Integrative taxonomy reveals two new cryptic species of Hyphessobrycon Durbin, 1908 (Teleostei: Characidae) from the Maracaçumé and middle Tocantins River basins, Eastern Amazon region

1,2,4,7 Universidade Federal do Maranhão, Programa de Pós-Graduação em Biodiversidade e Biotecnologia da Amazônia Legal. Av. dos Portugueses 1966, Cidade Universitária do Bacanga, CEP 65080-805, São Luís, MA, Brazil. 3 South African Institute for Aquatic Biodiversity, Private Bag 1015, Grahamstown, 6140, South Africa. 4 Universidade Estadual do Maranhão, Laboratório de Ictiofauna e Piscicultura Integrada, Centro de Ciências Agrárias, Campus Paulo VI, avenida Lourenço Vieira da Silva, n. 1000, bairro Jardim São Cristóvão, CEP 65055-310, São Luís, MA, Brazil. 5 Universidade Federal do Rio de Janeiro, Laboratório de Sistemática e Evolução de Peixes Teleósteos, Departamento de Zoologia, Instituto de Biologia, Cidade Universitária, CEP 21941-599. Rio de Janeiro, RJ, Brazil. 1,2,6 Universidade Federal do Maranhão, Departamento de Biologia, Laboratório de Genética e Biologia Molecular, Av. dos Portugueses 1966, Cidade Universitária do Bacanga, CEP 65080-805, São Luís, MA, Brazil. 1,2,7 Universidade Federal do Maranhão, Laboratório de Sistemática e Ecologia de Organismos Aquáticos, Centro de Ciências Agrárias e Ambientais, Centro de Ciências Agrárias e Ambientais (CCAA), campus de Chapadinha, BR-222, KM 04, S/N, Boa Vista, CEP 65500-000, Chapadinha, MA, Brazil.


Introduction
Hyphessobrycon Durbin, 1908 is one of the most species-rich and widespread characid genus, comprising about 160 valid species and occurring from southern Mexico to the Río de La Plata in Argentina (Fricke et al. 2020a). Despite its non-monophyletic status (Mirande 2010;Oliveira et al. 2011;Betancur-R et al. 2018;Guimarães et al. 2018;Mirande 2018;Guimarães et al. 2019), Hyphessobrycon is one of the genera within Characidae whose species diversity has increased signifi cantly in the past decades. Considering only the last decade, the number of species has risen from 109 to approximately 160, an increase of about 30% in the number of valid species (Fricke et al. 2020a), and currently comprising approximately 15% of all characid species (Fricke et al. 2020b).
Recent studies on Neotropical Characiformes, based on molecular data and species delimitation approaches, revealed a high cryptic diversity (e.g., Pereira et al. 2011;Melo et al. 2014;Benine et al. 2015;Benzaquem et al. 2015;Melo et al. 2016aMelo et al. , 2016bMelo et al. , 2016cGarcía-Melo et al. 2019). Similarly, three recently published papers (e.g., Castro-Paz et al. 2014;Guimarães et al. 2018Guimarães et al. , 2019 applied molecular species delimitation to investigate the species diversity within the "Rosy tetra" clade, and they all agreed that its taxonomy should be better investigated, since it includes cryptic species.
In addition to the wide distribution and the presence of cryptic species, the confuse taxonomic history of that clade is one of the main challenges when studying its diversity and relationships. The type species of Hemigrammus, H. unilineatus (Gill, 1858), and Pristella Eigenmann 1908, P. maxillaris (Ulrey, 1894, have been historically considered as closely related or belonging to the "Rosy-tetra" clade, which also includes the type species of Hyphessobrycon, H. compressus (Meek, 1904) (Weitzman & Palmer 1997;Carvalho & Malabarba 2015). Thus, Hyphessobrycon may be a putative synonym of Hemigrammus (Mirande 2010) or Pristella. However, further studies on the systematics of the species currently assigned to Hyphessobrycon and close related genera are still necessary to clarify this issue (Guimarães et al. 2018. The aim of this study is to describe two new cryptic species of Hyphessobrycon sensu Bickford et al. (2006), members of the "Rosy tetra" clade, from the Maracaçumé and middle Tocantins River basins, Eastern Amazon region, based on an integrative taxonomy approach. Furthermore, a new subclade, within the "Rosy tetra" clade, is herein proposed.

Taxa sampling, specimen collection and preservation
Specimens were captured with manual trail-net (2 m long × 1.8 m high; mesh size: 2 mm) and euthanized in a buffered solution of ethyl-3-amino-benzoat-methanesulfonate (MS-222) with a concentration of 250 mg/l until completely ceasing opercular movements, according to animal welfare laws and guidelines (Close et al. 1996(Close et al. , 1997Leary et al. 2013). Specimens selected for morphological analysis were fi xed in formalin and left for ten days, after which they were preserved in 70% ethanol. Molecular data were obtained from specimens that were euthanized, fi xed and preserved in absolute ethanol. Specimens used for morphological analysis are listed in type and comparative material lists. Specimens for molecular approaches are listed in Table 1. Sequences from other species of Hyphessobrycon and allied species were obtained from the Barcode of Life Database (BOLD) and the National Center for Biotechnology Information (NCBI) databases (Table 1).

Morphological analysis
Measurements and counts were made according to Fink & Weitzman (1974). Vertical scale rows between the dorsal-fi n origin and the lateral line do not include the scale of the median predorsal series situated just anterior to the fi rst dorsal-fi n ray. Counts of supraneurals, vertebrae, procurrent caudal-fi n rays, unbranched dorsal-and anal-fi n rays, branchiostegal rays, gill-rakers, premaxillary, maxillary, and dentary teeth, and other internal counts were taken only from cleared and stained (C&S) paratypes, prepared according to Taylor & Van Dyke (1985). The four modifi ed vertebrae that constitute the Weberian apparatus were not included in the vertebrae counts, and the fused PU1 + U1 was considered as a single element. Osteological nomenclature follows Weitzman (1962). Institutional abbreviations follow .

DNA extraction, amplifi cation and sequencing analyses
DNA was extracted from fi n clips using Wizard Genomic DNA Purifi cation kit (Promega) according to the manufacturer's protocol. Fragments of the mitochondrial cytochrome  (Invitrogen), 100 ηg of DNA template, and ultrapure water. The PCR cycles were as follows: 2 min at 94°C, followed by 35 cycles of 94°C for 30s, 54°C for 30s, and 72°C for 1 min, and 10 min at 72°C. Amplicons were purifi ed using Illustra GFX PCR DNA and Gel Purifi cation Kit (GE Healthcare Systems) and sequenced using the forward primer by an outsourced sequencing service at the University of São Paulo, using BigDye Terminator kit ver. 3.1 Cycle Sequencing kit in ABI 3730 DNA Analyser (Applied Biosystems).

Data partition, evolution models, and alignment
The dataset cytochrome c oxidase subunit 1 (COI) (639 bp). Sequences were aligned using ClustalW (Chenna et al. 2003). The DNA sequences were translated into amino acids residues to test for the presence of premature stop codons or indels using the program MEGA 7 (Kumar et al. 2016). Measure Substitution Saturation tests were performed in DAMBE5 (Xia 2013) according to the algorithm proposed by Xia et al. (2003). The best-fi t evolutionary model (GTR+G) was calculated, using the corrected Akaike Information Criterion (AICc) determined by the jModelTest 2.1.7 (Darriba et al. 2012).

Species concept, species delimitation, and diagnoses
The unifi ed species concept is herein adopted by expressing the conceptual defi nition shared by all traditional species concepts -"species are (segments of) separately evolving metapopulation lineages" -disentangling operational criterion elements to delimit taxa from species concepts (de Queiroz 2005(de Queiroz , 2007. According to this concept, species are treated as hypothetical units and could be tested (detected) by the application of distinct criteria (species delimitation methods) (de Queiroz 2005(de Queiroz , 2007. It allows for any method to independently provide evidence about species limits and identities (de Queiroz 2005(de Queiroz , 2007. Seven distinct and independent methods relying on different operational criteria for sp suzu65@gmail. com ecies delimitation, based on morphological and molecular data, were implemented here: PAA, Population Aggregation Analysis (Davis & Nixon 1992); DBC, DNA barcoding, as proposed by Hebert et al. (2003aHebert et al. ( , 2003bHebert et al. ( , 2004aHebert et al. ( , 2004b; ABGD, Automatic Barcode Gap Discovery (Puillandre et al. 2012 Desalle et al. (2005); as well as two coalescent species delimitation methods termed bPTP, the Bayesian implementation of the Poisson tree processes (Zhang et al. 2013), and GMYC, the General Mixed Yule Coalescent method , single-threshold version (Fujisawa & Barraclough 2013).
All species delimitation methods here adopted, except PAA that relies only on morphological characters, were performed using cytochrome c oxidase subunit 1 (COI) sequences, since it is a mitochondrial gene with a fast evolutionary rate, thus suitable and widely used for single locus species delimitation approaches (Avise 2000).

Population aggregation analysis (PAA)
The PAA (Davis & Nixon 1992) is a character-based method, in which species are delimited by a unique combination of morphological character states occurring in one or more populations (Costa et al. 2014). Basically, the occurrence of the same character state in individuals from different populations is evidence to the presence of a gene fl ow between them, indicating that all correspond to the same species (lineage), whereas the presence of exclusive character states not shared between individuals belonging to two distinct populations suggests that there is no gene fl ow between them, thus corresponding to distinct lineages.

Wiens and Penkrot analysis (WP)
WP is based on the direct inspection of haplotype trees generated by a phylogenetic analysis having as terminals at least two individuals (haplotypes) of each focal species. In this method, the term 'exclusive' is used instead of monophyletic, since the term monophyly is considered inapplicable below the species level (Wiens & Penkrot 2002). Clustered haplotypes with a concordant geographic distribution forming mutual and well supported clades (exclusive lineages) are considered as strong evidence for species discrimination (absence of gene fl ow with other lineages). When haplotypes from the same locality fail to cluster together, there is potential evidence for gene fl ow with other populations (Wiens & Penkrot 2002). Haplotype tree statistical support was assessed by the posterior probability value, considered as signifi cant at about 0.95 or higher (Alfaro & Holder 2006). When only one haplotype (specimen) from one putative population was available, the species delimitation was based on the exclusivity of the sister clade of this single haplotype, supported by signifi cant values, allowing us to perform the test in populations with only one haplotype (Wiens & Penkrot 2002). In addition, the method allows the recognition of nonexclusive lineages as species since their sister clades are exclusive and supported by signifi cant values (Wiens & Penkrot 2002).
A Bayesian inference-based phylogenetic (BI) tree was estimated in MrBayes (Huelsenbeck & Ronquist 2001) plugin in Geneious 9.0.5 to reconstruct the evolutionary relationships among terminals using General Time Reversible (GTR+G) as evolutionary model. The Bayesian tree inference was based on a chain length of 10 million, a burn-in length of 500 000 generations subsampling trees every 10 000 generations. We used a sequence of Hyphessobrycon fl ammeus Myers, 1924 as outgroup, and all the other haplotypes were considered as ingroup in our analysis..

Traditional DNA barcoding (DBC)
We used the Kimura-2-parameters model (K2P) (Kimura 1980) to estimate the pairwise genetic distances between species in MEGA 7 software (Kumar et al. 2016). We used DnaSP ver. 6. (Rozas et al. 2003) to estimate the number of variable sites and haplotypes. We considered a cutoff of 2% as suffi cient to discriminate species, since this threshold is commonly inferred by species delimitations among Neotropical fi sh species based on COI (Jacobina et al. 2018).

Character-based DNA barcoding (CBB)
The CBB is similar to the population aggregation analysis proposed by Davis & Nixon (1992), but directed to nucleotides as an alternative method for diagnosing taxa through DNA barcodes (De Salle et al. 2005), since the original method is based on subjective cut-off distance thresholds for the inference of species limits (Hebert et al. 2003a(Hebert et al. , 2003b(Hebert et al. , 2004a(Hebert et al. , 2004b. This method delimits species by the presence of a unique combination of nucleotides within a site shared by individuals of the same population or group of populations. In addition, species were diagnosed by nucleotide substitutions (Costa et al. 2014;Ottoni et al. 2019). Nucleotide substitutions among lineages were optimized in the Bayesian inference topology using PAUP ver. 4 (Swofford 2002). Each nucleotide substitution is represented by its relative numeric position determined through sequence alignment with the complete mitochondrial genome of Astyanax paranae Eigenmann, 1914 (KX609386.1:5503-7062), followed by the specifi c nucleotide substitution in parentheses. The results of this analysis are presented in Appendix 2 and the molecular diagnosis sections.

General Mixed Yule Coalescent (GMYC)
The GMYC is a single locus coalescent phylogeny-based species delimitation approach that relies on the branch length to establish a threshold between speciation and coalescent processes (Fujisawa & Barraclough 2013). Here we applied the single-threshold version of the method, which usually outperforms the multiple-threshold version (Fujisawa & Barraclough 2013 (Chenna et al. 2003). The best-fi t evolutionary model (GTR+G) for the reduced dataset was calculated using the corrected Akaike Information Criterion (AICc) determined by the jModelTest ver. 2.1.7 (Darriba et al. 2012). The input ultrametric phylogenetic tree was performed in BEAST ver. 1.8.4 (Drummond et al. 2012), with the following parameters: an uncorrelated relaxed clock with lognormal distribution, a Yule Process as tree prior with 10 million generations and sampling frequency of 1000. The GMYC analysis was performed in the Exelixis Lab's server https://species.h-its.org/gmyc/.

Bayesian implementation of the poisson tree processes (bPTP)
The bPTP is another single locus coalescent phylogeny-based species delimitation method, but it differs from other unilocus species delimitation coalescent approaches, such as GMYC, because it does not need an ultrametric tree (not relying on branch length to delimit species), thus avoiding errors and computer intensive processes (Zhang et al. 2013). The method assumes that more molecular variability (number of nucleotide substitutions) is expected between haplotypes from different species than within a species (Zhang et al. 2013), establishing a threshold between speciation and coalescent processes. The reduced dataset and the input phylogenetic tree for performing the bPTP was the same used in GMYC species delimitation method ,since this method also requires the use of only unique haplotypes, and previously detailed in the GMYC species delimitation method description. The bPTP analysis was performed in the Exelixis Lab's web server http://species.h-its.org/ptp/, following the default parameters except for a 20% burn-in.

Automatic Barcode Gap Discovery (ABGD)
The ABGD is a barcode species delimitation method that aims to establish a minimum gap that probably corresponds to the threshold between interspecifi c and intraspecifi c processes (Puillandre et al. 2012). The major advantage of ABGD when compared to the other barcode species delimitation methods is that the inference of the limit between interspecifi c and intraspecifi c processes (gap detection) is recursively applied to previously obtained groups to get fi ner partitions until there is no further partitioning, allowing a more refi ned search. Basically, the ABGD analysis indicates the number of groups (species) estimated relative to a large spectrum of p values (prior intraspecifi c values) in which a 0.1 value assumes a maximum intraspecifi c variability indicating that all sequences belong to only one species, whereas a 0.001 value assumes a very small intraspecifi c variability indicating that each distinct haplotype represents a different species. After running the ABGD, additional molecular, morphological or ecological characters are needed to infer the correct number of species, following an integrative taxonomy perspective. The ABGD analysis was performed on the ABGD server website https://bioinfo.mnhn.fr/abi/public/abgd/abgdweb.html following the default parameters.  (Géry, 1960), H. socolofi Weitzman, 1977, andH. sweglesi (Géry, 1961 (Eigenmann, 1915), H. megalopterus (Eigenmann, 1915), H. simulatus (Géry, 1960) and H. takasei Géry, 1964;

Etymology
The new species is named after the ichthyologist Ronald Fricke, in recognition of his contribution to ichthyology.

Type locality
Maracaçumé River basin, a coastal river in the Maranhão state, northeastern Brazil (Fig. 3).

Description
Morphometric data of holotype and paratypes are presented in Table 2. BODY. Small (larger specimen with 19.6 mm of SL), laterally compressed, moderately deep, greatest body depth slightly anterior to dorsal-fi n base; body profi le straight and downward directed from the end of dorsal fi n to adipose fi n, straight or slightly convex between latter point and origin of dorsal most procurrent caudal-fi n ray; dorsal profi le of head convex from upper lip to vertical through eye; predorsal profi le of body roughly straight, dorsal-fi n base slightly convex, posteroventrally inclined; ventral profi le of head convex from lower jaw to pelvic-fi n origin; straight and posterodorsally slanted along anal-fi n base; and slightly concave on caudal peduncle. Jaws equal, mouth terminal. Maxilla reaching vertical to anterior margin of pupil.

Sexual dimorphism
Bony hooks on fi ns are absent in all examined specimens. According to Malabarba & Weitzman (2003), the presence of bony hooks is a common dimorphic feature among characid species. Although this sexual dimorphism is not observed in all characid species, as in the case of the species described here.
Color pattern is not sexually dimorphic either.  Table 3 Morphological diagnosis (PAA)

Etymology
The new species is named after the ichthyologist Jacques Géry (in memoriam) in recognition of his scientifi c contribution on Characiformes.

Description
Morphometric data of holotype and paratypes are presented in Table 3. BODY. Small (with maximum SL of 23.6 mm), compressed, moderately deep, greatest body depth slightly anterior to dorsal-fi n base; body profi le straight and downward directed from end of dorsal fi n to adipose fi n, straight or slightly convex between latter point and origin of dorsal most procurrent caudal-fi n ray; dorsal profi le of head convex from upper lip to vertical through eye; predorsal profi le of body roughly straight, dorsal-fi n base slightly convex, posteroventrally inclined; ventral profi le of head convex from lower jaw to pelvic-fi n origin; straight and posterodorsally slanted along anal-fi n base; and slightly concave on caudal peduncle; jaws equal, mouth terminal. Maxilla reaching vertical to anterior margin of pupil.
FINS. Dorsal-fi n origin at midbody. Base of last dorsal-fi n ray at vertical through fi rst third of anal fi n. Dorsal-fi n rays ii + 9(50)* or iii + 9(2). First dorsal-fi n pterygiophore main body located behind neural spine of 4 th vertebrae. Adipose fi n present. Anal-fi n origin aligned with vertical line through middle of dorsal fi n, between 6 th and 8 th dorsal-fi n rays. Anteriormost anal-fi n pterygiophore inserting posterior to haemal spine of 11 th vertebrae. Anal-fi n origin aligned with vertical line through middle of dorsal fi n (between base of 6 th and 8 th dorsal-fi n rays). Anal fi n iii + 22(40)* or iii + 23(12); Anterior anal-fi n margin slightly convex, with anteriormost rays more elongate and slightly more thickened than remaining rays, forming a distinct lobe. Remaining rays smaller with straight distal margin. Pectoral-fi n rays 12(52) total rays. Tip of pectoral fi n surpassing pelvic-fi n base. Pelvic-fi n rays 8(52) total rays, surpassing anal-fi n origin. Caudal fi n forked, upper and lower lobes similar in size. Principal caudal-fi n rays 11+10(30) or 10+9(22); dorsal procurrent rays 8(8) or 9(2) and ventral procurrent rays 7(3) or 8(7). OSTEOLOGICAL COUNTS. Branchiostegal rays 10(10). First gill arch with 1(8), 2(2) hypobranchial, 11(2), 12(8) ceratobranchial, 1(10) cartilage between ceratobranchial and epibranchial, and 5(2) or 6(8) epibranchial gill-rakers. Supraneurals 4(8) or 5(2). Precaudal vertebrae 11 (10) and caudal vertebrae 19 (10). Total vertebrae 30(10). (Fig. 4) Ground coloration light yellowish brown. Humeral spot conspicuous, slightly vertically elliptical, with a thin vertical line, formed by the concentration of dark chromatophores, at middle of humeral spot, extending one to two scales above and below humeral spot. Flank with conspicuous chromatophores scattered mainly on middle region, from just after humeral spot, reaching caudal peduncle end. Ventral region lacking dark brown chromatophores. Dark brown chromatophores present on head,  (2020) 94 more concentrated on dorsal portion, tip of snout, and opercular region, becoming sparser on cheek. Dorsal-fi n ground coloration hyaline, with conspicuous black or dark brown spot located on anterior portion of fi n, reaching about 6 th ray, approximately between half to two thirds of fi n depth. Conspicuous pigmentation on dorsal-fi n base formed by concentration of conspicuous chromatophores. Anal and caudal fi ns hyaline. Anal and caudal fi n with darker, usually dark brown, posterior margin. Adipose fi n hyaline to light brown, with dark chromatophores at margin and posterior portion. Pectoral and pelvic fi ns hyaline; pelvic fi n with variable amounts of dark brown pigmentation remaining depending on specimen preservation state.

Coloration in alcohol
Color in life (Fig. 5) Color pattern similar to coloration of preserved specimens, with more translucent body. Ground coloration translucent, white to grey, with orange pigmentation on vertebrae region, and usually with orange chromatophores. Ventral region anterior to anal-fi n origin lighter.
Chromatophores on humeral region black. Head darker than fl ank, especially on opercular and dorsal regions; ventrally lighter. Conspicuous black spot on dorsal fi n, with yellow or white pigmentation on dorsal and ventral margins of spot; rest of dorsal fi n hyaline. Anal fi n base with light red chromatophores, with different degrees of intensity, with milk white pigmentation on anterior tip of anal fi n, reaching between fi rst to second rays. Adipose fi n hyaline, with light red chromatophores mainly at base, and black chromatophores at margin. Pectoral and pelvic fi ns hyaline, with some sparser dark brown chromatophores, more concentrated at pelvic fi n, and with milk white pigmentation on anterior tip of pelvic fi n. Caudal fi n with red pigmentation on almost fi n, with inconspicuous light brown margin.

Sexual dimorphism
Bony hooks on fi ns is absent in all examined specimens. According to Malabarba & Weitzman (2003) the presence of bony hooks is a common dimorphic feature among characids species. Although this sexual dimorphism is not observed in all characid species, as in case of species described here. Color pattern is not sexually dimorphic either.

DBC
The alignment comprised a total of 26 haplotypes. After trimming, the fi nal alignment yielded 639 base pairs with 159 polymorphic sites. Average genetic distances were 17%, with the highest values between H. epicharis and H. erythrostigma (23.4%), while the lowest value of (0.08%) was between H. pyrrhonotus and H. erythrostigma (Table 3). Hyphessobrycon frickei sp. nov. and H. geryi sp. nov. are divergent on average 17%, from the other taxa, with a minimum distance of 3.5% and 4.4%, respectively, to H. copelandi, and 4.7% of divergence between them (Table 4), thus both species being recovered by this genetic distance method.

ABDG
In the ABDG species delimitation method, the same ingroup taxa as delimited in the coalescent species delimitation methods have been defi ned comprising: H. caru, H. piorskii, H. geryi sp. nov., H. copelandi, H. frickei sp. nov. and H. aff. copelandi. In constrast to the coalescent methods results, all haplotypes from H. bentosi, the outgroup taxa, were recognized as a single species. The same seven groups (species) were delimited between P values ranging from 0.0017 and 0.0215 (Fig. 8).

Discussion
Recent DNA-based studies suggest that tropical faunas contain a large proportion of undescribed species (e.g., Roca et al. 2001;Hrbek et al. 2014;Poulakakis et al. 2015;Murphy et al. 2016;Bittencourt et al. 2019;Guimarães et al. 2018Guimarães et al. , 2019de Brito et al. 2019). This fact is more evident in freshwater fi shes, since freshwater systems are isolated, hampering the migration of species from one river system into another. Such a situation favors speciation, in many cases resulting in cryptic species, which is common among Neotropical fi sh lineages, as evidenced and revealed by work dealing with species delimitation molecular approaches (e.g., Costa & Amorim 2011;Costa et al. 2012Costa et al. , 2014Costa et al. , 2017Pereira et al. 2011;Benzaquem et al. 2015;Melo et al. 2014Melo et al. , 2016aMelo et al. , 2016bMelo et al. , 2016cAmorim 2018;Mattos & Costa 2018;Carvalho et al. 2018;Guimarães et al. 2018Guimarães et al. , 2019Rosso et al. 2018;Ottoni et al. 2019;de Santana et al. 2019). Recently, the presence of cryptic species and a greater biodiversity have been also evidenced for the "Rosy tetra" clade (e.g., Castro-Paz et al. 2014;Guimarães et al. 2018Guimarães et al. , 2019, which is herein corroborated, with the description of two new species belonging to this clade, based on seven different species delimitations methods. In the present research, we suggest a new clade within the "Rosy tetra", the Hyphessobrycon copelandi clade, comprising four species: Hyphessobrycon copelandi, with its type locality at Tabatinga, Amazonas, Brazil, close to the border with Peru and Colombia (Eigenmann 1908); the two new species herein described, Hyphessobrycon frickei Guimarães, Brito, Bragança, Katz & Ottoni sp. nov. and H. geryi Guimarães, Brito, Bragança, Katz & Ottoni sp. nov. from the Maracaçumé River basin in northeastern Brazil and the middle Tocantins River basin in east Amazon, Brazil, respectively; and an undescribed species from the lower Rio Negro basin (here called H. aff. copelandi). The Hyphessobrycon copelandi clade was corroborated by maximum node support (see Fig. 7), as well as, by 25 synapomorphic nucleotide substitutions (see Fig. 7 and Appendix 2).
Hyphessobrycon frickei sp. nov. and H. geryi sp. nov. were identifi ed as new species by seven different and independent methods of species delimitation (PAA, DBC, WP, CBB, bPTP, GMYC and ABGD), relying on different sources of characters (morphology and DNA), different criteria and assumptions; under an Integrative Taxonomy perspective (Sytsma & Schaal 1985;Dayrat 2005;Goldstein & Desalle 2010;Padial et al. 2010). Both species are distinguished from all their congeners mainly by the arrangement, shape and color pattern of the humeral and dorsal-fi n spots, as well as by other characters related to scale counts and body pigmentation (see PAA -Diagnosis). In our bayesian inference phylogenetic analysis (Fig. 7), haplotypes belonging to both species formed two exclusive clades with maximum posterior probability value (posterior probability = 1) (WP), as well as, the other two species within the Hyphessobrycon copelandi clade. Furthermore, the minimum COI genetic distance (DBC) of Hyphessobrycon frickei sp. nov. and H. geryi sp. nov. when compared to H. copelandi were 3.5% and 4.4%, respectively, and 4.7% of divergence between them (see Table 4). Considering this value, Hyphessobrycon frickei sp. nov. and H. geryi sp. nov. genetic divergence is greater than that commonly inferred by species delimitations among Neotropical fi sh species (2%) based on COI (Jacobina et al. 2018). Moreover, Hyphessobrycon frickei and H. geryi sp. nov. were also molecularly diagnosed by synapomorphic nucleotide substitutions, as well as by an accessory combination of other nucleotide substitutions (see CBB -molecular diagnosis sections, Fig. 7 and Appendix 2). At last, both coalescent species delimitation methods (GMYC and bPTP) corroborated these two new species, as well as, the Automatic Barcode Gap Discovery (ABGD) (Fig. 8). The same result found by all the methods applied here only reinforces the hypothesis that these species are in fact new to science.
Similar results from previous studies (e.g., Castro-Paz et al. 2014;Guimarães et al. 2018Guimarães et al. , 2019 agree with our hypothesis of the existence of cryptic species within the "Rosy tetra" clade, and also suggests that its species diversity is still underestimated. We would like to highlight the paper published by Guimarães et al. (2018Guimarães et al. ( , 2019 that also proposed and evidenced a new subclade within the "Rosy tetra" clade, the Hyphessobrycon micropterus clade, revealing the existence of cryptic species and a greater diversity than previously known within the group.

Conclusion
The present paper highlights the importance of a species delimitation through an integrative framework allowing to solve taxonomic and classifi cation problems in Hyphessobrycon taxonomy, and estimating the diversity of this group with accuracy, being crucial for futures studies on phylogeny, phylogeography, ecology, conservation and biogeography. We consider that diffi culties to recognize the boundaries at the species or genus levels can be overcome through the adoption of integrative studies and perspectives, dealing with multiple criteria and character sources to make taxonomic decisions.
of the data analyzed in this study; Wilson Costa (UFRJ) for the loan and donation of material; Andrew Williston (MCZ), James Maclaine and Kevin Swagel (BMNH), Mark Sabaj Pérez and Dave Catania (CAS), Riedel Bettina (NMW) and Sandra Raredon (USNM) for providing photographs, x-ray images, and information on the type material of some species; CAPES (Coordenação de Aperfeiçoamento de pessoal de nível Superior -Finance Code 001) and FAPEMA for providing the scholarship to PSB under the process 88887.159561/2017-00. All material was collected with permits 51540-3/ from SISBIO (Brazilian Institute of Environment and Natural Resources).