Onsins SE, et al. DnaSP 6: DNA sequence polymorphism analysis of large data sets. Molecular biology and evolution. 2017; 34: 3299–3302. pmid:29029172
- 40. Dupanloup I, Schneider S, Excoffier L. A simulated annealing approach to define the genetic structure of populations. Molecular ecology. 2002; 11: 2571–2581. pmid:12453240
- 41. Excoffier L, Lischer HE. Arlequin suite ver 3.5: a new series of programs to perform population genetics analyses under Linux and Windows. Molecular ecology resources. 2010; 10: 564–567. pmid:21565059
- 42. Bandelt HJ, Forster P, Röhl A. Median-joining networks for inferring intraspecific phylogenies. Molecular biology and evolution. 1999; 16: 37–48. pmid:10331250
- 43. Freeman PW, Lemen CA. Morphometrics of the family Emballonuridae. Bulletin of the American Museum of Natural History. 1992; 206: 54–60.
- 44. Josse J, Husson F. missMDA a package for handling missing values in multivariate data analysis. Journal of Statistical Software. 2016; 70: 1–31.
- 45.
R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. 2019. https://www.R-project.org/.
- 46. Phillips SJ, Anderson RP, Schapire RE. Maximum entropy modeling of species geographic distributions. Ecological modelling. 2006; 190: 231–259.
- 47. Boria RA, Olson LE, Goodman SM, Anderson RP. Spatial filtering to reduce sampling bias can improve the performance of ecological niche models. Ecological Modelling. 2014; 275: 73–77.
- 48. Aiello-Lammens ME, Boria RA, Radosavljevic A, Vilela B, Anderson RP. spThin: an R package for spatial thinning of species occurrence records for use in ecological niche models. Ecography. 2015; 38: 541–545.
- 49. Rodhouse TJ, Vierling KT, Irvine KM. A practical sampling design for acoustic surveys of bats. The Journal of Wildlife Management. 2011; 75: 1094–1102.
- 50. Bader E, Jung K, Kalko EK, Page RA, Rodriguez R, Sattler T. Mobility explains the response of aerial insectivorous bats to anthropogenic habitat change in the Neotropics. Biological Conservation. 2015; 186: 97–106.
- 51. Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A. Very high resolution interpolated climate surfaces for global land areas. International Journal of Climatology: A Journal of the Royal Meteorological Society. 2005; 25: 1965–1978.
- 52. Escobar LE, Lira-Noriega A, Medina-Vogel G, Peterson AT. Potential for spread of the white-nose fungus (Pseudogymnoascus destructans) in the Americas: use of Maxent and NicheA to assure strict model transference. Geospatial health. 2014; 9: 221–229.
- 53. Barve N, Barve V, Jiménez-Valverde A, Lira-Noriega A, Maher SP, Peterson AT, et al. The crucial role of the accessible area in ecological niche modeling and species distribution modeling. Ecological Modelling. 2011; 222: 1810–1819.
- 54. Peterson AT, Papeş M, Soberón J. Rethinking receiver operating characteristic analysis applications in ecological niche modeling. Ecological modelling. 2008; 213: 63–72.
- 55. Anderson RP, Lew D, Peterson AT. Evaluating predictive models of species’ distributions: criteria for selecting optimal models. Ecological modelling. 2003; 162: 211–232.
- 56. Warren DL, Glor RE, Turelli M. ENMTools: a toolbox for comparative studies of environmental niche models. Ecography. 2010; 33: 607–611.
- 57. Owens HL, Campbell LP, Dornak LL, Saupe EE, Barve N, Soberón J, et al. Constraints on interpretation of ecological niche models by limited environmental ranges on calibration areas. Ecological modelling. 2013; 263: 10–18.
- 58. Otto-Bliesner BL, Marshall SJ, Overpeck JT, Miller GH, Hu A. Simulating Arctic climate warmth and ice field retreat in the last interglaciation. Science. 2006; 311: 1751–1753. pmid:16556838
- 59. Brown JL, Hill DJ, Dolan AM, Carnaval AC, Haywood AM. PaleoClim, high spatial resolution paleoclimate surfaces for global land areas. Scientific data. 2018; 5: 1–9.
- 60. Cobos ME, Osorio-Olvera L, Peterson AT. Assessment and representation of variability in ecological niche model predictions. BioRxiv 603100 [Preprint]. 2019. [cited 2021 February]. https://www.biorxiv.org/content/10.1101/603100v1.full.pdf+html.
- 61. Cobos ME, Peterson AT, Barve N, Osorio-Olvera L. kuenm: an R package for detailed development of ecological niche models using Maxent. PeerJ. 2019; 7: e6281.
- 62. Srinivasulu C, Srinivasulu A, Srinivasulu B, Jones G. Integrated approaches to identifying cryptic bat species in areas of high endemism: The case of Rhinolophus andamanensis in the Andaman Islands. PloS one. 2019; 14(10), e0213562.
- 63. Goodman SM, Rakotondramanana CF, Ramasindrazana B, Kearney T, Monadjem A, Schoeman MC, et al. An integrative approach to characterize Malagasy bats of the subfamily Vespertilioninae Gray, 1821, with the description of a new species of Hypsugo. Zoological Journal of the Linnean Society. 2015; 173(4): 988–1018.
- 64. Pavan AC, Marroig G. Integrating multiple evidences in taxonomy: species diversity and phylogeny of mustached bats (Mormoopidae: Pteronotus). Molecular Phylogenetics and Evolution. 2016; 103: 184–198.
- 65. Clare EL, Lim BK, Engstrom MD, Eger JL, Hebert PD. DNA barcoding of Neotropical bats: species identification and discovery within Guyana. Molecular Ecology Notes. 2007; 7: 184–190.
- 66. Clare EL, Lim BK, Fenton MB, Hebert PD. Neotropical bats: estimating species diversity with DNA barcodes. PloS one. 2011; 6: e22648. pmid:21818359
- 67. Martins FM, Ditchfield AD, Meyer D, Morgante JS. Mitochondrial DNA phylogeography reveals marked population structure in the common vampire bat, Desmodus rotundus (Phyllostomidae). Journal of Zoological Systematics and Evolutionary Research. 2007; 45: 372–378.
- 68. Tobe SS, Kitchener AC, Linacre AM. Reconstructing mammalian phylogenies: a detailed comparison of the cytochrome b and cytochrome oxidase subunit I mitochondrial genes. PloS one. 2010; 5: e14156. pmid:21152400
- 69. Gutiérrez EE, Helgen KM. Outdated taxonomy blocks conservation. Nature. 2013; 495: 314.
- 70. Thomson SA, Pyle RL, Ahyong S, Alonso-Zarazaga M, Ammirati J, Araya JF, et al. Taxonomy based on science is necessary for global conservation. PLoS Biol. 2018; 16: e2005075. pmid:29538381
- 71. Baird AB, Braun JK, Engstrom MD, Holbert AC, Huerta MG, Lim BK, et al. Nuclear and mtDNA phylogenetic analyses clarify the evolutionary history of two species of native Hawaiian bats and the taxonomy of Lasiurini (Mammalia: Chiroptera). PloS one. 2017; 12: e0186085. pmid:29020097
- 72. Demos TC, Webala PW, Kerbis Peterhans JC, Goodman SM, Bartonjo M, Patterson BD. Molecular phylogenetics of slit‐faced bats (Chiroptera: Nycteridae) reveal deeply divergent African lineages. Journal of Zoological Systematics and Evolutionary Research. 2019; 57: 1019–1038. pmid:31894177
- 73. Bickford D, Lohman DJ, Sodhi NS, Ng PK, Meier R, Winker K, et al. Cryptic species as a window on diversity and conservation. Trends in ecology & evolution. 2007; 22: 148–155. pmid:17129636
- 74. Gutiérrez-Pinto N, Cuervo AM, Miranda J, Pérez-Emán JL, Brumfield RT, Cadena CD. Non-monophyly and deep genetic differentiation across low elevation barriers in a Neotropical montane bird (Basileuterus tristriatus; Aves: Parulidae). Molecular Phylogenetics and Evolution. 2012; 64: 156–165.
- 75. Castellanos-Morales G, Gámez N, Castillo-Gámez RA, Eguiarte LE. Peripatric speciation of an endemic species driven by Pleistocene climate change: the case of the Mexican prairie dog (Cynomys mexicanus). Molecular Phylogenetics and Evolution. 2016; 94: 171–181.