Molecular phylogeny of the subfamily Stevardiinae Gill, 1858 (Characiformes: Characidae): classification and the evolution of reproductive traits
© Thomaz et al. 2015
Received: 13 January 2015
Accepted: 2 June 2015
Published: 21 July 2015
The Erratum to this article has been published in BMC Evolutionary Biology 2015 15:269
The subfamily Stevardiinae is a diverse and widely distributed clade of freshwater fishes from South and Central America, commonly known as “tetras” (Characidae). The group was named “clade A” when first proposed as a monophyletic unit of Characidae and later designated as a subfamily. Stevardiinae includes 48 genera and around 310 valid species with many species presenting inseminating reproductive strategy. No global hypothesis of relationships is available for this group and currently many genera are listed as incertae sedis or are suspected to be non-monophyletic.
We present a molecular phylogeny with the largest number of stevardiine species analyzed so far, including 355 samples representing 153 putative species distributed in 32 genera, to test the group’s monophyly and internal relationships. The phylogeny was inferred using DNA sequence data from seven gene fragments (mtDNA: 12S, 16S and COI; nuclear: RAG1, RAG2, MYH6 and PTR). The results support the Stevardiinae as a monophyletic group and a detailed hypothesis of the internal relationships for this subfamily.
A revised classification based on the molecular phylogeny is proposed that includes seven tribes and also defines monophyletic genera, including a resurrected genus Eretmobrycon, and new definitions for Diapoma, Hemibrycon, Bryconamericus sensu stricto, and Knodus sensu stricto, placing some small genera as junior synonyms. Inseminating species are distributed in several clades suggesting that reproductive strategy is evolutionarily labile in this group of fishes.
The family Characidae is the largest family of freshwater fishes in the Neotropics, comprising around 1065 species in approximately 146 genera . Because of its considerable species richness and diversity, the relationships and limits of the main lineages in the family have been controversial: two-thirds of all species were considered incertae sedis in the family just a decade ago due to lack of consistent information on their relationships . Besides species richness, a primary challenge to establishing relationships based on morphological phylogenies has been associated with their conservative morphology though their long period of evolution, since characid fossils with essentially modern morphologies are known from Eocene-Oligocene deposits . Taken together, these factors may explain the high level of morphological homoplasy inferred across lineages within the family (e.g. [4, 5]). In recent years, an increasing understanding of the relationships among major lineages within Characidae is emerging on the basis of evidence provided by molecules, osteology, and primary and secondary sexual characters [4, 6–12].
Current classification of Stevardiinae
No. of species
Diapoma Cope, 1894
Planaltina Böhlke, 1954
Acrobrycon Eigenmann and Pearson, 1924
Glandulocauda Eigenmann, 1911
Lophiobrycon Castro, Ribeiro, Benine and Melo, 2003 
Mimagoniates Regan, 1907
Hysteronotus Eigenmann, 1911
Pseudocorynopoma Perugia, 1891
Landonia Eigenmann and Henn, 1914
Corynopoma Gill, 1858
Gephyrocharax Eigenmann, 1912
Pterobrycon Eigenmann, 1913
Phenacobrycon Eigenmann, 1922
Argopleura Eigenmann, 1913
Chrysobrycon Weitzman and Menezes, 1998
Iotabrycon Roberts, 1973
Ptychocharax Weitzman, Fink, Machado-Allison and Royero, 1994
Scopaeocharax Weitzman and Fink, 1985
Tyttocharax Fowler, 1913
Xenurobrycon Myers and Miranda Ribeiro, 1945
Attonitus Vari and Ortega, 2000
Aulixidens Böhlke, 1952 
Boehlkea Géry, 1966
Bryconacidnus Myers, 1929
Bryconadenos Weitzman, Menezes, Evers and Burns, 2005 
Bryconamericus Eigenmann, 1907
Caiapobrycon Malabarba and Vari, 2000
Carlastyanax Géry, 1972 
Ceratobranchia Eigenmann, 1914
Creagrutus Günther, 1864
Cyanocharax Malabarba and Weitzman, 2003
Cyanogaster Mattox, Britz, Toledo-Piza and Marinho, 2013 
Hemibrycon Günther, 1864
Hypobrycon Malabarba and Malabarba, 1994
Knodus Eigenmann, 1911
Lepidocharax Ferreira, Menezes and Quagio-Grassiotto, 2011 
Markiana Eigenmann, 1903 
Microgenys Eigenmann, 1913
Monotocheirodon Eigenmann and Pearson, 1924
Nantis Mirande, Aguilera and Azpelicueta, 2006 
Odontostoechus Gomes, 1947
Othonocheirodus Myers, 1927
Phallobrycon Menezes, Ferreira and Netto-Ferreira, 2009 
Piabarchus Myers, 1928
Piabina Reinhardt, 1867
Rhinobrycon Myers, 1944
Rhinopetitia Géry, 1964
Trochilocharax Zarske, 2010 
The presence of modified scales in the caudal fin of males has been the main character defining the Glandulocaudinae since the group was described . The complex morphologies of these caudal organs were explored in detail in phylogenetic studies to test the monophyly and internal relationships of the glandulocaudin tribe Xenurobryconini  and to diagnose and propose internal relationships within Glandulocaudinae. On this basis, Glandulocaudinae was divided into seven tribes : Corynopomini Eigenmann, 1927 (renamed Stevardiini in 2005  – see below), Diapomini Eigenmann, 1910, Glandulocaudini Eigenmann, 1914, Hysteronotini Eigenmann, 1914, Landonini Weitzman and Menezes, 1998, Phenacobryconini Weitzman and Menezes, 1998 and Xenurobryconini Myers and Böhlke, 1956 (Table 1). This classification was further modified in light of new histological evidence from the caudal organs of males . As a consequence, the subfamily Glandulocaudinae was restricted to the tribe Glandulocaudini sensu Weitzman and Menezes  and the six remaining tribes were placed in the resurrected subfamily Stevardiinae . The family group name Corynopomini is a junior synonym, being this tribe consequently also renamed as Stevardiini (see Additional file 1 for nomenclatural remarks).
Monophyly of “clade A”, including the Glandulocaudinae and Stevardiinae sensu Weitzman, Menezes, Evers and Burns  and the 18 incertae sedis genera listed previously , has been supported more recently on the basis of both morphological [4, 11, 20, 21] (Fig. 1b, c) and molecular data [7, 10, 12] (Fig. 2). Furthermore, later studies have added ten genera to “clade A” after the original definition , such that it currently includes 48 genera (Table 1). In a morphological phylogeny of 160 characiform species, including 23 species and 14 genera of “clade A”, the Stevardiinae sensu Weitzman and collaborators  was reported to be paraphyletic given that it also included Glandulocaudinae [4, 8] (Fig. 1b). Based on this inference, a more comprehensive concept of the name Stevardiinae was advanced embracing all members of “clade A”, with Glandulocaudinae being lowered in rank to a monophyletic tribe (Glandulocaudini) within Stevardiinae  (Table 1, see nomenclatural remarks in Additional file 1). The newly defined Stevardiinae sensu Mirande  was diagnosed on the basis of three synapomorphies: (i) the previously mentioned possession of eight branched dorsal-fin rays , (ii) the absence of the epiphyseal branch of the supraorbital canal, and (iii) the presence of nine dorsal-fin pterygiophores. For clarity, “clade A” was then named Stevardiinae and, as presently recognized, is widely distributed in the Neotropics on both sides of the Andes, from Costa Rica in Central America to central Argentina.
Insemination strategy and sperm morphology
Species analyzed for sperm morphology
D. speculiferum 
L. weitzmani 
H. megalostomus 
P. doriae 
L. latidens 
P. henna 
C. riise 
C. riisei 
Chrysobrycon sp. 
I. praecox 
P. rhyacophila 
S. rhinodus 
B. fredcochui 
B. tanaothoros 
B. tanaothoros 
B. ellisi 
B. pectinatus 
B. alpha, B. deuterodonoides, B. exodon, B. iheringii, B. pachacuti, Bryconamericus sp.
B. exodon 
C. obtusirostris 
C. affinis, C. britskii, C. changae, C. cochui, C. figueiredoi, C. holmi, C. menezesi, C. paralacus, C. taphorni Creagrutus sp.
C. meridionalis 
C. alburnus, C. alegretensis, C. dicropotamicus, C. itaimbe, C. lepiclastus, C. macropinna
C. alburnus 
H. dariensis, H. metae
Knodus sp. 
K. beta, K. breviceps, K. meridae, K. septentrionalis, K. turiuba, Knodus sp.
K. meridae 
O. lethostigmus 
R. negrensis 
Except for two species of Monotocheirodon that bear intromittent organs, all remaining species of Stevardiinae lack copulatory organs [24, 25]. Among these, insemination strategy has been documented by the presence of sperm in females through histological examination of the ovaries for all species of the tribes Diapomini, Glandulocaudini, Hysteronotini, Landonini, Phenacobryconini, Stevardiini and Xenurobryconini, all the species of Attonitus, Bryconadenos and Monotocheirodon and some species of Bryconamericus (B. pectinatus), Creagrutus (C. lepidus and C. melasma) and Knodus (Knodus sp.)  (Table 2). Inseminating species, however, also have been described in other lineages of Characidae, such as the tribe Compsurini within the subfamily Cheirodontinae  and in a clade formed by Hollandichthys plus Rachoviscus . Taking this pattern at face value, insemination seems to have at least three independent origins within Characidae .
The presence of insemination correlates with differences in sperm morphology. While in most externally fertilizing teleosts the spermatozoa are characterized by a spherical to ovoid nucleus and short midpiece, in inseminating fishes an elongated nucleus is the most frequently observed [23, 27]. This elongation may be advantageous for insemination over ovoid nucleus since it facilitates sperm movement through the female gonopore and within the female reproductive tract, increasing directional movement toward female gonopore, and facilitating the formation of sperm packets to be moved by the male urogenital papilla to the female urogenital pore .
Three unique morphotypes of sperm have been described among species of Stevardiinae (M1, M2, and M3; Table 2), based on arrangement of centrioles, flagellum, nucleus, and the midpiece  (Table 2). Based on the phylogenetic evidence available , it has been proposed that insemination is likely to have evolved only once within Stevardiinae and may constitute a synapomorphy shared by the most derived species of this subfamily . However, recent morphological phylogenies [4, 8, 20, 21] (Fig. 1) still have limited taxonomic representation to effectively test this hypothesis which is, likewise, not supported by molecular evidence given the absence of a monophyletic group with all inseminating stevardiines [10, 12] (Fig. 2). In terms of sperm morphology, it has been proposed that sperm type M1 is synapomorphic to Stevardiinae, M2 synapomorphic to all the inseminating species of the Stevardiinae, and M3 synapomorphic to the Xenurobryconini . Information about these traits related to reproduction, however, are just known for less than 1/3 of all stevardiines species, which hamper a better understading of the evolution of the reproductive strategy in this subfamily.
Most of the species-level diversity in Stevardiinae is contained in only four genera (out of 48) that include 200 out of 311 nominal species described for the subfamily (Table 1). Among these four genera, Creagrutus (71 species) is the only one supported as a monophyletic group based on apomorphic features associated to jaws and teeth . The remaining three genera Bryconamericus (78 species), Hemibrycon (31 species) and Knodus (20 species) have been traditionally and arbitrarily diagnosed using pre-cladistic criteria based on the number of teeth on the maxilla and on the extension of scales over the caudal-fin rays . Not surprisingly, recently published morphological- and molecular-based characid phylogenies found that Bryconamericus and Knodus are polyphyletic groups (Fig. 1b ; Fig. 2b–c [10, 12]), thus demonstrating the need of further study to diagnose monophyletic genera based on consistent phylogenetic evidence.
This study presents phylogenetic relationships for a large number of taxa of Stevardiinae based on analysis of a multi-locus data set to address two main goals: (i) test the monophyly of the Stevardiinae and the included putative tribes and genera, with emphasis on the species-rich genera Bryconamericus, Hemibrycon, and Knodus; and (ii) shed light on the evolution of insemination and secondary sexual dimorphism among stevardiines, specifically whether insemination had a single origin among members of this subfamily.
Summary information of molecular data analyzed in this study
Number of sequences
% present data
Number of variable sites
Overall mean genetic distance (p-value)
Clade support obtained with different phylogenetic methods
Diapomini + Lepidocharax
Bryconamericus, new circumscription
Diapoma, new circumscription
Eretmobrycon, new circumscription
Hemibrycon, new circumscription
Knodus, new circumscription
Piabarchus, new circumscription
Piabina, new circumscription
Topology tests for currently accepted taxonomic groups
Δ In L
Based on the results presented here, we proposed a new classification of Stevardiinae, (see Additional file 5 – New Stevardiinae classification), subdividing the subfamily into seven tribes, redefining some generic circumscriptions, and leaving only 11 of the 48 genera as incertae sedis, as discussed below.
In agreement with previous molecular [7, 10, 12] and morphological studies [4, 8, 21], the monophyly of Stevardiinae or “clade A”  was resolved with confidence in our results (Figs. 3–10). The most comprehensive morphological analysis was based on 91 species in 20 genera  and a large-scale molecular phylogenetic analysis of the Characidae included 23 stevardiine species in 21 genera and along with a broad taxonomic representation of other subfamilies . None of these previous studies, however, resolved internal relationships within Stevardiinae due to limited taxonomic sampling. Here we present analyses of a large and comprehensive sampling of Stevardiinae, with 32 genera and around 153 species/morphotypes, which identification was based on morphology and geographic location (see Additional file 2), and propose a new classification (Fig. 3, Additional file 5) based on monophyletic units (tribes and genera) supported by our results. The following sections address morphological, reproductive and geographic distributional data in relation to the major clades supported by this study, as well as limitations imposed by missing data and ambiguous phylogenetic resolution in subsections of the phylogeny.
Monophyly of the Stevardiinae
Our results are consistent with the definition of “clade A” based on the presence of four teeth in the inner row of the premaxilla  (reversed to five teeth in Nantis), but not with the placement of Bryconamericus scleroparius clade and Markiana (positioned in Astyanax clade) outside the Stevardiinae [4, 8, 21] (Fig. 1b). This alternative hypothesis requires the separate origin of this trait in Markiana and in a clade containing the subfamilies Aphyocharacinae, Aphyoditeinae, Cheirodontinae, Gymnocharacinae, and Stevardiinae plus the Bryconamericus scleroparius clade, with a reversal to five or more teeth (in a single series) in a less inclusive clade containing Aphyocharacinae, Aphyoditeinae, and Cheirodontinae. The inclusion of the Bryconamericus scleroparius clade and Markiana in the Stevardiinae herein contradicts this hypothesis and resolves the presence of four teeth in the inner series of the premaxilla as a diagnostic character for the Stevardiinae (reversed to five teeth in Nantis).
A dorsal fin with ii+8 fin rays  is another diagnostic character proposed for “clade A”; although some authors split this trait into two characters [4, 8, 21]: eight or fewer branched dorsal-fin rays and nine or fewer dorsal-fin pterygiophores. According to our results (Figs. 3 and 4), the ii+8 dorsal-fin rays would not be a diagnostic character for Stevardiinae, but for a less inclusive clade including all Stevardiinae except Markiana plus Bryconamericus scleroparius group (the new tribe Eretmobryconini - see below). The inferred order of appearance of these diagnostic characters along the tree is informative as to the placement of two fossil species of Paleotetra from the Eocene-Oligocene (P. entrecorregos and P. aiuruoca) to support the hypothesis that these fossils constitute a stem lineage of Stevardiinae  or potentially within the tribe Eretmobryconini, since these fossils have four teeth in the inner series of the premaxilla but ii+9 dorsal-fin rays. Secondary increases in the number of dorsal-fin rays are recorded in Pseudocorynopoma doriae (ii+9), Mimagoniates rheocharis (ii+8-12), Chrysobrycon myersi (ii+9-10), differing from the other species of these genera with ii+8 dorsal-fin rays .
Seven strongly supported monophyletic groups observed in the phylogeny (Fig. 3), herein designated tribe-level taxa are discussed in the following sections. Argopleura is the only genus that could not be included in any tribe with confidence but it is closely related to Glandulocaudini and Stevardiini (as newly defined herein).
Eretmobryconini, new tribe
The East Andean genus Markiana and the West Andean and Central American species of Bryconamericus (B. bayano, B. brevirostris, B. dahli, B. emperador, B. gonzalezi, B. miraensis, B. peruanus, B. scleroparius, and B. terrabensis) form a strongly supported clade (100 % boostrap) that is the sister group of the remaining Stevardiinae. They share with Stevardiinae the apomorphic presence of four teeth in the inner series of the premaxilla, but have ii+9 dorsal fin rays, which is considered a plesiomorphic state in Characidae.
Markiana is a monophyletic genus with two geographically disjunct species distributed in the Rio Orinoco and the Paraná, Paraguay and Mamoré river basins. Its relationships with other taxa were not clearly resolved and hence the genus was treated as incertae sedis in Characidae . Markiana was first proposed as belonging to Stevardiinae  based on spermiogenesis, sperm morphology, the possession of four teeth in the inner series of the premaxilla, and short triangular ectopterygoid. Consistent with our results, this genus has been found related with the Central American characid species Bryconamericus emperador , currently assigned to the Bryconamericus scleroparius group [8, 21] and this clade formed the sister group to the remaining Stevardiinae. Although Mirande and collaborators [4, 8, 21] favored a grouping of Markiana with Astyanax, separately from the Bryconamericus scleroparius group and from the Stevardiinae, they found this relationship variable and with low stability in their analyses, and noted that Markiana and the Bryconamericus scleroparius group “share the absence of an ossified rhinosphenoid, an overlap of the horizontal arm of the preopercle by the third infraorbital, the possession of only four teeth on the inner premaxillary row, and the presence of two uroneurals” . Under self-weighted parsimony optimization  and in other analyses , these characters support the monophyly of a group formed by these two taxa. Taken together, all the evidence strongly supports the recognition of Markiana plus the Bryconamericus scleroparius group as a monophyletic unit, and contradicts alternative hypotheses grouping Markiana and Astyanax.
The species of the Bryconamericus scleroparius group form a clade both in our hypothesis based on molecular data (node 1, Fig. 4: B. bayano, B. brevirostris, B. dahli, B. emperador, B. gonzalezi, B. miraensis, B. peruanus, B. scleroparius, B. terrabensis, and B. sp. from Río Patia) and on analysis based on morphological data [4, 8] (B. brevirostris, B. emperador, B. peruanus, B. guaytarae, B. scleroparius, and B. simus). These species are not closely related to Bryconamericus exodon, the type species of the genus, which is placed in a distant position in the phylogeny (node 8, Fig. 9), and must be recognized as a group separate from Bryconamericus. In addition, the monophyly of Bryconamericus is strongly rejected by our molecular data (Table 5). Among the species of Bryconamericus scleroparius group, B. bayano is the type species of the genus Eretmobrycon Fink, 1976, which was synonymized with Bryconamericus . Based on our results and in previous results from morphology and sperm data [8, 21, 31], we propose the revalidation of the genus Eretmobrycon and the inclusion of all Bryconamericus species present in node 1 (B. scleroparius group; Fig. 4) within this genus (E. bayano, E. brevirostris, E. dahli, E. emperador, E. gonzalezi, E. miraensis, E. peruanus, E. scleroparius, and E. terrabensis). Also, Bryconamericus guaytarae and B. simus, not examined here, were tentatively resolved within the B. scleoparius group  and, because of it, are also tentatively included in Eretmobrycon. Even though the number of ii+9 dorsal-fin rays is plesiomorphic, it can be used as a further character to distinguish the species of Eretmobrycon from Bryconamericus.
Xenurobryconini Myers and Böhlke, 1956, new usage
The tribe Xenurobryconini, as currently defined, contains seven genera: Argopleura, Chrysobrycon, Iotabrycon, Ptychocharax, Scopaeochrax, Tyttocharax and Xenurobrycon (Table 1). This group was first proposed for Xenurobrycon and Tyttocharax  based on morphological similarities of their caudal fins. Tyttocharax was later split in two genera: Tyttocharax and Scopaeocharax , and in this same study Argopleura and Iotabrycon were added as members of the Xenurobryconini. Ptychocharax and Chrysobrycon were described later [13, 34] and also added to the tribe.
The monophyly of this tribe is strongly rejected by our molecular data (Table 5). Instead, our results (Figs. 3 and 4 and Table 4) resolved a highly supported clade (Scopaeocharax (Tyttocharax + Xenurobrycon)) congruent with “Subgroup B xenurobryconins” sensu Weitzman and Fink , diagnosed by 20 morphological synapomorphies. Monophyly of this group is further supported by the apomorphic sperm morphotype M3 shared by these three genera  (Table 2). Iotabrycon and Ptychocharax (not examined here) were found as successive sister groups to (Xenurobrycon (Scopaeocharax + Tyttocharax)) , but further investigation is necessary to test their membership within Xenurobryconini since molecular or sperm ultrastructure information are currently unavailable for these two genera. We provisionally list them within Xenurobryconini (Additional file 5).
The other two genera, Argopleura and Chrysobrycon, are more closely related to the tribes Glandulocaudini and Stevardiini (Fig. 5). In most of our phylogenetic results, Argopleura is resolved as the sister group of Glandulocaudini, however in the ML tree reconstructed in RAxML it is placed as sister group of Glandulocaudini and Stevardiini (70 % bootstrap). Since affinities of Argopleura with Glandulocaudini and Stevardiini are not clearly resolved and Argopleura could be potentially included in Gladulocaudini (pending further investigation), this genus is temporarily placed as incertae sedis in Stevardiinae. The new circumscription of the tribe Xenurobryconini is restricted to the genera Scopaeocharax, Tyttocharax, and Xenurobrycon, and possibly Iotabrycon and Ptychocharax.
Glandulocaudini Eigenmann, 1914 sensu Menezes and Weitzman, 2009
The monophyly of the tribe Glandulocaudini including the genera Mimagoniates as sister group of Lophiobrycon and Glandulocauda is supported by our results (Fig. 5). The monophyly of Glandulocaudini was hypothesized  based on the presence of modified caudal peduncle squamation extending onto the caudal fin from the ventral region of the dorsal caudal-fin lobe and by the presence of modified club cells on the caudal organ, which probably secretes a pheromone during courtship. All the species in this group are inseminating. Relationships within Glandulocaudini, grouping Lophiobrycon as sister group to Glandulocauda receive high bootstrap support (Figs. 3 and 5), rejecting previous hypothesis that placed Lophiobrycon as the sister group of the other two genera in this clade [35, 36].
Stevardiini Gill, 1858, new usage
Monophyly of a clade composed by the tribes Hysteronotini, Phenacobryconini (not analyzed herein), Stevardiini and Xenurobryconini has been proposed [13, 19] based mostly on the morphology and histology of the caudal organ. The molecular data resolved a monophyletic group including Chrysobrycon (considered part of Xenurobryconini ) with other taxa previously assigned to Stevardiini and Hysteronotini (Fig. 5). The Stevardiini (= Corynopomini sensu Weitzman and Menezes ; Table 1) is resolved as a strongly supported monophyletic group that includes Corynopoma riisei and the four species of Gephyrocharax (Pterobrycon was not examined; Table 5). This clade is the sister group of a clade that includes Pseudocorynopoma (one of the two genera of the Hysteronotini), Chrysobrycon myersi, and a characid from the southwestern Amazon (Madre de Dios, Ucayali and Yuruá basins; see : figure of Gephyrocharax sp.; MUSM 33860, 38.3 mm SL) tentatively assigned to Gephyrocharax. Strong support of node 2 (Fig. 5) suggests that this form could be assigned to Chrysobrycon, with whom it also shares a distribution in western Amazonia, while other species assigned to Gephyrocharax are distributed in the Orinoco, Atrato, and Central American rivers, but further study may be necessary to fully resolve this issue.
In conclusion, we recognize an extended tribe Stevardiini that includes Chrysobrycon and Pseudocorynopoma, and possibly Hysteronotus, in addition to the three genera currently recognized in this tribe (Corynopoma, Gephyrocharax, and Pterobrycon). Further analysis of Pterobrycon and Hysteronorus is necessary to resolve of the final composition of this tribe.
Hemibryconini Géry, 1966, new usage
The Hemibryconini was first introduced  to refer to a large group of characids consisting of Boehlkea, Bryconacidnus, Bryconamericus, Ceratobranchia, Coptobrycon, Hemibrycon, Knodus, Microgenys, Nematobrycon, Piabarchus, Rhinobrycon, and Rhinopetitia. We adopt the name but modify the circumscription to recognize a well-supported clade that includes solely Acrobrycon, Boehlkea and Hemibrycon. These genera are characterized by the presence of teeth along more than one-half the length of the dentigerous margin of the maxilla [39, 40], which may constitute a diagnostic morphological character for the tribe. Although we did not examine Boehlkea in this study, it is provisionally included in Hemibryconini, based on this morphological evidence.
Previous studies that addressed relationships within Stevardiinae using sexually dimorphic characters inferred a sister-group relationship between Acrobrycon and Diapoma plus Planaltina . Other studies based on osteological and external morphological characters failed to support this relationship, suggesting instead that Acrobrycon is most closely related to Mimagoniates, Pseudocorynopoma, and Diapoma [4, 8]. Similarly, other morphological studies have suggested a close relationship between either Boehlkea and Hemibrycon  or Bryconamericus and Hemibrycon . The sister-group relationship between Hemibrycon and Bryconamericus lacks morphological or molecular support, which is not surprising since the monophyly of Bryconamericus has been strongly rejected by several independent studies [4, 10, 12]. In contrast to previous studies, we find strong support for the sister group relationship between Acrobrycon and Hemibrycon (Fig. 6).
Under the phylogenetic hypothesis presented in Fig. 6 (node 3), Hemibrycon is paraphyletic, including five species originally described in Bryconamericus (B. cristiani, B. caucanus, B. galvisi and B. plutarcoi). These species are not closely related to Bryconamericus exodon, the type species of the genus, which is placed in a distant position in the phylogeny (node 8, Fig. 9), and, therefore, should be recognized as a group separate from Bryconamericus. We reassign these five species to Hemibrycon (Additional file 5). Hemibrycon is found in both sides of the Andes, with at least two separations, showing a peripheral distribution pattern within the Amazon basin .
Our results also highlight issues with delimitation of species in Hemibrycon. For instance, the clade from the upper Río Cauca basin includes three nominal species with negligible genetic differentiation: H. boquiae, H. brevispini, and H. quindos (Fig. 6). Based on these results and taking into account the lack of strong morphological evidence diagnosing these nominal species [41, 42], it is likely that H. brevispini, and H. quindos are junior synonyms of H. boquiae. The specimen from the upper Cauca labeled Bryconamericus caucanus is likely misidentified and should be included in H. boquiae. Other specimens assigned to Bryconamericus caucanus form a differentiated and well-supported clade with H. jabonero. Further study of this group is warranted.
Creagrutini Miles, 1943, new usage
The affinities between Carlastyanax and Creagrutus were recently explored , and seven morphological synapomorphies were identified uniting these genera, with one of these being unique among characids (the presence of a ligament between the ascending process of the maxilla and the dorsal margin of the alveolar premaxillary ramus). This clade is well supported in our analysis (node 4, Fig. 7) that also confirms the re-validation of Carlastyanax at the rank of genus and as sister group to Creagrutus . However, according to our results Piabina is not the sister group of Creagrutus or Creagrutus plus Carlastyanax, rejecting previous proposals [16, 21]. As in Hemibrycon, Creagrutus has a distribution to the two sides of the Andes. Nested among the Amazonian species of Creagrutus, the molecular data include an enigmatic taxon collected from the Río Marañón (Perú), identified at this time as Characidae sp. n. that differs from Creagrutus in having flattened multicuspids teeth on the premaxilla, as opposed to the massive teeth typical of Creagrutus. This taxon is closely related to C. muelleri (Fig. 7), another species from the western Amazon basins, strongly suggesting that it should be assigned to Creagrutus.
Diapomini Eigenmann, 1909, new usage
A large clade including the remaining taxa in Stevardiinae is supported by the molecular data (bootstrap 83 %) and herein named Diapomini, differing radically from the current usage  (Table 1). Diapomini herein includes taxa assigned to Attonitus, Bryconacidnus, Ceratobranchia, Cyanocharax, Diapoma, Hypobrycon, Knodus, Nantis, Odontostechus, Piabina, Piabarchus and Rhinobrycon, in addition to a large number of species of Bryconamericus plus Hyphessobrycon guarani. Within this tribe there are three large clades of which two clades with an Amazon-Orinoco basin distribution are weakly supported and one clade with a southern South America distribution is highly supported (see below). In addition, Lepidocharax was found  as sister group of all former Stevardiinae , except Landonia and Glandulocauda. In our results, Lepidocharax is instead placed as sister group of Diapomini in all our phylogenetic results (except in STAR tree), although with low support (Table 4; Figs. 3 and 8). However, because of the consistency found among trees where Lepidocharax is sister group of the Diapomini, we tentatively assign Lepidocharax as member of this tribe, but we highlight the necessity of further investigation.
The species with Amazon-Orinoco distribution are split into two weakly supported clades (Figs. 8–9), both of which contain specimens assigned to Knodus interspersed with other taxa. The monophyly of this genus is strongly rejected by the molecular data (Table 5). The paraphyly of Knodus with some species of Bryconamericus was already pointed in a morphological study . Consistent with this hypothesis, our results support the separation of this genus in two clades. In the first Amazonian clade, a number of Bryconamericus species from the Amazon and Orinoco basins (B. caquetae, B. cinarucoenses, B. deuterodonoides, B. alpha and B. orteguasae) are grouped in a clade with nearly all of the species of Knodus analyzed here (node 5, Fig. 8), including its type species K. meridae, and the inseminating Bryconadenos tanaothoros from the Rio Xingu. Although the node subtending this group (node 5) received low bootstrap support (35 %), it is supported in all our results (Table 4), suggesting that Knodus sensu stricto could be circumscribed to this clade (Fig. 3). Bryconadenos was hypothesized as closely related to Attonitus , but our results support the inclusion of Bryconadenos in the genus Knodus as defined above. Some species definitions within this clade, most notably K. smithi from the Río Purus should be revised, since it does not group with other taxa from the Urubamba and Yuruá rivers assigned to the same species. Knodus sensu stricto is the sister group of Rhinobrycon negrensis (Fig. 8).
In the second Amazonian clade, a well-supported group (node 6, Fig. 9) includes the inseminating Bryconamericus pectinatus along with Knodus hypopterus, Knodus sp. from Madre de Dios and Urubamba (Perú) and two unidentified species of Bryconacidnus. This clade (“Bryconacidnus” clade) does not include the type species of either Knodus (node 5) or Bryconamericus (node 8, Fig. 9; see below), and the strong support for this group suggests that all these species could be assigned to Bryconacidnus, pending further study that includes the type species of this genus (B. ellisi).
Relationships of Bryconamericus pachacuti from the upper Amazon in Peru remain somewhat uncertain but our results place it as the sister group to Attonitus (Fig. 9) with low bootstrap support (38 %), and separated from B. exodon by five nodes. Generic assignment of B. pachacuti cannot be resolved with confidence, but it is tentatively retained in “Bryconamericus” pending further study. Ceratobranchia and Attonitus are resolved as monophyletic with confidence and placed with low support as closely related to the “Bryconacidnus” clade (Fig. 9).
The remaining members of the Diapomini form a strongly supported clade (bs = 99 %: node 7, Fig. 9) that occupies the southern portion of the range of the Stervardiinae and includes the species of Bryconamericus analyzed in our study and not discussed so far, including the type species of the genus, B. exodon, and the species of the remaining genera listed for the Diapomini.
The type species of Bryconamericus is grouped with node 8 (Fig. 9). Although this node receives relatively low bootstrap support (60 %), it is present in all results obtained in this study (Table 4). The sister group to Bryconamericus exodon is a clade (bs = 90 %, Fig. 9) formed by several species of Bryconamericus (B. iheringii, B. ikaa, B. lethostigma, B. microcephalus, B. patriciae, B. rubropictus, and B. uporas), and other taxa assigned to three small genera Hypobrycon, Nantis, and Odontostoechus. While describing the genus Hypobrycon, it has been hypothesized that some of the species described in Bryconamericus (e.g., B. iheringii) would be possibly more closely related to this new genus than to B. exodon, and that the definition of Hypobrycon could be expanded to include these species . Our finding corroborates the grouping of Hypobrycon with Bryconamericus iheringii and some other congeners but the recognition of Hypobrycon or of the monotypic genera Nantis and Odontostoechus as valid genera would demand the recognition of several small genera in node 8. Odontostoechus has been described originally in Cheirodontinae based on the presence of a single series of teeth in the premaxilla, but this seems to be an autapomorphy of the type species, O. lethostigmus. Similarly, the diagnostic characters of Nantis seem to constitute autapomorphies of the type species. Given that Bryconamericus has priority over all other nominal genera in this clade, our results provide the basis for defining a monophyletic genus Bryconamericus sensu stricto that includes all the species subtended by node 8 (Fig. 9). Further study of this group is warranted given the morphological diversity of the studied species. Other species of Bryconamericus not subtended by node 8 and not discussed in this study, would be retained in “Bryconamericus” until further study may justify shifting then to other genera (Additional file 5).
Piabarchus analis plus Bryconamericus stramineus (node 9, Fig. 10) and Bryconamericus thomasi plus Piabina (node 10, Fig. 10) form two well-supported groups. These results suggest that the two species assigned to Bryconamericus should be reassigned to Piabarchus and Piabina, respectively (Fig. 10), since they are only distantly related to the type species Bryconamericus exodon. A specimen from the Amazon basin in Peru assigned to Piabarchus analis [ROM 55742, 300 m E of Panguana camp, isolated pool of a Llulapichis river tributary, Peru, Huanuco, Ucayali River drainage, collected by Erling on 23 July 1988] was misidentified and corresponds to Gephyrocharax (Lopéz-Fernandéz, Taphorn, and Vanegas-Rios, pers. comm.). Therefore the distribution of Piabarchus is restricted to the Paraguay-Paraná and São Francisco basins.
Among the remaining genera included in this tribe (node 11, Fig. 10), there is a very well supported clade embracing Cyanocharax, Diapoma and “Hyphessobrycon” guarani. This clade is also supported by the apomorphic number of i+6 pelvic-fin rays, a count also shared with the species of Planaltina not analyzed herein, and differing from the other genera in Stevardiinae that have i+7 rays [6, 45–47]. The genus Diapoma (with all inseminating species) has been hypothesized as closely related to other inseminating species of Stevardiinae that also bear a caudal organ [13, 20, 35]. All these studies, however, did not include a comprehensive sampling of other species without caudal organs to test this relationship. It is of note that the caudal organs of the Diapomini  differ from those of taxa in other tribes in being present and identical in both males and females, while it is sexually dimorphic (only males) in other inseminating tribes (except in Acrobrycon ), suggesting the non-homology of the caudal organ of the Diapomini relative to other tribes. Our results place Diapoma as a clade inserted in Cyanocharax, making this genus as defined  paraphyletic. Similarities between Cyanocharax and Diapoma, such as tooth arrangement, general body shape, color pattern and number of pelvic-fin rays have been previously discussed , in a comparison of the morphology of Diapoma speculiferum and Diapoma terofali with Cyanocharax alburnus (therein named Astyanax hasemani). Based on the strong molecular support for this clade (node 11, Fig. 10), species of Cyanocharax and Hyphessobrycon guarani should be reassigned to the genus Diapoma (Additional file 5) in order to define monophyletic genera and to be consistent with a phylogenetic classification. Diapoma as herein defined includes five highly supported internal clades (bootstrap 99 %, 95 %, 85 %, 93 % and 99 %). Further study of this group is warranted given the morphological diversity of the included species and the presence of four well-defined internal lineages.
Monotypic tribes (not analyzed herein)
Two monotypic tribes in Stevardiinae were not analyzed here: Landonini Weitzman and Menezes, 1998 and Phenacobryconini Weitzman and Menezes, 1998. Landonia latidens was sampled in our study, but excluded from our analyses because of few genes amplified that produced instability in the phylogeny; Phenobrycon henni was not available. As already pointed in the literature , the large amount of monotypic genera being proposed in Characidae to accommodate autapomorphies is somehow arbitrary and, in many instances, lacking phylogenetic inference. In the case of Landonini and Phenacobryconini, monotypic tribes were proposed to accommodate the large amount of autapomorphies presented by these monotypic genera; however these propositions do not reflect phylogenetic relationships. Since the relationships for these two monotypic genera are unknown in Stevardiinae, we placed them as incertae sedis (Additional file 5) to emphasize the necessity of further investigation of their relationships.
External fertilization is the most common reproductive strategy across the Ostariophysi, including Characiformes. Insemination as a reproductive strategy is found only in some representatives of the Characidae, including several members of the Stevardiinae (Table 2). Historically, the study of insemination was prompted by the presence of modified glandular organs in the caudal or anal fins of some species and documented histologically by the presence of sperm in the ovaries of stevardiine species that bear this caudal organ. Insemination, however, has also been documented for a few species lacking modified caudal or anal fins (e.g., Bryconamericus pectinatus, Creagrutus lepidus and C. melasma), but information on insemination, or the lack of it, is absent for most of the remaining members of this subfamily, being available only for 92 of the 311 species of Stevardiinae (Table 2), and only for 46 species analyzed here (30 %). Apparently due to the fact that external fertilization is ancestral and widespread in Characidae, characid species whose reproductive strategy is unknown have been treated informally in the literature as externally fertilized, but absence of evidence should not be used as evidence for absence. A topology test constraining all known inseminating taxa into one clade, however, rejects a single origin for insemination (Table 5). All members of Xenurobryconini, Glandulocaudini, Stevardiini and scattered species in Hemibryconini, Creagrutini and Diapomini are inseminating. This phylogenetic distribution suggests multiple origins and/or multiple losses of the inseminating reproductive strategy in Stevardiinae. That is in agreement with the general phylogenies of the Characidae [4, 7, 10, 12, 21]. Further analyses should produce more conclusive results once the reproductive strategy of most species of stevardiines is known.
In term of sperm morphology, M3 spermatozoid from the xenurobryconins is the only sperm morphology recovered as monophyletic, while M1 and M2 are found scattered along the tree. Since M1 appears close to the root in Markiana and in several other terminal nodes, it could constitute a diagnostic character for the Stevardiinae. For M2, the sporadic position for the few taxa that present this sperm type along the tree may indicate that this trait has multiple origins in Stevardiinae. However, the proposition of these sperm morphotypes was based on grouping multiple characters (e.g., centriole and midpiece position and slightly to moderately elongated shape) into a single sperm morphotype . A reductionist analysis that creates artificial sperm types could lump independent characteristics, with distinguished evolutionary histories, into a single sperm group (e.g. M1, M2), not necessarily recovering phylogenetic relationships among taxa that share them.
Unfortunately, it is important to highlight that the lack of information about sperm morphology and reproductive strategy in several taxa precludes the use of probabilistic approaches for ancestral state reconstruction to test the hypotheses presented here.
The molecular phylogeny presented in this study provides a significant advance in our knowledge of the relationships among of characid fishes in the subfamily Stevardiinae. On the basis of well-supported monophyletic groups, we defined seven tribes (Figs. 3–10, Additional file 5) and propose new circumscriptions for historically problematic (polyphyletic) genera such as Bryconamericus and Knodus, splitting part of the species of Bryconamericus into different and not closely related genera (e.g., Eretmobrycon and Hemibrycon). Some key taxa not included in our study or with poor resolution in our phylogeny remain with uncertain classification, and point to necessary future studies. Reproductive traits among stevardiine species that include inseminating strategy and variable sperm morphology are interpreted under the new phylogenetic hypotheses to show that some are evolutionary labile characters while others are phylogenetically informative requiring additional documentation and further analyses to understand their origins and phylogenetic distributions.
A total of 330 specimens were obtained from species assigned to the subfamily Stevardiinae. These represent 153 species or morphotypes (49 % of 311 valid species in Stevardiinae), and 32 (67 %) of the 48 recognized genera, plus one species (Hyphessobrycon guarani) not currently included in Stevardiinae [6, 49]. These samples originated in all major Neotropical river basins. In addition, 25 samples from 17 characiform species were used as outgroup. All specimens used for this molecular work have tissue samples preserved in ethanol associated with voucher specimens in museum fish collections and were identified to species (or genus) based on diagnostic morphological traits. In cases that fieldwork was required, specimens collection was authorized by the Ministério do Meio Ambiente - MMA, Instituto Chico Mendes de Conservação da Biodiversidade - ICMBio, licence # 12038/2, in accordance with protocols in their ethical and methodological aspects for the use of fish. Specimens were euthanized with Eugenol after capture, and all efforts were made to minimize suffering. In addition, for two species sequences were obtained from GenBank (Pseudocorynopoma heterandria and Piabarchus sp. ). A complete list of tissues with their associated voucher identification, collection locality (basin) and contributing institutions (abbreviations as in ) is presented in Additional file 2.
Genomic DNA was extracted from muscle and fin tissues preserved in 96 % ethanol. For each specimen, DNA was extracted from approximate 20 to 30 mg of the tissue using the DNeasy tissue extraction kit (Qiagen) or Promega Wizard® SV 96 Genomic DNA Purification System. We collected DNA sequences from seven molecular markers (four nuclear: MYH6, PTR, RAG1 and RAG2; and three mitochondrial genes: 16S, 12S and COI).
PCR reactions to amplify mtDNA fragments used the Promega GoTaq® qPCR Master Mix in 30 μl reactions with the following concentrations: 10 to 50 ng genomic DNA, 0.16 mM of each primer, 0.9 mM of each dNTP, 1x PCR Buffer, 4 mM MgCl2 and 0.026 U Taq DNA polymerase. For the amplification of the nuclear loci, a nested-PCR approach was required using the Takara® LA PCR Kit. PCR reactions consisted of 10 to 50 ng genomic DNA, 2.5 μl of dNTP mix (1 mM each), 3.0 μl 10x buffer, 0.5 μl of each primer (10 μM), 1.2 μl of bovine serum albumin (BSA), 0.2 μl of Takara Taq (5 U/μl), and dH2O to a final volume of 30 μl. A list of the primers used as well as optimized PCR conditions for the seven markers are presented in the Additional file 6. PCR products were submitted for purification and sequencing in both directions to the High Throughput Sequencing facility, University of Washington, Seattle – Washington, USA. The forward and reverse chromatograms were assembled and visualized using the program Codon Code Aligner v3.7.1. (Codon Code Corporation). IUPAC ambiguity codes were applied when heterozygotes or uncertainty of the nucleotide identity was detected. All sequences produced for this study have been deposited [GenBank: KF209375 - KF209697 (12S), KF209698 - KF210029 (16S), KF210030 - KF210276 (COI), KF210277 - KF210567 (MYH6), KF210568 - KF210779 (PTR), KF210780 - KF210995 (RAG1), KF210996 - KF211258 (RAG2)] – see Additional file 2 for details.
The 16S and 12S sequences were aligned in SATé v1.4  using 50 iterations under default settings. Exon markers and COI were aligned individually using Muscle v3.6  software under the default parameters. All exons and COI alignments were unproblematic because these are conserved markers that exhibit very little or no length variation among species, but were, nevertheless, visually inspected using Mesquite v2.7  to verify that all sequences follow the correct reading frame and contain no stop codons. Because nested PCR is highly prone to cross contamination, a quality control step that involved gene tree estimation via NJ was included in our workflow to detect potential cases of sequencing errors due to contamination. Given the degree of redundancy designed in our taxonomy sampling, errors due to either sequencing or misidentification can be detected when sequences from putative conspecific specimens are not placed together in the tree. Sequences that were found misplaced in the resulting gene trees were re-checked and re-sequenced or removed from the alignments in cases of poor quality. Once individual gene alignments were finalized, these were used for gene tree estimation (for species tree analysis, see below), and concatenated in a single matrix of all genes for phylogenetic analysis.
Phylogenetic reconstructions were performed using the concatenated dataset as well as species-tree approach that reconcile discordant gene genealogies. Concatenated analyses were conducted under Maximum Likelihood (ML) and Maximum Parsimony (MP) criteria. For ML analyses, the concatenated data matrix was partitioned into data blocks to account for heterogeneity among sites and to select the most appropriate partitioning scheme and models. PartitionFinder v1.1.1  was implemented with data blocks defined a priori according to commonly used structural and functional criteria, separating the sequence data into 16 different blocks as follows: the mitochondrial 12S and 16S genes (one block) and each codon position of each protein-coding gene (COI, MYH6, PTR, RAG1, RAG2) for the other 15 blocks (3 codon positions x 5 genes). PartitionFinder tests combinations of the 16 initial subsets into a smaller number of data blocks to select the optimal partitioning scheme based on AIC scores and determine the best-fit model for each partition. Sequences of Serrasalmus, the most external characiform in our dataset, were used to root the phylogenetic analyses .
The ML analyses were run in RAxML v7.2.8  and Garli v2.0 . The evolutionary model used for all data blocks in RAxML was GTRGAMMA, while the Garli settings were adjusted according to the best-fit models selected with PartitionFinder. RAxML searches were conducted in the CIPRES portal v3.1  using ten parallel runs and starting with a randomly generated tree. Branch support was assessed using the rapid bootstrap algorithm with 1000 replicates. Garli searches used automatic termination (enthreshfortopoterm command), with eight parallel runs, eight search replicates and 1000 bootstrap replicates.
For comparison purposes, the concatenated dataset was analyzed under equally weighted parsimony in TNT v1.1 . The TNT analysis used a driven-search strategy combining several tree-search algorithms (e.g., ratchet, drift, sectorial searches and tree fusion). To maximize tree-space exploration, the final searches implemented the tree-bisection–reconnection (TBR) algorithm with 1000 independent replicates. Assessment of branch support was based on bootstrap search strategies using TBR and 1000 replicates.
Species tree analyses were conducted in the program STAR, as implemented in the STRAW web server [59–61]. Input gene trees for STAR were estimated in RAxML, using the same settings as explained above, and re-rooted with Serrasalmus. The genes CO1, 12S, and 16S were concatenated and analyzed as a single mitochondrial locus; the four nuclear gene alignments were run separately, to obtain a total five gene trees for the STAR analysis. Preliminary analyses on three genes (MYH6, RAG1, and PTR) resulted in the non-monophyly of Stevardiinae. The ingroup monophyly was thus enforced and new gene trees were estimated in RAxML. Species that had sequences from multiple individuals were annotated using the Species Allele Table Creator in STRAW. Additional species tree analysis was performed using *BEAST 1.8.0 [62, 63] with one billion MCMC iterations, however this analysis did not reach convergence and for this reason are not reported here.
Finally, topology tests were conducted to assess whether the monophyly of traditionally recognized tribes and genera of Stevardiinae (that were not obtained in our resulting trees) can be rejected by the new data. We used RAxML to obtain maximum likelihood phylogenies consistent with the alternative hypotheses. The unconstrained ML tree, commonly leading to non-monophyly of the group of interest, was compared to the ML topology consistent with enforcing the monophyly of each of the tribes and genera challenged by our results. For each of these analyses, the best tree of 10 independent searches was selected. To evaluate the differences in likelihood scores between constrained and unconstrained tree topologies, the site likelihood scores were extracted using RAxML and various topological tests were performed in Consel, including the AU (approximately unbiased), SH (Shimodaira and Hasegawa), and KH (Kishino and Hasegawa) tests .
Availability of supporting data
The data sets supporting the results of this article are available in the Dryad Digital Repository: http://datadryad.org/review?doi=doi:10.5061/dryad.7nd42 .
Financial support for this study was provided by the DeepFin Student Exchange Program to DA and ATT (NSF grant DEB-1004765), by NSF grant DEB-1019308 (Euteleost Tree of Life) to GO and by CNPQ, Brazilian Government (process 300705/2010-7; 477318/2012-6) to LRM. We thank all collections and people who kindly assisted with the loan of tissues and/or specimens: M. Sabaj Pérez (ANSP), N. Lujan and J. Armbruster (AUM), M. Mirande (FML), C. A. Lucena (MCP), R. Covain and S. Fisch-Muller (MHNG), P. A. Buckup and M. Britto (MNRJ), H. Ortega (MUSM), H. López-Fernández (ROM), R.G. Reina (STRI) and J. Albert (ULL). Thanks to the ichthyologists that helped in confirming the identification of vouchers: T. P. Carvalho, F. Carvalho, J. Anyelo and C. A. Lucena and to J. Burns for providing the unpublished data on the presence or absence of the insemination in some stevardiines. Finally, we are thankful to R. P. Vari and W. Fink for valuable comments on earlier versions of this manuscript.
- Eschmeyer WN, Fong, JD. California Academy of Sciences. Catalog of fishes: Species by family/subfamily. http://researcharchive.calacademy.org/research/ichthyology/catalog/SpeciesByFamily.asp. Accessed 18 June 2014.
- Lima FCT, Malabarba LR, Buckup PA, Pezzi da Silva J, Vari RP, Harold A, Benine R, Oyakawa OT, Pavanelli CS, Menezes NA, Lucena CAS, Malabarba MC, Lucena ZMS, Reis RE, Langeani F, Cassati L, Bertaco VA, Moreira C, Lucinda PHF. Genera Incertae Sedis in Characidae. In: Reis RE, Kullander SO, Ferraris CJJ, editors. Check list of the freshwater fishes of South and Central America. Porto Alegre: EDIPUCRS; 2003. p. 107–71.Google Scholar
- Weiss FE, Malabarba LR, Malabarba MC. Phylogenetic relationships of Paleotetra, a new characiform fish (Ostariophysi) with two new species from the Eocene-Oligocene of south-eastern Brazil. J Syst Palaeontol. 2012;10:73–86.View ArticleGoogle Scholar
- Mirande JM. Phylogeny of the family Characidae (Teleostei: Characiformes): from characters to taxonomy. Neotrop Ichthyol. 2010;8:385–568.View ArticleGoogle Scholar
- Malabarba LR, Bertaco VA, Carvalho FR, Litz TO. Revalidation of the genus Ectrepopterus Fowler (Teleostei: Characiformes), with the redescription of its type species, E. uruguayensis. Zootaxa. 2012;3204:47–60.Google Scholar
- Malabarba LR, Weitzman SH. Description of a new genus with six new species from Southern Brazil, Uruguay and Argentina, with discussion of a putative characid clade (Teleostei: Characiformes: Characidae). Comum Mus Ciênc Tecnol PUCRS. 2003;16:67–151.Google Scholar
- Calcagnotto D, Schaefer SA, DeSalle R. Relationships among characiform fishes inferred from analysis of nuclear and mitochondrial gene sequences. Mol Phylogenet Evol. 2005;36:135–53.View ArticlePubMedGoogle Scholar
- Mirande JM. Weighted parsimony phylogeny of the family Characidae (Teleostei: Characiformes). Cladistics. 2009;25:574–613.View ArticleGoogle Scholar
- Azevedo MA. Reproductive characteristics of characid fish species (Teleostei: Characiformes) and their relationship with body size and phylogeny. Iheringia Sér Zool. 2010;100:469–82.View ArticleGoogle Scholar
- Javonillo R, Malabarba LR, Weitzman SH, Burns JR. Relationships among major lineages of characid fishes (Teleostei: Ostariophysi: Characiformes), based on molecular sequence data. Mol Phylogenet Evol. 2010;54:498–511.View ArticlePubMedGoogle Scholar
- Baicere-Silva CM, Ferreira KM, Malabarba LR, Benine RC, Quagio-Grassiotto I. Spermatic characteristics and sperm evolution on the subfamily Stevardiinae (Ostariophysi : Characiformes : Characidae ). Neotrop Ichthyol. Porto Alegre, Brazil: Universidade Federal do Rio Grande do Sul; 2011;9:377–92.Google Scholar
- Oliveira C, Avelino GS, Abe KT, Mariguela TC, Benine RC, Ortí G, Vari RP, Castro RMC. Phylogenetic relationships within the speciose family Characidae (Teleostei: Ostariophysi: Characiformes) based on multilocus analysis and extensive ingroup sampling. BMC Evol Biol. 2011;11:275.PubMed CentralView ArticlePubMedGoogle Scholar
- Weitzman SH, Menezes NA. Relationships of the tribes and genera of the Glandulocaudinae (Ostariophysy: Characiformes: Characidae) with a description of a new genus, Chrysobrycon. In: Malabarba LR, Reis RE, Vari RP, Lucena ZMS, Lucena CAS, editors. Phylogeny and classification of neotropical fishes. Porto Alegre: EDIPUCRS; 1998. p. 171–92.Google Scholar
- Géry J. Characoids of the world. Neptune City: TFH publications; 1977. p. 672.Google Scholar
- Malabarba LR, Vari RP. Caiapobrycon tucurui, a new genus and species of characid from the rio Tocantins basin, Brazil (Characiformes: Characidae). Ichthyol Explor Freshwaters. 2000;11:315–26.Google Scholar
- Vari RP, Harold AS. Phylogenetic study of the Neotropical fish genera Creagrutus Gunther and Piabina Reinhardt (Teleostei: Ostariophysi: Characiformes), with a revision of the Cis-Andean species. Smithsonian Contrib Zool. 2001;613:1–252.View ArticleGoogle Scholar
- Eigenmann CH. The Glandulocaudinae (a new subfamily of characid fishes with innate potentialities for sexual dimorphism). Indiana University: South American Fishes; 1914. p. 32–42.Google Scholar
- Weitzman SH, Fink SV. Xenurobryconin phylogeny and putative pheromone pumps in glandulocaudine fishes (Teleostei, Characidae). Smithsonian Contrib Zool. 1985;421:1–121.View ArticleGoogle Scholar
- Weitzman SH, Menezes NA, Evers HG, Burns JR. Putative relationships among inseminating and externally fertilizing characids, with a description of a new genus and species of brazilian inseminating fish bearing an anal-fin gland in males (Characiformes: Characidae). Neotrop Ichthyol. 2005;3:329–60.View ArticleGoogle Scholar
- Ferreira KM, Menezes NA, Quagio-Grassioto I. A new genus and two new species of Stevardiinae (Characiformes: Characidae) with a hypothesis on their relationships based on morphological and histological data. Neotrop Ichthyol. 2011;9:281–98.View ArticleGoogle Scholar
- Mirande JM, Jerep FC, Vanegas-Ríos JA. Phylogenetic relationships of the enigmatic Carlastyanax aurocaudatus (Eigenmann) with remarks on the phylogeny of the Stevardiinae (Teleostei: Characidae). Neotrop Ichthyol. 2013;11:747–66.View ArticleGoogle Scholar
- Burns JR, Weitzman SH. Insemination in ostariophysan fishes. In: Uribe MC, Grier HJ, editors. Viviparous fishes. Homestead: New Life Publications; 2005. p. 107–34.Google Scholar
- Javonillo R, Burns JR, Weitzman SH. Sperm modifications related to insemination, with examples from the Ostariophysi. In: Jamieson B, editor. Reproductive biology and phylogeny of fishes (Agnathans and Bony fishes). Enfield: Science Publishers; 2009. p. 723–63.Google Scholar
- Burns JR, Weitzman SH. Intromittent organ in the genus Monotocheirodon (Characiformes: Characidae). Copeia. 2006;3:529–34.View ArticleGoogle Scholar
- Menezes NA, Weitzman SH, Quagio-Grassiotto I. Two new species and a review of the inseminating freshwater fish genus Monotocheirodon (Characiformes: Characidae) from Peru and Bolivia. Papéis Avulsos de Zoologia. 2013;53:129–44.View ArticleGoogle Scholar
- Burns JR, Weitzman SH, Malabarba LR. Insemination in eight species of cheirodontine fishes (Teleostei: Characidae: Cheirodontinae). Copeia. 1997;2:433–8.View ArticleGoogle Scholar
- Quagio-Grassiotto I, Malabarba LR, Azevedo MA, Burns JR, Baicere-Silva CM, Quevedo R. Unique derived features in spermiogenesis and sperm morphology supporting a close relationship between the species of Hollandichthys and Rachoviscus (Characiformes: Characidae). Copeia. 2012;4:609–25.View ArticleGoogle Scholar
- Azevedo MA. Análise comparada de caracteres reprodutivos em três linhagens de Characidae (Teleostei: Ostariophysi) com inseminação. Unpublished Doctoral dissertation, Porto Alegre, Brazil: Universidade Federal do Rio Grande do Sul.; 2004.Google Scholar
- Eigenmann CH. The american characidae. Memoirs Museum Comparative Zool. 1927;43:311–428.Google Scholar
- Mariguela TC, Ortí G, Avelino GS, Abe KT, Oliveira C. Composition and interrelationships of a large Neotropical freshwater fish group, the subfamily Cheirodontinae (Characiformes: Characidae): a case study based on mitochondrial and nuclear DNA sequences. Mol Phylogenet Evol. 2013;68:23–34.View ArticlePubMedGoogle Scholar
- Baicere-Silva CM, Benine RC, Quagio-Grassiotto I. Markiana nigripinnis (Perugia, 1891) as a putative member of the subfamily Stevardiinae (Characiformes : Characidae): spermatic evidence. Neotrop Ichthyol. 2011;9:371–6.View ArticleGoogle Scholar
- Román-Valencia C. Tres nuevas especies de Bryconamericus (Ostariophysi: Characidae) de Colombia y diagnóstico del género. Rev Biol Trop. 2000;48:449–64.PubMedGoogle Scholar
- Myers GS, Böhlke JE. The Xenurobryconini, a group of minute South American characid fishes with teeth outside the mouth. Stanford Ichthyol Bull. 1956;7:6–12.Google Scholar
- Weitzman SH, Fink SV, Machado-Allison A, Royero R. A new genus and species of Glandulocaudinae (Teleostei: Characidae) from southern Venezuela. Ichthyol Explor Freshwaters. 1994;5:45–64.Google Scholar
- Menezes NA, Weitzman SH. Systematics of the neotropical fish subfamily glandulocaudinae (teleostei: characiformes: characidae). Neotrop Ichthyol. 2009;7:295–370.View ArticleGoogle Scholar
- Castro RMC, Ribeiro AC, Benine RC, Melo ALA. Lophiobrycon weitzmani, a new genus and species of glandulocaudine fish (Characiformes: Characidae) from the rio Grande drainage, upper rio Paraná system, southeastern Brazil. Neotrop Ichthyol. 2003;1:11–9.View ArticleGoogle Scholar
- Carvalho TP, Tang SJ, Fredieu JI, Quispe R, Corahua I, Ortega H, Alber JS. Fishes from the upper Yuruá river, Amazon basin. Peru Check List. 2009;5:673–91.Google Scholar
- Géry J. A review of certain Tetragonopterinae (Characoidei), with the description of two new genera. Ichthyol Aqua J. 1966;37:211–35.Google Scholar
- Bertaco VA, Malabarba LR. A review of the Cis-Andean species of Hemibrycon Günther (Teleostei: Characiformes: Characidae: Stevardiinae), with description of two new species. Neotrop Ichthyol. 2010;8:737–70.View ArticleGoogle Scholar
- Arcila D, Vari RP, Menezes NA. Revision of the Neotropical Genus Acrobrycon (Ostariophysi: Characiformes: Characidae) with description of two new species. Copeia. 2013;4:604–11.View ArticleGoogle Scholar
- Román-Valencia C, Arcila-Mesa DK. Two new species of Hemibrycon (Characiformes, Characidae) from the Magdalena River Colombia. Anim Biodivers Consev. 2009;32:77–87.Google Scholar
- Román-Valencia C, Arcila-Mesa DK. Five new species of Hemibrycon (Characiformes: Characidae) from the Río Magdalena basin. Colombia. Rev Biol Trop. 2010;58:339–56.PubMedGoogle Scholar
- Ferreira KM. Análise filogenética e revisão taxonômica do gênero Knodus Eigenmann, 1911 (Characiformes: Characidae). Universidade de São Paulo: Unpublished Doctoral dissertation; 2007.Google Scholar
- Malabarba MC, Malabarba LR. Hypobrycon maromba, a new genus and species of characiform fish from the upper rio Uruguai, Brazil (Ostariophysi: Characidae). Ichthyol Explor Freshwaters. 1994;5:19–24.Google Scholar
- Malabarba LR. Redescrição e discussão da posição taxonômica de Astyanax hasemani Eigenmann, 1914 (Teleostei: Characidae). Comum Mus Ciênc Tecnol PUCRS. 1983;29:177–99.Google Scholar
- Mahnert V, Géry J. Deux nouvelles espèces du genre Hyphessobrycon (Pisces, Ostariophysi, Characidae) du Paraguay: H. guarani n. sp. et H. procerus n. sp. Bonn Zool Beitr. 1987;38:307–14.Google Scholar
- Menezes NA, Weitzman SH, Burns JR. A systematic review of Planaltina (Teleostei: Characiformes: Characidae: Glandulocaudinae: Diapomini) with a description of two new species from the upper rio Paraná, Brazil. Proc Biol Soc Wash. 2003;116:557–600.Google Scholar
- Vari RP, Siebert DJ. A new, unusually sexually dimorphic species of Bryconamericus (Pisces: Ostariophysy: Characidae) from the Peruvian Amazon. Proc Biol Soc Wash. 1990;103:516–24.Google Scholar
- Carvalho FR, Langeani F. Hyphessobrycon uaiso: new characid fish from the rio Grande, upper rio Paraná basin, Minas Gerais State (Ostariophysi: Characidae), with a brief comment about some types of Hyphessobrycon. Neotrop Ichthyol. 2013;11:525–36.View ArticleGoogle Scholar
- Sabaj Pérez MH (editor). Standard symbolic codes for institutional resource collections in herpetology and ichthyology: an online reference. Version 5.0 (22 September 2014). American Society of Ichthyologists and Herpetologists, Washington, DC. 2014. http://www.asih.org.
- Liu K, Raghavan S, Nelesen S, Linder CR, Warnow T. Rapid and accurate large-scale coestimation of sequence alignments and phylogenetic trees. Science. 2009;324:1561–4.View ArticlePubMedGoogle Scholar
- Edgar RC. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 2004;32:1792–7.PubMed CentralView ArticlePubMedGoogle Scholar
- Maddison WP, Maddison DR. Mesquite: a modular system for evolutionary analysis. Version 2.7. 2009. http://mesquiteproject.org.
- Lanfear R, Calcott B, Ho SYW, Guindon S. PartitionFinder: combined selection of partitioning schemes and substitution models for phylogenetic analyses. Mol Biol Evol. 2012;29:1695–701.View ArticlePubMedGoogle Scholar
- Stamatakis A, Hoover P, Rougemont J. A rapid bootstrap algorithm for the RAxML Web servers. Syst Biol. 2008;57:758–71.View ArticlePubMedGoogle Scholar
- Zwickl DJ. GARLI: genetic algorithm for rapid likelihood inference. 2006. Ph.D. dissertation, The University of Texas at Austin. http://garli.googlecode.com.
- Miller MA, Pfeiffer W, Schwartz T. Creating the CIPRES Science Gateway for inference of large phylogenetic trees. New Orleans: Proceedings of the Gateway Computing Environments Workshop (GCE); 2010. p. 1–8.Google Scholar
- Goloboff PA, Farris JS, Nixon KC. TNT, a free program for phylogenetic analysis. Cladistics. 2008;24:774–86.View ArticleGoogle Scholar
- Liu L, Yu L, Pearl DK, Edwards SV. Estimating species phylogenies using coalescence times among sequences. Syst Biol. 2009;58:468–77.View ArticlePubMedGoogle Scholar
- Liu L, Yu L, Edwards SV. A maximum pseudo-likelihood approach for estimating species trees under the coalescent model. BMC Evol Biol. 2010;10:302.PubMed CentralView ArticlePubMedGoogle Scholar
- Shaw TI, Ruan Z, Glenn TC, Liu L. STRAW: species TRee analysis Web server. Nucleic Acids Res. 2013;41:238–41.View ArticleGoogle Scholar
- Heled J, Drummond AJ. Bayesian Inference of species trees from multilocus data. Mol Biol Evol. 2010;27:570–80.PubMed CentralView ArticlePubMedGoogle Scholar
- Drummond AJ, Suchard MA, Xie D, Rambaut A. Bayesian phylogenetics with BEAUti and the BEAST 1.7. Mol Biol Evol. 2012;29:1969–73.PubMed CentralView ArticlePubMedGoogle Scholar
- Shimodaira H, Hasegawa M. CONSEL: for assessing the confidence of phylogenetic tree selection. Bioinformatics. 2001;17:1246–7.View ArticlePubMedGoogle Scholar
- Thomaz AT, Arcila A, Ortí G, Malabarba LR. Molecular phylogeny of the subfamily Stevardiinae Gill, 1858 (Characiformes: Characidae): classification and the evolution of reproductive traits. Dryad Digital Repository. 2015. doi:10.5061/dryad.7nd42.Google Scholar
- Mattox GMT, Britz R, Toledo-Piza M, Marinho MMF. Cyanogaster noctivaga, a remarkable new genus and species of miniature fish from the Rio Negro, Amazon basin (Ostariophysi: Characidae). Ichthyol Explor Freshwaters. 2013;23:297–318.Google Scholar
- Menezes NA, Ferreira KM, Netto-Ferreira AL. A new genus and species of inseminating characid fish from the rio Xingu basin (Characiformes: Characidae). Zootaxa. 2009;2167:47–58.Google Scholar
- Zarske A. Der Kolibrisalmler - Trochilocharax ornatus gen. et spec. nov. - ein neuer Salmler aus Peru (Teleostei: Characiformes: Characidae). Vertebrate Zool. 2010;60:75–98.Google Scholar
- Burns JR, Weitzman SH, Grier HJ, Menezes NA. Internal fertilization, testis and sperm morphology in glandulocaudine fishes (Teleostei: Characidae: Glandulocaulinae). J Morphol. 1995;224:131–43.View ArticleGoogle Scholar
- Burns JR, Weitzman SH, Lange KR, Malabarba LR. Sperm ultrastructure in characid fishes (Teleostei, Ostariophysi). In: Malabarba LR, Reis RE, Vari RP, Lucena ZMS, Lucena CAS, editors. Phylogeny and classification of Neotropical fishes. Porto Alegre: EDIPUCRS; 1998. p. 235–44.Google Scholar
- Pecio A, Rafinski J. Spermiogenesis in Mimagoniates barberi (Teleostei: Ostariophysi: Characidae), an oviparous, internally fertilizing fish. Acta Zool. 1999;80:35–45.View ArticleGoogle Scholar
- Pecio A, Burns JR, Weitzman SH. Comparison of spermiogenesis in the externally fertilizing Hemigrammus erythrozonus and the inseminating Corynopoma riisei (Teleostei: Characiformes: Characidae). Neotrop Ichthyol. 2007;5:457–70.View ArticleGoogle Scholar
- Burns JR, Quagio-Grassiotto I, Jamieson BGM. Ultrastructure of spermatozoa: Ostariophysi. In: Jamieson BGM, editor. Reproductive biology and phylogeny of fish (Agnatha and Osteichthyes). Enfield: Science Publishers; 2009. p. 287–388.Google Scholar
- Pecio A, Burns JR, Weitzman SH. Sperm and spermatozeugma ultrastructure in the inseminating species Tyttocharax cochui, T. tambopatensis, and Scopaeocharax rhinodus (Pisces: Teleostei: Characidae: Glandulocaudinae: Xenurobryconini). J Morphol. 2005;263:216–26.View ArticlePubMedGoogle Scholar
- Burns JR, Pecio A, Weitzman SH. Sperm and spermatozeugma structure in Xenurobrycon (Teleostei: Characidae: Stevardiinae: Xenurobryconini). Copeia. 2008;3:656–60.View ArticleGoogle Scholar
- Burns JR, Weitzman SH, Malabarba LR, Meisner AD. Sperm modifications in inseminating ostariophysan fishes, with new documentation of inseminating species. In: Norberg B, Kjesbu OS, Taranger GL, Andersson E, Stefansson SO, editors. Proceedings of the 6th International Symposium on the Reproductive Physiology of Fish. Bergen: Institute of Marine Resources and University of Bergen; 2000. p. 255.Google Scholar
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.