A neotype for Hyla x‐signata Spix, 1824 (Amphibia, Anura, Hylidae)

The uncertain identity of Hyla x‐signata Spix, 1824 has been a pervasive problem in the taxonomy of the genus Scinax. A species supposedly distributed from northeastern Brazil northwards to Colombia and Venezuela, described in a few lines without much information and with an accompanying figure, and its type specimen lost during World War II, combined to produce a curious situation. Twenty-one of the 39 species of the S. ruber Clade described in the last 50 years were considered to require a diagnosis from S. x‐signatus by their authors. In most cases these had no other alternative than to gather information about this species from indirect sources, frequently pointing out the problems associated with its uncertain identity. In this paper, we review the taxonomic history of Hyla x‐signata, designate a neotype, provide a redescription including advertisement call and sequence data, and diagnose it from all other species of the S. ruber Clade. Key-Words. Hylinae; Scinaxini; Scinax; Scinax ruber.


INTRODUCTION
Scinax currently includes 125 species in two major clades, the S. catharinae and the S. ruber Clades (species number from Frost, 2020; taxonomy of Scinaxini as recognized by Faivovich et al., 2018). The latter clade includes 75 species; three in the S. uruguayus species Group (Faivovich et al., 2005;Baldo et al., 2019), 10 in the S. rostratus species Group (Lima et al., 2005), and 62 unassigned to any species group. These are mostly the result of a phylogenetic analysis (Faivovich, 2002) that rejected the monophyly of most species groups recognized at that time for the species now included in the S. ruber Clade. Subsequent analyses (Wiens et al., 2010;Duellman et al., 2016) corroborated that result, always with a low density of exemplar species of the clade, and therefore without the possibility of recognizing monophyletic groups of taxonomic relevance. While this situation alone requires extensive diagnoses every time a new species is described, the efforts are magnified by the complicated taxonomy of these frogs, and some confusion persists regarding the identity of certain species. One of these is S. x-signatus (Spix, 1824), a putatively widespread species whose identity has been, for nearly 70 years, a specter of doubt hunting the taxonomy of several species of Scinax.
A species supposedly distributed from northeastern Brazil northwards to Colombia and Venezuela, described in a few lines without much information and with an accompanying figure, and its type specimen lost during World War II (Hoogmoed & Gruber, 1983;Glaw & Franzen, 2006), combined to produce a curious situation. Twenty-one of the 39 species of the S. ruber Clade described in the last 50 years were considered to require a diagnosis from S. x-signatus by their authors (e.g., Bokermann, 1968;Pombal et al., 1995a;Barrio-Amorós et al., 2004;Sturaro & Peloso, 2014;Araujo-Vieira et al., 2016;Acosta-Galvis, 2018). In most cases, these had no other alternative than to gather information about this species from indirect sources, frequently pointing out the problems associated with its uncertain identity. In this paper, we review the taxonomic history of Hyla x-signata, designate a neotype, provide a redescription of the species, and diagnose it from all other species of the S. ruber Clade.
Fingers were numbered II to V following Fabrezi & Alberch (1996). Webbing formula follows Savage & Heyer (1967) as modified by Myers & Duellman (1982). Dorsal and profile standards of snout shape follow Heyer et al. (1990). Nuptial pad terminology is that of . Sex was determined by examination of secondary sexual characters (nuptial pads, vocal slits, and vocal sacs) or dissections.
Muscles were studied with a Lugol solution to enhance contrast (Bock & Shear, 1972). We considered three morphologies of external, subgular vocal sacs: single, weakly bilobate, and bilobate (modified from Duellman, 1970). They correspond anatomically to three different configurations of the subgular portion of the m. interhyoideus. In single vocal sacs, the m. interhyoideus does not have a medial constriction in its posterior portion, forming a uniform lobe. In weakly bilobate vocal sacs, the m. interhyoideus has a slight medial constriction in its posterior portion, forming small but continuous contralateral lobes. In bilobate sacs, the m. interhyoideus has a conspicuous medial constriction in its posterior portion, forming a discrete lobe on each side of the subgular region.
Recordings were made with Marantz PMD660, PMD661MKII, or Tascam DR40 digital recorders, coupled with a Sennheiser ME66 directional microphone from a distance of 0.5-1.0 m from the individual. Calls were recorded and digitized in uncompressed PCM and WAV formats with a sample rate of 44.1 kHz or 48 kHz and 16-bit encoding. Call analyses were performed with Raven Pro If necessary, we filtered the frequencies above and below the bandwidth of the call, to reduce interference from environmental sounds (usually below 500 Hz and above 5,500 Hz). Temporal parameters were measured from waveforms, whereas spectral parameters were measured from spectrograms. We use the term "note" as defined by Köhler et al. (2017): subunits of a call that are separated by 100% amplitude modulation with only short intervals between them relative to length of note.
For the calls analyzed in this study, we selected 15-32 notes from each recorded male, comprising a total of 141 notes. We delimited the selection borders (i.e., beginning and end) of each note using the threshold of 10% of its maximum amplitude (see Littlejohn, 2001). From each selection we took the following Raven Pro auto measurements (see Charif et al., 2010 for definitions): note duration (delta time), dominant frequency (peak frequency), frequency 5% (freq5%), frequency 95% (freq95%), bandwidth 90% (BW90%). We also measured manually the interval between notes, note repetition rate (notes per minute), number of pulses per note, pulse repetition rate (pulses per second), and pulse duration (measured from the pulse with the highest amplitude in 10 notes from each individual using the threshold of 10% of its maximum amplitude). We did not measure the pulse duration of the call of MZUESC 20683 due to the recording's suboptimal waveform resolution.

Phylogenetic analysis
We performed a phylogenetic analysis with the sole goal of testing the identification of all sequences in GenBank that had been associated with Scinax x-signatus in different publications (Fouquet et al., 2007a, b;Lyra et al., 2016;Vacher et al., 2020). We included sequences of the 16S rRNA mitochondrial gene for 28 specimens of S. x-signatus, including the neotype, from some localities in the Brazilian states of Bahia, Ceará, Pernambuco (including the Island Fernando de Noronha, introduced), and Amapá, and also from French Guiana and Martinique. One sequence associated with S. x-signatus (KU495577; Lyra et al., 2016) from Itabuna, Bahia, was excluded from the analysis because the tissue sample (CFBHt 03433) is associated with a specimen of S. argyreornatus (CFBH 2890) that belongs to the S. catharinae Clade. We also included sequences for one related lineage to S. x-signatus (S. x-signatus "Scinax_64" in Vacher et al., 2020), specimens of S. eurydice, S. fuscovarius, S. nasicus, S. ruber, and S. similis -species that share some morphologically similarities with S. x-signatus -, and S. berthae, S. catharinae (S. catharinae Clade), and Sphaenorhynchus lacteus (Sphaenorhynchini) as outgroup taxa. The dataset includes 52 terminals, of which sequences of twelve were produced for this study; the remaining corresponds to sequences from GenBank. See Appendix 2 for a complete list of voucher specimens included in the analysis.
For sequences produced in this study, whole cellular DNA was extracted from frozen and ethanol-preserved tissues (liver or muscle) using either phenol-chloroform extraction methods or the Qiagen DNeasy isolation kit. Primers used in PCR amplification were 16sAR (5'-CGCCTGTTTATCAAAAACAT-3'; Palumbi et al., 1991) and 16sWilk2 (5'-GACCTGGATTACTCCGGTCTGA-3'; Wilkinson et al., 1996) or 16sBR (5'-GACCTGGATTACTCCGGTCTGA-3'; Palumbi et al., 1991). PCR amplification was carried out in 25 µl reactions using 0.2 µl Taq (Fermentas). The PCR protocol consisted of an initial denaturation step of 3 minutes at 94°C, 35 cycles of 30 s at 94°C, 40 s at 48°C, and 30-60 s at 72°C, and a final extension step of 10-15 minutes at 72°C. The PCR amplified products were cleaned with 0.5 µL of Exonuclease plus 1 µL of Alkaline Phosphatase per 20 µL of reaction. Sequencing was carried out on an automatic sequencer ABI 3730XL (Applied Biosystems) in both directions to check for potential errors and polymorphisms. The chromatograms obtained from the automated sequencer were read, contigs made using the sequence editing software Sequencher v5.3 (Gene Codes, Ann Arbor, MI, USA), and edited the complete sequences with BioEdit (Hall, 1999).
We performed a multiple sequence alignment of the 16S rRNA fragments employing MAFFT v.7 (Katoh & Standley, 2013;default parameters). For the parsimony analysis, we employed TNT v1.5 Willi Hennig Society Edition (Goloboff et al., 2008;Goloboff & Catalano, 2016; equal costs for all transformations, gaps treated as fifth state). The shortest trees were found using the option "Traditional Search" performing a large number of random addition sequences (RAS, usually 300-500), retaining five trees per replicate, and then submitting them to a round of tree bisection and reconnection (TBR) branch swapping. All searches were done under the collapsing option "minimum length", which collapses every node whose minimum length is 0. Parsimony jackknife absolute frequencies (Farris et al., 1996) were calculated in TNT v1.5 using traditional search requesting 100 RAS+TBR, retaining five trees per replicate for a total of 1,000 replicates. Trees were edited with FigTree v1.4.3 (Rambaut, 2016). Uncorrected pairwise distances were calculated in PAUP* (Swofford, 2002) for specimens of S. x-signatus and the related lineage S. x-signatus "Scinax_64". Most vouchers specimens of S. x-signatus were studied to corroborate their specific identity; exceptions were the specimens from sequences produced by Fouquet et al. (2007a, b) and Vacher et al. (2020) that were not available for examination (see Appendix 2).

Historical resume
Hyla x-signata was illustrated and briefly described by Spix (1824), based on a single specimen collected in "Provincia Bahiae" (part of the current State of Bahia, east of the Rio São Francisco, Brazil; see Vanzolini, 1981). The first reference to H. x-signata after the original de-scription was by Wagler (1830). In a footnote to the description of the genus Enydrobius, a replacement name for Hylodes Fitzinger, 1826 (currently considered a synonym of Hylodes), Wagler (1830) stated that H. x-signata was very similar to the two species included in this genus. Duméril & Bibron (1841) considered H. x-signata to be a synonym of H. rubra Daudin, 1802(which is H. rubra Laurenti, 1768; see León, 1969 andWiens, 1993). They based this conclusion on the variation in dorsal pattern observed in specimens identified as the latter species; some of them showed the two X-shaped marks, which they associated with the pattern described and illustrated by Spix (1824). This position was followed by Burmeister (1856) and Günther (1858). Cope (1870), instead, included H. x-signata in the genus Scytopis Cope, 1862 (currently a synonym of Trachycephalus Duellman, 1971;Faivovich et al., 2005).
However, Peters (1872) was the only herpetologist to explicitly report on the type specimen of Hyla x-signata. He stated that it was similar with H. rubra, but differed in the marbled reticulation of the anterior and posterior surfaces of thighs and the presence of granules on the dorsum, considering it a variety, H. rubra var. x-signata. Cope (1874) still considered H. x-signata as a valid species of Scytopis, distinct from H. rubra. Boulenger (1882) treated H. x-signata as a synonym of H. rubra and described the dorsal pattern with X-shaped marks as one of the two commonly occurring patterns in this species. Subsequently, the references in the literature to H. x-signata are scarce, including it as a synonym of H. rubra (e.g., Berg, 1896;Nieden, 1923), a variety (Baumann, 1912), or a subspecies (Müller, 1927) of the latter, or a related form of uncertain status (Lutz, 1951).

A neotype for Hyla x-signata
The need to designate a neotype for Hyla x-signata is well justified because it will clarify the status of this species, and of the several species to which the name has been applied throughout its wide distribution. The International Code of Zoological Nomenclature (ICZN, 1999) establishes (art. 75.1) that a neotype can be designated "when no name-bearing type specimen (…) is believed to be extant and an author considers that a name-bearing type is necessary to define the nominal taxon objectively. " Qualifying conditions established by the ICZN (1999) when describing a neotype include "the author's reasons for believing the name-bearing type specimen (s)… to be lost or destroyed, and the steps that had been taken to trace it…" (art. 75.3.4); the existence of "evidence that the neotype is consistent with what is known of the former name-bearing type from the original description and from other sources…" (art. 75.3.5); and "evidence that the neotype came as nearly as practicable from the original type locality…" (art. 75.3.6).
As for the first qualifying condition, Lutz (1973) commented that the holotype of Hyla x-signata was lost, considering it destroyed during World War II, as did Duellman (1977). This fact was corroborated by Hoogmoed & Gruber (1983) on their thorough study on the status of Spix's types, and more recently, by Glaw & Franzen (2006). A recent inquire to the Zoologischer Staatssamlung München corroborated that the holotype is still considered lost (Frank Glaw, pers. comm.).
As for the second qualifying condition, the available, direct evidence on the holotype of Hyla x-signata stems from the original description and figure by Spix (1824), and the comments on the type specimen by Peters (1872). The description of H. x-signata by Spix (1824) is not particularly informative, except for a reference to two X-shaped marks on the dorsum ("…dorsum maculis binis x-formibus signatum…"), and a curious mention of "digiti ranaeformes". Otherwise, it is similar in terms of the morphological description to other hylids described in the same book [e.g., H. nebulosa (currently Scinax nebulosus) and H. strigilata (currently S. strigilatus; see Pimenta et al., 2007)]. The figure (here reproduced as Fig. 1), however, clearly depicts the typical dorsal pattern showing a pair of inverted parentheses on the dorsum, that occurs in several species of Scinax (such as S. camposseabrai, S. eurydice, S. fuscovarius, S. granulatus, S. nasicus, some specimens of S. similis, and the populations that have been associated with S. x-signatus, from NE Brazil to Colombia).
If the description by Spix (1824) had some ambiguous reference, such as the raniform digits, the comments by Peters (1872) on the type specimen should dispel any doubt of its similarity with Hyla rubra, or at least with hylids that could be confused with that species. Peters (1872) stated that the type specimen is overall comparable with H. rubra, from which it differs by the reticulated pattern on the hidden surfaces of the thigh, and the occurrence of granules on the dorsum. He compared the type specimen with two specimens from the surroundings of Rio de Janeiro (ZMB 5922) that unfortunately could not be located in the Berlin Museum collection (Frank Tillack, pers. comm., 24 July 2020).
The comments by Wagler (1830), Müller (1927), and Cochran (1952Cochran ( , 1955 on Hyla x-signata are considered as indirect evidence since, although the first two very likely had contact with the type specimen, they made no direct reference to it. Wagler (1830)  , a comment that is certainly confusing. It could be assumed that Wagler had direct knowledge of the type specimen of Hyla x-signata given his involvement in the study of the collection amassed by Spix during his travels (Vanzolini, 1981). However, it is noticeable that Peters (1872), when studying the types of Hylodes ranoides and Rana binotata made no reference to Wagler's comment (1830) nor Hyla x-signata, that in the same paper he associated with H. rubra a few pages ahead.
Lorenz Müller (1868Müller ( -1953 was in charge of the herpetological collection of the Zoologischer Staatssamlung München. Although Müller (1927) did not explicitly state that his conclusion that Hyla x-signata is a subspecies of H. rubra was based on the type specimen of the former species, it is likely so (as also assumed by Hoogmoed & Gruber, 1983). Doris M. Cochran (1898Cochran ( -1968 visited ZSM in October 1938 (Heyer in Hoogmoed & Gruber, 1983), where she studied some of Spix's types. Unlike other cases (e.g., H. strigilata), Cochran (1952Cochran ( , 1955 did not list the type specimen in the list of specimens examined following the only mention of that species in her study -in the account of H. similis. This absence could indicate that she did not see the type specimen or that she did not consider this species to occur in the area of her study of 1955 (the Brazilian states of Minas Gerais, Rio de Janeiro, and São Paulo). In any case, she stated that H. x-signata was "apparently nearest to [Hyla] similis, after [Hyla] fuscovaria, but fresh Bahian material is needed before an exact comparison can be made." As a synthesis, with the exception of the statements by Wagler (1830) that we consider to be implicitly contested by Peters (1872), all  fig. 3). Note that the reproduced figure is from the 1839 reprint of the Ranarum section alone, housed in the MACN central library. The illustrations of the reprints were printed from the same metal plates of the first edition (Adler, 1981). There are some differences in the coloring of the figures among editions, and the one reproduced here has some subtle differences in the flanks and sides of the head. Please refer to the digital version of the 1824 edition in Biodiversity Heritage Library (https://www.biodiversitylibrary.org/ item/21828#page/119/mode/1up).
the available evidence indicates that the type specimen of H. x-signata corresponded to a hylid similar to Scinax ruber and other species of Scinax, and it had a dorsal pattern with two pairs of inverted parentheses, the typical X-shaped markings.
Regarding the third qualifying condition, Spix (1824) stated the type locality as "Provincia Bahiae", an expression that, according to Vanzolini (1981), refers to the ancient limits of the State of Bahia, which correspond to the current limits, east of the river São Francisco. The itinerary followed by Spix and Martius in Bahia is well known (Papavero, 1971;Vanzolini, 1981), having entered the limits of that province from the west, and covered several localities from Malhada to Salvador (e.g., Caetité, Rio de Contas, Maracás, Santa Teresinha, São Félix, Salvador), and Ilhéus and surrounding localities (e.g., Camamu, Itabuna, Itacaré). We designate a neotype for Hyla x-signata collected in Ilhéus. Furthermore, we have specimens of the same species collected in some other localities visited by Spix.
Our study of photographs of the holotype (ZSM 2945) indicates that the situation is uncertain. Only a study of the taxonomy of amazonian populations associated with Scinax x-signatus would allow to clarify the status of this nomen. Hyla coerulea Spix, 1824 -Considered a synonym of Hyla x-signata by Hoogmoed & Gruber (1983). Sturaro & Peloso (2014) questioned this association based on the description and figure provided by Spix (1824).
Our study of photographs of the lectotype designated by Hoogmoed & Gruber (1983) (ZSM 2710-0-1) indicates that the situation is uncertain. Only a study of the taxonomy of amazonian populations associated with Scinax x-signatus would allow to clarify the status of this nomen. Hyla rubra Daudin, 1802 First treatment as a synonym of Hyla rubra Laurenti, 1768(not Daudin, 1802; see León, 1969;Rivero, 1969;Duellman & Wiens, 1993

Diagnosis (based on neotype and referred specimens)
Scinax x-signatus is a species of Scinax, as it shares three synapomorphies of this genus: webbing between toes I and II that does not extend beyond the subarticular tubercle of toe I; origin of the m. pectoralis abdominalis through well-defined tendons; and m. pectoralis abdominalis overlapping m. obliquus externus (da Silva, 1998;Faivovich, 2002;Faivovich et al., 2005). A single synapomorphy is known for the S. ruber Clade: tadpoles with the vent tube above the margin of the lower fin (Faivovich, 2002;Faivovich et al., 2005). While tadpoles unequivocally associated to S. x-signatus remain unknown, this species was associated to the S. ruber Clade by having the unique combination of external vocal sac and presence of pectoral fold [internal vocal sac and pectoral fold absent in most species of the S. catharinae Clade; in few species where the vocal sac is external, the pectoral fold is absent (e.g., S. garibaldiae, S. rizibilis); otherwise, in the two cases where the pectoral fold is present, the vocal sac is internal (S. agilis and S. melanodactylus); J. Faivovich & K. Araujo-Vieira, pers. obs.; see also Bokermann, 1964;Cruz & Peixoto, 1982;Faivovich, 2002;Lourenço et al., 2014Lourenço et al., , 2019.
Scinax x-signatus can be differentiated from all other species of the S. ruber Clade by the combination of the following characters: (1) male SVL 34.5-38.4 mm, n = 13; (2) snout rounded in dorsal view and profile; (3) pointed tubercles on lower jaw absent; (4) vocal sac subgular, weakly bilobate; (5) spicule-shaped papillary epidermal projections on the nuptial pads and pectoral region present in males; (6) pectoral glands present in males; (7) dorsal color pattern with large irregular dark blotches, commonly with dark X-shaped mark composed of one or two pairs of inverted parenthesis-like blotches; (8) hidden surfaces of thighs dark with irregular pale blotches, yellow in living specimens; (9) iris yellowish golden or bronze with a median black streak; (10) physiological chlorosis absent; and (11) advertisement call composed of a single, multipulsed note, 0.11-0.25 s duration, 6-14 pulses/note, 52-64 pulses/s.
The snout rounded in dorsal view and profile differentiates Scinax x-signatus from S. alter, S. auratus, S. cretatus, S. crospedospilus, S. imbegue, S. juncae, and S. tymbamirim (sub-elliptical with a pointed tip in dorsal view and slightly acute in profile), S. fuscovarius (roundly acute in dorsal view and protruding in profile), S. caldarum, S. curicica, S. duartei, S. maracaya, S. rossaferesae, and S. tigrinus (sub-elliptical or subovoid in dorsal view and slightly acute in profile), S. squalirostris (pointed in dorsal view and acute in profile), and species of the S. rostratus group (elongate pointed in dorsal view and acute with or without a fleshy proboscis in profile). Furthermore, the absence of pointed tubercles on the lower jaw differentiates S. x-signatus from almost all species of the S. rostratus Group; exceptions are S. kennedyi and S. rostratus (e.g., Duellman, 1972aDuellman, , 1973Pyburn, 1973;Lescure & Marty, 2000;Lima et al., 2005;this study).
The presence of spicule-shaped papillary epidermal projections on the nuptial pad and pectoral region in males differentiates Scinax x-signatus from all other species of the S. ruber Clade, except for S. fuscovarius (see also   fig. 10A, C). The presence of pectoral glands in males differentiates S. x-signatus from most species of the S. ruber Clade, except for S. funereus, S. fuscovarius, S. nasicus, S. onca, and S. similis, and species of the S. uruguayus Group (e.g., Müller & Hellmich, 1936;Lutz, 1973;Cei, 1980;this study).

Description of the neotype
Head as wide as long, HL 35.1% and HW 33.2% of SVL (Fig. 2). Snout rounded in dorsal view and profile, with a low protuberance on the tip (Fig. 3A, B). Nostrils dorsolateral, elliptical, protruded; IND 39.7% of IOD. Canthus rostralis marked, convex. Loreal region slightly concave. Eyes large, protuberant, ED 94.7% of IOD and 92.3% of END. Pupil horizontal, subelliptical. Tympanum rounded, separated from eye by a distance almost half TD; TD 75.0% of ED. Tympanic annulus rounded, with the posterior upper portion hidden by the supratympanic fold. Supratympanic fold evident, from the posterior upper portion of the tympanum to the insertion of the forearm. Vocal sac subgular, weakly bilobate, externally evident by the loose skin, not occupying space between head and body, and ventrally not reaching the pectoral fold (Fig. 2B). Pectoral fold present, with pre-and postaxillar elements. Vocal slits present, nearly parallel to the mandible, originating laterally to the tongue and running towards the corner of the mouth. Tongue ovoid, free laterally and posteriorly, slightly notched posteriorly. Vomerine teeth in two slightly separated convex series, bearing five (right) and four (left) teeth. Choanae oval.
Axillary membrane absent. Upper arm more slender than forearm. A series of small, flat, ulnar tubercles on the forearm. Fingers short and slender, fringed (Fig. 3C). Relative finger length II<III≈V<IV. Discs elliptical, wider than long, 3FD 59.2% of TD; disc of Finger II smaller than others. Subarticular tubercles single, conical on fingers II and III; rounded and quadrangular on fingers IV and V. Supernumerary tubercles small, single, rounded. Inner metacarpal tubercle single, elliptical; outer metacarpal tubercle flat, nearly triangular, bilobate. Webbing absent between fingers II and III; basal between fingers III, IV, and V. Slightly thickened, light-colored nuptial pad, covering Metacarpal II dorsomedially, and ventrally extending from the base of inner metacarpal tubercle, obscuring its outer margin, to the subarticular tubercle (Fig. 4A, B). Glandular acini on inner margin of fingers II-III; on Finger II from the distal margin of nuptial pad to the disc. Spicule-shaped papillary epidermal projections on the nuptial pad, margins of fingers II-III (Fig. 4A, B), and dorsum of fingers II-V.   3D). Relative toe length I<II<III≈V<IV. Discs elliptical, wider than long, slightly smaller than discs of fingers, 4TD = 3FD. Subarticular tubercles single, conical, rounded; supernumerary tubercles small, single, rounded. Inner metatarsal tubercle single, elliptical; outer metatarsal tubercle single, slightly marked, two thirds smaller than inner tubercle. Webbing formula I 2-2⁺ II 1¹ / ⁴-2⁺ III 1-2¹ / ³ IV 2⁺-1 V. Fringe on lateral margin of Toe V extends along the margin of the sole by a poorly developed ridge that reaches the distal portion of the metatarsus. Ventrolateral margin of tarsus smooth; heel slightly granular.
Cloacal opening directed posteriorly at upper level of thighs. Skin on dorsum of head, upper eyelid, trunk, and limbs smooth, with scattered granules. Posterior corner of eyes, around tympanum and forearm insertion, supratympanic fold, and flanks granular. Vocal sac, hidden surfaces of limbs, and inguinal region smooth; other ventral surfaces and subcloacal area densely covered with rounded, flat granules. Pectoral region and inner margin of upper-and forelimbs with glandular acini, covered with spicule-shaped papillary epidermal projections.

Coloration in life:
The description is based on the freshly euthanized specimen (Fig. 5). Dorsal color dark brown, with two pairs of large, irregular, black blotches on the suprascapular and sacral regions, and scattered, small, round or irregular, light blotches; interocular region with an inverted triangle-shaped, black marking (Fig. 5A). Upper lip light with diffuse brown blotches anteriorly, and a white stripe on the infraorbital region extending to posterior margin of the tympanum. Loreal region brown with small, irregular, black dots; dark brown canthal line. Post-orbital dark brown line from anterior corner of the eyes, upper margin of tympanum, to the middle of the flanks. Flanks light with irregular, dark brown blotches. Dorsal surfaces of discs, fingers, toes, forearms, and tarsus brownish gray with transverse, brown bars; upper arms uniform; shanks and thighs with large dark brown blotches. Toe webbing covered by brown melanophores. Iris grayish bronze with thin black reticulations, thin yellow halo bordering the pupil, and a median black streak.
Soles and palms light brown; glandular pectoral region yellowish white; other ventral areas creamy white, immaculate, but margins of gular region, around forearm insertion and knees, tarsus, and shanks finely spotted with brown (Fig. 5B). Inguinal region yellow, with irregular dark blotches. Hidden surfaces of thighs and shanks brown, with small to medium-size, rounded or irregular yellow blotches (Fig. 5C). White bones.

Coloration in preservative:
Paler than in the fresh specimen. The coloration on the glandular pectoral region, inguinal region, and hidden surfaces of thighs and shanks faded to light beige or cream white.
Dorsal skin similar to that of neotype, with scattered or uniformly distributed granules. Ulnar and tarsal tubercles protuberant or inconspicuous. Three or four, low, rounded tubercles can be present next to the tibio-tarsal articulation; the distal one is elongated in some individuals.
In freshly euthanized specimens, dorsal color varies from beige or gray to dark brown, with large irregular brown to black blotches, and small white blotches (Fig. 6). In males, pectoral region light yellow and abdo-  Araujo-Vieira, K. et al.: Neotype for Hyla x-signata Pap. Avulsos Zool., 2020;v.60: e20206056 11/30 men yellowish beige; in females, cream white. Inguinal region light yellow or yellow, with irregular dark blotches; hidden surfaces of thighs, shanks, and tarsi brown to black, with small to large, rounded or irregular yellow blotches. Flanks light yellow or cream white; axillae yellow in some individuals. Iris yellowish golden or bronze, with many thin, dark reticulations, and a median black streak. In life, overall coloration similar to fleshly euthanized specimens. Still, dark and light tones are more intense and brighter, especially yellow coloration on inguinal region and hidden surfaces of hindlimbs (Fig. 7). Iris iridescent yellowish golden or bronze (Fig. 7A-E).
In preservative, dorsal pattern varies from light beige or light gray to dark brown, with large, irregular dark blotches, usually with X-shaped marks composed of one or two pairs of inverted parentheses-like blotches, with or without light blotches (Fig. 8). Small, dark brown, irregular blotches on the dorsum of all individuals. Interocular marking can be inverted triangle, T-shaped, or W-shaped, sometimes fragmented. Upper lip light with diffuse brown blotches, sometimes with a white infraorbital stripe that extends to the tympanum. Dark canthal line present in all specimens. Post-orbital line varies in extension, reaching posteriorly level of forearm insertion or middle of flanks. Dark blotches on flanks and inguinal region rounded or irregular. Small to large, rounded or irregular light blotches on hidden surfaces of hindlimbs. Ventral surfaces from cream white to light beige, finely or conspicuously covered with brown spots in some individuals. Dark coloration predominates on dorsum of specimen MHNJCH 1014 (Fig. 8L); however, this resulted from the fixation process in 10% formalin. In life, this specimen showed the X-shaped mark and other dorsal blotches common to the other specimens.
The occurrence of glandular tissue in the pectoral region has been considered a secondary sexually dimorphic character occurring in males (Lutz, 1973), and this is corroborated here (Fig. 9A). This is also the case of the spicule-shaped papillary epidermal projections on the nuptial pad (Fig. 9B, C), inner margin of upper-and forearms, and pectoral region. The glandular areas (acini) on the inner margins of upper-and forearms, and fingers II-III (excluding the nuptial pad) are absent in some specimens (e.g., MHNJCH 1014, 1698-1700); when present, spicule-shaped papillary epidermal projections also occur on these areas. Spicule-shaped projections can be  Araujo-Vieira, K. et al.: Neotype for Hyla x-signata Pap. Avulsos Zool., 2020;v.60: e20206056 13/30 present, scattered, and apparently not associated to acini on fingers II-V in some individuals. Although our sample of females is small (three individuals), females tend to be larger than males (Table 1).

Advertisement call
The advertisement call of Scinax x-signatus consists of a single multipulsed note, emitted at a highly variable repetition rate (2-105 notes/minute; Table 2; Fig. 10A), which is likely affected by conspecific chorus density, since the longer intervals between notes (up to 25.3 s; Table 2) were observed in the recording of the neotype, which was calling alone with no nearby conspecific. Notwithstanding, much longer intervals are the exception, and the notes are repeated at faster rates (29-105 notes/minute; Table 2), but never composing a stereotyped series.
Note duration is 0.11-0.25 s (Table 2; Fig. 10A-C); each note is composed of 6-14 pulses with modulating amplitude increasing from the first third, reaching the maximum amplitude around the middle of the note, and gradually decaying towards the last pulse (Fig. 10B, C). Pulse rate is 52-64 pulses/s and pulse duration 0.013-0.018 s (Fig. 10B, C).
Calls have a broadband spectrum (BW90% 2250-2799 Hz; Table 2; Fig. 10C). The power spectrum is usually biphasic, with two main emphasized frequency bands (Fig. 10C). The lower band (i.e., the low-frequency band) comprises most of the power of the spectrum, with the dominant frequency ranging between 904-1359 Hz (Fig. 10C), and also including part of the freq5% (861-991 Hz; Table 2). The upper band (i.e., the high-frequency band) has less power than the first one, surrounding the freq95% (3188-3704 Hz; Table 2   its peak frequency between 2885-3618 Hz (Fig. 10C). Between the two power spectrum bands, there is a low-power "valley" (or gap) around 2.0-2.5 kHz (Fig. 10C).
The dominant frequency does not alternate between the lower and upper bands, remaining in the lower band.

Notes on calling site and calling behavior
Males of Scinax x-signatus call near lentic water bodies, either natural or artificial (such as pools and tanks). They usually call from the ground, either uncovered or hidden among the vegetation. Less often, they call perched at low heights (below 1.5 m; rarely above that height) on the vegetation inside or at the margins of water bodies. Other species of Scinax found calling syntopically with S. x-signatus are Scinax sp. aff. hayii, S. auratus, S. eurydice, and S. pachycrus. Scinax x-signatus seems to tolerate some degree of light and sound disturbance; as we recorded the neotype inside the University Campus, with plenty of artificial light and crowd voices as background noise.

Phylogenetic analysis
The two most parsimonious trees (length 468) recovered all specimens considered in the literature as Scinax x-signatus closely related with the neotype and our referred specimens (100% jackknife; Fig. 11). These include specimens from Bahia, Ceará, and Pernambuco (NE Brazil) and Amapá (N Brazil), and from French Guiana and Martinique (Fig. 12). The selected fragment of the mitochondrial ribosomal gene 16S rRNA showed uncorrected pairwise distances of 0.2-1.7% among the 28 individuals of S. x-signatus (Table 3). The maximum value (1.7%) is between specimens from Kourou and Ile Royale (French Guiana), and those from Ubajara (Ceará, NE Brazil), Fernando de Noronha, and Sanharó (Pernambuco, NE Brazil); the geographic distances between these points are approx. 1,700 km (see distances between points 8-9 and 14-15 in Fig. 12).
Scinax x-signatus is moderately supported (85% jackknife) as sister taxon of Scinax sp. (as S. x-signatus "Scinax_64" in Vacher et al., 2020) from Parque Nacional Cavernas do Peruaçu, Januária, N Minas Gerais, Brazil. Uncorrected pairwise distances between S. x-signatus and Scinax sp. are 6.9-10.2%, with a sequence divergence of 8.0% between one specimen of S. x-signatus (UFMG 4787) from Sebastião Laranjeiras (Bahia, NE Brazil) only distant approx. 170 km ENE from the locality of this candidate species in N Minas Gerais (see Fig. 12). The voucher specimen of Scinax sp. (MTJ0578) was not available for morphological study, and therefore we are not aware of any diagnostic characters for this candidate species. The Figure 11. One of the two most parsimonious trees (length 468 steps) obtained from the analysis of the 16S rRNA mitochondrial gene in TNT with gaps as fifth state. Values above or below nodes are jackknife support values. Nodes without values indicate < 50% jackknife support; black dot indicates a node that collapses in the strict consensus; an asterisk (*) indicates groups with 100% jackknife support.

DISCUSSION
Scinax x-signatus has been considered to occur in Colombia, Venezuela, Guyana, Suriname, French Guiana, and extensive areas in Brazil, including also oceanic islands such as Fernando de Noronha (introduced), Guadeloupe, La Désirade, Margarita, Marie Galante, and Martinique (e.g., Lescure & Marty, 2000;Juncá, 2006;Henderson & Powell, 2009;Ugueto & Rivas-Fuenmayor, 2010;Ouboter & Jairam, 2012;Cole et al., 2013;Novaes & Zina, 2016;Barrio-Amorós et al., 2019). As reviewed earlier in this paper, several authors expressed concerns regarding the taxonomy of this species and the possibility that the name had been applied to several different species. Our designation of a neotype, and the analysis of DNA sequences allows to conclude that S. x-signatus occurs at least in Northeastern and Northern Brazil, French Guiana, and the Islands of Fernando de Noronha (Brazil) and Martinique, as previously reported (e.g., Fouquet et al., 2007a, b;Lyra et al., 2016;Novaes & Zina, 2016;Vacher et al., 2020;see Figs. 11, 12). However, our analysis lacks samples from Colombia, Venezuela, Suriname, Guyana, and many northeastern Brazilian states where several populations have been assigned to S. x-signatus. Our characterization of the neotype specimen and other referred specimens from some localities in the State of Bahia, Brazil, should be seen only as a partial characterization of this species. A more thorough study is now necessary, to understand its geographic variation along its extensive distribution.
The combination of several adult morphological and advertisement call characters differentiates Scinax x-signatus from the remaining 74 described species of the S. ruber Clade (see the Diagnosis section). However, two sexually dimorphic morphological characters present in males deserve comments, the pectoral glands, and the  -Vieira, K. et al.: Neotype for Hyla x-signata Pap. Avulsos Zool., 2020;v.60: e20206056 20/30 spicule-shaped papillary epidermal projections on the nuptial pad, inner margin of upper-and forearms, and pectoral region.
A white, slightly thickened area on the pectoral region was observed in males of some species of the Scinax ruber Clade (e.g., S. eurydice, S. haddadorum, and S. pachycrus; Araujo-Vieira et al., 2016;this study). However, we did not observe the presence of packed acini in this region with high magnification, and therefore considered that these species lack pectoral glands. Histological studies are necessary to corroborate our observations. Otherwise, glandular acini in the pectoral region were reported for S. goya (the S. catharinae Clade; Andrade et al., 2018). These acini seem to be scattered throughout this region and likely not form a delimited area; it is unclear if these acini are sexually dimorphic in S. goya, since Andrade et al. (2018) did not mention them as absent in females.
Spicule-shaped papillary epidermal projections on nuptial pads were first described for Scinax fuscovarius by , who find it similar to those previously described for the pipids Xenopus epitropicalis and X. tropicalis (Dolder, 1976;Evans et al., 2015). Our observations on several species of the S. ruber Clade showed the presence of these projections only on the nuptial pads of S. fuscovarius and S. x-signatus (see Fig. 9B, C; observations only missing for S. baumgardneri, S. blairi, S. castroviejoi, S. karenanneae, S. lindsayi, S. ruberoculatus, S. sateremawe, and S. strussmannae). These spicule-shaped projections are also present on the inner margins of upper-and forearms and fingers II-III, and in the pectoral region, always associated with glandular acini, but they might also be present on dorsum and outer margins of fingers II-V, where they are apparently not associated with acini. This distribution on the body suggests that further research is necessary to understand if these spicules represent sexually dimorphic characters that cover different parts including the nuptial pad, in which case they should not be considered as an intrinsic morphology of this structure (i.e., one of the many morphologies of papillary epidermal projections of the nuptial pad recognized by . The advertisement call parameters of Scinax x-signatus (note duration, dominant frequency, pulse duration, number of pulses/note, and pulse rate; see Table 2) from our recordings and those produced by Novaes & Zina (2016) mostly overlap the values from recordings from Venezuela (Rivero, 1969;Tárano, 2010). Rivero (1969) attributed two different calls to S. x-signatus with dif-ferent pulse rate: 40 and 80 pulses/s. The spectrograms from the call with pulse rate of 40 pulses/s (Rivero, 1969: figs. 2, 6) is similar to those of S. x-signatus described here (Fig. 10), whereas those from the call with pulse rate of 80 pulses/s are more similar to calls attributed to S. ruber by Rivero (1969;compare fig. 1 and fig. 3), as also noticed by De la Riva et al. (1994) and Novaes & Zina (2016). Our call values for S. x-signatus overlap, in part, with those provided by Sichieri et al. (2019: fig. 2) for specimens from Bahia (e.g., note duration 0.10-0.34 vs. 0.11-0.25 s in this study, number of pulses 5-27 vs. 6-14, pulses rate 32-90 vs. 52-64 pulses/s, and dominant frequency 1034-3790 vs. 904-1359 Hz). Otherwise, the advertisement call of S. x-signatus reported by Freitas & Toledo (2020: fig. 2E, table 1) is different from those reported here in note duration (0.13-2.43 vs. 0.11-0.25 s in this study), number of pulses (5-104 vs. 6-14 pulses in this study), and dominant frequency (950-4050 vs. 904-1359 Hz in this study). These differences likely result from calls of different species that were analyzed under the name S. x-signatus, as also suggested by Freitas & Toledo (2020: 9).
The advertisement call of Scinax x-signatus is similar to that of S. fuscovarius, in that both produce low frequency calls, with similar pulse structure, note duration, number of pulses per note, and pulse rate, and there are no obvious differences between the advertisement calls of these species (Pombal et al., 1995b;Bevier et al., 2008;Novaes & Zina, 2016; this study). These species also share the presence of pectoral glands and spicule-shaped epidermal projections in males, as mentioned above.

AUTHORS' CONTRIBUTIONS
Katyuscia Araujo-Vieira and Julián Faivovich conceived and designed the study, carried out the sequencing, analyzed the data, wrote the original draft, revised and edited the final document, prepared figures and/or tables. José P. Pombal Jr. and Ulisses Caramaschi analyzed the data and revised the original draft. Gabriel Novaes-e-Fagundes collected specimens, recorded calls analyzed the data, revised the original draft, and prepared figures. Victor G.D. Orrico collected specimens, analyzed the data, and revised the original draft.
List of specimens, collection numbers, localities, and GenBank accessions of sequences. The neotype of Scinax x-signatus is underlined. New sequences produced for this study are in bold.