Chrysobrycon, AND

MONOPHYLY OF CHRYSOBRYCON AND View in CoL PTEROBRYCON

This is the first phylogenetic study proposing the monophyly of these genera, with Chrysobrycon being the sister group of the remaining stevardiins. The monophyly of Chrysobrycon was supported by eight synapomorphies, most of which are associated with caudal-fin squamation of adult males. The majority of these synapomorphies present reversals and/or convergences with other stevardiines, especially with genera that have a hypertrophied caudal-fin squamation in the lower lobe of adult males (e.g. Acrobrycon , Gephyrocharax and Pterobrycon ). The synapomorphies of Chrysobrycon involving the pouch scale of adult males (state 1 of the characters 502 and 525) were not observed in any other examined stevardiine. Chrysobrycon mojicai Vanegas-Ríos & Urbano-Bonilla (2017) was recently described from the Amazon Basin in Colombia. In that study, the presence of an extensive contact between the frontals (rarely the parietals) along the midline was identified as a diagnostic characteristic of the genus. Based on the results found here, this characteristic, which was coded in two characters (26 and 40), supports the monophyly of Chrysobrycon . The phylogenetic placement of C. mojicai , which could not be analyzed here, will be tested in a later study.

Thomaz et al. (2015) found that an unidentified species of Gephyrocharax was more related to C. myersi (the single species of Chrysobrycon included) than to the Gephyrocharax clade. In the results, conversely, both Chrysobrycon and Gephyrocharax were resolved as monophyletic groups in the consensus topologies, independent of the weighting scheme used ( Fig. 2 View Figure 2 ). Furthermore, the support measures obtained for the Chrysobrycon clade were relatively high (>50) in the final consensus topology. The findings of the taxonomic revision of Gephyrocharax ( Vanegas-Ríos, 2016) suggest that the unidentified species of Gephyrocharax from the southwestern Amazon ( Thomaz et al., 2015) might correspond to G. major . Further examination of the specimens used by Thomaz et al. (2015) and molecular data for all Chrysobrycon species are needed to better understand the incongruences between both hypotheses. The two known species of Pterobrycon were resolved as a sister clade to Corynopoma and Gephyrocharax . This result differs from the traditional phylogenetic concept under which Pterobrycon and Corynopoma have been considered sister genera ( Weitzman & Menezes, 1998). The Pterobrycon clade was supported by four synapomorphies related to anal and pelvic fins and body squamation of adult males. Additionally, only two of these synapomorphies were optimised without homoplasy on the most parsimonious trees used to calculate the final consensus topology (characters 422, state 1: the middle pelvic-fin rays are longer than the remaining rays; character 494, state 1: the presence of one or two paddleshaped scales on the body in adult males). Even though the monophyly of Pterobrycon is not an unexpected result, it is indispensable for endorsing its current taxonomy ( Bussing, 1974).

COMMENTS ON THE MONOPHYLY AND INTERRELATIONSHIPS OF STEVARDIINI

Weitzman & Menezes (1998) carried out the first phylogenetic study that supported the monophyly of Stevardiini (=Corynopomini) consisting of the genera Corynopoma , Gephyrocharax and Pterobrycon . In subsequent morphology-based phylogenetic studies, including at least one stevardiin species, that definition of the tribe remained unchanged ( Castro et al., 2003; Weitzman et al., 2005; Ferreira et al., 2011). Mirande (2010) did not analyze any stevardiin species in his phylogenetic study of Characidae , but he tentatively assigned them to several nodes of his phylogenetic hypothesis (nodes 235–244 and 244) based on the placement of Stevardiini within the ‘clade A’ (sensu Malabarba & Weitzman, 2003) and the phylogenetic hypothesis of Glandulocaudinae (sensu Weitzman & Menezes, 1998). In my results, those nodes were not recovered with the same composition supposed by Mirande (2010) ( Figs 1 View Figure 1 , 2 View Figure 2 ).

Although Hysteronotus megalostomus Eigenmann View in CoL and Pseudocorynopoma heterandria Eigenmann View in CoL could not be coded in this work, the number of analyzed stevardiins is greater (20 species and 5 genera vs. 11 species and 4 genera) than that included in the most recent phylogenetic study of Stevardiinae ( Thomaz et al., 2015). The stevardiins are recovered as a monophyletic group in the final consensus topology [ Fig. 2 View Figure 2 : ( Chrysobrycon View in CoL ( Pseudocorynopoma View in CoL ( Pterobrycon View in CoL ( Gephyrocharax View in CoL , Corynopoma View in CoL ))))]. Such congruence between morphological and molecular data reinforces the monophyly of the tribe as currently defined. The monotypic genus Hysteronotus View in CoL is a putative member of Stevardiini ( Thomaz et al., 2015). Weitzman & Menezes (1998) and Ferreira et al. (2011) obtained a sister-group relationship between Hysteronotus View in CoL and Pseudocorynopoma View in CoL , a hypothesis also pointed out by Thomaz et al. (2015). In the results obtained here, Stevardiini was supported by ten synapomorphies, three of which can be observed in H. megalostomus ( Menezes, Weitzman & Teixeira, 2016) View in CoL : the middle dorsal-fin rays are longer than the anterior and posterior dorsal-fin rays, the anterior tip of the premaxilla is horizontally aligned with the upper half of the orbit, and the presence of well-developed grooves with neuromasts along the dorsal surface of the head. Although none of the synapomorphies defining Stevardiini were optimised without homoplasy on the final phylogenetic hypothesis, the majority of these had relatively high values in the retention indices (ranging from 0.6 to 1).

In the final consensus, Acrobrycon View in CoL was obtained as the sister group of Stevardiini , which disagrees with the traditional phylogenetic position of the genus as part of the Diapomini ( Weitzman & Menezes, 1998; Arcila et al., 2013). In a phylogenetic study using morphological, reproductive and spermatic characters, Ferreira et al. (2011) found Acrobrycon View in CoL sister to a clade consisting of Gephyrocharax View in CoL plus other xenurobryconin and hysterotonin genera (sensu Weitzman & Menezes, 1998). However, more recently, Acrobrycon View in CoL has been considered to be the sister group of Hemibrycon ( Thomaz et al., 2015) View in CoL .

Another stevardiine genus with a contentious position between the tribes allied to Stevardiini is Argopleura , which was obtained as the sister group of a clade including Scopaeocharax , Tyttocharax and Xenurobrycon . These genera have been grouped together in Xenurobryconini, a tribe related to the Stevardiini ( Weitzman & Fink, 1985; Weitzman & Menezes, 1998). In the DNA-based phylogenetic study by Thomaz et al. (2015), Argopleura was resolved as the sister group of Glandulocaudini in most of their phylogenetic results, but in their ML tree it was obtained as the sister group of Glandulocaudini and Stevardiini . Based on these results, Thomaz et al. (2015) placed Argopleura as incertae sedis in Stevardiinae . The phylogenetic position obtained for Argopleura in the final consensus topology agrees more with that found by Weitzman & Fink (1985) and Weitzman & Menezes (1998) than with that found by Thomaz et al. (2015). Despite this disagreement between the molecular and morphological data, which should be investigated further, Argopleura is tentatively considered the sister genus of Xenurobryconini based on the results of the present study.

COMMENTS ON THE INTERRELATIONSHIPS WITHIN STEVARDIINAE

Based on the type of cells constituting part of the glandular pocket (mucous vs. club), Weitzman et al. (2005) defined the stevardiines as a group consisting of the six tribes ( Stevardiini = Corynopomini, Diapomini, Hysteronotini, Landonini, Phenacobryconini and Xenurobryconini) that had been previously placed in Glandulocaudinae by Weitzman & Menezes (1998). Later, Mirande (2010) expanded the phylogenetic concept of the subfamily to include the species of ‘clade A’ of Malabarba & Weitzman (2003). Since then, the monophyly of Stevardiinae has been widely supported based on molecular data ( Javonillo et al., 2010; Oliveira et al., 2011; Thomaz et al., 2015). In the final tree topology ( Figs 1 View Figure 1 , 2 View Figure 2 ), the monophyly of Stevardiinae was resolved with 26 of the 44 genera recognised in this subfamily by Mirande (2010), Mirande et al. (2013) and Thomaz et al. (2015). In total, 73 stevardiine species were analyzed in the data matrix, whereas Mirande (2010) and Mirande et al. (2013) analyzed 27 and 41 stevardiines, respectively (excluding Creagrutus species added in their extended matrix). After comparing the results with those presented by Mirande (2010) and Mirande et al. (2013), most of the differences found among the final topologies are associated with the placement of the species of Bryconamericus , Diapoma and Knodus , which in all cases did not constitute monophyletic groups. Additionally, the final tree topology ( Fig. 1 View Figure 1 ) recovered the monophyly of a group consisting of Carlastyanax , Creagrutus and Piabina , which was proposed by Mirande et al. (2013).

Markiana View in CoL was obtained within the Astyanax Baird & Girard View in CoL clade instead of the stevardiine clade (Supporting Information, Appendix S6), which disagrees with recent molecular phylogenetic studies that placed the genus within Stevardiinae ( Oliveira et al., 2011; Thomaz et al., 2015). According to Baicere-Silva et al. (2011), the genus should be considered a putative member of Stevardiinae since the spermatozoa of its type species [ M. nigripinnis (Perugia) View in CoL ] share the characteristics of the non-inseminating members of the subfamily.

In respect to other stevardiines, the Eretmobrycon species were found within the Stevardiinae clade, which is consistent with the phylogenetic result proposed by Thomaz et al. (2015). This finding was not obtained in previous morphology-based phylogenetic studies ( Mirande, 2009; Mirande et al., 2011; Mirande et al., 2013). Other genera such as Phenacobrycon View in CoL and Landonia View in CoL were resolved as sister groups within the subfamily and, remarkably, they were not found to be closely related to the genera with which they have been traditionally placed in Glandulocaudinae (sensu Weitzman & Menezes, 1998), in the Stevardiinae (sensu Weitzman et al. 2005), or in the node 237 (similar to Glandulocaudinae in Weitzman & Menezes, 1998) by Mirande (2010).

Although the purpose of the present work may be considered as a reappraisal of the phylogenetic study of Stevardiinae by Mirande (2010) and Mirande et al. (2013), the primary object was the study of the phylogeny of Gephyrocharax and other stevardiins based on a large data matrix. The effect of adding Stevardiini (and other terminal taxa) to the data matrices of Mirande (2010) and Mirande et al. (2013) can be considered as a secondary result of the cladistic analysis presented herein, which represents an advance in the phylogenetic knowledge of the subfamily. It is evident that our understanding of the phylogenetic relationships of many stevardiines has improved in recent years ( Malabarba & Weitzman, 2003; Mirande, 2010; Mirande et al., 2013; Thomaz et al., 2015), but further research is still needed to achieve a more consensual view of the internal classification of this subfamily.