Description and phylogenetic analysis of a new firefly genus from the Atlantic Rainforest, with five new species and new combinations (Coleoptera: Lampyridae: Lampyrinae)

Here, based on phylogenetic analyses of 18 taxa and 57 morphological characters, we propose a new firefly genus, Costalampys gen. nov., to accommodate eleven species. Five new species are herein described: C. bella sp. nov., C. capixaba sp. nov., C delicata sp. nov. (designated as type species), C. joanae sp. nov. and C. minima sp. nov. In addition, six species are redescribed and transferred from other genera: C. bisbinotata (Pic) comb. nov., transferred from Platylampis Motschulsky; C. decorata (Olivier) comb. nov., transferred from Ethra Laporte; C. pauper (Olivier) comb. nov., transferred from Cladodes Solier; as well as C. klugii (Motschulsky) comb. nov., C. quadriguttata (Gorham) comb. nov., and C. tricolor (Gorham) comb. nov., transferred from Lucidota Laporte. Costalampys gen. nov. is tentatively placed in Lampyrinae, and is diagnosed by: antennae with 11 articles, III–X basally flabellate, lacking dense and upright bristles; clypeus connected to frons by membrane, pygidium rounded; sternum VIII mucronate; phallus with dorsal plate enlarged apically, projecting ventrally and partially embracing the internal sac. Our phylogenetic analyses supported both the monophyly of Costalampys gen. nov. and the new combinations proposed. However, the relationship among congeneric species was poorly resolved. Finally, we provide illustrations of diagnostic features, distribution maps, as well as a key to Costalampys gen. nov. species, based on males.


Introduction
Fireflies (Coleoptera: Lampyridae) are charismatic insects with a surprisingly poorly-defined taxonomy worldwide and across all levels, from subfamilies to species. On one hand, the foundational framework of Lampyrid higher-level classification was developed by the early twentieth century -i.e. before the advent of phylogenetic systematics. This classification was kept largely unchanged (e.g., McDermott 1964McDermott , 1966 until fairly recently, when comprehensive analyses brought about new phylogenetic hypotheses that challenged standing lampyrid taxonomy (Branham and Wenzel 2003, Stanger-Hall et al. 2007, Jeng 2008, Martin et al. 2017, 2019. On the other hand, earlier authors, notably Ernest Olivier and Maurice Pic (e.g., Olivier 1907, Pic 1938, relied on few traits, with overly simplistic definitions, that obscured important differences across species (e.g., antennae "ramose" in opposition to "simple" [McDermott, 1964]). Therefore, most subfamilies as traditionally defined (i.e. sensu McDermott 1966) lack phylogenetic support, and their definitions have been questioned and substantially changed (Branham and Wenzel 2003, Bocakova et al. 2007, Stanger-Hall et al. 2007, Jeng 2008, Martin et al. 2017, 2019. The same problems extend to genus-and species-level taxonomy, as the vast majority of genera have never had its monophyly tested under a phylogenetic approach (but see Vaz et al. 2020).
A recent study proposed an updated classification of the Lampyridae that reshaped significantly our understanding of the evolution of fireflies -particularly regarding the relationship among taxa where males have flabellate antennae (Martin et al. 2019). Importantly, Martin et al. (2019) -based on a comprehensive taxon sampling and a sound genomic dataset -confirmed that the taxa traditionally placed in Amydetinae were a polyphyletic assemblage of taxa in which males had convergently evolved complex, branched antennae (as previously supposed by earlier authors, e.g., McDermott 1964). For example: Ethra Laporte, Scissicauda McDermott, and Cladodes Solier were found scattered throughout a newly defined Lampyrinae, whereas Vesta was found nested within Photurinae. Despite being "branched", male antennal structure in the latter four taxa is distinct in many regards, including flabellum shape and insertion on the antennomere, as well as the presence of certain bristles (see below).
The overall similarities across sensor morphologies of distantly-related species has been proposed to be associated with changes in the major signal type used in sex communication (e.g., Stanger-Hall et al. 2018). On the other hand, Martin et al. (2019) found Memoan ciceroi Silveira & Mermudes, a species with serrate male antennae, to be sister to Amydetes fastigiata Illiger, with flabellate male antennae. Interestingly enough, Silveira and Mermudes (2013) had pointed out important similarities between Memoan and Amydetes, in spite of their differences in antennal morphology. Together, these observations call for the use of a more comprehensive set of traits in firefly taxonomy, as well as a review of character statements and hypotheses of homology based on recent improvements in phylogenetic epistemology (e.g., Sereno 2007, Vogt 2017. Phylogenetic studies at and above genus level are particularly lacking for Neotropical firefly fauna -the most species-rich on Earth (Branham 2010) -where taxa often have ill-defined, outdated and overlapping diagnoses. This tangled taxonomy has led to difficulties in classifying and identifying Neotropical firefly taxa. In contrast, phylogenetic studies at and above species level have proven very useful in redefining the taxonomic boundaries of Southeast Asian and Australopacific firefly taxa (reviewed in Ballantyne et al. 2019). Therefore, thorough systematic studies aimed at providing a solid foundation for the Neotropical firefly taxonomy, and facilitating identification, are sorely needed.
Preliminary studies of our group found that some species belonging to four Neotropical lampyrine genera -Cla do des Solier, 1849, Ethra Laporte, 1833, Lu cido ta La porte, 1833, and Platylampis Motschulsky, 1853didn't fit their current placement but had instead similar mor pho logy, which motivated the present study. These four genera lack comprehensive taxonomic reviews, and their diagnoses partially overlap, especially due to Lucidota, which includes over one hundred species with strikingly disparate phenotypes (McDermott 1964(McDermott , 1966. To further complicate the issue, some species involved were described based on sexually dimorphic traits, exclusively based on a single female (e.g., Pic 1938, see below). However, lampyrid taxonomy was largely built a upon few male traits that are usually sexually dimorphic, such as the form of antennal lamellae (e.g., McDermott 1964). These factors contribute to the tangled taxonomy of Neotropical fireflies, and stresses the need for revisionary work.
Here we explore the phylogenetic relationship among several Amydetinae and Lampyrinae taxa, in support of the identification and description of a new genus. After the examination of six species previously classified as Ethra, Cladodes, Platylampis and Lucidota, we conclude that these constitute -along with five new species -a previously unknown lineage herein described and named Costalampys gen. nov. We provide detailed morphological descriptions, illustrations of diagnostic features, and a detailed distribution map for species in this genus, and a key to Costalampys gen. nov. spp. based on males.

Taxon sampling
Lampyrid taxonomy is undergoing important changes regarding higher-level classification (e.g., Martin et al. 2019), which renders taxon sampling rather challenging in groups poorly represented in phylogenies. In addition to having unreliable subfamily definitions, relationships among genera and their monophyly remain largely unexplored across subfamilies. Therefore, we tried to include in our taxon sampling the type species of genera whose taxonomic placement is at stake (i.e. would involve nomenclatural acts), as well as early-branching taxa, outside Lampyrinae, that are considered more distantly related (e.g., Psilocladus, Amydetes). For the subfamilial placement of genera, we followed Martin et al. (2019). The whole list, includes taxa from three subfamilies -Lampyrinae Rafinesque, 1815, Amydetinae Olivier, 1907, andPsilocladinae McDermott, 1964 -as well as Lampyridae incertae sedis taxa: Araucariocladus Silveira & Mermudes, 2017, previously placed in Amydetinae (Silveira and Mermudes 2017) but never properly included in any phylogenetic analysis.

Morphology, illustrations and maps
For the type species, whole specimens were soaked in 10% KOH for 24h and then dissected. Tagmas were separated, then appendages, terminalia and genitalia were separated. For the remaining species, the same procedure was applied only to the terminalia and male genitalia. Dissected structures were preserved in glycerin (for pinned specimens) or in 92% ethanol (for specimens preserved in 92% ethanol). The general terminology follows Silveira & Mermudes (2014a) and Silveira et al. (2016aSilveira et al. ( , 2016b, except for ovipositor morphology, for which we follow Lawrence et al. (2011). Photographs and measurements were taken with the Leica DFC450 and Application Suite CV3 multifocus software. The photographs were edited using Adobe Photoshop and the plates were designed with Adobe Illustrator (Adobe Systems). Total length was given by the sum of pronotum and elytron lengths. Distribution maps were built using the program QGIS 2.18.10 (QGIS.org 2017).

Character coding
Characters were coded as binary or multistate, following the logical basis of Sereno (2007). The matrix was constructed in Mesquite v3.2 (Maddison and Maddison 2018). The data matrix can be found in the supplementary file 2. The proposition of homologies was based on direct observations of adult males and females. The following outgroup taxa are known only from adult males, therefore information about females is lacking: Araucariocladus, Amydetes, Cladodes, Dadophora and Ethra. This is also true for the ingroup species Costalampys joanae sp. nov. and Costalampys minima sp. nov., of which females are unknown. We only included one female-based character, which added only 0.6% of missing data in our analysis.

Phylogenetic analyses
We ran and compared the results of both Parsimony and Bayesian analyses, seeking consistency among them, as there is no consensus regarding which is the best method to infer phylogenies using morphological characters (Goloboff et al. 2008b, 2018, Wright and Hillis 2014, Puttick et al. 2017, O'Reilly et al. 2018, Sansom et al. 2018, Schrago et al. 2018, Smith 2019. Parsimony analyses were performed with TNT (Goloboff et al. 2008a), under equal weights (EW) and implied weights (IW) (Goloboff 1993). All analyses were conducted by heuristic searches, TBR branch swapping, 1000 replicates and 100 trees saved by replicate. For the IW analysis, we explored the topologies obtained under different concavity constant values (k = 1, 2, 3, 5, 10, 15). The k values were not chosen with regular intervals because high k values tend to generate uniform results, similar to the ones of EW analysis (Goloboff et al. 2008b). Therefore, analysis with regular k intervals may result in bias toward high k topologies (Mirande 2009). Support was assessed through Symmetric resampling (SR), which is not distorted by implied weights (Goloboff et al. 2003).
Absolute Bremer values were calculated to the EW analysis as well.
The topology to evaluate the character evolution was chosen by a sensibility analysis (Giribet 2003), which selects the most stable topology, that is, the one that shows up more times. In addition, generated trees with the synapomorphies were mapped in Winclada, version 1.00.08 (Nixon 2002).
Bayesian Inference was performed on MrBayes 3.2 (Huelsenbeck and Ronquist 2001). The analysis was performed with 10.000.000 generations, trees saved each 1.000 generations, and 10% burn-in. The analysis used the evolution model MKV, modified from MK (Lewis 2001), and was checked for convergence on Tracer v1.6 (Rambaut et al. 2014). Trees were read in FigTree version 1.4 (Rambaut 2009).

Morphological characters
The detailed study of the morphology of the taxa included in the analyses allowed us to identify 57 characters: 11 from the head; nine from the thorax; and 37 from the abdomen, 18 of which from the aedeagus. The characters were coded as binary (N=47) or multistate (N=10).
The parsimony analyses under IW found one most parsimonious tree in each k value. The relationship among outgroup taxa were consistent across analyses, showing minor differences only on the k1 topology. D. hyalina was always recovered as sister to Costalampys gen. nov., but with low support, and this clade was sister to L. banoni. In all analyses Costalampys gen. nov. was found monophyletic, highly supported by symmetric resampling (>71). The synapomorphies supporting the monophyly of the genus were the same as for the EW analysis in all topologies.
The second clade is composed by Lucidota tricolor + (C. joanae sp. nov. + (Cladodes pauper + (L. tricolor + (C. bella sp. nov. + C. delicata sp. nov.)))), well supported by symmetric resampling [65]. The internal clade of C. bella sp. nov. and C. delicata sp. nov. is highly supported in all analyses (by symmetric resampling [>95]).     The second topology was found on the analyses with k 10, 15 and is equal to trees found on EW analysis (with a minor change across outgroup taxa in one of EW trees). As both topologies were found the same number of times (three times, with just one with different outgroup topology), we discuss both of them here.
Etymology. Costalampys is proposed in honor of our dear Professor and friend, the entomologist Dr. Cleide Costa, which greatly influenced our lives and the study of beetles. Gender feminine.
Biology. Five species of Costalampys gen. nov. have been observed active during daytime, namely C. decorata (Olivier, 1888) comb. nov., C. delicata sp. nov., C. joanae sp. nov., C. pauper (Olivier, 1899) comb. nov., and C. tricolor (Gorham, 1880) comb. nov. Males and females were observed perching on leaves of understory bushes and growing trees (as in Fig. 10). Males will sustain the antennae erect and pointed out towards the wind current, very similar to other firefly taxa with flabellate antennae, like many Ethra Laporte, 1833, Amydetes Illiger, 1807 and Cladodes species (L. Silveira pers. ob.). Given the similarity in overall morphology, it is likely that other Costalampys gen. nov. species in the genus are also diurnal. Where observed, lanterns are functional, and individuals will glow if disturbed. Other roles of bioluminescence (e.g., in courtship) are unknown.

Diagnostic description.
Overall dark brown. Pronotal disc (Fig. 7F) dark brown, lacking vittae, pronotal expansions variable, from having anterior small vittae to almost entirely pale yellow, except by posterior margin, which is outlined in brown, blending with the dark brown disc color. Elytron (Fig.7A, C) dark brown or brown, with a pale-yellow longitudinal stripe about 1/2 as long as elytra, usually with a roundish pale yellow spot at posterior 3/4 of elytral length, reaching inner suture, as wide as about 1/3× elytral width. Legs (Fig.7B, D) dark brown, except for trochanters and, sometimes, metacoxa, which are light brown. Sternum VIII with lateral vitreous spots. Pygidium (Fig. 7H) with medial 1/5 brown, with lateral pale-yellow stripes in males, entirely brown in females.
Antennomeres III-X with apical corners projected, pointed, sterna VI and VII lacking lanterns.
Remarks. Platylampis Motschulsky, 1853 is currently a consortium of morphologically very disparate species, with no clear-cut definition to accommodate all its species (cf. Silveira et al. 2019). We did not include the type species P. latiuscula Motschulsky, 1853 in our phylogenetic analyses because we had no permission to dissect any specimen of that species -internal traits constitute an important part of our character list (see section 3.1). However, we did study the external morphology of P. latiuscula Motschulsky, 1853, and were able to contrast its morphology with that of P. bisbinotata Pic, 1943. It becomes evident that the latter is more similar to Costalampys spp. than to P. latiuscula Motschulsky, 1853. Specifically, P. latiuscula Motschulsky, 1853 males have serrate antennae, terga VI and VII with posterior corners well-developed, projected posteriorly and pointed, sternum VII with posterior margin medially notched, sternum IX with posterior margin strongly acuminate, pygidium with well-developed posterior corners, and with central portion acuminate. However, C. bisbinotata (Pic, 1943) comb. nov. males have flabellate antennae with basally branching lamellae, terga VI and VII with pos- terior corners just slightly produced, sternum VII with posterior margin medially straight, sternum IX with posterior margin rounded, pygidium with weakly developed posterior corners, and with central portion rounded -all of which are defining characters of Costalampys gen. nov. (see above). Along with our phylogenetic analyses, these observations support the transfer of P. bisbinotata Pic, 1943 to Costalampys gen. nov., therefore C. bisbinotata (Pic, 1943). Pic (1943) has not given an etymology for P. bisbinotata, but it could mean "with four spots" (i.e two times double spots) in Latin, referring to its color pattern.
Costalampys bisbinotata (Pic, 1943) comb. nov. is similar to C. capixaba sp. nov. and C. quadriguttata (Gorham, 1880) comb. nov. in the pronotal color pattern (overall dark brown; pronotal disc dark brown, lacking vittae, pronotal expansions somewhat translucent anteriorly, almost entirely pale yellow, except by posterior margin, which is outlined in brown, blending with the dark brown disc color). It differs from both species by the lack of lanterns in females (present on sternum VI in the others). Costalampys bisbinotata (Pic, 1943) comb. nov. differs from C. capixaba sp. nov. by the shorter elytral pale-yellow stripe in both sexes (about 1/2 as long as elytra), and pygidium with lateral translucent spots in males (elytral stripe as long as about 4/5× elytral length, lanterns present in females, and male pygidium entirely brown in C. capixaba sp. nov.). Costalampys bisbinotata (Pic, 1943) comb. nov. differs from C. quadriguttata (Gorham, 1880) comb. nov. by the sternum VIII and pygidium medially dark brown in both sexes (entirely translucent in C. quadriguttata (Gorham, 1880) comb. nov.).

Remarks.
Costalampys capixaba sp. nov. is similar to C. bisbinotata (Pic, 1943) comb. nov. and C. quadriguttata (Gorham, 1880) comb. nov. in the pronotal color pattern (overall dark brown; pronotal disc dark brown, lacking vittae, pronotal expansions somewhat translucent anteriorly, almost entirely pale yellow, except by posterior margin, which is outlined in brown, blending with the dark brown disc color), but can be distinguished from both species by the more elongate elytral pale yellow stripe (as long as 4/5× elytral length in C. capixaba sp. nov., versus 1/2 as long, in the other mentioned species), and the pygidium entirely brown in males (at least partially translucent in the others). Other traits that distinguish C. capixaba sp. nov. from C. quadriguttata (Gorham, 1880) comb. nov. are the elytron without a roundish pale-yellow spot at posterior 3/4, reaching inner suture (present, and about as wide as a 1/3 of elytra in C. quadriguttata (Gorham, 1880) comb. nov) and the sternum VII entirely dark brown in both sexes (with a central translucent spot in C. quadriguttata (Gorham, 1880) comb. nov). Finally, C. capixaba sp. nov. can also be discriminated from C. bisbinotata (Pic, 1943) comb. nov. by the presence of a rounded lantern on sternum VI in females (absent in C. bisbinotata (Pic, 1943) (Figs 9F, H, 10) entirely pink or orangish, sometimes as two almost contiguous spots occupying most disc, pronotal expansions from entirely black to entirely testaceous, mildly translucent anteriorly. Elytron (Fig. 9A, C) black, with a pale-yellow longitudinal stripe about 4/5× as long as elytra. Legs (Fig. 9B, D) dark brown to black, except for trochanters, which are brown. Sternum VIII entirely dark brown, without lateral vitreous spots. Pygidium (Fig. 9H) entirely dark brown.
Remarks. Olivier (1888) described Ethra decorata based on specimens from Brazil (Petrópolis, Rio de Janeiro State), and mentioned that the specimens had testaceous pronotal expansions. Later on, Olivier (1909) described a variant of Ethra decorata with darker pronotal expansions, which he named "var. Circumcincta", but didn't mention any type locality -the holotype examined is from Brazil, Minas Gerais. The latter variant must be deemed as a subspecies according to ICZN article ICZN Art. 45.6.4, therefore Ethra decorata circumcincta Olivier, 1909 (according to Art. 5.2). After analyzing a larger series of C. decorata (Olivier, 1888) comb. nov. (including type material of both subspecies), we found out that some individuals in the same population had pronotal expansions of variable color -some were testaceous, and others were dark brown to varying degrees. Therefore, we synonymize here both subspecies, Ethra decorata decorata Olivier, 1888, and Ethra decorata circumcincta Olivier, 1909, the former having priority over time. Pic (1938) published a study about mimicry across firefly species from Brazil, including some new taxa. He noted that a new variety (i.e. a subspecies, see above), Lucidota oculata diversicollis Pic, 1938 was very similar to Ethra decorata Olivier, 1888, but with different antennae, among other traits. The holotype of Lucidota oculata diversicollis Pic, 1938 is a female specimen, cospecific of Ethra decorata. Hence, after the study of the type material, we synonymize Lucidota oculata diversicollis Pic, 1938 with Ethra decorata, which has priority over time. After our phylogenetic analyses (see section 3.2), we transfer Ethra decorata to Costalampys gen. nov., therefore Costalampys decorata (Olivier, 1888) comb. nov.
In the same aforementioned work on mimicry, Pic (1938) described a second subspecies of E. decorata, which he named Ethra decorata schneideri Pic, 1938. After the study of the holotype, we synonymize it with Lucidota tricolor Gorham, 1880, which we transfer to Costalampys gen. nov. (see below) in this paper.
Olivier (1888) does not give an explicit etymology for decorata, but we assume it was based on the Latin word for "decorated" or "adorned".
Costalampys decorata (Olivier, 1888) comb. nov. is similar to C. delicata sp. nov., C. joanae sp. nov., C. tricolor (Gorham, 1880) comb. nov., and C. bella sp. nov., in the dorsal color pattern (overall brown or dark brown, with pronotal disc broadly pink, and with an elongate pale-yellow spot on elytron). C. decorata (Olivier, 1888) comb. nov. is unique among the aforementioned species by having more well-developed lanterns -both in males and females -which occupy the medial 1/3 of the sternum, reaching or almost reaching the anterior margin of the sternum. The sternum VI lantern is never longer than 2/3× sternum length, or wider than 1/5× sternum width in the other similar species given above.
Costalampys decorata (Olivier, 1888) comb. nov. was previously reported for the Brazilian states of Rio de Janeiro and Minas Gerais. Here, we expand the geographic range to the states of Espírito Santo and São Paulo, and provide for the first time detailed information of temporal occurrence of the said species.

Remarks.
The type species Costalampys delicata sp. nov. is similar to C. bella sp. nov., Costalampys joanae sp. nov., C. tricolor (Gorham, 1880) comb. nov., and C. decorata (Olivier, 1888) comb. nov., in the dorsal color pattern (overall brown or dark brown, with pronotal disc broadly pink, and with an elongate pale-yellow spot on elytron). C. delicata sp. nov. is unique among the aforementioned species by the pale-yellow legs in both sexes (overall brown or dark brown in the others), and antennomere III with lamella short, as long as core antennomere in males (at least 1.5 longer in the others).

Remarks.
Costalampys joanae sp. nov. is similar to C. delicata sp. nov., C. decorata (Olivier, 1888) comb. nov., C. tricolor (Gorham, 1880) comb. nov., and C. bella sp. nov., in the dorsal color pattern (overall brown or dark brown, with pronotal disc broadly pink, and with an elon- gate pale-yellow spot on elytron). Costalampys joanae sp. nov. is unique among the aforementioned species by having a longer antennomere III lamella, which is about 4× longer than core antennomere, as well as sides of pygidium strongly convergent posteriorly (homogeneously rounded in the other species).
Remarks. Lychnuris klugii was described by Motschulsky  (1853: 4), who wrongly attributed the species authorship to Dejean. Indeed, Dejean (1833) cited Lychnuris klugii in his second catalog, but without a description, making it therefore a nomen nudum (cf. Bousquet and Bouchard 2013). The correct authorship of the species belongs to Motschulsky (1853), who in first provided a description for Lychnacris klugii. Motschulsky (1853) lists the species for Southern Brazil, without further details.

Remarks.
Costalampys minima sp. nov. is unique among its congenerics by the color pattern (described above).

Etymology.
Minima is a Latin adjective that means tiny.
Remarks. Olivier (1899) described Cladodes pauper from Brazil -the syntype we studied is from Teresópolis, Rio de Janeiro, Brazil. In addition to Teresópolis, in the Serra dos Órgãos mountain range, we found previously unreported populations at the Serra da Mantiqueira and Maciço da Pedra Branca mountain ranges, in the states of Rio de Janeiro and Minas Gerais. Our study provides detailed geographic information for the first time, and expands considerably the known range of the species. After our phylogenetic analysis (see section 3.2), we transfer Cladodes pauper to Costalampys gen. nov., therefore Costalampys pauper (Olivier, 1899). Olivier (1899) didn't provide an etymology for pauper, but we assume it was based on the Latin word for "poor". Because Oli vier (1899) originally placed it in Cladodes Solier, 1849, Cladodes pauper would be relatively smaller than the remaining species in this genus, therefore "poor".
Costalampys pauper (Olivier, 1899) is unique among Costalampys spp. by the pronotum entirely dark brown to black and sternum VI without lantern, in both sexes. C. pauper is also unique in being restricted to relatively cooler climates, either at sites of relatively higher elevation (>1200m) during the Austral Spring season, or at lower elevations (~600m), during Austral Winter. These observations suggest that this species might be adapted to climates substantially cooler than its congenerics, which deserves further scrutiny.

Diagnostic description.
Overall dark brown. Pronotal disc (Fig. 20A, F) dark brown, lacking vittae, pronotal expansions somewhat translucent anteriorly, almost entirely pale yellow, except by posterior margin, which is outlined in brown, blending with the dark brown disc color. Elytron (Fig. 20B) brown, with a pale yellow longitudinal stripe about 1/3× as long as elytra, with a roundish spot at posterior 3/4, reaching inner suture, about a 1/3× as wide as elytra. Legs brown, except for trochanters and metacoxa, which are light brown. Sternum VIII (Fig. 20G) entirely translucent. Pygidium (Fig. 20H) entirely translucent. Pronotum (Fig. 20F) with sides rounded, divergent posteriorly. Male. Total length = 11.1-11.9 mm (aver. 11.5 mm); Pronotal length = 2.5 mm; Pronotal width = 4.0-4.2 mm (aver. 4.1 mm); Elytral length = 9.0 mm; Elytral width = 2.9-3.0 mm (aver. 2.95 mm). Antennomere III (Fig.  20E) with lamella about 2× longer than core antennomere. Sternum VI (Fig. 20B) with lantern of moderate size, occupying the medial 1.4 of the sternum, almost reaching its anterior margin. Sterna VI and VII (Fig. 20B)  Remarks. Gorham (1880) Lucidota quadriguttata from Bahia, Brazil: The author didn't give an explicit etymology for the species, but we assume it was based on the Latin word for "with four spots", referring to its color pattern. Costalampys quadriguttata (Gorham, 1880) comb. nov. is similar to C. bisbinotata (Pic, 1943) comb. nov. and C. capixaba sp. nov. in the pronotal color pattern (overall dark brown; pronotal disc dark brown, lacking vittae, pronotal expansions somewhat translucent anteriorly, almost entirely pale yellow, except by posterior margin, which is outlined in brown, blending with the dark brown disc color), but differs from both in having the pygidium, as well as sterna VIII and IX, entirely translucent in both sexes (dark brown in both sexes of C. capixaba sp. nov., partially translucent in C. bisbinotata (Pic, 1943) comb. nov. males, entirely dark brown in females), and sternum VII with a medial translucent spot in males (absent in the other species. Costalampys quadriguttata (Gorham, 1880) comb. nov. was previously reported for the Brazilian State of Bahia. Olivier (1886) mentioned specimens collected by E. Gounelle in "Salobro", currently Canavieiras, BA (cf. Aguiar et al. 2014), which we found in his collection at MNHN (see below). Pierre Émile Gounelle was an engineer and entomologist who worked for the French Ministry of the Colonies, and traveled to Brazil on several occasions (Papavero 1971, Lourteig 1986). On his first trip, among other places, Gounelle visited "Salobro" between June and July, 1885 (Papavero 1971), which was a pop-ular destination for its diamond mines at that time (e.g., Aguiar et al. 2014). Here, we expand the known range of the species to the State of Pernambuco, and also give more detailed locality data within the state of Bahia.
Remarks. Gorham (1880) described Lucidota tricolor from Brazil, but provided no further provenance data, based on a male specimen provided by the French naturalist Chevrolat. Gorham (1880) didn't give an explicit etymology for Lucidota tricolor, but we assume it was based on the Latin word for "with three colors", referring to its color pattern. Olivier (1886) described Lucidota oculata based on a single female specimen from Matosinhos, Minas Gerais, Brazil: The said author didn't give an explicit etymology for Lucidota oculata, but we assume oculata comes from the Latin adjective that means "with eyes", referring to the pronotal spots. Pic (1938) described Ethra decorata schneideri from Brazil, without further details. The holotype of Ethra decorata schneideri Pic, 1938 is presumed here based on a specimen in M. Pic's collection (MNHN) that was labeled as "Lucidota oculata E Ol" and "var. Schneideri".
We based our inference on the combination of the following observations: (i) the aforementioned specimen in Pic's collection was labeled as "type"; (ii) we found no description of a species or variety named L. oculata var. Schneideri Pic; and (iii) Pic described Ethra decorata var. schneideri in a paper about mimetism among some lampyrid species with similar dorsal color pattern, including both Lucidota oculata Olivier and Ethra decorata Olivier (Pic 1938), and his description given for his new variety fits the latter, but not the former. Furthermore, Pic (1938) regarded Lucidota spp. as having antennae long and compressed but with no lamellae, whereas Ethra spp. had flabellate antennae (a trait found in the presumed holotype). Together, these observations support our inference that our presumed holotype of Ethra decorata var. schneideri is the holotype, that nevertheless was mislabeled. The latter must be deemed as a subspecies according to ICZN article ICZN Art. 45.6.4, therefore Ethra decorata schneideri Pic, 1938 (according to Art. 5.2).
After the study of the type materials, we synonymize here Lucidota oculata Olivier, 1886 and Ethra decora- ta schneideri Pic, 1938 with Lucidota tricolor Gorham, 1880, which has priority over time. After our phylogenetic analysis, we transfer Lucidota tricolor Gorham, 1880 to Costlampys gen. nov. (see section 3.2).
Costalampys tricolor (Gorham, 1880) comb. nov. is most similar to C. bella sp. nov., as the two species share a similar dorsal color pattern (overall brown or dark brown, with pronotal disc with paired pink vittae to broadly pink, and with an elongate pale-yellow spot on elytron). It differs from C. bella sp. nov. by the rounded sides of pronotum in both sexes (almost straight in C. bella sp. nov.), the pygidium with posterior margin rounded to almost straight in males (mucronate in C. bella sp. nov.), as well as by the presence of lanterns in sternum VI in the females (absent in C. bella sp. nov.).
Costalampys tricolor (Gorham, 1880) comb. nov. was previously known from Brazil, without further details. Here, we provide more detailed information on the geographic and temporal occurrence of the species.

Discussion
Here we proposed a new genus of lampyrid fireflies, based on a phylogenetic analysis of five new and six previously described species -all of them are reviewed for the first time since their original descriptions.

Phylogeny of Costalampys gen. nov.
Costalampys gen. nov. is monophyletic, although the relationship among tips was sensitive to the parameters selected (Figs 1-5; see Results). The relationship between species in the genus changed moderately between analyses. The sister species to all other Costalampys gen. nov. was C. klugii, and the clades (C. bella sp. nov. + C. delicata sp. nov.) and (C. capixaba sp. nov. + (C. minima sp. nov. + (C. quadriguttata + C. bisbinotata)). The position of C. decorata, C. joanae sp. nov., C. pauper and C. tricolor was different in EW and IW (low K values topology) trees, suggesting that the relationships between these species and the others in the genus were mostly based on homoplastic features (as IW downweights homoplastic traits (Goloboff et al. 2018)).
As expected, the Bayesian analysis (Fig. 5) generated results slightly more similar to the EW parsimony, considering that MKv model considers equal weights across characters. The topologies generated by parsimony (Figs  1, 3) had better resolution compared with the Bayesian tree (Fig. 5). Indeed, Bayesian analyses tend to be more accurate, but with lower resolutions -especially in small datasets, as the present study (Puttick et al. 2017, O'Reilly et al. 2018, Schrago et al. 2018). Since accuracy without any resolution is not helpful, parsimony analyses may sometimes be preferred over Bayesian analyses for small datasets (Sansom et al. 2018).
All the three methods we used (Bayesian, EW and IW parsimony analyses) are useful in inferring phylogenies, the difference between their respective results being sensitive to the parameters used and character to taxon ratio (Smith 2019). The debate around the most appropriate method of inferring phylogenies is ongoing, and the lack of studies employing empirical analyses instead of simulations stresses that we might be far from reaching a consensus (exceptions are works by Schrago et al. (2018) and Sansom (2018), which have found opposing results).
Even though we had a rather limited taxon sampling family-wise, we included all taxa morphologically similar to Costalampys gen. nov. (particularly Scissicauda McDermott, 1964, Ethra Laporte, 1833and Lucidota Laporte, 1833. Interestingly, Costalampys species were never found to be sister to any of the latter three genera. Instead, the poorly known Dadophora hyalina Olivier, 1907 was recovered as sister to Costalampys gen. nov. in all our analyses. Surprisingly, males of D. hyalina Olivier, 1907 are superficially very different from those of Costalampys gen. nov. species. For example, males of D. hyalina Olivier, 1907 have antennae with twelve antennomeres, III-X being cup-shaped, and subparallel-sided pronotum and elytra. On the other hand, males in Costalampys gen. nov. species have basally flabellate antennae with eleven antennomeres, and pronotum and elytra with rounded sides (except for C. bella sp. nov., which has pronotum with sides almost parallel-sided). Yet, D. hyalina Olivier, 1907 and Costalampys gen. nov. share several traits, particularly in their aedeagus, for example: dorsal plate curved dorsally, shorter than paramere, the latter bearing a well-developed ventral projection and a subapical tooth. It is noteworthy that genitalic characters are currently unknown for most Neotropical taxa, or may show little difference between species in lampyrid genera (e.g., Amydetes (cf. Silveira and Mermudes 2014a)). In Costalampys gen. nov., however, these characters influenced the analysis considerably, which shows the importance of including firefly genitalia in taxonomic phylogenetic studies (as seen, for example, in Ballantyne et al. (2019)). Unfortunately, females of D. hyalina remain unknown, which precludes comparisons between these and females of Costalampys species is not possible.
Costalampys gen. nov. is roughly circumscribed within the domains of the Atlantic Rainforest (but see below, and Fig. 22), a much-threatened biome and biodiversity hotspot (Myers et al. 2000, Ribeiro et al. 2009, Mittermeier et al. 2011, Scarano and Ceotto 2015. In fact, less than 11.7% of its original extent remains (of which only 9% is protected), and even these fragmented patches of forest are very often embedded in a matrix of intense agriculture or densely inhabited cities (Ribeiro et al. 2009, Scarano andCeotto 2015). Although native Brazilians had already mildly transformed the landscape by using food resources (Gaspar et al. 2008), these currently isolated patches were mostly connected until early 1500s, when European settlers started extensive logging and conversion of forests into pasture (Dean 1996). The replacement of native forest for agriculture or urban areas was intensified in the nineteenth and twentieth centuries (Scarano and Ceotto 2015). Two species, C. quadriguttata and C. minima sp. nov., had populations recorded in forested patches of the Caatinga biome, in addition to areas of Atlantic Rainforest. C. quadriguttata was recorded in areas of Agreste formation (e.g., Caruaru, Pernambuco State) -ecotones between Caatinga and Atlantic rainforest, characterized by a warm and sub-humid forest climate (Alvares et al. 2013), although increasingly drier (Pereira et al. 2017) -but also in typical areas of Atlantic Rainforest, with more humid climate (Camacan and Canavieiras, both in Bahia State) (Alvares et al. 2013). On its turn, C. minima sp. nov. was collected in a forested area at Sete Cidades National Park, in the State of Piauí (within the Caatinga biome), in addition to coastal sites of Rio Grande do Norte state (National Heritage Private Reserve [RPPN, acronym in Portuguese] Mata Estrela, Baia Formosa), within the Atlantic Rainforest biome.
All our analyses recovered a weakly supported monophyletic clade consisting of all four species occurring from the margins of the Doce River and northwards (C. capixaba sp. nov. + (C. minima sp. nov. + (C. quadriguttata + C. bisbinotata))). This area encompasses a major refugium within the Atlantic Rainforest that remained evergreen throughout the Pleistocene (Carnaval andMoritz 2008, Martins 2011). If true, this finding would provide meaningful hints on the evolutionary biogeography of this lineage.
Interestingly enough, Costalampys species have rather narrow geographic ranges for fireflies with winged females. For example, the South American Cratomorphus cossyphinus is common throughout continental marshes Eastern to the Andes (Campos et al. 2018), and the Nearctic Photinus pyralis and Ellychnia corrusca are common throughout Southern and Eastern North America (Faust 2017. In sharp contrast, Costalampys species are limited to narrow ranges, either certain mountain slopes Southwards to the Doce River (e.g., C. delicata, C. decorata), or coastal plains Northwards to the Doce River (C. quadriguttata, C. bisbinotata). These restricted ranges are more common among firefly taxa with presumed larviform females, like Amydetes and Ybytyramoan (cf. Mermudes 2014a, 2014b). Assuming the distribution of Costalampys species is not limited by dispersal, stenotopic niches (i.e., narrow environmental preferences) might explain their restricted ranges -although limited sampling can not be ruled out.

Conclusions
Our contribution is part of a series of revisionary studies aiming at obtaining a deeper knowledge about neotropical fireflies, by combining museum research, new collections and field data, along with phylogenetic analyses. We redescribed, after studying the type material, then transferred certain previously described species to Costalampys gen. nov. from genera as disparate as Ethra, Cladodes, Lucidota, and Platylampis, underlining the poor state of neotropical lampyrid taxonomy. Although we might be far from having a complete picture about the species-level diversity within Costalampys gen. nov., our study sets a solid framework for future studies in the genus. Finally, our study paves the way to the use of phylogenetic analysis in comprehensive systematic studies of Neotropical firefly taxa.

Acknowledgements
We thank J. P. Botero, M. Geiser and V. Nunes for critically reviewing the manuscript. This research was funded by Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro (FAPERJ 101.476.2010)