The M. rufa neotype presented a karyotype constitution of 2n = 52 and FN = 56. The chromosome measurement by relative length classified pair 1 belonging to Group A, pairs 2 to 4 to Group D, pairs 5 to 25 to Group E. The X chromosome was submetacentric (Figure 3). Two to six supernumerary chromosomes (Bs) were observed at metaphases. This autosomal pattern observed for the neotype corresponds to the Paraná cytotype of M. americana (Abril et al., 2010), which from now on corresponds to the species M. rufa. Thus, the description of the sex chromosomes was also performed for a male specimen (T308). The sexual system was XY1Y2 due to an X-autosomal fusion. The male presented a karyotypic constitution of 2n = 53, FN = 56, Y1 was the smallest chromosome, and Y2 a medium acrocentric.

C-banding showed constitutive heterochromatin blocks in the centromeric region of all autosomal chromosomes, two interstitial bands in the q arm of pairs 1 and 3, and three discrete interstitial bands in pair 2. The sex chromosome X had a strong interstitial band in the proximal region of the q arm and a discrete interstitial band in the distal region of the same arm. Chromosome Y₁ is completely euchromatic and chromosome Y₂ has a discrete interstitial band in the medial region of its arm. The B chromosomes showed variation in their heterochromatin content. While some B were completely heterochromatic, others had only interstitials heterochromatin bands. The nucleolus organizer regions were localized in the distal region of the long arms of both chromosomes from pairs 5 and 6. Figure 3 shows the graphical representation of the G and C-banded M. rufa chromosomes, which is of considerable importance in comparative cytogenetics.

The comparative cytogenetics analysis with G-band and cattle (BTA) WCP probes (Figure 4) showed that M. rufa (MRU) and M. americana (sensu stricto) (MAM, 2n = 45, FN = 51) accumulated different rearrangements during their karyotype evolution. The p arm of the submetacentric MRU1 is homologous to the proximal region of MAM4 (Figure 3, 4a), and the q arm to the proximal region of MAM3 (Figure 3, 3a) and to the MAM5 chromosome. The large acrocentric MRU2 showed homologies with the terminal region of MAM1 (Figure 3, 1d) and the distal region of MAM4 (Figure 3, 4b). The MRU3 pair corresponded to the distal regions of MAM1 (Figure 3, 1c) and MAM2 (Figure 3, 2d) and to the proximal region of MAM3 (Figure 3, 3b). MRU4 and MRU12 are homologous to the proximal region of the MAM1 q arm (Figure 3, 1b) and to the p arm (Figure 3, 1a), respectively. MRU5 corresponded to the distal region of MAM3 (Figure 3, 3c). MRU8 and MRU15 is homologous to the proximal region of the q arm of MAM2 (Figure 3, 2b and 2c). Finally, acrocentric pairs MRU 6, 7, 9, 11, 13, 14 and MRU 16 to 25 were homologous to one MAM acrocentric chromosome.

The composition of sex chromosomes also differed between the two species. Both species have a multiple sex chromosome system XY₁Y₂ formed by X-autosomal fusions during the divergence from a common ancestor. However, the distal parts of X and Y₂ were not homologous between the two species. In M. rufa, the distal regions of X and Y₂ correspond to the p arm of MAM2 (Figures 3, 2a), while the distal part of X and Y₂ of M. americana corresponds to MRU10. The Y₂ chromosomes of these species were partially homologous, demonstrating that the first X-autosomal fusion probably occurred in an ancestral karyotype of the group and the second fusion was formed later in each species (Figure 3). Thus, at least 15 rearrangements separate M. rufa and M. americana karyotypes.

Furthermore, the M. americana Carajás cytotype (2n = 50, FN = 54) showed the fixation of a tandem fusion between two acrocentric chromosomes of Mazama rufa. This fusion occurred between MRU5 and MRU8, forming CA3, while the other chromosomes corresponded between these species. The location of the hybridization signals for each bovine probe used is specified in SM-05.

In the body biometric dataset, none of the variables presented extreme outliers (>3SD in relation to the mean) and the total proportion of lost data was 5.16%. In the craniometry dataset, only one variable for a single animal showed extreme outlier and was removed, and the total proportion of lost data was 4.08%. Descriptive analyses (mean, standard deviation, maximum and maximum) of both datasets are presented in SM-03.

The UPGMA distance tree from the two morphometry datasets subdivided the specimens into two groups (Figure 5). The first group corresponds to the M. americana complex (sensu lato), including the neotype from French Guiana, the neotype proposed for M. rufa and specimens from Brazilian cytotypes. The second group was represented by the M. nana and M. temama. Finaly, M. bororo was subdivided with two individuals in each group. All specimens were positioned in a mixed manner within each main group where neither species nor cytotypes were grouped together (Figure 5). In the PCA analysis, the two datasets presented similar results to those of the UPGMA with a high overlap among species and among cytotypes, but some distinctions were observed in the craniometry dataset (Figure 6). It was possible to discriminate M. bororo from the other animals, characterizing the species as an intermediate morphological taxon between the M. americana complex, M. nana and M. temama. The M. americana complex, represented by the neotypes and by the Paraná (M. rufa), Carajás, Rondônia, and Juína cytotypes, presented the greatest overlap, reinforcing their cryptic complex aspect. In both datasets (body and skull) it is noteworthy that despite the large number of variables, all showed a high association with PCA1, which was able to concentrate respectively 76 and 64% of the data variation in each dataset.

The qualitative characteristics of coat color and hair length showed subtle differences between the amended description of the M. rufa (Illiger 1815) neotype and the M. americana (Cifuentes-Rincón et al., 2020) neotype. The M. rufa species presented a reddish-brown color pattern in the anterobasal region of the ears and in the inner proximal region of the hind limbs, with the presence of long hairs in the basal ears regions and medial to hind limbs. This contrasts with the M. americana neotype, which showed a whiter pattern in each of these areas, and an absence of hair at the base of the ears, and the inner proximal region of both limbs (Cifuentes-Rincón et al., 2020).

The phylogenetic hypothesis obtained by BI in the tissue dataset presented most clades with good support (PP > 0.95 Figure 7A). Molecular species delimitation analysis showed the GMYC model significantly higher than the null model for the single threshold (p = 0.0018), while the GMYC model was not significantly higher for multiple thresholds (p = 0.0959) (SM-04). Thus, the single threshold model identified nine MOTUs (Figure 7D) that were compared with delimitations based on the current taxonomy of the analyzed species (Merino and Rossi, 2010; Figure 7B) and on the cytotypes taxonomic hypothesis (Abril et al., 2010; Cursino et al., 2014; Salviano et al., 2017; Galindo et al., 2021; Figure 7C).

The phylogenetic hypothesis recovered the sampled Mazama species as a well-supported monophyletic group (clade I) in relation to genus Odocoileus (Figure 7). Mazama temama was positioned as a basal species in the group and was recovered as a monophyletic clade and single MOTU. The rest of the genus (clade II) was divided into two main clades, one with M. americana Juína and Rondônia cytotypes, M. nana, and M. bororo (III), and another with only M. americana (sensu lato) lineages (IV). In clade III, we observed a reciprocal monophyly relationship between M. nana and M. bororo, and each one was indicated as a distinct MOTU. Also in clade III, a sub-clade composed of the Juína and Rondônia M. americana (sensu lato) cytotypes was positioned as a sister group to M. nana-M. bororo. Those did not show reciprocal monophyly because one Rondônia sample (T269) was grouped in the Juína clade. However, the GMYC analysis indicated the clades of each cytotype as two distinct MOTUS and recognize this sample as belonging to a third potential species. In the other main clade (IV), only with M. americana (sensu lato) samples, two MOTUs were delimited, each one with a distinct neotype. The M. americana neotype (T358) was grouped with another sample from French Guiana and with animal T424 from the extreme north of Brazil, forming an exclusive clade (V) that represents M. americana (sensu stricto). This clade, in turn, was the sister group of the M. rufa neotype (T385) clade (VI), whose MOTU was composed of the Carajás and Paraná cytotypes. The animals of the Paraná cytotype were recovered in a well-supported monophyletic clade, but nested to the two clades composed by the Carajás cytotype animals (VII and VIII), with no reciprocal monophyletic relationship observed among the two cytotypes.

Non-invasive sampling resulted in 241 fecal samples that, coupled with the previously red brocket samples collected, resulted in 15 sites with M. rufa confirmed presence (Figure 1; SM-02). Our distribution model had a high AUC value (AUC = 0.908, SD 0.018), low omission errors (0.05) for a fixed cumulative value 10 Logistic threshold and were significant (p < 0.05) for the binomial test. It indicated a potential distribution of M. rufa across southeast Brazil extending to eastern Paraguay and Argentina, northward to eastern Bolivia and southern Amazon (Figure 1). The high suitability core areas were close to the neotype capture site encompassing forests fragments in the Misiones province in Argentina, and other fragmented areas in the border between Paraguay and Bolivia. This analysis also showed the habitat suitability connection between the neotype locality and the original type locality described by Illiger (1815) in the Asunción city region.

The difference between M. americana (sensu stricto) and M. rufa (former Paraná cytotype) was suggested to envolve two tandem fusions, two Robertsonian translocations and a pericentric inversion based on the hypothetical M. americana ancestral karyotype (Abril et al., 2010; Cifuentes-Rincón et al., 2020). However, the association of the G-Band with cattle WCP probes demonstrated that M. americana (sensu stricto) (2n = 45, FN = 51) and M. rufa (2n = 52/53, FN = 56) underwent surprisingly distinct chromosomal rearrangements during their karyotypic evolutions and diverge in at least 15 rearrangements. These findings support the hypothesis that M. rufa is a distinct species from M. americana, given that smaller karyotypic differences among other M. americana cytotypes resulted in complete sterility of their hybrids (Cursino et al., 2014; Salviano et al., 2017). Accumulation of chromosomal rearrangements leads to errors in meiosis and chromosome pairing, recombination suppression, errors in meiotic segregation, and subsequent germ cell or zygote death (Rieseberg, 2001; Villagómez and Pinton, 2008; Dobigny et al., 2017).

On the other hand, the difference of one tandem fusion between M. rufa and the Carajás cytotype (MRU 5/8, forming CA3) has been described (Abril et al., 2010; Galindo et al., 2021). The potential hybridization between both populations would lead to the formation of individuals carrying one tandem fusion in heterozygosis, considering the conservative observation of absence of a prezygotic barrier in captivity (Carranza et al., 2018). Such rearrangement is considered deleterious and with a hypothetical 50% reduction in the production of balanced gamete (King, 1993), resulting in reduced fertility of the carrier (Moritz, 1986; Pillay et al., 1995; Long, 1996; Galindo et al., 2021). Nevertheless, spermiogram and testicular histology analysis failed to demonstrate a significant reduction in the reproductive fitness of hybrids carrying one tandem fusion in heterozygosis when compared to purebred animal. These were classified as sub-fertile, which would reduce the possibility of establishing a post-zygotic reproductive barrier between populations (Salviano et al., 2017). In this context, a recent study evaluated the meiotic segregation of hybrids with one tandem fusion in heterozygosis in the M. americana complex, including a hybrid between the Carajás and Paraná (M. rufa) cytotypes, obtaining a rate of ∼30% gametic unbalance (Galindo et al., 2021). Considering the chance of successful reproduction in a backcross breeding with animals from the parent populations (∼35% with the presence or ∼35% with the absence of tandem fusion in balanced gametes) this reduction in hybrid fertility might be even greater (∼65% of gametic unbalance). Thus, the difference of one tandem fusion among populations is suggested as an efficient post-zygotic reproductive barrier, similar to that observed for Otomys irroratus (Pillay et al., 1995). Furthermore, it is important to indicate the apparent ongoing process of karyotypic evolution of the Carajás cytotype, demonstrated by the presence of chromosomal polymorphisms such as centromeric fusion in heterozygosis and homozygosis (Galindo et al., 2021). The possibility of crossing these latter cytogenetic variants with M. rufa could result in the production of hybrids carrying tandem fusion and centromeric fusion, both in heterozygosis. This interaction among different chromosomal rearrangements could result in increased rates of unbalanced gametes, as already observed in a hybrid of the M. americana complex carrying both rearrangements (43.05% of sperm with aneuploidy), with a subsequent gametic unbalance of ∼70%, regarding any parent population (Galindo et al., 2021). This evidence supports the hypothesis of considering M. rufa and the animals of the Carajás cytotype as distinct species.

The external morphological similarity (body biometry) among M. americana lineages and other Mazama species has been previously demonstrated (Duarte et al., 2008). A similar result was also obtained in a craniometry analysis of M. americana lineages, but the comparison with other species of the genus was limited (Cifuentes-Rincón et al., 2020). Our results showed a level of distinction among different species in the body biometric data, specifically between M. nana, M. temama and the other evaluated species. In the craniometry data, M. bororo was isolated, with only one specimen of M. americana overlapping it. Nevertheless, this distinction of M. bororo was not statistically supported in Rossi (2000) assessment, nor is it confirmed in the broad overlap of forest deer craniometry reported by Cassini and Toledo (2021).

The positioning of the M. americana neotype was close to the M. rufa neotype and quite overlapping with other M. rufa specimens (Paraná cytotype) and also Carajás cytotype specimens in both datasets. Qualitative analysis, however, identified coat features that appreciably distinguish the neotypes of the two species. This represents a first step towards finding diagnostic morphological characters, which must be tested within a broader sample. The detailed analysis of deposited specimens using a post hoc approach to compare the M. americana genetic lineages may contribute to the identification of diagnostic characters. This type of analysis is also called “reverse taxonomy” (Markmann and Tautz, 2005) and can be an interesting way to identify morphological differences, reinforce the delimitations observed by genetic data and organize scientific collections where a high primary identification error in their Mazama vouchers have been observed (Michaloudi et al., 2018; Mantellatto et al., 2020).

Mitochondrial genes have been systematically used in several works that aimed to recover the phylogenetic relationships among species of the tribe Odocoileini in the last 2 decades (Gilbert et al., 2006; Duarte et al., 2008; Gutiérrez et al., 2015, 2017; Escobedo-Morales et al., 2016; Heckeberg et al., 2016; Cifuentes-Rincón et al., 2020; Heckeberg, 2020). Some studies tested the use of nuclear regions (a-lactalbumin; protein kinase C iota; satellite DNA) but they were not informative in recovering phylogenetic relationships between recently divergent species (Gilbert et al., 2006; Heckeberg, 2020; Vozdova et al., 2021). Although the present work brought the widest sampling (number of vouchers and mitochondrial sequence size) in the M. americana complex, it still represents a single-locus analysis that limits interpretation of the results, given that mtDNA gene-tree may underestimate introgression and hybridization processes (Shaw, 2002). Given that results can diverge depending on the analyzed regions, a multi-loci analysis involving nuclear markers and species-trees would be a more precise approach to recover phylogenetic relationships, divergence times, and MOTU delimitation in an unbiased analysis (Igea et al., 2015; Dool et al., 2016).

The polyphyletic status of Mazama americana was revealed when the species was recovered into two clades that formed a polytomy with Odocoileus virginianus and were nested with M. nana and M. bororo (Duarte et al., 2008). Different studies observed the same results in phylogenetic analyses involving more species from the tribe Odocoileini (Escobedo-Morales et al., 2016; Heckeberg et al., 2016; Gutiérrez et al., 2017; Heckeberg, 2020). The problem is that studies restricted to Cytb gene did not present sufficient resolution, and recovered the specimens in a mixed form, without clades consistent with the region of origin, chromosomal lineage, or cytotype (Duarte et al., 2008; Gutiérrez et al., 2017). On the other hand, more informative alignments recovered each chromosomal lineage in distinct and well-supported clades (Abril et al., 2010). In these analyses, it was possible to observe the sister-group relationship between the M. americana lower diploid number lineage and M. bororo, and between the higher diploid number lineage and M. americana (sensu stricto) (Cifuentes-Rincón et al., 2020). Nevertheless, previous works have not actually tested reciprocal monophyly between closely related cytotypes, such as Rondônia x Juina (lower 2n) or Paraná x Carajás (higher 2n). Our results were consistent with previously phylogeny regarding chromosomal lineages and species, and for the first time tested the relationship between cytotypes of close divergence. The absence of reciprocal monophyly between both pairs of close cytotypes would invalidate the hypothesis that they represent distinct species, taking the strict phylogenetic species concept into account (Cracraft, 1983). However, this result should be evaluated with caution as it may be the result of insufficient polymorphism in the analyzed sequence or the bias from incomplete lineage sorting in an analysis restricted to a single locus.

Modern taxonomy proposes an integrative method in which different data sets could result in a greater accumulation of evidence to support taxonomic revisions (Padial et al., 2010; Fujita et al., 2012). More interesting than searching for consistency in the different sources of evidence it is necessary to search for the information that is biological meaningful for the taxon in question (Padial et al., 2010). The definition of a taxonomic hypothesis in view of contrasting data can be equally relevant if well-grounded in the evolutionary processes that drive the speciation of the analyzed group (Fišer et al., 2018).

The incongruity between morphological and genetic data should not represent a problem given that morphological similarity in cryptic species is understood as a result of three possible evolutionary mechanisms: recent divergence, niche phylogenetic conservatism, or morphological convergence (Fišer et al., 2018). The niche phylogenetic conservatism hypothesis is interesting for the M. americana complex as it considers specialist species under strong selective pressure (Fišer et al., 2018). In this regard, brocket deer species are considered forest specialists (Weber and Gonzalez, 2003; Duarte et al., 2017; Oliveira et al., 2019; González and Duarte, 2020) and the selective pressure of this environment has already been identified as a determinant feature in the morphological convergence of different lineages of the genus Mazama and Pudu (Gilbert et al., 2006; Duarte et al., 2008; Heckeberg, 2020). In these taxa, for example, spiked antlers are understood as characters that emerged independently throughout the evolution of neotropical deer (Gilbert et al., 2006; Duarte et al., 2008; Heckeberg, 2020). In the case of the red brocket lineage (M. temama, M. nana, M. bororo, M. rufa and M. americana complex), a group of close-sister species, it would not be the case of convergence, but of morphological stasis resulting from the maintenance of the niche under similar environmental pressure of forests habitats. Thus, the absence of morphological differentiation does not necessarily represent persistence of gene flow among M. americana cytotypes and therefore does not invalidate the recognition of distict species such as the present description of M. rufa.

Molecular data and cytogenetic data converged in indicating Mazama rufa (Illiger, 1815) as a distinct species in relation to Mazama americana (Erxleben, 1777). The role of chromosomal divergence is related to the gene flow reduction between divergent populations of mammals (Rieseberg, 2001), which can lead to a process of isolation and speciation as already discussed above. The operational biological species concept is the core of this hypothesis, in which reproductive isolation would unequivocally identify the existence of different species (Mayr, 1942). It is clear that both species represent independent evolutionary lineages within the ontological concept of Lineage Species Concept from De Queiroz (2007). Cytogenetic data also provide strong evidence that Mazama rufa is composed only by the Paraná cytotype, even considered the close divergence between the Paraná and Carajás cytotypes. They would be in isolation since the reproductive fitness of a potential hybrid could be greatly aggravated by chromosomal imbalance, and thus being practically sterile (Galindo et al., 2021). This is observed in hybrids or populations with the presence of heterozygous tandem fusion, which are considered highly deleterious chromosomal rearrangements, rapidly removed or fixed at meiosis during speciation (King, 1993; Yang et al., 1997; Dobigny et al., 2017; Mudd et al., 2020). Although molecular data indicated the absence of reciprocal monophyly between Paraná and Carajás and suggested both as a single MOTU in the GMYC analyses, this needs to be further explored. The phylogenetic analysis and species delimitation presented here have important limitations, as they assessed only one locus, represented by part of the mitochondrial DNA. Additionally, literature strongly recommends that GMYC results should not be considered alone to guide taxonomic reviews, but should be discussed in the light of other evidences and biological information (Talavera et al., 2013; Tang et al., 2014).

The elucidation of the phylogenetic relationships and species delimitation in the M. americana complex depends not only of an advance in molecular analysis, but also an advance in Amazonian populations sampling, which are still underrepresented. Future analyses should include other chromosomal variants associated with the higher diploid number lineage described for this region (Abril et al., 2010). Additionally, the evaluation and nominal description of new species related to the lower diploid number lineage (Juína and Rondônia) also depends on the evaluation of names in the synonymy of M. americana in neighboring Amazonian areas such as Mazama sarae (Thomas, 1925) described in Bolivia and Mazama whitelyi (Gray, 1873) described in Peru.

The names and informal descriptions of brocket deer in South America by Azara in “Apuntamientos Para La Historia Natural De Los Quadrúpedos del Paraguay Y Rio De La Plata” based Illiger (1815) formal description. He named the species “Gouazoupitá” (red brocket deer) and gouazoubirá” (brown brocket deer) as Cervus rufus and Cervus simplicicornis, respectively. These descriptions, according to Allen (1915), were the basis for the nomenclature of the species in the genus and represent a historical landmark. It is necessary, however, to mention that Fischer (1814) also formally described the Azara deer (Cervus gouazoupita, the red one and Cervus gouazoubira, the brown one). However, Fisher’s work is summarily disregarded in relation to the red brocket deer forest species and the name Cervus gouazoupita Fischer, 1814 disappears completely from the literature, never again being mentioned or listed as synonymous in taxonomic reviews of the genus Mazama. For example, publications that bring a broad taxonomic organization of the genus, such as that conducted by Lydekker (1898), Allen (1915), Cabrera (1960), and more recently Merino and Rossi (2010), do not even mention Cervus gouazoupita Fischer, 1814 and always list Mazama rufa (Illiger, 1815) as the valid species. It is important to mention that Illiger makes his work public at the Berlin Academy of Sciences in 1811, the publication of 1815 makes it clear on its back cover that the works described there were presented between 1804 and 1811. Perhaps a previous reading is one of the reasons for his historical preference, given that Illiger’s oral presentation is highlighted as a priority by Allen (1915), and Lydekker (1915) mentions that separate copies were issued in 1811.

The International Code on Zoological Nomenclature is clear that any work published after 1757 [Art. 11.1] must comply with the requirements of providing a public and permanent scientific record and be easily obtainable soon after publication to be considered valid [Art. 8.1], it does not consider public speech or materials issued primarily to participants of scientific meetings [Art. 9.10]. In this regard, a historical recovery would be necessary to prioritize the name Mazama goazoupita (Fischer, 1814). However, the priority principle does not aim to change a name already widely used by introducing a new one that is an older synonym or homonym [Art. 23.2]. Given the frequent use and total dominance of the name Mazama rufa (Illiger, 1815) by the scientific community, to the detriment of Mazama gouazoupita (Fischer, 1814) in the last 200 years, we decided to maintain the use of the first for the sake of clarity and taxonomic stability in the specific case of the red brocket deer of Paraguay-Brazil.

The results indicate that the species is distributed over a wide area, from the south of the continent, in the Atlantic Forest; passing the dry diagonal through the Cerrado and to the southern limit of the Amazon. In addition to occurring in Brazil, it is likely to occur in the Atlantic Forest and Humid Chaco regions of Argentina and Paraguay, as well as possibly in the Dry Forest of Bolivia, on the western edge of the original distribution of M. americana. The eastern portion of the red brocket deer distribution, which extends to the coastal Atlantic Forest and the central Cerrado region, needs to be evaluated for the presence of any red brocket deer, whether M. americana, M. bororo or M. rufa.

Despite the wide distribution, this species must be under severe anthropogenic pressure. The forest formations where it occurs in the Atlantic Forest (Araucaria and Interior formations) are the most devastated in the biome, characterized by intense fragmentation (Ribeiro et al., 2009). The conversion of native vegetation and intense fragmentation also impact the Cerrado, which has lost more than 50% of its original cover (Klink and Machado, 2005; Reynolds et al., 2016). The biome also lacks conservation units, with only 3% of its area under strict protection (Françoso et al., 2015). The species was only marginally detected in the Pantanal, an area that did not show suitable habitat. It is important to confirm this information as the presence of a red brocket deer species is known in the area and the region is a refuge for large populations of various mammals, especially deer (Tomas et al., 2010). Finally, the local reality in the southern range of the Amazon where the species was found is of severe deforestation (Santos et al., 2021) and is the most worrying region for M. rufa decline.

The present work demonstrated the revalidation of Mazama rufa (Illiger, 1815) as a distinct species, another cryptic species to be separated from the Mazama americana complex. It is composed by the former M. americana Paraná cytotype and widely distributed throughout South America in the Atlantic Forest, Cerrado, and the south of Amazon. The information presented here should serve as a basis for a detailed assessment of the species extinction risk by the IUCN Red List since it occupies areas of high anthropogenic pressure.

The Mazama americana complex begin to be unraveled after a century of divergent taxonomic arrangements. The next steps in the group’s taxonomic review should prioritize: 1) a comprehensive review of museum specimens, accompanied by molecular characterization, and including the analysis of holotypes; 2) sampling and collection of vouchers for cytogenetic characterization with live tissue, especially in the Amazon; 3) a multi-locus molecular analysis, with nuclear markers or population approaches, to better understand the historical and current isolation between close divergence cytotypes.