Combining target enrichment and Sanger sequencing data to clarify the systematics of the diverse Neotropical butterfly subtribe Euptychiina (Nymphalidae, Satyrinae)

The diverse, largely Neotropical subtribe Euptychiina is widely regarded as one of the most taxonomically challenging groups among all butterflies. Over the last two decades, morphological and molecular studies have revealed widespread paraphyly and polyphyly among genera, and a comprehensive, robust phylogenetic hypothesis is needed to build a firm generic classification to support ongoing taxonomic revisions at the species level. Here, we generated a dataset that includes sequences for up to nine nuclear genes and the mitochondrial COI ‘barcode’ for a total of 1280 specimens representing 449 described and undescribed species of Euptychiina and 39 out‐groups, resulting in the most complete phylogeny for the subtribe to date. In combination with a recently developed genomic backbone tree, this dataset resulted in a topology with strong support for most branches. We recognize eight major clades that each contain two or more genera, together containing all but seven Euptychiina genera. We provide a summary of the taxonomy, diversity and natural history of each clade, and discuss taxonomic changes implied by the phylogenetic results. We describe nine new genera to accommodate 38 described species: Lazulina Willmott, Nakahara & Espeland, gen.n., Saurona Huertas & Willmott, gen.n., Argentaria Huertas & Willmott, gen.n., Taguaiba Freitas, Zacca & Siewert, gen.n., Xenovena Marín & Nakahara, gen.n., Deltaya Willmott, Nakahara & Espeland, gen.n., Modica Zacca, Casagrande & Willmott, gen.n., Occulta Nakahara & Willmott, gen.n., and Trico Nakahara & Espeland, gen.n. We also synonymize Nubila Viloria, Andrade & Henao, 2019 (syn.n.) with Splendeuptychia Forster, 1964, Macrocissia Viloria, Le Crom & Andrade, 2019 (syn.n.) with Satyrotaygetis Forster, 1964, and Rudyphthimoides Viloria, 2022 (syn.n.) with Malaveria Viloria & Benmesbah, 2020. Overall, we revised the generic placement of 79 species (74 new generic combinations and five revised combinations), and as a result all but six described species of Euptychiina are accommodated within 70 named, monophyletic genera. For all newly described genera, we provide illustrations of representative species, drawings of wing venation and male and (where possible) female genitalia, and distribution maps, and summarize the natural history of the genus. For three new monotypic genera, Occulta gen.n., Trico gen.n. and Xenovena gen.n. we provide a taxonomic revision with a review of the taxonomy of each species and data from examined specimens. We provide a revised synonymic list for Euptychiina containing 460 valid described species, 53 subspecies and 255 synonyms, including several new synonyms and reinstated species.


INTRODUCTION
Over the last decade, high-throughput sequencing has revolutionized systematics and phylogenetic studies by helping resolve many difficult relationships, including those likely resulting from bursts of rapid radiation with short internal branches (e.g. Espeland et al., 2019a;Espeland, Breinholt, Willmott, Warren, Vila, Toussaint, Maunsell, et al., 2018;Hackett et al., 2008;Kawahara et al., 2019;Leebens-Mack et al., 2019). For studies of large radiations, however, high per-sample sequencing costs are an issue, and for many it is still too expensive to sequence thousands of samples using such approaches. Furthermore, it is sensible to make use of existing sequence data in groups where obtaining tissue samples for new genomic studies may be difficult. As a result, several studies have combined high-throughput methods with Sanger sequencing to generate more taxonomically comprehensive phylogenetic hypotheses (e.g. Azevedo et al., 2022;Chamberland et al., 2018;Leaché et al., 2014; and natural history. A growing number of phylogenetic studies are helping to build a better understanding of butterfly biogeography and diversification in the Neotropics (e.g. Chazot et al., 2014Chazot et al., , 2019Condamine et al., 2012;Ebel et al., 2015;Elias et al., 2008;Kozak et al., 2015;Matos-Maraví et al., 2013, 2021Nadeau et al., 2013;Peña et al., 2010). One of the most diverse radiations of Neotropical butterflies is the satyrine subtribe Euptychiina, a group estimated to contain more than 500 species, of which 460 valid species (see Appendix A) are recognized. Aside from a single species of Megisto in East Asia, euptychiines are exclusively distributed throughout the temperate and tropical Americas, in both lowland and montane habitats, and are common inhabitants of both grasslands as well as forests. Community diversity peaks in the western Amazon, where 100 species can coexist (Brown, 1997;Lamas et al., 1991), while the Atlantic Forest of south-eastern Brazil also has a diverse, endemic fauna. As a result of their diversity, abundance and the strong attraction of adults to rotting fruit, euptychiines are common components of ecological studies using bait traps (e.g. Lourenço et al., 2019;Melo et al., 2019;Santos et al., 2018;Uehara-Prado & Freitas, 2019).
These phylogenetic studies revealed a number of cases of generic paraphyly or polyphyly, implying the need for extensive revision of the generic classification. Substantial progress has been made by focusing more intensively on particular clades and species groups (Barbosa et al., 2022;Matos-Maraví et al., 2013). Furthermore, a recent phylogeny using target enrichment data helped resolve earlier issues with long-branch attraction (Peña et al., 2010(Peña et al., , 2011 and confidently define the limits of the subtribe (Espeland et al., 2019a). Nevertheless, until now, phylogenetic studies that have included the entire subtribe have had limited taxon sampling, and a much more comprehensive approach is needed to build a firm generic classification across the group.
We here combine target enrichment and Sanger sequencing data to infer the largest phylogeny of the Euptychiina to date. First, we expand the backbone phylogeny of Espeland et al. (2019a) to include additional Euptychiina genera and additional closely related out-group species, using the same target enrichment approach employed in that study (Espeland, Breinholt, Willmott, Warren, Vila, Toussaint, Maunsell, et al., 2018). This backbone was used as a constraint tree to add as many Euptychiina species as possible for which data were available for up to 10 genes. Our inferred phylogeny includes 1280 specimens, representing 69 of the 70 described genera (all except the monotypic Llorenteana) and 359 out of the 460 described species plus 88 undescribed species (see Supplementary material). We use this phylogeny to divide the subtribe into major clades, building on those identified by Espeland et al. (2019a), and for each clade we summarize the natural history and taxonomic diversity, and discuss future work needed. Although not the focus of this study, for each clade we also identify morphological characters that are relatively distinctive in comparison to other members of the subtribe, or among closely related clades. These characters have not been exhaustively compared across all euptychiine species, and in some cases (e.g. immature stage characters), such a comparison is not yet possible, but we highlight them as possibly informative characters for future study. In addition, we describe nine new genera revealed by this study as needed to preserve the monophyly of existing genera, and, where necessary, we discuss why combining smaller genera to form larger genera is a less suitable solution. Our phylogeny provides a foundation for the future description and revision of additional genera across the subtribe.

Taxon sampling and taxonomy
To generate a stable backbone phylogeny, we added sequence data for 25 taxa to those included in Espeland et al. (2019a), with the total backbone dataset including 131 samples. These extra taxa include additional key genera not included in those analyses, such as Zischkaia, additional specimens of genera with unstable placement, such as Amiga and Hermeuptychia, as well as more out-groups to confirm the monophyly of the subtribe.
In addition, we compiled a Sanger sequencing dataset that included 1149 samples in order to represent as many Euptychiina species as possible. A total of 1775 individual sequences were obtained from GenBank, and identifications were confirmed from images, voucher specimens or by clustering of sequences with those of specimens with voucher images that could be examined. A total of 1118 sequences of four genes, cytochrome c oxidase subunit 1 (COI, mainly the barcode region only), elongation factor 1-alpha (EF1a), Glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and ribosomal protein S5 (RPS5), were newly generated during the course of this study. Tissue samples for molecular study were obtained from specimens collected during long-term inventories of several Neotropical countries by multiple authors, particularly in Brazil, Peru and Ecuador, in addition to specimens deposited in museums and private collections, following international and local permitting requirements. Information for each sample, including GenBank accession numbers, can be found in Table S1.
The taxonomy used in this paper follows Lamas (2004), updated with subsequent publications (e.g. see ) and work (unpublished) by G. Lamas, except for North American taxa which follow Pelham (2022) and Asian taxa which follow an unpublished global butterfly names list in preparation by G. Lamas. A synonymic list for Euptychiina is presented in Appendix A, which includes authors for all names, and the latter information is thus not included in the main text unless necessary for clarity. We continue to include the monotypic genus Llorenteana

Morphological methods
Legs, labial palpi and abdomens were soaked in hot 10% KOH solution for 5-10 minutes, dissected and subsequently stored in glycerine.
Drawings of external morphology were done using a camera lucida attached to a Leica MZ 16 stereomicroscope, as were most drawings of genitalia. Terminology for wing venation follows the Comstock and Needham (1898) system and nomenclature of genitalic structures mostly follows Klots (1956), although we also follow Muschamp (1915) in using the term 'brachia'. We refer to the eyes of adults as being either naked or setose, depending on whether they lack or possess hair-like setae, respectively. In females of some taxa, we refer to the intersegmental membrane between the seventh and eighth abdominal segments as being wrinkled or pleated and expandable, in which the eighth segment may be retracted inside the seventh segment and only barely visible, or expanded to completely expose the lateral parts of the eighth segment.
We examined Euptychiina specimens in collections in South America, USA and Europe to study variation, locate specimens for morphological examination, and record locality data from specimen labels. When geographic coordinates were not provided on labels, localities were georeferenced using Google Earth, gazetteers, and other publications. More than 60,000 specimens were studied, and was done at Rapid Genomics, Gainesville, Florida. The final concatenated dataset included 365 loci, a total of 181,110 bps (base pairs) and 131 samples.

Sanger data
Samples were extracted using the Qiagen DNeasy Blood & Tissue kit. PCR conditions, cleanup and sequencing for the four included genes, COI, EF1a, GAPDH and RPS5 follow Nakahara, Hall, et al. (2015); Nakahara, Janzen, et al. (2015); Nakahara, Willmott, et al. (2018); . DNA extractions and PCR work was performed at the McGuire Center for Lepidoptera and Biodiversity, Gainesville, Florida, USA, and at the Departmento de Biologia Animal, Universidade de Campinas, Campinas, Brazil. DNA Sequencing was done at Eurofin Genomics or the Interdisciplinary Center for Biotechnology Research (ICBR) at the University of Florida. Raw Sanger sequence data were trimmed and corrected using various versions of Geneious (Biomatters Ltd., [Kearse et al., 2012]).

Data processing and analyses
In addition to the data from the four genes sequenced for this study mentioned above, we added available data from these genes, as well as six additional nuclear genes based on sequences from GenBank and data from ; Espeland et al. (2019b): carbamoyl-phosphate synthetase 2, aspartate transcarbamylase and dihydroorotase (CAD), dopa decarboxylase (DDC), isocitrate dehydrogenase (IDH), malate dehydrogenease (MDH), ribosomal protein S2 (RPS2) and wingless (WG); see Table S1 for further information. All 10 genes used are also present in the BUTTERFLY1.0 kit to assure data overlap. The total concatenated dataset (Dataset FULL) included 1280 samples, 10 genes and 10,296 bps. Since only a subset of the genes are available for most taxa, the final dataset contained 65.7% missing data. To investigate the impact of missing data we generated a second dataset (Dataset FOUR_GENES) only including the four genes for which we had the most data (COI, EF1a, GAPDH, RPS5, a total of 4395 bps). This reduced the degree of missing data to 35%.
Sequences were aligned using MAFFT v. 7.130b with the L-INS-i algorithm (Katoh & Standley, 2013), and visually inspected in Aliview v.1.26 (Larsson, 2014). A backbone tree based on the target enrichment data only was inferred using IQ-TREE 1.6.11 (Nguyen et al., 2015). Maximum likelihood analysis was carried out on the concatenated dataset including 364 nuclear loci + COI. The data were partitioned according to loci and model selection performed in Model-Finder (Kalyaanamoorthy et al., 2017) with rcluster set to 10. A total of 250 likelihood searches were run and the tree with the highest loglikelihood (LnL) was selected as the best tree. Ultrafast bootstrap (Hoang et al., 2018) with 1000 replications and the -bnni option was used to assess branch support. This tree was then collapsed to only include branches with at least 95% ultrafast bootstrap (UFB) support, and used as a constraint backbone tree in the remainder of the analyses. This approach was chosen to keep the backbone topology stable since it has been shown in previous studies (Peña et al., 2010(Peña et al., , 2011) that datasets for this group based only on few genes are struggling with long-branch attraction artefacts, which were no longer apparent in the phylogenomic approach by Espeland et al. (2019a).
Alignments of the 10 genes were concatenated using AMAS (Borowiec, 2016) and model selection was performed using ModelFinder. Partition files with the selected models can be found on Zenodo (See DOI below). Three hundred and fifty likelihood searches were performed in IQ-TREE 2.1.1 (Minh et al., 2020) with 1000 ultrafast bootstrap replicates and the -bnni option to reduce model violation.
As an additional measure of branch support, we applied the SH-like approximate likelihood ratio test (SH-aLRT support) (Guindon et al., 2010) with 1000 replications. The tree with the highest loglikelihood was selected as the best tree. To assess topological stability of the trees we used the tipdiff and plotTreeDiff functions in the R package treescape (Jombart et al., 2017) to vizualize the differences between the trees with the highest and second highest LnL for both datasets, and the best trees from both datasets. Figures were produced using the R packages ggtree 3.2.1 (Yu et al., 2017) and ggplot2 3.3.5 (Wickham, 2016). All analyses were performed on the ZFMK-Rocks cluster available at the Leibniz Institute for the Analysis of Biodiversity Change, Museum Koenig.

Raw data have been uploaded to the NCBI Sequence Read
Archive (Bioproject PRJNA483787, accession numbers SRR23171752-SRR23171754). Sequences of the individual genes have been uploaded to GenBank. Accession numbers can be found in Table S1. Alignments, partition files with inferred models and tree files are available on Zenodo (DOI: 10.5281/zenodo.7584717).

RESULTS AND DISCUSSION
The backbone hybrid enrichment dataset contained 131 taxa and 181,110 base pairs, and the combined dataset included 1280 terminals and 10,344 bps. The general topology of the expanded backbone phylogeny ( Figure S1) is identical to that found by Espeland et al. (2019a). As in that tree, the subtribe is well supported as monophyletic (UFB = 100). The additional out-group taxa furthermore confirm the sister-group relationships between Euptychiina and the Pronophilina, the other taxonomically diverse Neotropical satyrine subtribe, here shown to be monophyletic. The monophyly of the latter has previously been questioned in a study utilizing Sanger sequencing data (Matz & Brower, 2016). Espeland et al. (2019a) found the enigmatic Caribbean genus Calisto Hübner, 1823 to be the sister group to the otherwise Old World subtribe Ypthimina. Including an additional Ypthimina taxon (Strabena smithii Mabille, 1877) to this analysis did not change this hypothesis of relationships.
We recognize eight major, well-supported clades of Euptychiina containing two or more genera, corresponding to clades identified in Chloreuptychia. A summary tree showing the relationships between clades (FULL dataset) can be found in Figure 1. The full trees inferred from both datasets can be found in the Supplementary information (Figures S2 and S3). The overall topology is the same as that found in previous molecular phylogenetic studies, namely (Euptychia('Megisto clade'('Cyllopsis clade'(remaining Euptychiina)))) (Murray & Prowell, 2005;Peña et al., 2010), although Euptychia tended to move around in those studies due to long-branch attraction artefacts (Peña et al., 2010(Peña et al., , 2011. Comparisons of the best and second best trees for the two datasets ( Figure S4) and between the best trees from the two datasets ( Figure S5)  The relationships between the genera Hermeuptychia + Saurona gen.  Espeland et al. (2019a). Support for major clades is shown. Interpretation of support values can be found in the legend.

Goda d d rti t t a i i na
n., Pindis and Lazulina gen.n., and the remainder of this larger clade, are not well supported in any dataset.

Clade summaries
Euptychia Euptychia is strongly supported as a sister to the remainder of the subtribe (Figures 2, S2 and S3) as also found in Espeland et al. (2019a).

MGCL_LOAN_137_Pharneuptychia_pharnabazos MGCL_LOAN_137_Pharneuptychia_pharnabazos
YPH0416_Pharneuptychia_sp3_EPB_2020 YPH0416_Pharneuptychia_sp3_EPB_2020 NW127_18_Pharneuptychia_pharnabazos NW127_18_Pharneuptychia_pharnabazos PM10_01_Pharneuptychia_sp2_EPB_2020 PM10_01_Pharneuptychia_sp2_EPB_2020  (Nakahara, Janzen, et al., 2015), and various Cyllopsis species have been reported feeding on bamboo and other grass genera (Beccaloni et al., 2008), but otherwise there are few comprehensive, detailed descriptions or hostplant records for most species in the clade.

Megisto clade
The 'Megisto clade' currently comprises 10 genera, Megisto (with 2 species), Pharneuptychia (5 species, in addition to 'Pharneuptychia' Megisto, but is placed deeper within the clade as sister to Graphita in the second best FULL tree, and as sister to Moneuptychia + Pharneuptychia in the best 4GENES tree. This genus was not included in the backbone tree, and only COI sequences are available for the included samples, which likely explains the instability of this taxon. The recently described Yphthimoides kinyoni (Nakahara, Barbosa, Nakamura, et al., 2021) is found within the genus Yphthimoides, as expected, in the best FULL tree, but in the second best FULL tree and in the best two 4GENES trees it is placed as sister to Stegosatyrus within Cissia. This instability is likely also caused by its lack of very close relatives within the genus and only short COI sequences being available, but a placement within Yphthimoides is also justified morphologically (Nakahara, Barbosa, Nakamura, et al., 2021).
It is estimated that the diversity of this clade will increase with the description of new species for Moneuptychia and Pharneuptychia (Barbosa et al., 2022;Freitas et al., in preparation) in forthcoming papers.

Hermeuptychia clade
The 'Hermeuptychia clade' is well-supported (FULL dataset 98.8/99) and composed of two genera, Hermeuptychia and Saurona gen.n. 18 species of Hermeuptychia and two species of Saurona gen.n. are described, but the actual diversity of the clade is likely to be at least twice as high. Several species of Hermeuptychia have been described recently (Cong et al., 2021;Cong & Grishin, 2014;Nakahara, Tan, et al., 2016), and there have been multiple recent taxonomic rearrangements (Cong et al., 2021;Viloria, 2021;Zacca et al., 2021). DNA barcoding data (COI) further suggest the existence of a number of cryptic species (Seraphim et al., 2014;Tan et al., 2021). Members of the 'Hermeuptychia clade' are relatively small, with the forewing length often less than 20 mm. Apart from a few species, they have sexually monomorphic wing patterns with the females being slightly paler than males. Members of Hermeuptychia are exceptionally uniform and drab in coloration, while Saurona gen.n. contains two colourful described species with rather modified ventral hindwing markings.
Possible synapomorphies and distinctive characters for the clade include the following: (a) absence of cephalic horns in all instars of the larvae (Cong & Grishin, 2014;Cosmo et al., 2014;, remaining to be confirmed in Saurona; (b) presence of a short, dark dash on VFW between submedial and medial lines, along discocellular veins (not well-marked in some Saurona males).
Members of the 'Hermeuptychia clade' range from southern USA to Argentina, and from sea level to almost 3000 m, in a wide variety of habitats, from grasslands to dry forest, rain forest to cloud forest.
There is an apparent slight centre of diversity in the western Amazon and east Andean foothills, where half-a-dozen species may be found in sympatry. As is typical for most members of Euptychiina, with the exception of Euptychia, known hostplants for members of this clade are either bamboos or other grasses in the family Poaceae, as well as species of Cyperaceae and Marantaceae (Beccaloni et al., 2008;Cong & Grishin, 2014;Cosmo et al., 2014;Murray, 2001b). A number of species also likely utilize non-native grasses as larval hostplants, as evidenced by the abundance of some Hermeuptychia species in highly disturbed habitats, garden lawns and cattle pastures (see also Cong & Grishin, 2014).

Amphidecta clade
The 'Amphidecta clade' is well-supported (FULL dataset, 97.9/97, Figure 6) as containing three described genera: Amphidecta (4 species), Zischkaia (12 species), Argentaria gen.n. (14 described species), two species whose generic status is under study (Nakahara et al., in preparation) and which remain for the moment in the genus in which they were described, Euptychia ('Euptychia' ordinata and 'Euptychia' insignis, listed by Lamas [2004] as incertae sedis), and one species currently placed in Pharneuptychia (P. innocentia), closely related to Pharneuptychia innocentia_01 in Figure 6, based on research by Barbosa et al. (unpublished data). As mentioned above, the placement of Zischkaia in this clade was not supported in the backbone analysis ( Figure S1), but the increased taxon sampling within the FULL dataset seems to have helped in placing the genus (Figures 6 and S2). The same relationship is found in the 4GENES dataset ( Figure S3), but LEP_58282_Forsterinaria_dif ficilis support is slightly lower (96.5/87). The relationships among genera in the group, is however, still largely unresolved in all trees, with only the relationships between Amphidecta and the undescribed genus containing 'E.' ordinata and 'E.' insignis well-supported. Peña et al. (2010) similarly found these five generic-level clades to cluster together, except that Hermeuptychia was also placed in the same clade, as sister to Amphidecta + ordinata/insignis. In our tree (Figure 1 (Nakahara, Zacca, et al., 2019), and multiple Argentaria gen.n. species require description. As noted above, the generic classification of this clade also requires additional work, with the description of two new genera in forthcoming papers (Barbosa et al., in preparation; Nakahara et al., in preparation). Currently, 33 species are recognized, but with the inclusion of undescribed species the clade is estimated to contain approximately 46 species. Members of the 'Amphidecta clade' range from being relatively small to large, and, in general, females present a paler wing colour pattern than males.
Some species of Amphidecta, 'Euptychia' ordinata and 'E.' insignis have the uncus of the male genitalia inflated at its base in lateral view.
Zischkaia species in the 'saundersii clade' and Amphidecta pignerator have a small but recognizable dome-like structure located posteriorly on the tegumen. Amphidecta clio comb.n. also has reduced brachia and males of 'Euptychia' ordinata and 'E.' insignis have a developed eighth tergite with a weakly sclerotized region in the middle. Many species in the clade show rather notable modifications of the VHW ocelli, which may be elongated, split or partially fused with those in adjacent cells. The clade ranges from Mexico to south-eastern Brazil, with its highest diversity in the south-western Amazon (southern Peru and adjacent areas of Brazil). As is common for most members of Euptychiina, hostplants for members of this clade are either bamboo or grasses in the family Poaceae (Beccaloni et al., 2008;Freitas, 2004Freitas, , 2022Nakahara, Zacca, et al., 2019).

Archeuptychia clade
The 'Archeuptychia clade' is a relatively small clade recognized in Espeland et al. (2019a), and that work showed its placement as a weakly supported sister group to the much larger 'Splendeuptychia clade', which is also found in the extended backbone tree in this study ( Figure S1). In the FULL dataset it is, however, weakly supported as sister to a clade including Chloreuptychia and the Pareuptychia clade ( Figure S2), while in the 4GENES dataset ( Figure S3) it is sister to Chloreuptychia, and these two again are sister to the 'Splendeuptychia clade'. Previous molecular phylogenetic studies only sampled a few representatives and thus did not result in recognition of this clade (Marín et al., 2017;Murray & Prowell, 2005)   Olyreae) (Nakahara, Rodríguez-Melgarejo, et al., 2022;Tejeira et al., 2021), and additional data will soon be published (Corahua-Espinoza et al., in preparation).

Splendeuptychia clade
The 'Splendeuptychia clade' was proposed by Peña et al. (2010), and was partially recovered by Espeland et al. (2019a) with the addition of Magneuptychia and Cepheuptychia; the clade also contains those species included within the 'Cissia clade' of Murray and Prowell (2005). In our present study, this clade comprises the genera Splendeuptychia Scriptor (Figure 10), in the second best FULL tree it is sister to Capronnieria, and in both 4GENES trees it is sister to Scriptor + Colombeia, but no placement is well-supported.Emeryus similarly moves around, and in the best FULL tree it is sister to Malaveria + Colombeia (on a very short branch) and in the second best FULL tree it is sister to Modicagen.n. Both 4GENES trees again give different placements of this genus; in the best tree it is sister to Parypthimoides and in the second best tree it is instead sister to a clade consisting of Scriptor, Deltayagen.n. and Modicagen.n. (again on a very short branch). Three genes or more were available for at least some samples of both Deltayagen.
n. and Emeryus, but none of these genera were included in the backbone, so long-branch-attraction might be one of the reasons why these genera move around. More data is needed to possibly resolve these relationships, but the very short branches in the backbonemight make this task difficult. show that species to be a member of a strongly supported clade containing two other species, which we here also transfer to Colombeia (Colombeia nossiscomb.n. and Colombeia hotchkissicomb.n.). The genus Malaveria was recently described with Euptychia nebulosa as type species (Benmesbah et al., 2021), with three other described species moved into the genus and a further four species described. These last species descriptions highlighted hitherto unrecognized important morphological and molecular variation, but all of the described species are allopatric with respect to their closest relatives, or with only limited evidence for sympatry, and further research is needed to confirm their status. Barbosa et al. (2022) showed that five additional species also form a clade with Malaveria species and are best placed in that genus, and a number of other new species require description. We also provisionally transfer Euptychia argyrospila Butler, 1867 from Yphthimoides to Malaveria (comb.n.) since partial DNA barcode data support this placement, while the male genitalia illustrated by Forster (1964)  described Splendeuptychia, and since morphological data also support a close relationship among these species (Huertas, 2014), we believe the best taxonomic solution is to regard them as a single genus, so we therefore synonymize Nubilasyn.n. with

Splendeuptychia.
Considering the high diversity in the wing pattern, venation, and  (Benmesbah et al., 2018;. Known larvae of the species in this clade feed on various genera of Poaceae (e.g. Beccaloni et al., 2008;Brown, 1992  Etymology. The generic name is a feminine noun in the nominative singular, derived from the Latin adjective 'lazulinus', meaning something that is blue, in reference to the distinctive blue dorsal wing markings of this genus. Description (Figures 11 and 12). Some notable characters include: Distribution and natural history (Figure 16). Saurona gen.n. is confined to lowland rainforest in the south-western Amazon, where both described species can be found in the same localities. The immature stages and hostplants have not been described.
Discussion. Weymer (1911, p. 194, pl. 46  These ventral characters are evident in the putative syntype specimen examined at the SMTD, and distinguish the species from the only other described congener. Aurivillius (1929) described Euptychia aurigera var. triangula based on three males from Rio Purus, Hyutanahan (= Hiutanaã) (Brazil, Amazonas), and accurately described several characters that distinguish the species from S. aurigera comb.n. and which are evident in the syntype specimens examined at the RMS and figured by Warren et al. (2022). We selected Euptychia triangula as the type species for this genus since it was the species for which we had most complete DNA sequence data. Neither species was mentioned by   Systematic placement and diagnosis. Argentaria gen.n. is a member of the 'Amphidecta clade', in which its monophyly is strongly supported in all datasets (FULL dataset, SH-aLRT 100, UFB 100, Figure 6). However, its relationships to other members of the clade, which include Notable features include, in the male, the lack of cornuti in the aedeagus (similar to many other euptychiine genera), and in the female, the pleated, expandable intersegmental membrane between the seventh and eighth abdominal segments, antrum and lamella antevaginalis membranous, and small, round corpus bursae with slender, elongate, converging signa. The other two described genera within the 'Amphidecta clade', namely Amphidecta and Zischkaia, each have distinctive male and female genitalia differing in numerous respects from Argentaria gen.n., with these distinctive characters apparently representing generic autapomorphies (Marín et al., 2017;Nakahara, Zacca, et al., 2019;Nakahara pers. obs.).
Etymology. The generic name is a feminine Latin noun in the nominative singular, meaning a silver-mine, in reference to the distinctive silver scales arranged as spots on the ventral hindwing and forewing of species in this genus.
F I G U R E 1 6 Saurona gen.n. Species diversity mapped on a 2 degree grid. Colours ranging from dark green to red represent increasing diversity.
Description (Figures 17-19). Some notable characters include: Guadua and various species of weeping bamboos (unpublished data), and many species are known from only a handful of widely scattered localities, with a number of species still undescribed. The immature stages were recently described for A. quadrina comb.n. by See et al. (2018), which was recorded feeding on the climbing bamboo Rhipidocladum racemiflorum with several other species having been reared in south-eastern Brazil (Freitas, 2022) and Ecuador (Willmott & Hall, unpublished data).
Discussion. The type species for this genus, Euptychia itonis, was described by Hewitson (1862)  which is unique within the genus to A. itonis comb.n.
Butler (1867) figure 163, 164), which are rather different to specimens we have examined, most notably in showing an aedeagus with cornuti, which was not observed in any dissected Argentaria gen.n.
In general,  separated numerous taxa from the genus Euptychia into different genera, based on wing pattern and male genitalia, but he also frequently grouped species by overall appearance.
Lamas (2004) followed Forster's arrangement, and retained in Splendeuptychia the species placed here in Argentaria gen.n., but shortly afterwards Murray and Prowell (2005) showed that, based on DNA sequence data, A. itonis comb.n. and S. ashna were distantly related.
Subsequent molecular (Espeland et al., 2019a;Peña et al., 2010) and morphological (Huertas, 2014;Marín et al., 2017) phylogenetic studies have corroborated that discovery, including this study, which contains sequence data for all but three of the described species.
The relationships of Argentaria gen.n. to other members of the 'Amphidecta clade' are not strongly resolved, but the numerous morphological differences that separate the genus from other genera within the clade, and the fact that its constituent species have universally been recognized as closely related for more than 150 years, support its recognition as a distinct genus.
F I G U R E 2 0 Argentaria gen.n. Species diversity mapped on a 2 degree grid. Colours ranging from dark green to red represent increasing diversity. Etymology. The generic name is based on that of the Taguaîba, an evil spirit from the mythology of the Brazilian indigenous people Tupinambá (Métraux, 1950), whose original distribution includes the core distribution of the species of this new genus. Taguaiba should be treated as a neuter noun in the nominative singular.
Description Distribution and natural history (Figure 22). Siewert et al. (2013) summarized the distribution (also shown here in Figure 22), biology and taxonomic history of members of this genus. species was subsequently described (Siewert et al., 2013). Taguaiba ypthima comb.n. and Taguaiba rectifascia comb.n. also share morphological similarities in their wing pattern elements, venation and male and female genitalia (see illustrations in Siewert et al., 2013), which are shared with Taguaiba drogoni comb.n., Taguaiba fulginia comb.n.

Discussion
and Taguaiba servius comb.n, all placed in the 'Taygetis ypthima species group' by Siewert et al. (2013). In our study, the 'T. ypthima species group' and respective subgroups as proposed in Siewert et al. (2013) were recovered as monophyletic with the following relation- The absence of signa is also a rare character state, although it is known to occur in a few euptychiine species (Nakahara, Llorente-Bousquets, et al., 2015). Phenotypically, Xenovena gen.n. resembles evolutionary distantly related taxa such as species in the genus Hermeuptychia The generic name should be regarded as a feminine noun in the nominative singular.
Description (Figures 23 and 24). Some notable characters include: Discussion. Xenovena gen.n. is established for Magneuptychia murrayae based primarily on molecular data, but this hypothesis is also supported by its distinctive morphological features, as documented above. Our molecular data presented herein, coupled with previous phylogenetic analyses, clearly shows that the type species of Magneuptychia, Papilio libye, is a distantly related taxon recovered as a member of the 'Splendeuptychia clade' (Espeland et al., 2019a). Thus, the question is whether to describe another monotypic euptychiine genus for Magneuptychia murrayae, or include M. murrayae in a large genus that accommodates species in Pareuptychia, Euptychoides, Megeuptychia and other genera. As discussed in the relevant section of other monotypic genera described here, we believe that such a broader classification should not be adopted. For example, Pareuptychia is a morphologically compact genus with unique male genitalic features (Nakahara, Marin, & Neild, 2016), can closely resemble some euptychiine taxa, perhaps phenotypically being most similar to Satyrotaygetis iris comb.n., despite being distantly related. The male of S. iris is easily distinguished from female specimens of Trico gen.n. by its androconial scales at the distal side of DHW discal cell extending slightly along M 3 and Cu 1 , as well as presence of greyish long setiform scales in the discal cell, which are both absent in the female of Trico gen.n. Female specimens of Trico gen.n.
are distinguished from females of S. iris comb.n. by having a smoother VHW submarginal band in cell Cu 2 , which usually bends inwards in S.
iris comb.n. The females of these two taxa can be further distinguished by the membranous lamella antevaginalis of specimens of Trico gen.n. (Figure 28c; sclerotized in S. iris) and ductus seminalis exiting from the ductus bursae closer to the corpus bursae in specimens of Trico gen.n. (Figure 28d; origin of ductus bursae close to ostium bursae in S. iris comb.n.).
Etymology. The generic name is based on the Latin word 'trico', which is a masculine noun in the nominative singular, meaning 'mischief-maker' or 'trickster', in reference to the remarkable sexual dimorphism of this species.
Euptychia fulgora Butler (1869, p. 7, pl Taxonomy. Euptychia tricolor was described by Hewitson (1850) based on an unspecified number of specimens from the river Amazon, Brazil, most likely in the vicinity of Pará. Although the sex was not stated, the characters in the original description are clearly those of the male, and the male syntype located at the NHMUK matches the original description. Butler (1869) described Euptychia fulgora based on an unspecified number of specimens from Pebas, Peru, and the syntype in the NHMUK matches the original description. To ensure future nomenclatural stability we thus designate these two specimens as lectotypes for E. tricolor and E. fulgora (lectotype designations).
is known to date from the Amazon basin, the Guianas and Trinidad.  under Occultagen.n.Euptychia fulgora was introduced as a species close to Euptychia tricolor by Butler (1869). As stated in the original description, there are consistent differences in wing shape and pattern between Euptychia fulgora and Euptychia tricolor: the male forewing is more elongated in Euptychia tricolor, whereas the forewing is broader in male Euptychia fulgora; E. tricolor lacks iridescent scales in the DFW discal cell, whereas a lilac bluish streak is present in the male DFW discal cell and extends further into adjacent cells distal of discal cell in Euptychia fulgora. Weymer (1911) recognized Euptychia fulgora as a form of Euptychia tricolor, and this taxonomy has been followed by subsequent authors such as Lamas (2004). Although there appear to be no genitalic differences between Euptychia tricolor and Euptychia fulgora, the rather stable aforementioned male wing characters can be considered as evidence for treating these two taxa as distinct subspecies. We were unable to obtain molecular data for Euptychia tricolor, so the decision might be somewhat subjective. Nevertheless, it must be noted that male specimens of the nominate race from the state of Amazonas, Brazil, do exhibit a slight blue stripe in the DFW discal cell, as a trace in some specimens, suggesting that there is gene flow between these taxa and that they are thus best treated as conspecific.
Type species: Euptychia ocnus Butler, 1867, by present designation. The male genitalia also differ between Occultagen.n. (Figure 31) and the type species of Deltayagen.n. and close relatives (referred to as 'core' Deltayagen.n.) (Figure 34), at least, by the lack of a developed 'hump' on the dorsal margin of the valva, and instead only have a slightly serrated region at the dorsal margin distal of the costa. In lateral view, the costa appears as a narrow plate in Occultagen.n., whereas the costa appears as a somewhat trapezoidal plate in lateral view in 'core' Deltayagen.n. See description of Deltayagen.n. below for details about 'core' Deltayagen.n.
Etymology. The generic name is a Latin feminine adjective treated as a noun in the nominative singular, 'occulta', meaning a 'hidden' or 'secret' thing, in reference to the former concealment of this taxon within Magneuptychia. The generic name is also coined in alliteration with the species-group name.
Taxonomy. Butler (1867, p. 467)   Systematic placement and diagnosis. Deltaya gen.n. is a member of the 'Splendeuptychia clade' (Figure 10), and three of its four 'core' described species, Deltaya ocypete comb.n, Deltaya louisammour comb.n and Deltaya opima comb.n. (Figures 33 and 34), form a clade with low support (FULL dataset SH-aLRT = 67.9, UFB = 53). Deltaya pallema comb.n. is closely related to D. ocypete and is also considered a 'core' species (Benmesbah et al., 2018). This last character also occurs in Scriptor, some Paryphthimoides, and some other Euptychiina (e.g. Vanima labe, Vanima palladia), but is otherwise a relatively uncommon character. Characters that differ among the genera within the clade in which Deltaya gen.n. is placed are summarized in Table 1. D. ocypete comb.n., which may also be locally common in secondary habitats, including in the drier interandean valleys and extra-Amazonian regions from Bolivia across to south-eastern Brazil. Deltaya gen.n. species have also been recorded in fruit-baited traps (Benmesbah et al., 2018;Oliveira et al., 2021;Zacca, Casagrande, et al., 2017). The immature stages have not been described for any species to date, and while there are records of Cyperaceae and Poaceae as hostplants of D.
ocypete comb.n. (Singer & Ehrich, 1993), these require confirmation given prior confusion over the identification of this species.
Discussion. Benmesbah et al. (2018) discussed in detail the taxonomic history of the type species of this genus, Papilio ocypete, and designated a specimen with a DNA barcode as the neotype. This species and three other described species, D. opima comb.n., D. louisammour comb.n. and D. pallema comb.n., form a clade based on DNA sequences ('core' Deltaya gen.n.), and share several morphological synapomorphies. All these species were previously placed in Magneuptychia by Lamas (2004), but, for the same reasons as discussed under Modica gen.n., they cannot reasonably be accommodated in any described genus, except perhaps Scriptor. Nevertheless, monophyly with Scriptor is only weakly supported in our results, and we thus believe that description of a new genus is the best solution. As mentioned above under Systematic Placement and Diagnosis, we also provisionally include two species in Deltaya that are weakly supported based on molecular data as members of the same clade as 'core' Deltaya gen.n., namely D. probata comb.n. (Figure 34d) and D. andrei comb.n. (Figure 34c). Although both species share some wing pattern and genitalic similarities with 'core' Deltaya gen.n., none of these are convincing synapomorphies for Deltaya gen.n. Deltaya andrei comb.n.
has a very distinctive autapomorphy, namely a pocket of dense black scales adjacent to the valvae in the male genitalia, while the wing pattern of D. probata comb.n. is rather different from the remaining Deltaya species, with a VHW marginal line that is thickened throughout. Only COI sequences are available for D. probata comb.n., and COI + RPS5 for D. T A B L E 1 Comparison of characters for distinguishing Deltaya gen.n., Modica gen.n. and related genera.  sclerotized plate present on the ventral intersegmental membrane of the seventh and eighth abdominal segments, unlike Scriptor and some species of Deltaya gen.n. Characters that differ among the genera within the clade in which Modica gen.n. is placed are summarized in Table 1.
Etymology. The generic name is a feminine Latin adjective treated as a noun in the nominative singular, meaning something that is modest, ordinary, average, in reference to the 'typical' euptychiine morphology of this genus and its lack of obvious distinguishing characters.
Description (Figures 37-39 Distribution and natural history (Figure 40). Modica gen.n. contains five described species and several undescribed species (Zacca et al., unpublished data), which occur in rainforest from sea level to 1300 m, ranging from southern Mexico to western Ecuador and throughout the Amazon and Guianas to south-eastern Brazil. The genus reaches its peak diversity in the western Amazon, where both sexes may be common throughout the understory of both disturbed and undisturbed forest, with some species also occurring along forest edges and in overgrown, shady plantations. Males of some species perch from 1 to 3 m in the forest understory in the morning and late afternoon, sometimes on hilltops, and both sexes are attracted to rotting fruit (DeVries, 1987;Zacca et al., pers. obs.). Notes on the immature stages of M. myncea comb.n. and M. confusa comb.n. were provided by Singer et al. (1983), with hostplants (natural and in captivity) including Cyperaceae, Palmae, Poaceae and Marantaceae (see also Beccaloni et al., 2008;Singer & Ehrich, 1993).
Discussion. The type species for this genus, Euptychia confusa, was described by Staudinger (1884) Lamas, 2004;Brévignon, 2005) and M. fugitiva comb.n. and M. kamel comb.n. were placed in Magneuptychia (Benmesbah et al., 2018;Lamas, 2004). As discussed by Zacca, Casagrande, et al. (2018), most species placed in Cissia prior to that paper were presumably considered to be related because of their possession of a yellowish patch on the VFW, but that character is clearly homoplasious, and the Cissia of Lamas (2004)

CONCLUSIONS
We provide the most comprehensive phylogeny to date for the diverse, largely Neotropical butterfly subtribe Euptychiina. To achieve high taxonomic coverage, our dataset necessarily included a large amount of missing data, which might be responsible for the low support found in parts of the tree and some of the differences found between analyses, although Talavera et al. (2022) showed that it is largely beneficial to include samples with less data (e.g. barcodes only) in analyses clarifying butterfly relationships. Some samples where only COI was included did move around between datasets, but in the best tree of the FULL dataset they were all in a reasonable position within F I G U R E 3 9 Modica confusa comb.n. (a-d) male genitalia (dissection KW-21-64), lateral (a) with posterior view juxta, dorsal (b), aedeagus lateral (c) and aedeagus dorsal (d); (e-h) female genitalia (dissection KW-21-65), lateral view exterior tip abdomen (e), ventral view exterior tip abdomen (f), dorsal view interior abdomen (g), corpus bursae perpendicular to signa (h). Scale bars 1 mm.
their expected genus and clade. In the best tree of the 4GENES dataset some of these, such as Forsterinaria quantius and Ypthimoides kinyoni, remained in unexpected positions, causing morphologically rather unproblematic genera to be non-monophyletic. This could indicate that the six additional genes included in the FULL dataset possibly had positive influence on the COI-only samples, even though more missing data were introduced. In addition, through our methodological approach, we also attempted to minimize the impact of missing data by using the genomic backbone as a constraint for the Sanger sequencing datasets, rather than by combining the data for a single analysis with a much higher fraction of missing data. Using the genomic backbone as a constraint tree further reduced issues with longbranch attraction (Espeland et al., 2019a), which has been shown to be problematic for Sanger sequence analyses of the Euptychiina (Peña et al., 2010(Peña et al., , 2011. Collapsing the backbone constraint tree to only those relationships that were well-supported assured that we did not over-constrain our FULL 10 gene tree, but kept those relationships that were stable. Furthermore, we performed 350 likelihood searches and visually compared the best trees from two datasets to make sure that the topology was stabilizing, although some differences were apparent. The placements of some genera and species in the 'Splendeuptychia clade' (e.g. Deltaya gen.n, Emeryus and Caeruleuptychia pilata/scripta) were shown to be the most unstable. For the last of these cases, this is likely due to only rather short COI sequences being present for these species combined with their rather distant relationships with other species. For the first two cases, up to four genes were available for multiple specimens, but no species from these genera were included in the backbone, so long-branch attraction effects might be causing the topological instability with these taxa.
We recognize eight major clades and 73 provisional generic-level clades of Euptychiina, of which nine are described here, bringing the total number of named, valid Euptychiina genera to 70. The remaining three clades are under study with additional morphological and molecular data to determine the most appropriate generic classification. We also make several taxonomic changes that reflect our improved understanding of species relationships, even as work continues to corroborate these changes and revise species taxonomy. For example, our study follows a number of previous studies, as well as ongoing research (Nakahara et al., in preparation), in showing the genus Chloreuptychia as classified by Lamas (2004) not to be monophyletic, and we thus transferred four species to Pseudeuptychia to restore monophyly. As common species that appear frequently in ecological publications and citizen science applications such as iNaturalist.org, we feel that attempts to expediently produce as natural and stable a generic classification as possible are valuable to many users of butterfly taxonomy. Several species still lack sufficient molecular or morphological data to determine their relationships with confidence, notably Stephenympha arius, Malaveria argyrospilacomb.n., Deltaya probatacomb.n., Deltaya andreicomb.n., Paryphthimoides joyceaecomb.n. (known only from a single specimen) and Llorenteana pellonia. Indeed, Viloria and

Luis-Martínez (2019) suggested that the monotypic Mexican genus
Llorenteana might actually belong in the subtribe Ypthimina, which is an otherwise Old World group not known from the Americas. Nevertheless, the Caribbean satyrine genus Calisto does seem to be related to Ypthimina (Espeland et al., 2019a), so a similar relationship in the case of Llorenteana would not be unprecedented. The relationships of L. pellonia thus could clearly benefit from further study with the inclusion of molecular data. Overall, our study further underlines the work needed to develop a natural generic-level classification for Euptychiina, in contrast to the other highly diverse lineage of Neotropical satyrines, the Pronophilina, in which the generic classification has remained rather stable. The lack of strong morphological characters to resolve relationships among genera in diverse clades such as the 'Splendeuptychia clade' is mirrored by limited support in the molecular dataset, and our understanding of euptychiine relationships will continue to evolve.
The synonymic list included here contains 460 senior specific names of Euptychiina (Appendix A), but ongoing research suggests that an additional approximately 140 species remain to be described.

Recent unexpected discoveries of cryptic species in genera such as
Zischkaia and Pseudodebis (e.g. Nakahara, Matos-Maraví, Nakahara, Zacca, et al., 2019) suggest that additional species await discovery, even among groups that appear to be taxonomically straightforward. Furthermore, a number of cases of COI barcode 'splits' where corresponding morphological characters have yet to be identified may also represent unrecognized species (e.g. Tan et al., 2021). Conversely, some species provisionally recognized here based on a combination of morphological, ecological and molecular data might not be diagnosable based on molecular data alone, such as several species of Taygetis. Work to clarify the taxonomy of these species using multiple data sources, as well as broader genomic data, is ongoing (e.g. Dietz et al., 2023).
Notwithstanding the need for continuing taxonomic work in Euptychiina, the substantial progress that has been made over the last few decades in understanding the phylogeny and species diversity of F I G U R E 4 0 Modica gen.n. Species diversity mapped on a 2 degree grid. Colours ranging from dark green to red represent increasing diversity.
the group should facilitate research into euptychiine evolution, ecology and biogeography. One fascinating outcome of improved understanding of euptychiine phylogeny is evidence for wing pattern convergence in a number of cases, including, for example, the dorsal blue coloration, ventral hindwing white bands and ventral forewing yellow-orange markings of species formerly placed respectively in the genera Chloreuptychia, Euptychoides and Cissia, sensu Lamas (2004) (see Willmott et al., 2019). Whether this convergence is coincidental, a response to some environmental cue, or serves some purpose in sexual selection or predator avoidance, remains to be studied. Improving knowledge of euptychiine immature stage biology, especially the documentation of natural hostplants, should also help to better understand euptychiine population dynamics and community ecology. In particular, the dependence of many species on spatially isolated patches of bamboo offers opportunities for studying how hostplant distribution constrains adult movement and distribution at multiple scales. With their high species diversity, in both forest and non-forest habitats, and radiation largely within the Americas, the group offers numerous opportunities as a model for studying diversification in the Neotropics, the world's most biodiverse region.

ACKNOWLEDGMENTS
We are grateful to the Ecuadorian Ministerio del Ambiente, Agua, y Transici on Ecol ogica, the Instituto Nacional de Biodiversidad, and Nacional de Pesquisa e Conservação de Lepid opteros' SISBIOTABrasil/CNPq (563332/2010-7). Open Access funding enabled and organized by Projekt DEAL.

CONFLICT OF INTEREST STATEMENT
The authors declare no conflict of interest.

Raw sequence data have been uploaded to the NCBI Sequence Read
Archive under BIOPROJECT number PRJNA483787, accession numbers SRR23171752-SRR23171754. All other raw data files have been uploaded as part of other projects. Sequences for 10 genes were submitted to NCBI GenBank. Accession numbers can be found in Table S1. Alignments, tree files and model selection files are freely available on Zenodo, doi: 10.5281/zenodo.7584717.

SUPPORTING INFORMATION
Additional supporting information can be found online in the Supporting Information section at the end of this article.