A revision of the new genus Amiga Nakahara, Willmott & Espeland, gen. n., described for Papilioarnaca Fabricius, 1776 (Lepidoptera, Nymphalidae, Satyrinae)

Abstract We here propose a new, monotypic genus, Amiga Nakahara, Willmott & Espeland, gen. n., to harbor a common Neotropical butterfly, described as Papilioarnaca Fabricius, 1776, and hitherto placed in the genus Chloreuptychia Forster, 1964. Recent and ongoing molecular phylogenetic research has shown Chloreuptychia to be polyphyletic, with C.arnaca proving to be unrelated to remaining species and not readily placed in any other described genus. Amigaarnacagen. n. et comb. n. as treated here is a widely distributed and very common species ranging from southern Mexico to southern Brazil. A neotype is designated for the names Papilioarnaca and its junior synonym, Papilioebusa Cramer, 1780, resulting in the treatment of the latter name as a junior objective synonym of the former. A lectotype is designated for Euptychiasericeella Bates, 1865, which is treated as a subspecies, Amigaarnacasericeella (Bates, 1865), comb. n. et stat. n., based on molecular and morphological evidence. We also describe two new taxa, Amigaarnacaadela Nakahara & Espeland, ssp. n. and Amigaarnacaindianacristoi Nakahara & Marín, ssp. n., new subspecies from the western Andes and eastern Central America, and northern Venezuela, respectively.


Introduction
Butterflies are considered to have the best-studied taxonomy of any insect group, but the nymphalid subfamily Satyrinae includes some of the remaining groups still in most need of research, in part owing to its high diversity. Many new satyrine taxa have recently been discovered and described, including some that are among the most common elements of the butterfly fauna around us (e.g., Cong and Grishin 2014; Freitas et al. 2016).
Chloreuptychia Forster, 1964 is a genus in the Satyrinae subtribe Euptychiina, currently (Lamas 2004) containing some of the most brightly coloured euptychiine species, which possess a bluish-lilac reflection on their wings. Butler (1867) made one of the first attempts to classify those species later placed in Chloreuptychia by Forster (1964) in a systematic context. The "Division IV" proposed by Butler (1867) was distinguished from other closely related species by "Wings dorsally fuscous, mostly with a violet sheen; ventrally with ocellated spots, those of the hindwings often with their centers elongated and silver", and he further separated this "Division IV" into two subdivisions. The "subdivision 1" was characterized by "Wings ventrally with regularshaped ocelli, not elongated" and included Euptychia sericeella Bates, 1865 and Papilio ebusa Cramer, 1780, "subdivision 2" was characterized by "Hindwings ventrally with ocelli centered with irregular and elongated spots" and included P. chloris Cramer, 1780, P. herse Cramer, 1775, E. callichloris Butler, 1867, E. hewitsonii Butler, 1867, E. agatha Butler, 1867, and E. tolumnia Cramer, 1777. Both subdivisions also contained some species currently (Lamas 2004) not placed in Chloreuptychia. Subsequently, Butler (1877) grouped species related to species currently placed in Chloreuptychia under his "arnaea [sic] group", although without providing a diagnosis, including Euptychia arnaea [sic] Fabricius, 1776, E. sericeella, E. chloris, E. herse, E. callichloris, E. hewitsonii, E. agatha, and E. tolumnia, in addition to three unrelated species. Unlike in Butler (1867), Euptychia arnaea [sic] and E. sericeella were apparently considered as conspecific in Butler (1877). Weymer (1911) recognized E. arnaea [sic], E. sericella [sic], E. chloris, E. marica Weymer, 1911, E. herse, andE. callichloris in his "Arnaea [sic] group", andE. tolumnia, E. catharina Staudinger, [1886], and E. hewitsonii in his "Tolumnia group", in which the latter group was characterized by the presence of forewing androconial scales (scent-scales) in males. Forster (1964) of Chloreuptychia. These are available on Mendeley Data (https://doi.org/10.17632/ m7gc59vnp3.1). Sequences generated by Sanger sequencing were assembled using Geneious 10 (Biomatters), and aligned using MAFFT v. 7 (Katoh 2013). Phylogenies were inferred for each gene separately (Sanger data only) as well for the concatenated data. The genes were partitioned to codon position and partitions and models were selected using ModelFinder (Kalyaanamoorthy et al. 2017) in IQ-Tree 1. 6.7 (Nguyen et al. 2015). Thereafter 200 tree searches were performed in IQ-tree and the tree with the highest likelihood was selected. Support was calculated based on 2000 ultrafast bootstrap replicates with the -bnni option to reduce the risk of overestimating branch support (Hoang et al. 2017). The trees were rooted with Cyllopsis hedemanni R. Felder, 1869 based on prior information (Espeland et al. 2019). The hybrid enrichment data were cleaned, assembled and aligned according to Espeland et al. ( , 2019. A phylogeny was inferred using IQ-tree as above, but with 1000 ultrafast bootstrap replicates, also rooting with Cyllopsis hedemanni. All nodes with a support lower than 75 were collapsed and this tree was subsequently used as a constraint tree for the concatenated Sanger data, which was analyzed in IQ-Tree as above, leading to the final dataset containing 31 taxa. GenBank accession numbers for sequences used in this study can be found in Table 1. For species delimitation, a dataset consisting only of COI sequences with unique haplotypes was used. Sequences were aligned as above. The alignment was shortened to minimize the amount of missing data at both ends and the final dataset consisted of 30 sequences with a length of 615 bps. This included 16 Amiga arnaca comb. n. specimens and multiple outgroups. Sequences were not available for one of the proposed taxa (A. arnaca indianacristoi ssp. n.). Single threshold GMYC (Generalized Mixed Yule Coalescent;Pons et al. 2006;Fujisawa and Barraclough 2013), bPTP (bayesian Poisson Tree Processes; Zhang et al. 2013) and ABGD (Automated Barcode Gap Discovery; Puillandre et al. 2012) were used to assess the threshold between infra-and interspecific relationships in A. arnaca comb. n. GMYC assigns branching events to either a yule process (interspecific) or the coalescent (intraspecific). The reduced COI dataset was partitioned into codon positions, and model selection and phylogenetic inference was done as above. The resulting tree was rendered ultrametric using semiparametric penalized likelihood (Sanderson 2002) by applying the chronos function in the APE package (Paradis et al. 2004, Popescu et al. 2012) in R. The fit of four clock models (strict clock, discrete clock with 10 rate categories, correlated clock and relaxed clock) was tested using the Φ information criterion by Paradis (2013). For all four clock models three different values (0.01, 0.1, 1) of the smoothing parameter (lambda) were tested. A strict clock model was found to be the best fit and the smoothing parameter did not affect the species delimitation result, so only GMYC results for the strict clock tree with lambda = 1 are shown below. Support values of species clusters delimited by GMYC were calculated using information-theoretic multimodel inference (Fujisawa and Barraclough 2013). GMYC was performed using the SPLITS package v. 1.0-19 (from http://r-forge.r-project. org/projects/splits) in R. bPTP models branching events were based on the number of substitutions (Zhang et al. 2013), and conse-quently do not require an ultrametric input tree. We used the ML tree inferred above as input for bPTP analyses on the bPTP webserver (https://species.h-its.org/ptp/). The Markov Chain Monte Carlo (MCMC) was run for a total of 500,000 generations with thinning set to 1000 and burnin to 0.1. ABGD tries to find a barcode gap in the distribution of pair wise differences (Puillandre et al. 2012) and does not require an input tree. The reduced COI alignment was used as input for ABGD analyses on the ABGD webserver (http://wwwabi.snv.jussieu.fr/public/abgd/abgdweb.html) using the Kimura distance with transition/transvertion ratio set to 2. Pmin, Pmax, Steps and Nb bins were kept as default, and the relative gap width (X) was set to 1. The genetic distances were calculated based on the Tamura-Nei model using Geneious version 11.1.5 (Biomatters Ltd.) based on COI data ( Table 2).

Molecular phylogeny and species delimitation
The ML tree based on four genes and with the 368 gene hybrid enrichment tree used as a constraint tree is shown in Figure 1A. Although deeper relationships are mostly not well supported it is clear that C. arnaca is not closely related to the remaining Chloreuptychia species, as also found by Espeland et al. (2019) based on 368 loci, as well as phylogeny inferred based on hybrid enrichment data generated for this study (see Suppl. material 1).
All three species delimitation methods delimited the same three "species" within Amiga gen. n., with allopatric distributions: eastern Central America to western Ecuador (A. arnaca adela ssp. n.), east of the Andes (A. arnaca arnaca comb. n.), and Mexico to Honduras (A. arnaca sericeella comb. n. et stat. n.), although support values are not very high (Fig. 1B). We here decide to consider these taxa as subspecies since A. arnaca comb. n. is a morphologically very distinct species itself, and although morphological differences exist between the subspecies they are not very prominent in comparison with other groups of sympatric, closely related euptychiine species. This taxonomic arrangement is reinforced by three subspecies being recovered only in the gene tree inferred using COI and one nuclear gene (GAPDH), but not in any of the trees inferred using EF1a, RPS5 or in the combined tree (Fig. 1A, B; Suppl. material 1). Because nuclear genes evolve slowly and lineage sorting is slower compared to mitochondrial genes, we consider these data to provide evidence for the subspecific status of the taxa discussed below. Braby et al. (2012) defined subspecies in butterflies as partially isolated, allopatric, lineages within a species that are phenotypically distinguishable with at least one fixed diagnosable character state correlated with genetic structure, which matches exactly what we find here. In addition to the three subspecies delimited above we consider another allopatric lineage (A. arnaca indianacristoi ssp. n.) as a subspecies based on morphological differences only, as we could not obtain sequence data from this particular taxon. Table 2. Genetic distances calculated based on the Tamura-Nei model.  Wing venation (Fig. 3): Basal half of forewing subcostal vein swollen; base of cubitus swollen; forewing recurrent vein present as small projection slightly above origin of M 2 ; hindwing humeral vein developed; origin of M 2 slightly towards M 1 than M 3 .
Wing shape: Forewing triangular, apex rounded, costal margin slightly convex, outer margin almost straight below M 2 , inner margin almost straight, but curving inwards towards thorax near base; hindwing somewhat rectangular, slightly elongate, costal margin almost straight, angled inwards at base, outer margin slightly undulating, inner margin slightly curved near tornus, anal lobe convex, slightly rounded.
Dorsal forewing: Ground color light brownish, slightly translucent, thus subtly revealing ventral bands and ocelli in cell M 1 .
Dorsal hindwing: Ground color similar to forewing, iridescent bluish lilac reflection covering most of dorsal hindwing, area near costa and area distal to marginal band revealing ground color; slightly translucent, thus subtly revealing ventral bands and ocelli.
Ventral forewing: Ground color pale grayish brown; pale reddish-brown discal band extending from radial vein, crossing discal cell, passing origin of Cu 2 , terminating at 2A; concolorous scales present along discocellular vein; pale reddish-brown postdiscal band extending from radial vein towards inner margin until reaching vein 2A, slightly wider than discal band; broad, faint, iridescent bluish-lilac reflection extending just distal of postdiscal band towards outer margin, prominent between postdiscal band and umbra; sinuate, narrow submarginal band, almost concolorous to basal two bands, extending from apex towards tornus, jagged above Cu 1 , almost straight below this vein; concolorous marginal band, not jagged, appearing narrower than submarginal band, traversing along marginal area from apex to tornus; fringe brownish; ocellus in cell M 1 , spilling out from veins M 1 and M 2 , black with two whitish pupils in center, ringed in yellowish orange, ocelli present in cell M 2 and M 3 , appearing as two slate-gray patches surrounded by yellowish-orange ring; umbra appearing as broad rather faint band concolorous to four ventral bands, visible around ocelli, extending to cell Cu 2 .
Ventral hindwing: Ground color similar to forewing; general wing pattern similar to forewing except as follows: iridescent bluish-lilac reflection extending from base of wing, towards outer and inner margin, especially area where iridescent bluish scales are present on dorsal surface; discal band passing cubital vein area basal to origin of Cu 2 ; postdiscal band passing origin of Cu 1 , bent inwards in cell Cu 2 ; submarginal band broadening towards tornus after passing Cu 1 ; five submarginal ocelli, those in cells M 1 and Cu 1 similar to that in VFW cell M 1 but with single pupil, those in cells M 2 and M 3 similar to those in VFW cells Rs, M 2 and M 3 but slate grayish patch appearing as single patch in middle.
Genitalia ( Fig. 4c-g): Tegumen appears semi-circular in lateral view, anteriorly and dorsally convex, ventral margin rather straight; uncus longer than tegumen in lateral view, apparently without setae, middle section somewhat broadening in dorsal view, tapering posteriorly and terminating in single point; brachium tapering towards apex, apical point positioned above uncus in lateral view, parallel to uncus with apical edge curving inwards in dorsal view; combination of ventral arms from tegumen and dorsal arms from saccus slightly curved distally; appendices angulares present; saccus narrow, concavity at base of ventral margin, anteriorly rounded, similar or shorter than uncus in length; juxta present as plate with deep concavity at dorsal margin in posterior view; valva distally setose, valva appearing roughly parallelogram in lateral view, ventral margin convex, dorsal margin distal of costa curved, costa curved inwards, apical process somewhat curving upwards; phallus roughly straight, similar in length to valva plus saccus, phallobase about one-third of phallus, ductus ejaculatorius visible as illustrated, posterior portion of aedeagus somewhat curved upwards, manica covering more than half of aedeagus, cornuti absent. FEMALE: forewing length: 19-21 mm (n = 6) Similar to male except as follows: Five tarsomeres present in foretarsus, with spines along some tarsomeres; forewing appearing somewhat rounded and broad; dorsal hindwing submarginal band somewhat more prominent; bluish lilac reflection appearing more purplish, extending to origin of M 1 or further anteriorly; feeble pearly reflection present on dorsal forewing (but see below for further information). Female genitalia and abdomen (Figs 4h-k): Inter-segmental membrane between 7 th and 8 th abdominal segments pleated and expandable, with weakly sclerotized region present; lamella antevaginalis membranous; lateral side of 8 th abdominal segment sclerotized (referred to as "lamella postvaginalis" by , this sclerotized plate fused to lamella antevaginalis at anterior margin; ductus bursae membranous, somewhat inflated around origin of ductus seminalis, located at approximately one-fifth distance from ostium bursae to corpus bursae, ductus bursae apparently weakly sclerotized at region posterior to origin of ductus seminalis; corpus bursae roughly oval in dorsal view, with two relatively narrow signa, together with ductus bursae extending to juncture of 4 th and 5 th sternite. Variation. This species exhibits geographic variation in wing pattern, some of which is recognized here with subspecific names. A broad, faint, iridescent bluish-lilac reflection between the ventral forewing postdiscal band and the umbra is present in specimens from the eastern Andes, whereas it is absent in many specimens from west of the Andes. The absence of this bluish-lilac reflection on the ventral forewing seems rather stable in specimens from western Colombia and western Ecuador. However, this character appears in a few specimens from Panama, and is present in some specimens from Costa Rica and Nicaragua, although the degree of reflection is variable. The bluish-lilac reflection seen mainly on both the dorsal forewing and hindwing is variable in color, varying from light blue to purple. Especially in female specimens, the extent of bluish-lilac reflection on the dorsal forewing is variable, being absent in some specimens, whereas covering most of the discal cell and cells Cu 1 and Cu 2 in others (see below for further information). There exists a feeble pearly reflection on the ventral forewing from the base to the postdiscal band in specimens from southeastern Brazil (see below for further information), although this is absent in a few specimens. The presence or absence of a ventral forewing ocellus in cell M 3 is variable, appearing as a trace in some specimens. The size of the ventral hindwing ocellus in cell M 1 is variable in comparison with the ventral hindwing ocellus in cell Cu 1 , ranging from similar in size to almost twice as large. The genitalia appear not to be informative in separating specimens from east and west of the Andes, although specimens from Central America, including A. arnaca sericeella comb. n. et stat. n., seem to have a rather curved dorsal margin of the uncus in lateral view (Fig. 4e).
Etymology. The new generic name is derived from the feminine Spanish noun "amiga", meaning "a (female) friend", alluding to the fact that this is a common, familiar butterfly. The generic name is regarded as feminine.
Biology. Janzen and Hallwachs (2018) report four grass species (Poaceae), Ichnanthus nemorosus, Ichnanthus pallens, Lasiacis ruscifolia, and Paspalum decumbens, as hostplants for Amiga gen. n. in Costa Rica. In addition, DeVries (1987) reported three grass genera, namely Eleusine, Ichnanthus, and Oplismenus, as hostplants in Costa Rica. Singer and Ehrlich (1993) reported Ichnanthus pallens as a hostplant in Trinidad and Tobago. The egg, mature larva, and pupa of Amiga gen. n. were described and the latter two stages illustrated in DeVries (1987). Various images of the penultimate instar, ultimate instar, prepupa and pupa are figured by Janzen and Hallwachs (2018), based on material reared in Costa Rica. In Costa Rica, adult females of Amiga gen. n. were seen ovipositing late in the afternoon, and some eggs were observed to be parasitized by trichogrammatid wasps (DeVries 1987). The species occurs from sea level to at least 1850 m, and it is common, indeed ubiquitous, in undisturbed to heavily disturbed rain and cloud forest. Both sexes fly low (0.5 m) along shady trails throughout the middle of the day (09:00-15:00), and males are often observed perching singly on tops of leaves, maintaining apparent territories, and patrolling for ca 10 m along a trail.
Distribution (Fig. 6). This genus ranges from southern Mexico throughout virtually all of tropical Central and South America, where its southernmost distribution appears to be southern Brazil.

Taxonomy
Amiga gen. n. is regarded as monotypic, with total of four subspecies recognized, of which two are named and described herein.
Amiga Nakahara, Willmott & Espeland, gen. n. (-denotes a subspecies, --denotes a synonym) arnaca (Fabricius, 1776)  Identification and taxonomy. Papilio arnaca Fabricius, 1776 was described based on an unspecified number of specimens from Suriname, in Johann Dominicus Schulze's collection (Fabricius 1776). Fabricius' description was not accompanied by any illustration of this species, and he did not specify either the sex nor the number of specimens he examined. However, his Latin description is somewhat precise and the identity of the species may be guessed from the description, given the mention of the following wing pattern characters: "forewing, towards the apex there are three ocelli: the distalmost ("exteriori") bi-pupilled"; "hindwings bluish; under surface with five ocelli"; "Hindwings bluish above, iridescent; below bluish with two oblique dark stripes. Submargin with five ocelli, the first and fourth the largest and black, the remainder dark". The mention of multiple ocelli on the (ventral) forewing excludes the possibility of this specimen being other "Chloreuptychia" species, which also possess bluish iridescent coloration, but have only a single ocellus on the ventral forewing. Fabricius (1793) considered P. lea Cramer, 1777 and P. arnaca as probably being conspecific, and these two names were associated by some subsequent authors (e.g., God-  Forster, 1964, also match the aforementioned wing pattern characters provided in Fabricius' original description. These two species do occur in Suriname, and based on the description provided for P. arnaca, it is difficult to exclude the possibility of Fabricius having examined one of these two species. Regardless of this fact, the name arnaca has been applied in numerous publications and collections to the species as it is identified here (e.g., Whittaker 1983). Considering this situation, stabilizing the nomenclature as currently perceived by many others is crucial regarding the specific epithet arnaca and a neotype is therefore designated for this name below. A worn specimen (whose sex cannot be confidently determined) in William Hunter's entomological collection is at the UMG, and it was photographed by GL as a potential type specimen of P. arnaca (see . GL assumed that William Hunter may have received a "duplicate" from Schulze through Fabricius, who did in-deed supply Hunter with duplicates, resulting in many Fabrician type specimens being found in Hunter's insect collection (Hancock 2015;Tuxen 1967). However, we have found no evidence to support that this specimen was originally in Johann Dominicus Schulze's collection, and the Surinamese provenance of the specimen is questionable given its rather narrow ventral bands, which are typical of A. arnaca indianacristoi ssp. n. rather than of specimens from Suriname. The specimen in Glasgow is missing its head and abdomen, in addition to having worn and faded wings, thus somewhat obscuring its true identity. Given this situation, combined with the fact that no authentic Schulze specimens appear to be in existence (e.g., Benmesbah et al. 2018), in addition to the explanation above, we therefore designate a male specimen from Suriname (type locality) as a neotype for P. arnaca following Article 75.3 of the ICZN (1999) (neotype designation): //Suriname Brokopondo Brownsberg, rainforest km 6-12, 30.1.1982, Olle Pellmyr // USNMENT 00913953// (USNM). (Fig. 2a).
After introducing this species to science, the specific epithet was misspelled as "Arnaea" by Fabricius himself in 1781 and 1787, and as "Aranea" in 1793. Subsequently, the specific epithet arnaca has been erroneously spelled in various ways in a disturbingly high number of publications (e.g., Butler and Druce 1874; Butler 1877; Godman and Salvin 1881; Sharpe 1890; Kaye 1904;Weymer 1911;D'Almeida 1922;Gaede 1931;Forster 1964;Brown 1975;Whittaker 1983;DeVries 1987; see also above), including some influential works on the classification of this group. This confusion surrounding its species-group name adds a special urgency for a neotype designation for this common butterfly.
Papilio ebusa Cramer, 1780 was described in Pieter Cramer's De uitlandsche Kapellen voorkomende in de drie Waereld-Deelen Asia, Africa en America. The original description describes the bluish-lilac reflection on both wing surfaces, although no further description of any wing element was provided in Cramer's Dutch and French description. Instead, Cramer compared P. ebusa to P. junia Cramer, 1780, an immediately preceding species described and named in Cramer (1780), but regarded as a junior subjective synonym of P. lea Cramer, 1777 by Lamas (2004). Evidently, P. ebusa and P. junia are not conspecific judging from the illustrations in Cramer (1780: 9; pl. CCXCII: figs D-G), and the illustrations of P. ebusa combined with Cramer's description enable this taxon to be confidently identified. Papilio chloris Cramer, 1780 (now known as Chloreuptychia chlorimene (Hübner, [1819])), is perhaps the only taxon known from Suriname which might have resulted in a similar illustration; however, the illustration of P. chloris provided by Cramer (1780: CCXCIII: figs A, B) excludes this possibility. Based on the Dutch and French description provided for P. ebusa, Cramer based his illustration on what he thought was a female specimen, although the illustration of the dorsal surface (Fig. F) showing the bluish-lilac reflection only on the hindwing indicates that this illustrated specimen is likely to be a male (but see also above for further information). In addition, whether the original description was based on a single specimen or several specimens cannot be unambiguously determined. During our attempt to locate syntype(s) of P. ebusa, two specimens with rounded labels indicating "[Johan] Calkoen" with the locality "Brasilia" were found in RMNH. Along with the collection of Joan Raye Heer van Breukelerwaard, Johan Calkoen's collection includes Cramer types, although given the locality "Brasilia", these two specimens are most likely not syntypes of P. ebusa. Considering that we were unable to find any additional possible syntype(s) of P. ebusa, we here designate a neotype for this name. Although treated as a valid species in the past (e.g., Butler 1867; Salvin 1880-1881), in order to maintain its status as a junior synonym of P. arnaca, first recognized by Kirby (1871) and followed by most subsequent authors (e.g., Weymer 1911; Gaede 1931; Lamas 2004), we designate the specimen designated as the neotype of P. arnaca as the neotype of P. ebusa as well and retain its synonymy as a junior objective synonym (neotype designation).
D'Almeida (1922) described Euptychia arnaea [sic] form priamis based on a single male from Três Rios, Jacarepaguá, Rio de Janeiro, Rio de Janeiro, Brazil, currently housed at the DZUP. Following Article 73.1.2. of the ICZN (1999), we consider this male specimen to be the holotype fixed by monotypy based on the statement of "one male collected at the type locality" provided in the original description (D'Almeida 1922). Lamas (2004) regarded this taxon as a junior subjective synonym of Papilio arnaca without providing any justification. D'Almeida's (1922) original description provides some wing pattern characters which he considered to separate f. priamis, namely "Underside, feeble pearly reflections extending from the base to the line of ocelli"; "Underside, the two rays in the middle are narrow". These two wing pattern characters are seen in the holotype male, and indeed, the overall phenotype of specimens from the Brazilian states of Minas Gerais, Espírito Santo, and Rio de Janeiro does look somewhat different compared to the neighbouring nominotypical subspecies. Although the feeble pearly reflection extending from the base of the ventral forewing is not seen in the nominotypical subspecies, a few specimens from the aforementioned states in southeastern Brazil appear to lack this reflection (e.g., FLMNH-MGCL-1036218). The narrow ventral bands of many specimens from southeastern Brazil resemble those of A. arnaca indianacristoi ssp. n., although the ventral bands are slightly variable in width and a few specimens (e.g., FLMNH-MGCL-262982, 263014) possess bands that are similar in width to the nominotypical subspecies. Thus, the majority of the specimens from Minas Gerais, Espírito Santo, and Rio de Janeiro are distinguishable from the nominotypical subspecies based on the aforementioned characters except for specimens from Bahia consistently possessing wider ventral bands and/or lacking the feeble pearly reflections on the ventral surface. Nevertheless, we decided not to treat A. arnaca from Minas Gerais, Espírito Santo, and Rio de Janeiro as a distinct subspecies because, based on molecular data, this taxonomy would result in the nominotypical subspecies being paraphyletic. Whether subspecies should simply represent geographical variation or should also represent an evolutionary unit (i.e. a monophyletic group) is not a focus of this study and this question merits further in-depth discussion and more data. To be consistent in terms of the subspecies concept used in this study, we consider that subspecies should ideally represent clades, unless there is a strong counter-argument, and thus retain the synonymy introduced in Lamas (2004).
Distribution (Fig. 6). The nominotypical subspecies occurs from eastern Colombia south to Bolivia, and in Brazil, southern Venezuela and the Guianas, where it is typically common and widespread in lowland to submontane forest.
Variation. As explained under the variation section under the genus, the absence of the bluish-lilac reflection on the ventral forewing is consistent in specimens from western Colombia and western Ecuador. However, this character appears in a few specimens from Panama, Costa Rica, and Nicaragua (e.g., FLMNH-MGCL 208036, 257145, 263067), although the degree of reflection is variable.
Other examined specimens (244 ♂, 104 ♀). See Appendix for the data of these specimens. These specimens are not included in the type series as labelling will likely not be completed for over 300 specimens.
Etymology. This species-group name is based on the Greek adjective "adelos", meaning "indistinct" or "inconspicuous", in reference to its lack of bluish reflection on the ventral forewing. This species-group name is treated as a Latinized feminine adjective in accordance with the feminine generic name. Distribution (Fig. 6). This subspecies occurs from Nicaragua to western Ecuador. Identification and taxonomy. Bates (1865) described Euptychia sericeella based on an unspecified number of "male" specimen(s) from Vera Paz, Guatemala. Nevertheless, the only syntype specimen that we have located, in the NHMUK, is a female, a fact also noted by Godman and Salvin (1880: 89-90), who had this specimen in their possession and referred to it as "our specimen, marked as the type". This specimen (B.M. TYPE No. Rh 3181) was also referred to as a type by Riley andGabriel (1924: 53) andD'Abrera (1988: 770-771). Because of the complexity of euptychiine taxonomy, to provide an unambiguous reference for this name we here designate this specimen as the lectotype of Euptychia sericeella (lectotype designation). Bates explicitly stated that the taxon was closely allied to Eu [ptychia]. ebusa (= Amiga arnaca arnaca comb. n.), indicating that he clearly regarded Euptychia sericeella and Amiga arnaca arnaca as two different species. Despite the ambiguous diagnosis provided in the original description ("the fore-wing having a narrow costal border, the apex, and a broader outer border of a brown hue"), the lectotype, figured in Warren et al. (2018), exhibits several rather distinctive phenotypic differences compared to specimens from South America and Nicaragua to western Ecuador. Presumably due to these wing pattern differences, described further below, subsequent authors treated E. sericeella as a species-level taxon (e.g., Butler 1867Butler , 1877Weymer 1911;Gaede 1931;Forster 1964;Lamas 2004). Specimens from Zelaya department, Nicaragua (e.g., FLMNH-MGCL-263066, 263067, 263072) are phenotypically similar to specimens known from Costa Rica to western Ecuador, and A. a. sericeella thus seems to replace A. a. adela n. ssp. with little or no obvious gap or dispersal barrier separating these taxa, and conversely with no known area of sympatry. Furthermore, the presence of an iridescent bluish-lilac reflection between the postdiscal band and umbra in some specimens of the subspecies from Nicaragua, Costa Rica, and Panama might be a result of introgression with A. a. sericeella. Finally, divergence in the DNA barcode between a specimen from Honduras (LEP-16997) phenotypically similar to the lectotype of sericeella and other Central American A. arnaca from further south is no greater than between east and west Andean A. arnaca (Table 2). Based on these considerations, we downgrade E. sericeella to subspecific rank, proposing, for the first time, what we believe to be the most parsimonious hypothesis of a single species. This null hypothesis remains to be more strongly tested when new evidence, such as distributional, behavioral, or genetic data, becomes available. Both sexes of A. arnaca sericeella comb. n. et stat. n. differ from the nominotypical subspecies in the following respects: DHW feeble pearly reflection restricted to posterior one-third of hindwing and not extending into discal cell; ventral ground color somewhat paler; discal band, postdiscal band, submarginal band on the ventral forewing and ventral hindwing narrower; ventral hindwing postdiscal band not bent inwards in cell Cu 2 and gently curving towards inner margin (but see also below); ventral ocelli smaller (but see also below); bluish-lilac reflection on ventral surface appearing purplish.
Variation. The ventral ocelli are variable in terms of size; while many specimens seem to possess ventral ocelli smaller than the nominotypical subspecies and A. arnaca indianacristoi ssp. n., some appear to have ocelli that are similar to the aforementioned two taxa in terms of size. The VHW postdiscal band is bent inwards in cell Cu 2 in some specimens, whereas it gradually curves towards the inner margin in other specimens. Distribution (Fig. 6). This subspecies ranges from southern Mexico to Honduras, where it appears to be uncommon.
Examined specimens (46 ♂, 23 ♀). See Appendix for the data of these specimens. Etymology. This new species-group name is proposed in recognition of our friend and colleague, Indiana Cristóbal Ríos-Málaver, known as "Indiana Cristo", who studied the butterflies of the area where this taxon occurs. Indiana Cristo has contributed to Neotropical lepidopterology in various ways, especially through social media, where he is bringing lepidopterology to the public. This species-group name is treated as a latinized masculine noun in the genitive case.

Amiga arnaca indianacristoi
Distribution (Fig. 6). This taxon occurs in the Venezuelan Cordillera de la Costa and northwestern Cordillera de Mérida, and possibly also into the Serranía de Perijá.

Remarks.
We have examined a single specimen from Monagas, Venezuela (FLM-NH-MGCL-264682), with a phenotype that corresponds to this subspecies. Thus, this taxon's range may extend further east than that indicated above. However, given that we have examined only a single specimen from the area, combined with the fact that specimens from the island of Trinidad, whose butterfly fauna has strong biogeographic affinities with Monagas (e.g., Neild 1996), clearly represent the nominotypical subspecies, there still remains the possibility of this specimen being mislabeled.

Appendix. Data of specimens examined in this study
Abbreviations: W&H = Willmott and Hall unpublished data; DABD = Darwin Andean Butterfly Database. The locality of some specimens will appear the same as its state, province, department, etc., if no further information was available on the label.