Poeciliopsis (Cyprinodontiformes: Poeciliidae) is a genus comprised of 25 species of freshwater fishes. Several well-known taxonomic uncertainties exist within the genus, especially in relation to the taxonomic status of Poeciliopsis pleurospilus and P. gracilis. However, to date, no studies have been conducted to specifically address the taxonomic status of these two species. The goal of this study was to examine the taxonomic validity of P. pleurospilus and P. gracilis using genomic data (ddRADseq) in phylogenetic, population genetic, and species delimitation frameworks. Multiple analyses support the recognition of both taxa as distinct species and also permits us to revise their respective distributions. A species delimitation analysis indicates that P. pleurospilus and P. gracilis are distinct species, each of which consists of two distinct lineages that are geographically structured. Phylogenetic and population genetic analyses provide clear evidence that individuals of P. gracilis are distributed north and west of the Isthmus of Tehuantepec in both Pacific and Atlantic river systems in Mexico, whereas individuals of P. pleurospilus are distributed in both Atlantic and Pacific river systems south and east of the Isthmus of Tehuantepec, from southern Mexico to Honduras.
Citation: Ward SJ, McMahan CD, Khakurel B, Wright AM, Piller KR (2022) Genomic data support the taxonomic validity of Middle American livebearers Poeciliopsis gracilis and Poeciliopsis pleurospilus (Cyprinodontiformes: Poeciliidae). PLoS ONE 17(1): e0262687. https://doi.org/10.1371/journal.pone.0262687
Editor: Windsor E. Aguirre, DePaul University, UNITED STATES
Received: July 29, 2021; Accepted: January 2, 2022; Published: January 31, 2022
Copyright: © 2022 Ward 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: Data Dryad accession (doi:10.5061/dryad.mkkwh711m).
Funding: This study was supported by a Theodore Roosevelt Memorial Grant to Sarah J. Ward from the American Museum of Natural History and an Edward Schlieder Professorship to Kyle R. Piller. April M. Wright and Basanta Khakurel were supported by the National Institute of General Medical Sciences of the National Institutes of Health under grant number P2O GM103424-20 and NSF 2113424. 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.
Middle America is a biologically and geologically diverse region that includes Mexico and the seven countries of Central America as well as the islands of the Greater Antilles [1, 2]. Various volcanic and tectonic events have shaped this dynamic landscape and have resulted in the formation of many geographic barriers, which have, in turn, affected the biodiversity of the region [3, 4]. For example, the closure of the narrow land-bridge at the Isthmus of Panama during the Pliocene had direct consequences for global ocean currents, thereby affecting global atmospheric circulation patterns and altering the tropical ecosystems of the region [5–9]. Another example is the formation of the Trans-Mexican Volcanic Belt (TMVB) beginning in the early Miocene and continuing into the present day, with large changes in topography occurring over the past 3 Myr, including the formation of some presently-active volcanos . The TMVB has been demonstrated to be a substantial barrier to gene flow in populations of freshwater fishes in Mexico [11–13]. Additionally, various events such as stream capture, artificial modifications to stream flow, and sea-level rise have disconnected streams and directly limited or enhanced the dispersal ability of freshwater fishes [14–22]. These issues have been highlighted by studies of freshwater fish diversity, which have played an important role in understanding the biogeographic history of Middle America [21–29].
One of the most ubiquitous groups of freshwater fishes in Middle America is the Poeciliidae , a family of New World cyprinodontiform fishes . Poeciliids have been broadly studied in vicariance biogeography due to parallel patterns of distributions being noted for some fishes, reptiles, and plants . In addition, much is known about the phylogenetic relationships within Poeciliidae [21, 32–38]. The family currently consists of 276 species in 27 recognized genera [34, 39, 40]. One of the more diverse genera in the family, Poeciliopsis , consists of 25 species [36, 42]. The genus inhabits a wide range of environments from lowlands to highlands, in rivers and streams, lakes, and springs, and it is primarily distributed in Pacific-slope drainages of Middle America [26, 43–45], but several species have been documented in Atlantic basin streams [44–46]. Although the genus has been heavily studied from biogeographic and phylogenetic perspectives [40, 42, 46–50], many species have been excluded from phylogenetic analyses. These exclusions are due to several well-known taxonomic uncertainties within Poeciliopsis , especially in relation to the status of Poeciliopsis gracilis  and P. pleurospilus .
Poeciliopsis gracilis was originally described as Xiphophorus gracilis from the Río Orizaba on the Atlantic Slope of Veracruz, Mexico . Historically, P. gracilis has been treated as a synonym of P. pleurospilus [43, 53, 54]; however, Miller et al.  identified a distinction between the species based on the separation of their native distributions as well as variation in pigmentation. Individuals of P. gracilis are recognized by lateral markings that are small spots no larger than the diameter of the pupil, or more frequently, horizontal dashes that are sometimes doubled or fused (Fig 1A) . Native populations of P. gracilis are hypothesized to be restricted to the Atlantic-slope rivers in the states of Veracruz and Oaxaca, Mexico, as well as having been introduced in the Ríos Pánuco (Atlantic) and Balsas (Pacific) basins within Mexico .
A) Female (upper) and male (lower) specimens of P. gracilis from the Río Ajal in Oaxaca, Mexico. B) Female (upper) and male (lower) specimens of P. pleurospilus from the Río Ostuta in Oaxaca, Mexico.
Poeciliopsis pleurospilus was originally described as Girardinus pleurospilus from Lago Dueñas in Guatemala (Pacific Basin). Individuals of P. pleurospilus are recognized by lateral markings that are large, dark, oval spots, which are larger than the diameter of the pupil, in addition to sometimes possessing one to three crescent-shaped bars along the lateral flank of the body (Fig 1B) . In the current literature, native populations of P. pleurospilus are hypothesized to inhabit Pacific-slope systems from Mexico to Honduras, as well as Atlantic-slope systems including the Río Grijalva (Mexico and Guatemala), upper Río Motagua (Guatemala), and Río Ulua (Honduras) [26, 45, 55].
Due to the fact that P. gracilis and P. pleurospilus are morphologically and ecologically similar species, most phylogenetic studies have only included one or the other, but not both [11, 42, 46, 56]. To date, no studies have been conducted to specifically address the taxonomic and phylogenetic status of these two species. The goal of this study was to examine the taxonomic validity of P. gracilis and P. pleurospilus using genomic-scale DNA sequence data. We tested the monophyly of each species by collecting double digest restriction-site associated DNA sequencing data (ddRADseq)  from multiple populations and localities across their putative ranges. The ddRADseq data provided a robust multilocus data set from which we were able to perform Maximum Likelihood (ML) phylogenetic inference, which allowed us to infer the evolutionary relationships among populations of each species sampled across their respective ranges in Mexico, Guatemala, El Salvador, and Honduras [26, 45, 55, 58]. We also performed several population genetic analyses to examine the amount of genetic diversity and genetic structure within and among the recovered lineages. In addition, we conducted a species delimitation analysis to inform our conclusions on diversity among the various populations. The results from this study provide much-needed insight into the long-standing taxonomic issue of the taxonomic distinctiveness of these two morphologically similar congeners.
Materials and methods
One-hundred forty-eight individuals were used in this study after filtering samples for low call quality, including 134 individuals currently recognized as either P. gracilis or P. pleurospilus (S1 Table), following the characters presented in Miller et. al. . Thirteen samples from three other species of Poeciliopsis (P. fasciata, n = 5; P. infans, n = 4; P. turrubarensis, n = 4), and a single outgroup Brachyrhaphis rhabdophora, also were included in the study (S2 Table). Tissue samples (fin clips) were collected from various localities in Oaxaca and Veracruz, Mexico in November, 2019 using a standard 10’ x 6’ seine, and immediately preserved in 95% ethanol (approved by Southeastern Louisiana University IACUC Committee Protocol #0074). Additional specimens from river systems in Mexico, Guatemala, El Salvador, and Honduras were obtained from the tissue collections at Southeastern Louisiana University (SLU), the Field Museum of Natural History (FMNH), Louisiana State University Museum of Natural Science (LSUMZ), and the Universidad de Ciencias y Artes de Chiapas Museo de Zoología (UNICACH). Specific locality information for each specimen currently recognized as P. gracilis and P. pleurospilus is included in S1 Table.
DNA was extracted from fin clips using the Qiagen DNeasy Tissue Extraction Kit following the manufacturer’s recommendations. Three double digest restriction enzyme DNA (ddRAD) libraries (batch information available in S3 Table) were prepared following a modified version (S1 Appendix) of the protocol from Peterson et al. . Each sample was digested with MspI and PstI restriction enzymes and ligated to common (5’- GTGACTGGAGTTCAGACGTGTGCTCTTCCGATCT—3’) and unique oligos (S3 Table).
We then used a BluePippin machine to size select for 300–500 bp fragments, then libraries were sent to the University of Oregon’s Genomic and Cell Characterization Core Facility (GC3F) for Illumina Sequencing on the Hiseq 4000 for 100 bp single-end reads.
The raw data files returned from GC3F were run through FastQC v0.11.3  to check the overall quality of the reads from the Illumina run. The FastQ file output from the previous step was input into the ipyrad  pipeline for assembly and initial filtering (parameters in S4 Table). Reads that contained more than 5 bases with a low-quality Phred score (<33) were excluded. Reads were then clustered based on an 85% similarity threshold and reads with less than 6x coverage were filtered out. A maximum of 5 ambiguous base calls and 5 heterozygous sites per read were allowed during filtering. Summary statistics of the ipyrad assembly are available in S2 Appendix.
Additional filtering using VCFtools  excluded individuals with more than 95% missing data, and single nucleotide polymorphism (SNP) loci with a 60% call rate or lower. We retained one dataset that consisted of all specimens that was used for our phylogenetic inference. A separate dataset that consisted only of ingroup individuals was used for the population genetic analyses. Aligned datasets, tree file, laboratory protocols, and bioinformatic workflow available in a Data Dryad accession (doi:10.5061/dryad.mkkwh711m).
Concatenated ddRAD loci were analyzed under Maximum Likelihood using IQ-TREEv1.6.12 . The filtered VCF file was converted to FASTA format using PGDSpider v126.96.36.199  and input into Model Finder [57, 64] within IQ-TREE to compute the log-likelihood of an initial parsimony tree for many different models under the Akaike information criterion (AIC), corrected Akaike information criterion (AICc), and the Bayesian information criterion (BIC). The AIC, AICc, and BIC output all selected GTR+F+R2 as the best fit model for our data. Ultrafast bootstraps were run for 10,000 generations and a Shimodaira-Hasegawa-like approximate likelihood ratio test (SH-aLRT) was run for 1,000 replicates in IQ-TREE and the resulting tree was visualized using FigTree v1.4.4 .
We used VCFtools to exclude individuals with more than 95% missing data, loci (SNPs) with a 60% call rate or lower, along with excluding all outgroup individuals to produce a VCF file of raw reads for 134 samples of P. pleurospilus and P. gracilis. We then used RStudio 4.0.0  to attach relevant population information and conduct population genetic analyses using the R package adegenet v 2.0.0 . The first analysis was a Discriminant Analysis of Principal Components (DAPC)  following the methods of Grünwald et al.  and the vignette provided by Jombart and Collins 2015 . We defined groups a priori and determined the optimal number of Principal Components (PCs) to retain using the a-score metric (S1 Fig) built into adegenet , which led us to include six PCs and three discriminant functions. We used the posterior assignment of each sample assigned within the DAPC object to visualize a composite stacked bar plot following the methods of Grünwald et al. . Finally, we calculated an unweighted Weir and Cockerham  pairwise FST between clades of ingroup individuals using VCFtools. Bioinformatic pipeline for all phylogenetic and population genetic analyses available in S3 Appendix.
Species delimitation analyses
In order to assign a marginal probability to our hypotheses of species delimitation, we used the software BPP [72, 73]. We performed this analysis on the concatenated SNP file obtained in the Bioinformatics section above, with one outgroup sequence, Brachyrhaphis rhabdophora. BPP uses the multi-species coalescent to estimate divergence times and population sizes for both extant and ancestral sequences. This information is used to derive the probability that different groups of organisms are actually separate populations. We used the algorithm referred to as “A10”, in which the researcher provides a guide species tree to be tested. We tested three population structure hypotheses including: 1) the one-species model, in which all in-group taxa are one species, 2) the two-species model in which the PM + AM populations are one species (P. gracilis) and the PSM + MCA populations are a separate species (P. pleurospilus) and 3) the four-species model, in which PM, AM, PSM, MCA are all separate species.
In order to estimate a marginal probability for the data under each model (i.e., one not biased by MCMC hill-climbing), we used the stepping stone method . We did this following the stepping stone procedure described in Rannala and Yang  as implemented in the bppr R package . Each stepping stone for each of the three models was run to convergence, as assessed in the bppr package, and Bayes Factor model comparisons were performed in bppr.
The full dataset including outgroups retained 178,130 out of a possible 1,817,209 SNPs. Additional read information is available in S4 Table. This dataset was used to infer a Maximum Likelihood phylogenetic tree under the GTR+F+R2 model. A collapsed cladogram rooted with Brachyrhaphis rhabdophora inferred seven major clades, four of which represented ingroup taxa (Fig 2). We found that the populations represented in the four major ingroup clades of our Maximum Likelihood inference (Fig 2) clustered into cohesive, allopatric groups that were geographically separate from other such groups. Specifically, we identified a ‘Pacific Mexico’ (PM) clade consisting of samples from populations from the Pacific-draining basins of southwestern Mexico; an ‘Atlantic Mexico’ (AM) clade from the Atlantic-draining basins of eastern Mexico; a ‘Pacific Southern Mexico’ (PSM) clade comprised of individuals collected from the Río Ostuta of southern Mexico; and an ‘Mexico and Central America’ (MCA) clade of individuals from Mexican drainage systems south of the Isthmus of Tehuantepec (including the Río Grijalva and nearby basins) and populations extending southward into El Salvador, Guatemala, and Honduras in Central America (Fig 3). Additional information on sampling localities is provided in S1 Table. Tree file with SH-aLRT and ultrafast bootstrap values available in supplementary material.
Collapsed ddRADseq cladogram with SH-aLRT (left) and ultrafast bootstrap (right) values. Colorized clades indicate P. gracilis (PM; AM) and P. pleurospilus (PSM; MCA).
Red circles represent Pacific Mexico (PM); orange circles represent Atlantic Mexico (AM); green circles represent Pacific Southern Mexico (PSM); and blue circles represent Mexico and Central America (MCA). The white star; represents the type locality of P. gracilis; the black star represents the type locality of P. pleurospilus. Map made in ArcMap 10.8.1 using the World Dark Gray Base map .
Four individuals of Poeciliopsis infans sampled from Manantial Mintzita, Michoacán, Mexico were sister to all other samples of Poeciliopsis included in this study, with 100% ultrafast bootstrap and 100% SH-aLRT support values (Fig 2). Four individuals of P. turrubarensis collected from the Ríos de la Virgen, Tehuantepec, Tequesistlan, and Totolapan, from the Pacific coast of Oaxaca, Mexico and five individuals of P. fasciata collected from the Río Ajal in Oaxaca, Mexico were sister to one another with 100% SH-aLRT and 100% ultrafast bootstrap support and sister to P. gracilis and P. pleurospilus (Fig 2).
Within the four ingroup clades, branch lengths were short and little geographic structure was recovered within any of the four clades, although clear geographic structure was seen among clades (Fig 4). The ultrafast bootstrap values are lower at terminal nodes within the ingroup clades compared to deeper nodes. Deeper splits among the four clades are supported by higher SH-aLRT (99.6–100) and higher ultrafast bootstrap values (91–100 UFBoot), signifying more robust clade assignments.
Red indicates individuals from Pacific basin streams north and west of the Isthmus of Tehuantepec (PM); orange indicates individuals from Atlantic basin streams north and east of the Isthmus of Tehuantepec (AM); green indicates individuals from a Pacific basin stream (Río Ostuta) in Oaxaca, Mexico (PSM); and blue indicates individuals from both Pacific and Atlantic streams south of the Isthmus of Tehuantepec in Mexico, Guatemala, El Salvador, and Honduras (MCA). Branch support is indicated by SH-aLRT (left) and ultrafast bootstrap (right) values. Individuals identified in the field as P. gracilis based on pigmentation patterns are indicated by a “G” in the tip label.
The PM clade consisted exclusively of specimens collected from Pacific-slope river systems (n = 64) (Fig 4). Individuals identified in the field as P. pleurospilus were collected from six localities in Oaxaca, Mexico, including the Ríos Coapa, de la Virgen, Octlan, Tehuantepec, Tequesistlan, and Totolapan. Also included in this clade were individuals identified in the field as P. gracilis that were collected from four localities in Oaxaca, Mexico, including the Ríos Huatulco, Las Flores, and Puente Tierra Azul, and Flor de Café; as well as an unnamed stream in Guerrero, Mexico. Branch lengths were extremely short within this clade, with ultrafast bootstrap support values ranging from 19%– 100% and SH-aLRT values ranging from 3.4–100%, indicating varying degrees of phylogenetic resolution in the data.
The AM clade consisted exclusively of specimens collected from Atlantic basin river systems (n = 44) (Fig 4) and were identified in the field as P. gracilis. Specimens were obtained from seven localities in Oaxaca, Mexico including three specimens from an unnamed arroyo, and specimens from the Ríos Ajal, Amapa, Barranca, Papaloapan, and two localities on the Río Tolosita; as well as from three localities in Veracruz, Mexico including the Ríos Actopan, Blanco, and Jaltepec. The type locality of P. gracilis is the Río Orizaba in Veracruz, Mexico (indicated by a white star in Fig 3), which is a tributary to the Río Blanco, indicating that this clade most likely represents individuals of topotypic P. gracilis. Ultrafast bootstrap support values ranged from 43% - 100% and SH-aLRT values ranged from 17.9–100%, indicating varying degrees of phylogenetic resolution within the clade.
The PSM clade consisted exclusively of specimens collected from the Río Ostuta (n = 13) in Oaxaca, Mexico (Pacific basin) (Fig 4). All individuals were identified in the field as P. pleurospilus. Branch lengths were short in this clade, and the bootstrap support values ranged from 33% - 100% and SH-aLRT values ranged from 27.5–100%, indicating varying degrees of phylogenetic resolution.
The MCA clade consisted of specimens collected from both the Pacific (n = 9) and Atlantic (n = 4) basin river systems (Fig 4). The Pacific basin specimens included individuals from Lago Güija and Laguna Metapán in El Salvador; individuals from the Río Achiguate and Río Negro in Guatemala; individuals from the Ríos Choluteca and Nacaome in Honduras; and the Río Margaritas in Chiapas, Mexico. The Atlantic basin specimens included individuals from the Ríos Amarillo and Tepemechin in Honduras; and the Ríos Ojo de Agua and Paso Hondo, in Chiapas, Mexico. The type locality of P. pleurospilus is Lago Dueñas on the Pacific slope of Guatemala. This lake no longer exists; however, the historical location of the lake is indicated by a black star (Fig 3). The historical location of Lago Dueñas is just under 12 km from the headwaters of the Río Achiguate, indicating that this clade most likely represents individuals of topotypic P. pleurospilus. Ultrafast bootstrap support values ranged from 87% - 100% and SH-aLRT values ranged from 70.2–100%, indicating relatively robust phylogenetic resolution within the clade.
Various population genetic metrics were used to assess genetic variability among individuals and clusters of P. gracilis and P. pleurospilus. For these analyses we used a second dataset that excluded outgroup specimens. This dataset retained 213,593 out of a possible 1,817,209 sites. The first method employed was a DAPC, which recovered four distinct clusters. The 95% confidence ellipses showed no overlap among DAPC clusters, with the clustering showing varying distances from one another in the DAPC space (Fig 5A). The MCA and PSM clusters corresponded to samples collected south and east of the Isthmus of Tehuantepec and were the same individuals obtained in the MCA and PSM clades of our Maximum Likelihood phylogeny. The AM and PM clusters corresponded to samples collected from the Isthmus of Tehuantepec and north and east, and clustered more closely with one another than with either of the other populations.
A) Discriminant Analysis of Principal Components (DAPC) of populations. Ellipses represent 95% confidence intervals. B) Composite Stacked Bar Plot of all individuals from the ingroup dataset. In panels A and B, inferred genetic clusters of individuals were correlated with our phylogenetic results (Figs 2 and 4) and assigned colors of the corresponding clade, as follows: PM cluster, red; AM cluster, orange; PSM cluster, green; MCA cluster, blue.
We then used the DAPC object to visualize the population assignment of each sample and created a composite stacked bar plot (Fig 5B). In a stacked bar plot, the probability of population membership is illustrated on the y-axis, from 0 to 100% probability of belonging to a specific population. The x-axis contains a bin of each sample. Colors represent the pre-assigned populations by clade, whereas the groupings represent the populations inferred from the DAPC object. The composite stacked bar plots also recover the four distinct clusters revealed by the DAPC analysis. The PM and AM clusters are completely pure, with no admixture with any of the other clusters. The PSM cluster showed one individual admixed with the PM cluster and the MCA cluster showed one individual admixed with the PSM cluster.
An unweighted Weir & Cockerham pairwise FST was calculated between clusters of ingroup individuals, where values closer to zero represent little to no genetic differentiation and values of one represent fixation of populations. The pairwise FST values calculated between the four major clusters ranged from 0.1125 to 0.24835, with the lowest pairwise FST being between the MCA and PSM clades and the highest being between the MCA and PM clades (Table 1).
Using BPP [72, 73], we recovered support for P. gracilis and P. pleurospilus as two separate species. The marginal likelihood of the data under the single-species model was -1,249,746 (Fig 6). The marginal probability of the data under the two-species model was -1,249,643, which per Bayes Factor comparison was significantly better than the one-species model (BF = 1.000). The marginal probability of the data under the four-species model was -1,249,656, which was significantly worse than the two-species model (BF = -1.000).
Marginal likelihoods of the number of species in the ingroup sample calculated using the software BPP, and the R package bppr. Each point represents the mean log-likelihood calculated per stepping stone in BPP. The annotated likelihood in the lower right corner indicates the total marginal probability of the data. We used Bayes factor analysis to compare these marginal probabilities, concluding that the two-species model was more probable than the one-species model (BF = 1.00) and the four-species model (BF = 1.00).
The family Poeciliidae represents one of the most ubiquitous groups of freshwater fishes in Middle America . Species in the family have been heavily studied from both taxonomic and phylogenetic perspectives [21, 34, 36, 42, 46], using morphological and mitochondrial DNA data sets. Despite this flurry of activity, there remains much taxonomic uncertainty within the family. In many groups of organisms, inclusion of genomic sequence data has proven to be an informative approach to resolve relationships among taxa and identify evolutionary lineages among taxonomically difficult or unstable groups and recently diverged lineages [77–79]. Therefore, the overall goal of this study was to examine the taxonomic validity of P. pleurospilus and P. gracilis using analyses of genomic-scale sequence data in both phylogenetic and population genetic frameworks.
In the current literature, native populations of P. pleurospilus are hypothesized to inhabit Pacific Basin systems from Mexico to Honduras, as well as Atlantic Basin systems including the Río Grijalva (Mexico), upper Río Motagua (Guatemala), and Río Ulua basins (Honduras) [26, 45, 55]. Native populations of P. gracilis are hypothesized to be restricted to the Atlantic Basin in the states of Veracruz and Oaxaca, Mexico, as well as having been introduced in the Rios Pánuco (Atlantic) and Balsas (Pacific) within Mexico . Based on the results of our phylogenetic, population genetic, and species delimitation analyses, we propose a distributional revision. Poeciliopsis gracilis consists of two lineages (AM and PM groups/clades) and their recognized distribution should be restricted to Atlantic basin streams from the Isthmus of Tehuantepec in Mexico and north in the states of Oaxaca and Veracruz and in Pacific basin streams from the Isthmus of Tehuantepec in Mexico, and north and west in the states of Oaxaca and Guerrero. Sister to these were a clade of P. pleurospilus, also with two lineages (PSM and MCA groups/clades). One lineage (PSM) was restricted to the Río Ostuta, a Pacific basin stream from Oaxaca, Mexico. The other lineage of P. pleurospilus (MCA) consists of individuals from both Pacific and Atlantic streams south and east of the Isthmus of Tehuantepec, from Chiapas, Mexico to Honduras, and include all specimens from all rivers in Guatemala and El Salvador.
As stated earlier, P. gracilis was described from the Río Orizaba on the Atlantic Slope of Veracruz, Mexico . The Río Orizaba is a tributary to the Rio Blanco, which is the sampling location of nine individuals used in this study. These samples grouped in the AM cluster/clade in the results of all of our phylogenetic and population genetic analyses, indicating that this cluster/clade represent topotypic P. gracilis. The other samples that grouped in this cluster/clade were all north and west of the Isthmus of Tehuantepec, on both the Pacific and Atlantic slopes of Mexico. The samples from the AM clade were sister to the PM clade with 91% ultrafast bootstrap support and 99.6% SH-aLRT support. The population genetic analyses both obtained two distinct clusters north and west of the Isthmus of Tehuantepec.
Poeciliopsis pleurospilus was originally described from Lago Dueñas on the Pacific slope of Guatemala . The historical location of this lake is approximately 12 km from the headwaters of the Río Achiguate, which was the sampling location of two samples used in this study. These samples grouped in the MCA cluster/clade in all phylogenetic and population genetic analyses, indicating that this cluster/clade represents individuals of topotypic P. pleurospilus. The other samples that were sister to this clade in the phylogenetic analysis were all south of the Isthmus of Tehuantepec, in southern Mexico, El Salvador, Guatemala, and Honduras. The samples from the PSM cluster/clade consistently were sister to the MCA clade in all of the phylogenetic and population genetic analyses, with the phylogenetic analysis having 92% ultrafast bootstrap support and 100% SH-aLRT support values for these two clades. The population genetic analyses also support the existence of two distinct clusters within this species.
Despite the fact that both P. gracilis and P. pleurospilus were described more than 150 years ago [51, 52], these two species have long remained taxonomically ambiguous and multiple reasons account for this. First, no comprehensive morphological or molecular study of either P. gracilis or P. pleurospilus has been published to date. Miller et al. 2005  describes different pigmentation patterns to differentiate these species, but our observations suggested that these may not be reliable indicators. Additional research is needed to quantify pigmentation patterns between P. gracilis and P. pleurospilus. Although both species are widespread and abundant, the lack of studies attempting to refute or support their taxonomic validity has led to continued taxonomic uncertainty for these species. Next, throughout Middle America, non-native introductions of freshwater fishes have been rampant, further leading to taxonomic uncertainty. For example, several poeciliids (i.e., Gambusia, Poecilia, and Xiphophorus spp.) have been introduced for mosquito control and have successfully colonized over 40 countries [53, 79]. Additionally, poeciliids have been incidentally released during the stocking of tilapia (Oreochromis and Tilapia spp.), which have been introduced into multiple water bodies throughout Mexico and Central America [29, 80]. Native poeciliids often inhabit stock ponds with tilapia and then are inadvertently introduced into other drainage basins as tilapia are stocked and their presence often confounds the process of making robust taxonomic identifications in the field(K. Piller and C. McMahan, pers. obs.).
Finally, an additional regional aspect to consider is the active history of inter-basin hydrological exchanges from headwater stream capture or across the flood plains of lowlands in Middle America. The area near the Isthmus of Tehuantepec represents one of the lowest elevation points between the Atlantic and Pacific basins in the New World. Many of the headwater reaches of Atlantic and Pacific basin streams come into close proximity in the region, and this situation may have permitted inter-basin exchanges of populations during rainy seasons or flooding events, from modern to recent geological times. In addition, it is well known that some species of freshwater fishes can tolerate low salinities and are therefore able to disperse along the flood plains of lowland regions during ecologically appropriate time periods [15, 17] and this is particularly true for some species of poeciliids, which can often be found in low salinity habitats . All of these factors together increase the likelihood that the neotropical freshwater ichthyofauna may have much wider distributions than originally proposed, especially for generalist species such as Poeciliopsis, and leaves room for distributional revisions for many other taxa as new data are gathered.
The results of this study are consistent with multiple other studies that have demonstrated the Isthmus of Tehuantepec to be an important biogeographical region where major changes occur in the distributional patterns of many groups [14, 82–86]. These distributional changes may stem in part from the fact that the Isthmus of Tehuantepec represents a geologically complex zone that has been subjected to various tectonic events, which in turn have changed the environmental conditions available to organisms, such as sea-level changes connecting or isolating various aquatic systems . In addition to being a region that may have facilitated headwater river capture events during high-water periods or under changing erosional conditions, the Isthmus may also have acted as a geologic barrier that separated P. gracilis and P. pleurospilus populations, as has been demonstrated in a variety of other taxa [83, 87–89].
Poeciliopsis gracilis and P. pleurospilus are morphologically and ecologically similar species of live-bearing freshwater fishes, which have had a widely debated taxonomic history and geographic distribution. This study presented comprehensive genomic evidence that was analyzed in both a phylogenetic and population genetic framework to shed light on the taxonomic status of both species. The results from phylogenetic, population genetic, and species delimitation analyses showed clear evidence that individuals of P. gracilis are distributed from the Isthmus and Tehuantepec and north in Atlantic basin systems in Mexico, whereas individuals of P. pleurospilus are distributed in both Atlantic and Pacific basin systems south and east of the Isthmus of Tehuantepec, from southern Mexico to Honduras.
S1 Appendix. Laboratory protocol.
S2 Appendix. Ipyrad summary statistics.
S3 Appendix. Bioinformatic workflow.
S4 Appendix. Treefile of phylogenetic inference.
S1 Fig. A-score optimization–spline interpolation.
S1 Table. Sampling localities of ingroup individuals.
S2 Table. Sampling localities of outgroup individuals.
S3 Table. Barcodes and batch information.
S4 Table. Parameters used for assembly of concatenated dataset.
We would like to thank Dr. Wilfredo Matamoros at the Universidad de Ciencias y Artes de Chiapas Museo de Zoología (UNICACH) and Dr. Prosanta Chakrabarty at the Louisiana State University Museum of Natural Science (LSUMZ) for providing tissue samples. In addition, we would like to thank Kearstin Findley and Pablo Gesundheit for field assistance.
- 1. Winker K. Middle America, not Mesoamerica, is the accurate term for biogeography. Condor. 2011; 113: 5–6.
- 2. Matamoros WA, McMahan CD, Chakrabarty P, Albert JS, Schaefer JF. Derivation of the freshwater fish fauna of Central America revisited: Myers’s hypothesis in the twenty-first century. Cladistics. 2015; 31: 177–188. pmid:34758579
Mann P. Overview of the tectonic history of northern Central America. In: Mann P, editor. Geologic and tectonic development of the Caribbean plate boundary in northern Central America. Geological Society of America Special Paper 428. 2007; 1–19. https://doi.org/10.1130/2007.2428(01)
Marshall JS. The geomorphology and physiographic provinces of Central America. In: Bundschuh J and Alvarado GE, editors. Central America: geology, resources, and hazards. London: Taylor and Francis Group. 2007; 1–51.
Schmidt DN. The closure history of the Central American seaway: evidence from isotopes and fossils to models and molecules. In: Williams M, Haywood AM, Gregory FJ, Schmidt DN, editors. Deep-time perspectives on climate change: marrying the signal from computer models and biological proxies. Geo Soc. 2007; 427–442.
- 6. O’Dea A, Lessios HA, Coates AG, Eytan RI, Restrepo-Moreno SA, Cione AL, et al. Formation of the Isthmus of Panama. Science Advances. 2016; 2; e1600883. pmid:27540590
- 7. Bacon CD, Silvestro D, Jaramillo C, Smith BT, Chakrabarty P, Antonelli A. Biological evidence supports an early and complex emergence of the Isthmus of Panama. National Academy of Sciences. 2015; 112; 6110–6115.
Coates AG, Obando JA. 1996. The geologic evolution of the Central American Isthmus. In: Evolution and Environment in Tropical America (eds Jackson J. B. C., Budd A. F. and Coates A. G.), pp. 21–56. University of Chicago Press, Chicago. 1996.
- 9. Coates AG, Collis LS, Aubry MP, Berggren WA. The geology of the Darien, Panama, and the late Miocene-Pliocene collision of the Panama arc with northwestern South America. Geol Soc Am Bull. 2004; 116; 1327–1344.
- 10. Mastretta-Yanes A, Moreno-Letelier A, Piñero D, Jorgensen TH, Emerson BC. Biodiversity in the Mexican highlands and the interaction of geology, geography and climate within the Trans-Mexican Volcanic Belt. J Biogeogr. 2015; 42; 1586–1600.
- 11. Mateos M. Comparative phylogeography of livebearing fishes in the genera Poeciliopsis and Poecilia (Poeciliidae: Cyprinodontiformes) in central Mexico. J Biogeogr. 2005; 32: 775–780.
- 12. Pérez-Rodríguez R, Domínguez-Domínguez O, Pérez-Ponce de León G, Doadrio I. Phylogenetic relationships and biogeography of the genus Algansea Girard (Cypriniformes: Cyprinidae) of central México inferred from molecular data. BMC Evol Biol. 2009; 223: 1–28. pmid:19735558
- 13. Zúñiga-Vega JJ, Ingley SJ, Unmack PJ, Johnson JB. Do freshwater ecoregions and continental shelf width predict patterns of historical gene flow in the freshwater fish Poecilia butleri? Biol J Linn Soc Lond. 2014; 112: 399–441.
- 14. Miller RR. Geographic distribution of Central American freshwater fishes. Copeia. 1966; 4: 773–802.
- 15. Myers GS. Derivation of the freshwater fish fauna of Central America. Copeia. 1966; 4: 766–773.
- 16. Rosen DE. Vicariance model of Caribbean biogeography. Syst Zool. 1975; 24: 431–64.
- 17. Briggs JC. Freshwater fishes and biogeography of Central America and the Antilles. Syst Zool. 1984; 33: 428–435.
Bussing WA. Patterns of the distribution of the Central American ichthyofauna. In: Stehli FG, Webb SD, editors. The Great American Biotic Interchange. Plenum Press, New York. 1985: 453–473.
- 19. Rauchenberger M. Historical biogeography of poeciliid fishes in the Caribbean. Syst Zool. 1988; 37: 356–365.
Albert JS, Reis RE. Historical biogeography of neotropical freshwater fishes. Berkeley, California: University of California Press. 2011.
- 21. Bagley JC, Alda F, Breitman MF, Bermingham E, van den Berghe EP, Johnson JB. Assessing species boundaries using multilocus species delimitation in a morphologically conserved group of neotropical freshwater fishes, the Poecilia sphenops species complex (Poeciliidae). PLoS One. 2015;10: e0121139. pmid:25849959
- 22. McMahan CD, Ginger L, Cage M, David K, Chakrabarty P, Johnston M, et al. Pleistocene to Holocene expansion of the black-belt cichlid in Central America, Vieja maculicauda (Teleostei: Cichlidae). PLoS One. 2017;12: e0178439. pmid:28558052
Lundberg JG, Marshall LG, Guerrero J, Horton B, Malabarba MCSL, Wesselingh F. The stage for neotropical fish diversification: a history of tropical South American rivers. In: Malabarba LR, Reis RE, Vari RP, Lucena ZMS, Lucena CAS, editors. Phylogeny and Classification of Neotropical Fishes. Edipucrs, Porto Alegre; 1998. pp.13–48.
- 24. Arbogast BS, Kenagy GJ. Comparative phylogeography as an integrative approach to historical biogeography. J Biogeogr. 2001; 28: 819–825.
- 25. Perdices A, Bermingham E, Montilla A, Doadrio I. Evolutionary history of the genus Rhamdia (Teleostei: Pimelodidae) in Central America. Mol Phylogenet Evol. 2002; 25:172–189. pmid:12383759
- 26. McMahan CD, Matamoros WA, Alvarez Calderon FS, Yamileth Henriquez W, Recinos HM, Chakrabarty P, et al. Checklist of the inland fishes of El Salvador. Zootaxa. 2013; 3608: 440–456. pmid:24614482
- 27. Albert JS, Crampton WGR, Thorsen DH, Lovejoy NR. Phylogenetic systematics and historical biogeography of the Neotropical electric fish (Teleostei: Gymnotidae). Syst Biodivers. 2014; 2: 375–417.
- 28. Martínez-Aquino A, Ceccarelli FS, Eguiarte LE, Vázquez-Domínguez E, de León GPP. Do the historical biogeography and evolutionary history of the Digenean Margotrema spp. across Central México mirror those of their freshwater fish hosts (Goodeinae)? PLoS One. 2014; 9: e101700. pmid:24999998
- 29. Beltrán-López RG, Domínguez-Domínguez O, Perez-Rodriguez R, Piller KR, Doadrio I. Evolving in the highlands: The case of the Neotropical Lerma live-bearing Poeciliopsis infans (Woodman, 1894) (Cyprinodontiformes: Poeciliidae) in Central México. BMC Evol Biol. 2018; 18: 56. pmid:29678141
- 30. Bonaparte CL. Saggio di una distribuzione metodica degli animali vertebrati. 78 pp. Giornale Arcadico di Scienze Lettere ed Arti, 1831; 52: 155–189.
- 31. Bragança PHN, Amorim PF, Costa WJEM. Pantanodontidae (Teleostei, Cyprinodontiformes), the sister group to all other cyprinodontoid killifishes as inferred by molecular data. Zoosyst Evol. 2018; 94: 137–145.
- 32. Hrbek T, Seckinger J, Meyer A. Molecular phylogeny of the Poeciliidae (Teleostei, Cyprinodontiformes): biogeographic and life-history implications. Mol Phylogenet Evol. 2007; 43: 986–98. pmid:17185005
- 33. Alda F, Reina RG, Doadrio I, Bermingham E. Phylogeny and biogeography of the Poecilia sphenops species complex (Actinopterygii, Poeciliidae) in Central America. Mol Phylogenet Evol. 2013; 66: 1011–1026. pmid:23277160
- 34. Reznick DN, Furness AI, Meredith RW, Springer MS. The origin and biogeographic diversification of fishes in the family Poeciliidae. PLoS One. 2017; 12: e0172546. pmid:28278162
- 35. Barts N, Purpleway R, Passow CN, Arias-Rodriguez L, Kelly JL, Tobler M. Molecular evolution and expression of oxygen transport genes in livebearing fishes (Poeciliidae) from hydrogen sulfide rich springs. Genome. 2018; 61: 273–286. pmid:29227751
- 36. Conway KW, Mateos M, Vrijenhoek RC. A new species of the live-bearing fish genus Poeciliopsis from northern México (Cyprinodontiformes, Poeciliidae). Zookeys. 2019; 88: 91–118. pmid:31719775
- 37. Thomaz AT, Carvalho TP, Malabarba LR, Knowles LL. Geographic distributions, phenotypes, and phylogenetic relationships of Phalloceros (Cyprinodontiformes: Poeciliidae): Insights about diversification among sympatric species pools. Mol Phylogenet Evol. 2019; 132: 265–274. pmid:30528083
- 38. Rodriguez-Silva R, Weaver PF. A new livebearing fish of the genus Limia (Cyprinodontiformes: Poeciliidae) from Lake Miragoane, Haiti. J Fish Biol. 2020; 96: 1360–1369. pmid:32112652
Lucinda PHF. Family Poeciliidae (Livebearers). In: Reis R.E., Kullander S.O., Ferraris C.J.J. (eds). Check List of the Freshwater Fishes of South and Central America. Edipucrs, Porto Alegre; 2003. pp. 555–81.
- 40. van Kruistum H, Guernsey MW, Baker JC, Kloet SL, Groenen MAM, Pollux BJA, et al. The genomes of the livebearing fish species Poeciliopsis retropinna and Poeciliopsis turrubarensis reflect their different reproductive strategies. Mol Biol Evol. 2020; 37: 1376–1386. pmid:31960923
- 41. Regan CT. A revision of the cyprinodont fishes of the subfamily Poeciliinae. Proc Zool Soc Lond B. 1913; 4: 977–1018.
- 42. Mateos M, Sanjur OI, Vrijenhoek RC. Historical biogeography of the livebearing fish genus Poeciliopsis (Poeciliidae: Cyprinodontiformes). Evolution. 2002; 56: 972–84. pmid:12093032
- 43. Rosen DE, Bailey RM. The poeciliid fishes (Cyprinodontiformes), their structure, zoogeography, and systematics. Bull Am Mus Nat Hist. 1963; 126: 1–176.
Bussing WA. Peces de las aguas continentales de Costa Rica/Freshwater Fishes of Costa Rica, 2nd ed. University de Costa Rica, San Jose. 2002.
Miller RR, Minckley WL, Norris SM. Freshwater fishes of México. The University of Chicago Press, Chicago and London. 2005.
- 46. Mateos M, Domínguez-Domínguez O, Varela-Romero A. A multilocus phylogeny of the fish genus Poeciliopsis: solving taxonomic uncertainties and preliminary evidence of reticulation. Ecol Evol. 2018; 9: 1845–1857.
- 47. Lucinda PHF, Reis RE. Systematics of the subfamily Poeciliinae Bonaparte (Cyprinodontiformes: Poeciliidae), with an emphasis on the tribe Cnesterodontini Hubbs. Neotrop Ichthyol. 2005; 3: 1–60.
- 48. Ho ALFC, Pruet CL, Lin J. Phylogeny and biogeography of Poecilia (Cyprinodontiformes: Poeciliinae) across Central and South America based on mitochondrial and nuclear DNA markers. Mol Phylogenet Evol. 2016; 101: 32–45. pmid:27129899
- 49. Weaver PF, Cruz A, Johnson S, Dupin J, Weaver KF. Colonizing the Caribbean: biogeography and evolution of livebearing fishes of the genus Limia (Poeciliidae). J Biogeogr. 2016; 43: 1808–1819.
- 50. Larson AD, Scarborough PL, Sayre CL, Johnson JB. What can phylogeographic breaks in the livebearer Poeciliopsis turrubarensis tell about fish community boundaries in western Costa Rica? J Fish Biol. 2012; 80; 728–736. pmid:22380567
Heckel JJ. Eine neue Gattung von Poecilien mit rochenartigem Anklammerungs-Organe (P. xiphophorus). In Band E., editor. Sitzungsberichte der Kaiserlichen Akademie der Wissenschaften. Mathematisch-Naturwissenschaftliche Classe v. 1 (pt 1–5), Pls. 8–9. 1848: 289–303.
Günther A. Catalogue of fishes in the British Museum. Catalogue of the Physostomi, containing the families Salmonidae, Percopsidae, Galaxidae, Mormyridae, Gymnarchidae, Esocidae, Umbridae, Scombresocidae, Cyprinodontidae, in the collection of the British Museum. London: British Museum. 1866; 6: 368 pp.
Contreras BS, Escalante MA. Distribution and known impacts of exotic fishes in México. In: Courtenay W.R. and Stauffer J.R., editors. Distribution, Biology, and Management of Exotic Fishes. London: The Show Hopkins University Press. 1984: 102–129.
- 54. Contreras-MacBeath T, Ramirez Espinoza H. Some aspects of the reproductive strategy of Poeciliopsis gracilis (Osteichthyes: Poeciliidae) in the Cuautla River, Morelos, México. J Freshw Ecol. 1996; 11: 327–338.
- 55. Matamoros WA, Schaefer JF, Kreiser BR. Annotated checklist of the freshwater fishes of continental and insular Honduras. Zootaxa. 2009; 2307: 1–38.
- 56. Pollux BJA, Pires MN, Banet AI, Reznick DN. Evolution of placentas in the fish family Poeciliidae: An empirical study of macroevolution. Annu Rev Ecol Evol Syst. 2009; 40: 271–289.
- 57. Peterson BK, Weber JN, Kay EH, Fisher HS, Hoekstra HE. Double digest RADseq: an inexpensive method for de novo SNP discovery and genotyping in model and non-model species. PLoS One. 2012; 7: e37135. pmid:22675423
- 58. Gómez-González AE, Velázquez-Velázquez E, de Jesús Anzueto Calvo M, Fabiola Maza-Cruz M. Fishes of the Grijalva River basin of México and Guatemala. Check List. 2015; 11: 1726.
Andrews, S. 2010 [cited 28 July 2021]. In: FastQC: A Quality Control Tool for High Throughput Sequence Data. Available from: http://www.bioinformatics.babraham.ac.uk/projects/fastqc/
- 60. Eaton DAR, Overcast I. ipyrad: interactive assembly and analysis of RADseq datasets. Bioinformatics. 2020; 36: 2592–2594. pmid:31904816
- 61. Daneck P, Auton A, Abecasis G, Albers CA, Banks E, DePristo DA, et al. The variant call format and VCFtools. Bioinformatics. 2011; 27: 2156–2158. pmid:21653522
- 62. Nguyen LT, Schmidt HA, von Haeseler A, Minh BQ. IQ-TREE: A fast and effective stochastic algorithm for estimating maximum likelihood phylogenies. Mol Biol Evol. 2015; 32: 268–274. pmid:25371430
- 63. Lischer HEL, Excoffier L. PGDSpider: an automated data conversion tool for connecting population genetics and genomics programs. Bioinformatics. 2012; 28; 298–299. pmid:22110245
- 64. Kalyaanamoorthy S, Minh BQ, Wong TKF, von Haeseler A, Jermiin LS. ModelFinder: fast model selection for accurate phylogenetic estimates. Nature Methods. 2017; 14: 587–589. pmid:28481363
Rambaut A. 25 Nov. 2018 [cited 28 July 2021]. “FigTree.” Molecular Evolution, Phylogenetics and Epidemiology. Available from: tree.bio.ed.ac.uk/software/figtree/
RStudio Team. 2015 [cited 28 July 2021]. RStudio: Integrated Development for R. RStudio, Inc., Boston, MA. Available from: https://www/rstudio.com/
Jombart T, Collins C. 2015 [cited 28 July 2021]. A tutorial for Discriminant Analysis of Principal Components (DAPC) using adegenet 2.0.0. Available from: https://adegenet.r-forge.r-project.org/files/tutorial-dapc.pdf
- 68. Jombart T, Devillard S, Balloux F. Discriminant analysis of principal components: a new method for the analysis of genetically structured populations. BMC Genetics. 2010; 11: 94. pmid:20950446
Grünwald NJ, Kamvar ZN, and Everhart SE. 2016a [cited 28 July 2021]. Grünwaldlab/Population Genetics and Genomics in R: First release [Tabima JF, Knaus BJ, Grünwald NJ. GBS Analysis]. Available from: https://grunwaldlab.github.io/Population_Genetics_in_R/gbs_analysis.html
Grünwald NJ, Kamvar ZN, Everhart SE. 2016b [cited 28 July 2021]. Grünwaldlab/Population Genetics and Genomics in R: First release [Knaus BJ, Grünwald NJ. Clustering Plots]. Available from: https://grunwaldlab.github.io/Population_Genetics_in_R/clustering_plot.html
- 71. Weir BS, Cockerham CC. Estimating F-statistics for the analysis of population structure. Evolution. 1984; 38: 1358–1370. pmid:28563791
- 72. Rannala B, Yang Z. Efficient Bayesian species tree inference under the multispecies coalescent. Syst Biol. 2017; 66: 823–842. pmid:28053140
- 73. Yang Z. The BPP program for species tree estimation and species delimitation. Current Zool. 2015; 61: 854–865.
- 74. Xie W, Lewis PO, Fan Y, Kuo L, Chen MH. Improving marginal likelihood estimation for Bayesian phylogenetic model selection. Syst Biol. 2011; 60: 150–160. pmid:21187451
dos Reis M. 2021 [cited 28 July 2021]. bppr: Functions to work with the multi-species coalescent program BPP, for example, functions to calibrate BPP trees to geological time. Available from: https://github.com/dosreislab/bppr
Esri. World Dark Gray Base. Scale Not Given. https://www.arcgis.com/home/item.html?id=a284a9b99b3446a3910d4144a50990f6. Credits: Esri, HERE, Garmin, (c) OpenStreetMap contributors, and the GIS user community. 2021.
- 77. Lal MM, Southgate PC, Jerry DR, Zenger KR. Fishing for divergence in a sea of connectivity: the utility of ddRADseq genotyping in a marine invertebrate, the black-lip pearl oyster Pinctada margaritifera. Marine Genomics. 2016; 25: 57–68. pmid:26545807
- 78. Harris RB, Alstrom P, Odeen A, Leache AD. Discordance between genomic divergence and phenotypic variation in a rapidly evolving avian genus (Motacilla). Mol Phylogenet Evol. 2018; 120: 183–195. pmid:29246816
- 79. Tonzo V, Papadopoulou A, Ortego J. Genomic footprints of an old affair: single nucleotide polymorphism data reveal historical hybridization and the subsequent evolution of reproductive barriers in two recently diverged grasshoppers with partly overlapping distributions. Molec Ecol. 2020; 29: 1–15. pmid:32418257
- 80. Welcomme RA. History of international introductions of inland aquatic species. ICES Marine Sci Symp. 1992; 194: 3–14.
- 81. Fricke R, Eschmeyer WN, van der Laan R. Eschmeyer’s Catalog of Fishes: Genera, Species, References. 2021. [cited 28 July 2021]. Available from: http://researcharchive.calacademy.org/research/ichthyology/catalog/fishcatmain.asp
- 82. Contreras-MacBeath T, Mejia-Mojica H, Carrillo-Wilson R. Negative impact on the aquatic ecosystems of the state of Morelos, Mexico from introduced aquarium and other commercial fish. Aquarium Sci Conserv. 1998; 2: 67–78.
- 83. Huidobro L, Morrone JJ, Villalobos JL, Alvarez F. Distributional patterns of freshwater taxa (fishes, crustaceans and plants) from the Mexican Transition Zone. J Biogeogr. 2006; 33: 731–741.
- 84. Mulcahy DG, Morrill BH, Mendelson III JR. Historical biogeography of lowland species of toads (Bufo) across the Trans Mexican Neovolcanic Belt and the Isthmus of Tehuantepec. J Biogeogr. 2006; 33: 1889–1904.
- 85. Choudhury A, Garcia-Varela M, Perez-Ponce de Leon G. Parasites of the freshwater fishes and the Great American Biotic Interchange: a bridge too far? J Helminthol. 2016; 91:1–23. pmid:26781336
- 86. Morcillo F, Ornelas-Garcia CP, Alcaraz L, Matamoros WA, Doadrio I. Phylogenetic relationships and evolutionary history of the Mesoamerican endemic freshwater fish family Profundulidae (Cyrpinodontiformes: Actinoptergii). Mol Phylogenet Evol. 2016; 95: 242–51.
Ferrusquia-Villafranca I. Geology of Mexico: A synopsis. Biological diversity of Mexico origins and distribution. In: Ramamoorthy TP, Bye R, Lot A, Fa J, editors. Oxford University Press, Oxford. 1993. pp. 3–107.
- 88. Warren DL, Glor RE, Turelli M. Environmental niche equivalency versus conservatism: quantitative approaches to niche evolution. Soc Stud Evol. 2008; 62: 2868–2883. pmid:18752605
- 89. Rodriguez-Gomez F, Gutierrez-Rodriguez C, Ornel JF. Genetic, phenotypic and ecological divergence with gene flow at the Isthmus of Tehuantepec: the case of the azure-crowned hummingbird (Amazilia cyanocephala). J Biogeogr. 2013; 40: 1360–1373.