We describe through integrative taxonomy a new species of snouted treefrog of the genus Scinax from white-sand forests of the Rio Negro Sustainable Development Reserve in Central Amazonia, Brazil. The new species is phylogenetically related to other Scinax with striped eyes and pulsed advertisement calls. It differs from other Amazonian species mainly by having snout–vent length 21.6–25.4 mm (n= 11) in adult males and 24.8–27.0 mm (n= 9) in females, snout subacuminate in dorsal view, a dark brown lateral stripe on each flank (fading posteriorly), brown tadpoles with labial keratodont row formula 2(2)/3 and keratodont row P-2 longer than P-1 and P-3, and an advertisement call consisting of a single pulsed note with a call duration of 502–652 ms, 79–105 pulses/note and a dominant frequency of 3,811–4,543 Hz. The new species clusters within a major, well-supported phylogenetic clade grouping several candidate and recently described species as well as species previously included in the former Scinax staufferi species group (viz., S. cruentomma, S. fuscomarginatus, S. staufferi, and S. wandae). The phylogenetic relationships and structural pattern in the advertisement calls of these species highlight the need for a redefinition and reevaluation of the monophyly of the S. staufferi species group.
INTRODUCTION
Diversity and taxonomy of the Amazonian snouted treefrogs of the genus Scinax Wagler, 1830 have attracted the attention of biologists for decades (e.g., Duellman, 1972a; Fouquette and Delahoussaye, 1977; Duellman and Wiens, 1993; Faivovich, 2002; Fouquet et al., 2007; Ferrão et al., 2016; Acosta-Galvis, 2018; Lopes et al., 2020). Nevertheless, results of the most recent studies show that species richness of the genus remains surprisingly underestimated (Vacher et al., 2020).
Analysis of the regional diversity of Scinax in central-to-southwestern Amazonia revealed the existence of at least seven molecularly and phenotypically distinct clades, each representing a candidate species (Ferrão et al., 2016). Later, three of these clades were described as new species (Ferrão et al., 2017, 2018a, b), including S. strussmannae Ferrão, Moravec, Kaefer, Fraga and Lima, 2018b which is closely related to S. cruentomma (Duellman, 1972b) and S. wandae (Pyburn and Fouquette, 1971). Two other candidate species related to the latter two species await formal description. In addition, a recent wide-scale DNA barcoding study documented seven additional candidate species related to S. cruentomma and S. wandae (Vacher et al., 2020). These findings indicate that taxonomic knowledge of this group of small-sized Scinax is still incomplete and that underexplored Amazonian regions harbor a very rich and mostly unknown anuran diversity.
White-sand ecosystems (hereafter, WSE) of Central Amazonia are a good example of poorly explored environments (Adeney et al., 2016). The WSE represent patchily distributed habitats with nutrient-poor sandy soils covered by grasslands, shrubs, or forests with low-stature canopies (Eiten, 1978). Usually, these habitats are embedded in dense forests with nutrient-rich soils (Eiten, 1978). Although WSE cover a small portion (approximately 5%) of the Amazon territory, the species assemblages occupying them are unique (Adeney et al., 2016; Costa et al., 2020). Rapid biological inventories of WSE have documented new species in various taxonomic groups (e.g., Vriesendorp et al., 2006; Cohn-Haft et al., 2013). More recently, several endemic or specialist WSE species have been reported for plants (Fine and Baraloto, 2016; Vicentini, 2016), insects (Lamarre et al., 2016), birds (Capurucho et al., 2013; Borges et al., 2016), snakes (Fraga et al., 2018) and frogs (Rojas-Zamora et al., 2015; Carvalho et al., 2018; Ferrão et al., 2019).
The largest regions of WSE lie in the Negro-Solimões interfluve—one of the most inaccessible and consequently unexplored parts of Brazilian Amazonia. Aiming to improve biological knowledge of WSE in the Negro-Solimões interfluve, the Brazilian Research Program on Biodiversity (PPBio - INPA) installed a long-term sampling module in the Rio Negro Sustainable Development Reserve (RDS Rio Negro), which crosses a heterogeneous landscape covered by a mosaic of dense forests and WSE patches. Recent herpetological surveys conducted along this sampling module by our research team resulted in the rediscovery of the WSE treefrog Osteocephalus vilarsi (Melin, 1941) (Ferrão et al., 2019), the recorded association of the gecko Hemidactylus palaichthus Kluge, 1969 with the bromeliad Aechmea huebneri (Ferreira et al., 2019), and the discovery of candidate species belonging to several anuran genera including Scinax (M. Ferrão and A. P. Lima, personal communication).
Herein, using an integrative taxonomy approach, we describe a new species of Scinax from these WSE patches which has morphological affinities to S. cruentomma and S. wandae.
MATERIALS AND METHODS
Sampling
Specimens of the new species were collected between June 2018 and April 2020 in the PPBio sampling module of the RDS Rio Negro (03°03′31″S, 60°45′42″W; 73 m a.s.l.), municipality of Novo Airão, state of Amazonas, Brazil (Fig. 1). Adults were euthanized with 2% benzocaine topical solution, fixed in 10% neutral-buffered formalin, and preserved in 70% ethanol. Tissue samples were collected before fixation and stored in 100% ethanol. Seventeen tadpoles were euthanized with 2% benzocaine liquid solution (lot INPA-H 42855), while six others (lot INPA-H 42854) were reared in the laboratory until approximately stage 45 (Gosner, 1960). Both adults and tadpoles were deposited in the Amphibians and Reptiles Collection of the Instituto Nacional de Pesquisas da Amazônia, Manaus, Brazil (INPA-H).
Figure 1.
Geographic distribution of Scinax albertinae sp. nov. and closely related species with a red- or brown-striped iris based on samples included in our molecular analyses (Fig. 2). Country acronyms: GUF, French Guiana; GUY, Guyana; SUR, Suriname.
Calls emitted by three adult males (INPA-H 42859, 42872, and an unvouchered individual) were recorded at the RDS Rio Negro sampling module at 1830 h on 25 April 2020. Calls were recorded with a Sennheiser K6/ME66 unidirectional microphone (Sennheiser, Germany) coupled to a Marantz PMD660 digital recorder (sampling rate 44.1 kHz, sample size 16 bits; Marantz, Japan). The microphone was positioned approximately 50 cm from each individual. Air temperature during recording was 25°C. Files were stored in WAV format and deposited in the Fonoteca Neotropical Jacques Vielliard (FNJV 50546–53) of the University of Campinas (UNICAMP), Campinas, Brazil.
Molecular phylogenetics
We extracted genomic DNA of four specimens of the new species by using a Wizard® Genomic DNA Purification Kit (Promega Corp., Madison, WI, USA) according to the manufacturer′s protocol. A fragment of the 16S rRNA mitochondrial gene was amplified through polymerase chain reaction (PCR) using 16Sar and 16Sbr primers (Palumbi, 1991) and following protocols described in Ferrão et al. (2016). PCR products were purified using Exonuclease I and Thermosensitive Alkaline Phosphatase (Thermo Fisher Scientific, Waltham, MA, USA) and submitted to sequencing reactions using standard protocols of the Big DyeTM Terminator Kit (Applied Biosystems, Waltham, MA, USA). Amplicons were sequenced in an ABI 3730XL (Macrogen Inc., Seoul, Korea) using the forward primer. The software Bioedit (Hall, 1999) was used to edit the sequences, which ranged from 552 to 624 base pairs (bp). Newly generated sequences were deposited in the online repository GenBank under accession numbers MW853693–MW853696.
To infer phylogenetic relationships of the new species, the above newly generated sequences were inserted into a data set containing homologous sequences retrieved from GenBank. These sequences represent 35 nominal and 16 candidate species of Scinax, covering all major species groups delimited in the taxonomic history of this genus (Faivovich, 2002; Faivovich et al., 2005). Nine other sequences of six species from the two most closely related genera Ololygon Fitzinger, 1843 and Julianus Duellman, Marion and Hedges, 2016 were included; a sequence of Sphaenorhynchus surdus (Cochran, 1953) was used as outgroup. Sequences retrieved from GenBank (Appendix 1) were published by Araujo-Vieira et al. (2019), Bell et al. (2012), Brusquetti et al. (2014), Faivovich et al. (2005), Ferrão et al. (2016), Fouquet et al. (2007), Frost et al. (2006), Guarnizo et al. (2015), Jansen et al. (2011), Lyra et al. (2017), Ron et al. (2018), Salducci et al. (2005), Vacher et al. (2020) and von May et al. (2019).
Sequences were aligned using the MAFFT online server with default parameters, except for the use of E-INS-i strategy (Katoh and Standley, 2013). The final matrix was composed of 146 terminals with 501 bp each. A comparison of Bayesian Information Criterion (BIC) values in a PartitionFinder (Lanfear et al., 2017) analysis indicated GTR + G + I as the most adjusted evolutionary model for this data set. Phylogenetic relationships were reconstructed using this model under Maximum Likelihood (ML) and Bayesian (BI) optimality criteria. The ML tree was inferred using IQTREE (Nguyen et al., 2015) as implemented in the webserver http://iqtree.cibiv.univie.ac.at (Trifinopoulos et al., 2016). Clade support was estimated with 10,000 ultrafast bootstrap replications (Hoang et al., 2018), 1,000 maximum iterations, and a minimum correlation coefficient of 0.99. The BI tree was inferred in MrBayes 3.2.6 (Ronquist et al., 2011) using four runs of 10 million generations with a Markov chain Monte Carlo algorithm. Each run had four Markov chains, with probabilities sampled every 1,000 generations. The BI analysis was implemented in the CIPRES online server https://www.phylo.org/. Stationarity of the posterior distributions and effective sample size (ESS >200) were examined in Tracer 1.7 (Rambaut et al., 2018) and the maximum clade credibility tree was extracted after discarding the first 25% of trees.
Pairwise genetic distances (uncorrected p-distance and Kimura-two-parameter distance; Kimura, 1980) between the new species and other Scinax species with a red- or brown-striped iris were calculated in MEGA 6 (Tamura et al., 2013). Intraspecific p-distances were also calculated for the new species and its closest relatives. Genetic distances were calculated using the method “pairwise deletion″.
Morphology
Gender of adult specimens was determined according to presence (males) or absence (females) of vocal slits. Sixteen morphometric measurements were taken following Duellman (1970; horizontal eye diameter [ED], foot length [FL], head length [HL], head width [HW], internarial distance [IND], interorbital distance [IOD], snout–vent length [SVL], horizontal tympanum diameter [TD], tibia length [TL], and upper eyelid width [UEW]), Heyer et al. (1990; hand length [HAL], tarsus length [TAL], and thigh length [THL]), and Napoli (2005; eye–nostril distance [END], finger III disc diameter [3FD], and toe IV disc diameter [4TD]). These measurements were obtained to the nearest 0.1 mm using a digital caliper and an ocular micrometer coupled to a stereomicroscope. Snout shape terminology followed the definitions of Heyer et al. (1990). Webbing formulae followed Savage and Heyer (1967) as modified by Myers and Duellman (1982). Descriptions of coloration patterns in life were based on photographs and field notes. Descriptions of adult morphology are formatted as in Ferrão et al. (2018b).
Larval developmental stages of 17 tadpoles (lot INPA-H 42855) were determined according to the table of Gosner (1960) and ranged from stages 34 to 38. Morphological description was based on 10 tadpoles at stage 37. Eighteen morphometric measurements were taken following Altig and McDiarmid (1999; body length [BL], internarial distance [IND], interorbital distance [IOD], maximum tail height [MTH], tail length [TAL], tail muscle height at body–tail junction [TMH], tail muscle width at the same level as TMH [TMW], and total length [TTL]), Haas and Das (2011; body height [BH], maximum body width [BW], eye diameter [ED], eye–nostril distance [END], head width at the level of spiracle [HWLE], nostril–snout distance [NSD], oral disc width [ODW], distance of snout to center of spiracle [SS]), and Randrianiaina et al. (2011; ventral tube length [VTL] and spiracle length [STL]). Measurements were obtained to the nearest 0.1 mm with an ocular micrometer coupled to a stereomicroscope. Marginal papillae row (MPRL) and labial keratodont row (LKRF) formulae, as well as the formatting of tadpole description, followed Schulze et al. (2015). Descriptions of coloration patterns in life of tadpoles and metamorphs were based on specimens raised in the laboratory.
Vocalization
Two types of calls of an unvouchered and two vouchered (INPA-H 42859 and 42872) males were analyzed. The most commonly emitted vocalization was classified as advertisement, whereas the sporadic one was classified as territorial call because it was mainly emitted by spatially closest males and spatially distant males only sporadically emit this call type (sensu Toledo et al., 2015). The following 12 acoustic parameters were measured from advertisement calls: call duration (CD); intercall interval (ICI); call period (CP); call repetition rate (CR); number of pulses per call (PN); pulse duration (PD); interpulse interval (IPI); pulse period (PP); pulse rate (PRR); and low (LF), high (HF), and dominant frequency (DF). We also analyzed the second type of vocalization emitted by the same three males. The following parameters of territorial calls were measured: call duration (CDA); intercall interval (ICIA); call period (CPA); call rate (CRA); and low (LFA), high (HFA), and dominant (DFA) frequencies. Acoustic parameters were measured following the recommendations of Köhler et al. (2017).
Calls were analyzed using RAVEN 1.5 (Bioacoustics Research Program, 2014) configured as follows: Blackman window; 3 dB Filter Bandwidth, 80 Hz; overlap, 80%; hop size, 4.1 ms; and Discrete Fourier Transform size, 1,024 samples. Call repetition rate was calculated as 1 min divided by the call period, and pulse repetition rate as 1 sec divided by the pulse period. Dominant frequency was measured using the peak frequency function. Low and high frequencies were measured 20 dB below the peak frequency to avoid overlap with background noise. Call terminology follows Köhler et al. (2017) and advertisement call description follows Ferrão et al. (2018b). Graphic representations of calls were generated in the Program R environment (R Core Team, 2019) through the seewave package 2.0.5 (Sueur et al., 2008) using a Hanning window, 256 points of resolution (Fast Fourier Transform) and an overlap of 85%.
Interspecific comparisons
We compare the new species with all nominal species of Scinax distributed throughout Amazonia (Appendix 2): S. baumgardneri (Rivero, 1961); S. blairi (Fouquette and Pyburn, 1972); S. boesemani (Goin, 1966); S. chiquitanus (De la Riva, 1990); S. cruentomma; S. danae (Duellman, 1986); S. exiguus (Duellman, 1986); S. funereus (Cope, 1874); S. fuscomarginatus (Lutz, 1925a); S. fuscovarius (Lutz, 1925b); S. garbei (Miranda-Ribeiro, 1926); S. ictericus Duellman and Wiens, 1993; S. iquitorum Moravec, Tuanama, Pérez and Lehr, 2009; S. jolyi Lescure and Marty, 2000; S. karenanneae (Pyburn, 1993); S. lindsayi Pyburn, 1992; S. madeirae (Bokermann, 1964); S. nebulosus (Spix, 1824); S. onca Ferrão, Moravec, Fraga, Almeida, Kaefer and Lima, 2017; S. oreites Duellman and Wiens, 1993; S. pedromedinae (Henle, 1991); S. proboscideus (Brongersma, 1933); S. rostratus (Peters, 1863); S. ruber (Laurenti, 1768); S. ruberoculatus Ferrão, Fraga, Moravec, Kaefer and Lima, 2018a; S. sateremawe Sturaro and Peloso, 2014; S. strussmannae Ferrão, Moravec, Kaefer, Fraga and Lima, 2018b; S. villasboasi Brusquetti, Jansen, Barrio-Amorós, Segalla and Haddad, 2014; S. wandae; and S. x-signatus (Spix, 1824). Detailed comparisons are provided for closely related species as indicated by molecular phylogenetic relationships (S. cruentomma, S. exiguus, S. wandae, and S. strussmannae) and for species with similar morphology and advertisement calls (S. baumgardneri, S. blairi, S. exiguus, and S. lindsayi).
RESULTS
Phylogenetic relationships and genetic distances
Phylogenetic trees inferred through BI and ML recovered similar topologies (Fig. 2, Appendix 3). Individuals of the new species form a well-supported clade within a major clade containing three nominal (Scinax cruentomma, S. strussmannae, and S. wandae) and nine undescribed species. All nominal species in this major clade have a red- or brown-striped iris. The new species is recovered as sister to a clade composed of specimens from Suriname and Guyana (Scinax sp. clade 72), but pairwise genetic distances to them are low (mean p-distance = 1.4%) compared with the distance between other clades. We lack morphological data for these specimens, so we do not consider them as conspecific to the new species until additional data are available. Scinax wandae from Colombia was recovered as sister to the clade comprising the new species and the Suriname and Guyana specimens (Fig. 2).
Figure 2.
Bayesian phylogenetic inference focused on the major clade containing Scinax albertinae sp. nov. and its most closely related species, based on a fragment of the 16S rDNA mitochondrial gene. The figure at lower left depicts the complete phylogenetic tree (Appendix 3). Support values along branches represent bootstrap values of a Maximum Likelihood analysis (above) and Bayesian posterior probabilities (in %; below).
Overall interspecific p-distance within the aforementioned major clade is 6.6%. Pairwise p-distances between the new species and other nominal species with a red- or brown-striped iris range from 2.7% (S. wandae) to 10.1% (S. strussmannae; Table 1). Intraspecific p-distances within each of these Scinax species range from 0% (Scinax sp. clade 47, S. strussmannae and S. wandae) to 0.7% (Scinax sp. clade 48). Average intraspecific p-distance within the new species is 0.4% and within the Scinax sp. clade 72 is 0.2% (Table 1).
Table 1.
Interspecific and intraspecific genetic distances between Scinax albertinae sp. nov. and closely related taxa included in our phylogenetic analyses. Uncorrected p-distances (lower diagonal) and Kimura two-parameter distances (K2p; upper diagonal) are based on a fragment of the 16S rRNA mitochondrial gene and expressed as percentages. Numbers in bold represent intraspecific p-distance values.
TAXONOMIC ACCOUNT
Scinax albertinae sp. nov.
LSID: urn:lsid:zoobank.org:act:2D465FEC21EC-4D93-829A-9730D2B7F028
White-sand′s snouted treefrog
Figures 3–4, 5A, B, 6–11, 13A; Tables 1–4
Scinax sp. - Ferrão et al. (2019)1, Lima et al. (2021).
Holotype. INPA-H 42872 (field number APL 23058; Figs. 3A, B, 4, 5A, B, 6), adult male, collected 25 April 2020 in Rio Negro Sustainable Development Reserve (03°03′31″S, 60°45′42″W; 73 m a.s.l.), municipality of Novo Airão, Amazonas, Brazil, by M. Ferrão and A. S. Ferreira.
Paratopotypes. Nineteen adult specimens, same locality as the holotype: three males INPA-H 42867, 42863, 42857 (field numbers APL 22249, 22245, and 22243, respectively) and four females INPA-H 42861, 42868–69, 42874 (field numbers APL 22248, 22247, 22244, and 22246, respectively) collected 14 September 2018 by M. Ferrão, J. Moravec, and A. S. Ferreira; a male INPA-H 42866 (field number APL 22264) and four females INPA-H 42864, 42876, 42871, 42870 (field numbers APL 22260–63) collected 15 September 2018 by M. Ferrão, J. Moravec, and A. S. Ferreira; three males INPA-H 42862, 42865, 42875 (field numbers APL 22635–37) and a female INPA-H 42856 (field number APL 22638) collected 17 January 2019 by M. Ferrão and A. S. Ferreira; a male INPA-H 42858 (field number APL 22923) collected 10 January 2020 by A. S. Ferreira, J. Dayrell, and R. Pereira; two males INPA-H 42873, 42859 (field numbers APL 23056–57) collected 25 April 2020 by M. Ferrão and A. S. Ferreira.
Referred specimens. Same locality as the holotype. Adult female INPA-H 42968 (field number APL 22239) collected 8 June 2018 by L. J. C. L. Moraes and I. Y. Fernandes; one tadpole, Gosner stage 42, and five newly metamorphosed specimens, Gosner Stage 45 (lot INPA-H 42854, field number APL 23061), collected 25 April 2020 by M. Ferrão and A. S. Ferreira; 17 tadpoles, Gosner stages 34–38 (lot INPA-H 42855, field number APL 23059), collected 25 April 2020 by M. Ferrão and A. S. Ferreira.
Etymology. The specific epithet honors Albertina Pimentel Lima (INPA, Brazil) for her outstanding contributions to the ecology and taxonomy of Amazonian frogs and to the professional careers of many students, including most authors of this study.
Diagnosis. The new species is assigned to the genus Scinax based on the position of the cloacal tube in tadpoles, molecular phylogenetic relationships, and morphological similarities among adults (see comparisons). Scinax albertinae sp. nov. is distinguished from other congeners by the following combination of characters: (1) SVL 21.6–25.4 mm (n = 11) in adult males and 24.8–27.0 mm (n = 9) in females; (2) head longer than wide; (3) snout subacuminate in dorsal view; (4) canthus rostralis straight in dorsal view; (5) nuptial pad present in breeding males; (6) small ulnar tubercles; (7) subarticular tubercle subconical on fingers I and II; (8) vocal sac bilobate; (9) outer metatarsal tubercle distinct and subconical; (10) small tubercles on subcloacal region; (11) dark brown lateral stripe on flanks, fading posteriorly; (12) posterior surfaces of thigh uniformly brown; (13) horizontal bar on iris reddish-brown; (14) vocal sac with dark blotches on the lateral region; (15) absence of dark bars on thigh; (16) white bones; (17) advertisement call consisting of a single pulsed note with call duration of 502–652 ms, a call rate of 32–68 calls/min, 79–105 pulses/note, and a dominant frequency of 3,811–4,543 Hz; (18) tadpoles with brown dorsum, labial keratodont row formula 2(2)/3, keratodont row P-2 longer than P-1 and P-3, and total length reaching ∼28.6 mm at Gosner stage 37.
Comparisons. Characters of compared species are enclosed in parentheses if not otherwise stated. A subacuminate snout in dorsal view and a uniformly brown posterior thigh distinguish Scinax albertinae sp. nov. from species of the Scinax rostratus group, viz., S. garbei, S. jolyi, S. nebulosus, S. pedromedinae, S. proboscideus, and S. rostratus (snout pointed or elongate in dorsal view and posterior portion of thigh spotted, marbled, or brindle; Duellman, 1972a; Duellman and Wiens, 1992, 1993; Lescure and Marty, 2000).
A SVL of 21.6–25.4 mm in adult males distinguishes S. albertinae sp. nov. from the large species S. boesemani (28.4–31.8 mm; Duellman, 1986), S. chiquitanus (27.9–33.3 mm; Duellman and Wiens, 1993), S. funereus (29.8–36.9 mm; Duellman, 1971; Duellman and Wiens, 1993), S. fuscovarius (36–54 mm; Goldberg et al., 2018), S. iquitorum (35.0–38.5 mm; Moravec et al., 2009), S. karenanneae (26.6–28.9 mm; Pyburn, 1993); S. onca (31.3–34.5 mm; Ferrão et al., 2017), S. oreites (28.4–33.5 mm; Duellman and Wiens, 1993), S. ruber (29.4–41.2 mm; Duellman and Wiens, 1993), S. sateremawe (35.2–38.1 mm; Sturaro and Peloso, 2014), and S. x-signatus (34.5–38.4 mm; Araujo-Vieira et al., 2020).
A reddish-brown horizontal bar on the iris and the subacuminate snout in dorsal view distinguish S. albertinae sp. nov. from S. ruberoculatus (upper portion of iris red and horizontal dark brow absent, snout truncate in dorsal view; Ferrão et al., 2018a). In addition, tadpoles of S. albertinae sp. nov. differ from those of S. ruberoculatus by the presence of a dark brown horizontal stripe on the iris (horizontal stripe on iris absent; Ferrão et al., 2018a).
Presence of a bilobate vocal sac in males distinguishes S. albertinae sp. nov. from S. fuscomarginatus, S. madeirae, and S. villasboasi (single vocal sac; Brusquetti et al., 2014). In addition, a wider head in males and females, HW/SVL = 34 ± 1%, distinguishes S. albertinae sp. nov. from S. fuscomarginatus (28 ± 2%), S. madeirae (30 ± 1%), and S. villasboasi (30 ± 1%; Brusquetti et al., 2014). The advertisement call of S. albertinae sp. nov. also differs from those of S. madeirae in call duration of 502–652 ms (720–1,160 ms; Brusquetti et al., 2014).
Scinax albertinae sp. nov. can be distinguished from S. baumgardneri by SVL in males, 21.6–25.4 mm, and females, 24.8–27.0 mm; snout subacuminate in dorsal view; canthus rostralis straight; and outer metatarsal tubercle distinct and subconical (SVL 29.0 mm in the male holotype and 29.5–32.0 mm in females, snout truncate in dorsal view, canthus rostralis rounded, outer metatarsal tubercle indistinct; Rivero, 1961).
Scinax albertinae sp. nov. can be distinguished from S. blairi by SVL in males, 21.6–25.4 mm, and females, 24.8–27.0 mm, and by the absence of dark bars on the thigh (SVL 27.8–30.1 mm in males and 32.1–32.5 mm in females, 1–3 dark bars on thigh; Fouquette and Pyburn, 1972). In addition, the advertisement call of S. albertinae sp. nov. consists of a single note with a call duration of 502–652 ms and 79–105 pulses/note (call duration 140–160 ms, 18–22 pulses/note; Fouquette and Pyburn, 1972).
Scinax albertinae sp. nov. can be distinguished from S. lindsayi by the subconical subarticular tubercle on fingers I and II; the dark lateral stripe on the dorsum, which fades posteriorly; and the white nuptial pad in breeding males (subarticular tubercle rounded on all fingers, dark lateral stripe and nuptial pad absent; Pyburn, 1992). In addition, the advertisement call of S. albertinae sp. nov. lasts 502–652 ms and has a dominant frequency of 3,811–4,543 Hz (call duration 80–100 ms, dominant frequency 2,695–3,235 Hz; Pyburn, 1992).
Scinax albertinae sp. nov. can be distinguished from S. cruentomma by the maximum SVL in males, 25.4 mm, and females, 27.0 mm; canthus rostralis straight; nuptial pad present in breeding males; bilobate vocal sac; small subcloacal tubercles; reddish-brown horizontal bar on iris; and white bones (maximum SVL 27.7 in males and 30.6 in females, canthus rostralis rounded, nuptial pad absent in breeding males, single vocal sac, subcloacal tubercles absent, bright red horizontal bar on iris, green bones; Duellman, 1972b). In addition, the advertisement call of S. albertinae sp. nov. differs from that of S. cruentomma in call duration, 502–652 ms, and dominant frequency, 3,811–4,543 Hz (call duration 350–370 ms, dominant frequency 3,200–3,400 Hz at the type locality in Ecuador; call duration 216–336 ms, dominant frequency 2,156–2,438 Hz at the upper Negro River in Brazil; Duellman, 1972b; Carvalho et al., 2015). Tadpoles of S. albertinae sp. nov. differ from those of S. cruentomma in having a brown dorsum, labial keratodont row formula 2(2)/3, and maximum total length ∼28.6 mm only at Gosner stage 37 (pale olive-tan dorsum, labial keratodont row formula 2(2)/3(1), maximum total length 28.2 mm at Gosner stage 30; Duellman, 1972b).
Scinax albertinae sp. nov. can be distinguished from S. exiguus by the SVL in males, 21.6–25.4 mm, and females, 24.8–27.0 mm; ulnar tubercles present; moderate-sized vocal sac; and TD/ED = 37–44% in males (SVL 18–20.8 mm in males and 20.1–24.5 mm in females, ulnar tubercles absent, large vocal sac, and TD/ED = 48–60% in males; Duellman, 1986). The advertisement call of S. albertinae sp. nov. has a call duration of 559 ± 44 ms, 89 ± 7 pulses/call, and a call rate of 54 ± 8 calls/min (call duration 1,148 ± 333 ms, 71 ± 8 pulses/call, and call rate 26 ± 8 calls/min at the type locality; Duellman, 1986; Carvalho et al., 2017).
Scinax albertinae sp. nov. can be distinguished from S. strussmannae by the SVL in males, 21.6–25.4 mm; snout subacuminate in dorsal view; canthus rostralis straight; reddish brown horizontal bar on iris; and vocal sac with dark blotches restricted to the lateral region (SVL 20.2–22.5 mm in males, snout truncate in dorsal view, canthus rostralis curved, bright red horizontal bar on iris, vocal sac immaculate bright yellow; Ferrão et al., 2018b). The advertisement call of S. albertinae sp. nov. is distinct from that of S. strussmannae in its call duration of 502–652 ms, call rate of 32–68 calls/min, 79–105 pulses/note, and dominant frequency of 3,811–4,543 Hz (call duration 97–115 ms, call rate 85–175 calls/min, 23–27 pulses/note, and 2,541–3,015 Hz; Ferrão et al., 2018b).
Scinax albertinae sp. nov. can be distinguished from S. wandae by the absence of a well-defined and continuous dark dorsolateral stripe, maximum SVL 25.5 mm in males, white nuptial pad present in breeding males, canthus rostralis straight and bilobate vocal sac (well-defined and continuous dark dorsolateral stripe always present, maximum SVL 26.9 mm in males, nuptial pad absent, canthus rostralis rounded, and single vocal sac; Pyburn and Fouquette, 1971). The advertisement call of S. albertinae sp. nov, with a call duration of 502–652 ms (average 559 ± 44 ms) and a dominant frequency of 3,811–4,543 Hz, differs from the calls of S. wandae at its type locality (call duration 653–696 ms, dominant frequency 4,800–5,050 Hz; Pyburn and Fouquette, 1971) and at San Martín, Meta, Colombia (average call duration 624 ± 17 ms, dominant frequency 4,800–5,050 Hz; Table 3).
Holotype description. INPA-H 42872 (field number APL 23058), an adult male (Fig. 3A, B), SVL 23.8 mm. Head longer than wide (HL/HW = 1.1); head length and width 36% and 34% of SVL, respectively. Snout subacuminate in dorsal view, rounded in lateral view; eye–nostril distance 35% of HL; nostrils protruding, nostril opening directed dorsolaterally; internarial region barely concave; canthus rostralis well-defined, straight in dorsal view; loreal region concave. Eyes large, protruding laterally; eye diameter equals eye–nostril distance (ED/END = 1.0) and about 35% of HL; upper eyelid 110% of IOD. Interorbital region flattened; interorbital distance about 28% of head width. Tympanum distinct and vertically ovoid; upper portion obscured by the supratympanic fold; tympanum diameter 40% of eye diameter. Supratympanic fold well-defined, extends from anterior portion of tympanum to upper arm insertion. Vocal slits present, extend from the angle of the jaw to the lateral base of the tongue, not covered by the lateral margin of the tongue. Tongue ovoid, entirely attached to the floor of the mouth. Dentigerous processes of vomers triangular and pronounced, separated from each other by 1/3 of their width, each with three evident teeth. Choanae oval. Vocal sac of moderate size, bilobate and subgular, when inflated extends from the subgular region to the upper arm insertion.
Arms slender, forearm more robust than upper arm. Axillary membrane absent (Fig. 3A, B). Ulnar tubercles low, distributed from elbow to lateral base of finger IV. Hand length about 26% of SVL; relative length of fingers I < II < IV < III; vestigial webbing only between fingers II and III (Fig. 4A); lateral fringes present on all fingers, but poorly developed; finger discs elliptical, wider than long; finger III disc wider than tympanum diameter (3FD/TD = 1.1); finger I disc smaller than others. Subarticular tubercles single and subconical on fingers I and II, single and rounded on finger III, and bifid on finger IV; supernumerary tubercles flattened on the base of fingers II–IV, subconical on the base of finger I; palmar tubercle bifid and flat; thenar tubercle oval and flat. Nuptial pad white, covers the thumb dorsolaterally from the level of the subarticular tubercle to the proximal base of the thenar tubercle.
Hind limbs slender and long; THL + TL about 97% of SVL; tibia longer than thigh (TL/THL = 1.1); thigh and tibia length 47% and 50% of SVL, respectively. Tubercles on knee absent. Foot length about 40% of SVL and 140% of tarsus length. Tarsal fold absent; tarsal tubercles small and low, less evident than ulnar tubercles. Toes slender (Fig. 4B); relative length of toes I < II < III < V < IV; toe discs elliptical, wider than long; disc on toe IV slightly larger than that on finger III (4TD/3FD = 1.03); discs on toes I and II wider than those on other toes; subarticular tubercle bifid on toe V, single and subconical on toes I–IV; toe webbing formula I vestigial II 2-–3- III 11/2–3- IV 3–11/2 V; lateral fringes present on all toes; supernumerary tubercles rounded and evident; inner metatarsal tubercle oval and protuberant; outer metatarsal tubercle small, subconical and about two times smaller than the inner tubercle. Cloacal opening at the upper level of thighs; small subcloacal tubercles present.
In life, skin on upper eyelids, flanks, and dorsal surfaces of body, forearm, and tibia finely granulate with several scattered tubercles; dorsal surface of the head mostly smooth with few scattered tubercles; upper arm and thigh smooth (Fig. 5A). Vocal sac smooth. Chest, belly, and ventral surface of thigh areolate. In preservative, dermal tubercles barely evident.
Holotype measurements (in mm). SVL, 23.8; HL, 8.6; HW, 8.0; ED, 3.0; TD, 1.2; UEW, 2.5; IOD, 2.3; IND, 2.1; END, 3.0; HAL, 6.3; 3FD, 1.3; THL, 11.2; TL, 11.9; FL, 9.5; TAL, 6.9; 4TD, 1.3.
Coloration of holotype in life. At night (Fig. 5A, B), the dorsal surface of the snout is brownish cream, with a longitudinal medial dark brown line that extends from the internarial area to the level of two dark brown tubercles that lie between anterior corners of the eyes; a brown W-shaped mark at the anterior edge of the interorbital region is bordered laterally by a light grey stripe; a dark brown canthal stripe extends from the tip of the snout to the anterior corner of each eye; loreal region dark brown; infra-ocular region cream. Iris tan with a horizontal reddish-brown stripe. Dorsum light yellowish brown with scattered dark brown tubercles; a dorsolateral brownish cream stripe that runs from the posterior margin of the eye is conspicuous anteriorly but fades posteriorly; a lateral dark brown stripe starts as a narrow supratympanic stripe from the posterior corner of each eye, widens above the upper arm insertion, and fades posteriorly; flanks pale yellow. Dorsal surfaces of forearm, tibia, and tarsus light brown; ulnar tubercles cream; two transverse brown stripes on forearm and three on tibia; dorsal surface of thigh yellowish brown, anterior and posterior surfaces brown; cloacal region yellowish cream. Vocal sac bright yellow with dark brown melanophores distributed laterally; chest and belly yellowish cream; ventral surfaces of hand, forearm, thigh, tarsus, and foot light brown; ventral surface of tibia translucent ventrally, pale brown laterally.
Coloration of holotype in preservative. Dark stripes and marks on head, dorsum, forearm, and thigh remain visible (Fig. 3A, B). The dorsolateral stripe is clear and more conspicuous. Cream and yellow surfaces on venter are white.
Vocalization. The advertisement call of Scinax albertinae sp. nov. consists of a single pulsed note emitted in series of 11 ± 4 calls (5–20; n = 13; Fig. 6A, B). Calls have an average duration of 559 ± 44 ms (502–652 ms; n = 30), an intercall interval of 590 ± 207 ms (342–1,346 ms; n = 30), and a call period of 1,149 ± 210 ms (878–1,878 ms; n = 30). Calls are emitted at an average rate of 54 ± 8 calls/min (32–68 calls/min; n = 30). On average, calls are composed of 89 ± 7 pulses (79–105 pulses; n = 30). Pulses have an average duration of 3.4 ± 0.7 ms (2–6 ms; n = 77) and an interpulse duration of 2.9 ± 0.8 ms (2–7 ms; n = 77). An average pulse lasts for 6.3 ± 0.6 ms (5–9 ms; n = 77). Pulses are emitted at an average rate of 160 ± 14 pulses/sec (111–200 pulses/sec; n = 77). Calls have three distinct frequency bands; the dominant frequency always corresponds to the central one, which averages 4,121 ± 149 Hz (3,811–4,543 Hz; n = 30). The average low frequency is 2,787 ± 138 Hz (2,551–3,052 Hz; n = 30) and the average high frequency is 5,396 ± 93 Hz (5,218–5,569 Hz; n = 30). See Table 3 for complete values of acoustic parameters measured for the three recorded males.
The territorial call consists of a single tonal note with a call duration of 115 ± 26 ms (72–168 ms; n = 15; Fig. 6C). Calls are usually emitted within an advertisement call series (n = 11), rarely succeeding (n = 3) or preceding (n = 1) it. Territorial calls have an average intercall interval of 550 ± 228 ms (299–1,091 ms; n = 15), an average call period of 666 ± 237 ms (405–1,219 ms; n = 15), and are emitted at a rate of 99 ± 31 calls/min (49–148 calls/min; n = 15). On average, 3–4 distinct frequency bands (3.2 ± 0.4; n = 15) were observed in the territorial calls. The dominant frequency, 3,748 ± 307 Hz (3,143–4,522 Hz; n = 15), usually corresponds to the second frequency band (n = 14). The average low frequency is 2,736 ± 216 Hz (2,323–3,083 Hz; n = 15) and the average high frequency is 4,792 ± 256 Hz (4,124–5,233 Hz; n = 15).
Variation. Morphometric variation of the paratopotypes is summarized in Table 2. External morphology mostly agrees with the holotype. Subarticular tubercle on finger IV is bifid in 74% of specimens, slightly bifid in 21%, and rounded in 5%. Subarticular tubercle on toe V is bifid in 42% of specimens, slightly bifid in 26%, and rounded in 32%. Expanded vocal sac and nuptial pads are absent in adult males collected in the nonbreeding season (56%). Number and presence of teeth varies from 1 to 5 on the right odontophore and from 1 to 6 on the left odontophore; male specimen INPA-H 42873 lacks teeth on both odontophores.
Although dark brown marks are evident in specimens during both day and night, overall external color changes markedly between these periods (Fig. 7). The nocturnal yellow-brownish color changes to darker brown tones during the day (Fig. 7G–J). Variation in overall coloration is also evident between distinct rainfall periods, likely correlated with the reproductive cycle of the species. During the rainy season, when the species reproduces, specimens show a more yellowish dorsum (Figs. 5A, B, 7A). During the dry season, specimens became paler, with a yellowish-grey or yellowish-brown coloration (Fig. 7B–F). Skin texture also varies between dry and rainy seasons, with granules more evident during the former season (Fig. 7B, C) and a smoother appearance during the latter (Fig. 7D–J).
In preservative, light grey dorsal spots are present in 63% of specimens (Fig. 8). Interorbital marks are present in all specimens except INPA-H 42876, 42856, and 42859. In 42% of specimens, this interorbital mark is formed by a dark brown mark bordered anteriorly by a light grey mark, whereas only a light grey mark is evident in 37% of specimens and only a dark brown mark is evident in 5% of specimens. One-to-three dark brown bars are present on the tibia in 68% of specimens; they are absent in 32% of specimens. A dark brown dorsolateral stripe formed by blotches and spots is present in 37% of specimens but absent in the remaining ones. The vocal sac is cream to white in the most recently collected males (INPA-H 42873, 42859); in five females, dark melanophores are randomly distributed in the central portion of the gular region (INPA-H 42856, 42861, 42864, 42868, and 42876). Chest and belly are immaculate cream in all specimens except two females that have small dark brown blotches on the belly (INPA-H 42868 and 42876).
Tadpoles. Description was based on 17 tadpoles (lot INPA-H 42855) at stages 34 to 38 (Figs. 9, 10). Total length 26.8 ± 1.3 mm (24.4–28.6 mm; n = 10; Table 4). Body triangular in lateral view, ovoid in dorsal view; body length 33% ± 1% (32–35%) of total length; body longer than high (BL/BH=1.6 ± 0.0, 1.6–1.7) and as wide as high (BW/BH = 1.0 ± 0.0, 1.0–1.1). Snout rounded in dorsal and lateral views. Nostrils rounded, dorsally positioned and directed; internarial region flattened; internarial distance 54% ± 3% (51–59%) of interorbital distance and 39% ± 2% (36–41%) of body width. Eyes positioned and directed dorsolaterally, marginally visible in ventral view; eye diameter 20% ± 1% (19–22%) of body length and 31% ± 2% (30–34%) of snout–spiracle distance; interorbital region flattened, interorbital distance 72% ± 4% (68–79%) of body width. Spiracle sinistral, positioned at an angle of approximately 30°, opening directed posterodorsally, inner wall free from body. Tail long and unflagellated, twice as long as the body (TAL/BL = 2.0 ± 0.1, 1.8–2.1; n = 10); tail length 67% ± 1% (65–68%) of total length; tail higher than body height (MTH/BH = 1.3 ± 0.1, 1.1–1.4); upper fin higher than lower fin. On average, tail muscle as high as wide (TMH/TMW = 1.0 ± 0.1, 0.9–1.1); tail muscle height 34% ± 3% (30–38%) of maximum tail height and 26% ± 1% (24–27%) of body height (Table 4).
Oral disc situated and directed anteroventrally, protuberant; oral disc width 53% ± 5% (44–61%) of body width. Upper lip unemarginated laterally; lower lip emarginated laterally. Papillae rounded and long; uniseriate row of marginal papillae on upper lip, with a moderate gap anteriorly; biseriate and uniseriate rows of marginal papillae on the lateral fold and on the central portion of lower lip, respectively; marginal papillae row formula (1)/2/1. Submarginal papillae present on the lateral fold. Jaw sheaths serrated with conical cusps; upper jaw sheath M-shaped; lower jaw sheath V-shaped. Labial keratodont row formula 2(2)/3; A-1 slightly longer than A-2; a narrow gap in A-2; P-1 and P-3 approximately same length, P-2 longer than P-1 and P-3.
In life, tadpoles of Scinax albertinae sp. nov. at stage 37 have a light brown snout that is translucent anteriorly; the dorsum is brown with a dark brown interorbital mark (Fig. 10). On the snout, a dark brown lateral stripe runs from the level of the lower labium to the eye; the infraorbital region is white. The iris is brownish orange with a horizontal dark brown stripe; there are dark brown blotches on the upper and lower edges of the iris, but they do not form a complete vertical stripe. A dark brown lateral stripe extends from the posterior eye corner to the level of the tail–muscle insertion. Tail muscle is brown with several tiny light spots (Fig. 10); a thin dark brown lateral line runs from the base of the tail muscle to the tail tip and is more conspicuous posteriorly. Dorsal and ventral fins are translucent cream with dark brown vermiculation, dark melanophores, and orange blotches; there are dark brown blotches on the upper and lower edges of the central region of the fins (Fig. 10). Venter is translucent with several silvery spots on the intestinal region (Fig. 10).
In life, metamorphs at stage 45 (lot INPA-H 42854) have a tan dorsum with light to dark brown marks, including interorbital stripes and blotches (Fig. 11). There is a dark brown canthal stripe; the loreal region is dark brown. A dark brown lateral stripe extending from the posterior corner of the eye to the groin is more conspicuous anteriorly. Lower portions of the flanks are greyish white. Dorsal surfaces of limbs are light tan; one-to-three grey bars are present in some individuals (Fig. 11). The iris is reddish orange with a dark brown horizontal stripe. Ventral surface is translucent with silver blotches.
Geographic distribution and habitat. Scinax albertinae sp. nov. is known only from the type locality in the RDS Rio Negro. All individuals were recorded exclusively in white-sand forest (campinarana), which is characterized by a canopy height below 20 m (Fig. 12A). Extensive sampling efforts were conducted in other habitats that contribute to the landscape mosaic of the RDS Rio Negro, including open white-sand forest (campina, with canopy height below 10 m; Fig. 12B) and dense forests with rich soils, but the new species was never observed in them. Thus S. albertinae sp. nov. appears to be typical of WSE but likely inhabits only a single habitat within these ecosystems. This hypothesis is supported by the fact that S. albertinae sp. nov. also has never been recorded in dense forests geographically close to RDS Rio Negro, in Manaus and northern Purus–Madeira interfluve, despite our long-term sampling effort in those regions.
During the dry season, we observed individuals of Scinax albertinae sp. nov. hidden in bromeliads (Guzmania brasiliensis; Fig. 12D) or in plastic tubes used to mark sampling trails in the RDS Rio Negro sampling module. This indicates that phytotelmata may represent important refuges for this species. During the rainy season, individuals were recorded active and more exposed, usually over shrub vegetation near large temporary pools (Fig. 12C). Although the closest relatives of S. albertinae sp. nov. are explosive breeders (e.g., S. strussmannae; Ferrão et al., 2018b), based on our field observations this new species is likely a prolonged breeder (sensu Wells, 1977): we recorded males calling for several consecutive days in the rainy season and females were not observed arriving synchronously with males. Males call while perched horizontally or vertically on leaves and small trunks near temporary pools not connected to streams, where we also recorded tadpoles. These pools were filled with dark, acidic water formed by the accumulation of organic material following incomplete decomposition in the surrounding sandy soil.
Two other hylid species were heard near the temporary pools used as breeding sites by Scinax albertinae sp. nov.: Osteocephalus vilarsi and Trachycephalus cunauaru Gordo, Toledo, Suárez, Kawashita-Ribeiro, Ávila, Morais and Nunes, 2013. Yet, the only other anuran species actually observed using the same pools to breed is the microhylid Chiasmocleis hudsoni Parker, 1940, whose tadpoles were recorded with those of S. albertinae sp. nov.
Figure 3.
Dorsal and ventral views of the (A, B) male holotype INPA-H 42872 and (C, D) female paratopotype INPA-H 42871 of Scinax albertinae sp. nov. from Rio Negro Sustainable Development Reserve, municipality of Novo Airão, Amazonas, Brazil. Photographs: I. Y. Fernandes.
Figure 4.
Ventral views of the hand (A) and foot (B) of the holotype of Scinax albertinae sp. nov., INPA-H 42872. Scale bar = 3 mm. Photographs: M. Ferrão.
Figure 5.
Live specimens of Scinax albertinae sp. nov. (A–B) and closely related species (C–F). (A–B) Male holotype, INPA-H 42872, RDS Rio Negro, Amazonas, Brazil. (C) Scinax cruentomma, QCAZ 43772, adult female from Francisco de Orellana, Orellana, Ecuador. (D) Scinax exiguus, unvouchered male from Tepequém, Roraima, Brazil. (E) Scinax strussmannae, INPA-H 34688, male holotype from Tapauá, Amazonas, Brazil. (F) Scinax wandae, IAvH-Am 15107, male from La Macarena, Meta, Colombia. Photographs: A. S. Ferreira (A–B), S. Ron, bioweb.bio (C), I. Y. Fernandes (D), R. de Fraga (E), and A. R. Acosta-Galvis (F).
Figure 6.
Vocalization of the holotype of Scinax albertinae sp. nov. (A) Call series with 12 advertisement calls, detailed in (B), are interleaved with 7 territorial calls (red stars), detailed in (C).
Figure 7.
Coloration in life of Scinax albertinae sp. nov. paratopotypes and the referred specimen. Nocturnal pattern: (A) Male, INPA-H 42873, SVL = 24.4 mm. (B, C) Male, INPA-H 42857, SVL = 25.4 mm. (D) Male, INPA-H 42867, SVL = 23.3 mm. (E, F) Female, INPA-H 42869, SVL = 24.8 mm. Diurnal pattern: (G) INPA-H 42867. (H, I) INPA-H 42869. (J) Female, INPA-H 42968, SVL unmeasured. (K) INPA-H 42869. (L) Female, INPA-H 42868, SVL = 26.6 mm. Photographs: A. S. Ferreira (A), J. Moravec (B–I, K and L) and L. J. C. L. Moraes (J).
Figure 8.
Dorsal and ventral views of Scinax albertinae sp. nov. paratopotypes. (A) Male, INPA-H 42873, SVL = 24.4 mm. (B) Female, INPA-H 42861, SVL = 26.2 mm. (C) Female, INPA-H 42870, SVL = 27.0 mm. (D) Male, INPA-H 42867, SVL = 23.3 mm. (E) Male, INPA-H 42863, SVL = 23.4 mm. (F) Female, INPA-H 42869, SVL = 24.8 mm. (G) Male, INPA-H 42857, SVL = 25.4 mm. (H) Male, INPA-H 42866, SVL = 23 mm. (I) Male, INPA-H 42857. (J) Male, INPA-H 42873. (K) INPA-H 42869. (L) Male, INPA-H 42868, SVL = 24.9 mm. Photographs: M. Ferrão.
Figure 9.
Preserved tadpole of Scinax albertinae sp. nov. (lot INPA-H 42855), Gosner stage 37. (A) Lateral and (B) ventral views, and (C) lateroventral and (D) ventral views of the oral disc. Scale bars = 5 mm (A, B) and 1 mm (C, D).
Figure 10.
Coloration in life of tadpoles of Scinax albertinae sp. nov. (lot INPA-H 42855), Gosner stage ∼ 37. Photographs: M. Ferrão.
Figure 11.
Color in life of metamorphs of Scinax albertinae sp. nov. (lot INPA-H 42854), Gosner stage 45. Photographs: M. Ferrão.
Figure 13.
Advertisement calls of Scinax albertinae sp. nov. and close relatives. Call vouchers and localities: (A) FNJV50546, Rio Negro Sustainable Development Reserve, Novo Airão, Amazonas, Brazil. (B) ML218321, near junction of Ariari and Guaviare rivers, Vaupés, Colombia. (C) AM1TRC_AAGm671, Cucuí, São Gabriel da Cachoeira, Amazonas, Brazil. (D) ML194693, km 144 of the El Dorado–Sta. Elena de Uairén Road, Bolívar, Venezuela. (E) FNJV50555, Paiva Falls, Tepequém, Roraima, Brazil. (F) ML194674, Belém, Pará, Brazil. (G) Unvouchered, Juruena River, Cotriguaçu, Mato Grosso, Brazil. (H) ML196980, El Porvenir, Beni, Bolivia. (I) ML210707, Cordoba, Veracruz, Mexico. (J) Unvouchered, Nascentes do Lago Jari National Park, Tapauá, Amazonas, Brazil. (K) FNJV50554, Cantá, Roraima, Brazil. (L) BSA15980, San Martín, vereda Montebello, finca Tocancip, Meta, Colombia. Scale bars = 100 ms (A–I, K–L) and 40 ms (J).
Table 2.
Morphometric measurements of the type series of Scinax albertinae sp. nov. from Rio Negro Sustainable Development Reserve, Municipality of Novo Airão, state of Amazonas, Brazil. Values depict average ± standard deviation (minimum–maximum). Measurements are explained in the text.
Table 3.
Comparisons of spectral and temporal parameters of the advertisement calls of Scinax albertinae sp. nov. and four closely related species. Measurement acronyms are defined in the text. Values depict average ± standard deviation (minimum–maximum).
Table 4.
Morphometric measurements, in mm, of Scinax albertinae sp. nov. tadpoles (lot INPA-H 42855) at distinct Gosner developmental stages. Values depict average ± standard deviation (minimum–maximum); acronyms are defined in the text.
Figure 12.
White-sand ecosystems in the Rio Negro Sustainable Development Reserve, municipality of Novo Airão, state of Amazonas, Brazil, the type locality of Scinax albertinae sp. nov. (A) White-sand forest with canopy below 20 m—campinarana—with several bromeliads. (B) Open white-sand forest with canopy below 10 m—campina. (C) Temporary pool filled with dark, acidic waters used as breeding sites by Scinax albertinae sp. nov. (D) An unvouchered specimen of Scinax albertinae sp. nov. hidden in a bromeliad during the dry season. Photographs: A. S. Ferreira.
DISCUSSION
Scinax albertinae is the third hylid frog associated with WSE of Central Amazonia. The other two are Trachycephalus venezolanus (Mertens, 1950) and Osteocephalus vilarsi (Carvalho et al., 2018; Ferrão et al., 2019). Moreover, molecular, morphologic, and bioacoustic data recover at least two additional candidate species of Adenomera (Leptodactylidae) and Rhinella (Bufonidae) that are associated with this environment at RDS Rio Negro (A. P. Lima, unpublished data). These associations, together with results of other studies of species diversity of WSE in Central Amazonia that consider anurans (Carvalho et al., 2018; Ferrão et al., 2019), reptiles (Fraga et al., 2018; Ferreira et al., 2019), birds (Borges et al., 2001; Almeida, 2016), dipterans (Ale-Rocha and Vieira, 2008), and plants (Terra-Araujo et al., 2015; Gaem et al., 2019), highlight the extreme biological significance and uniqueness of this area, which until recently has been largely overlooked.
Scinax albertinae is known only from its type locality in the RDS Rio Negro. Based on the distribution of other WSE hylids (Carvalho et al., 2018; Ferrão et al., 2019), we expect the new species to have a considerably wider distribution throughout WSEs in the Negro-Solimões Interfluve. Moreover, the surprisingly low genetic divergence (K2P and p-distances= 1.4%) between S. albertinae and Scinax sp. clade 72 from Guyana and Suriname (Vacher et al., 2020) indicates the possible occurrence of the new species in the Guiana Shield. We were not able to examine the Guyana and Suriname individuals morphologically, but if future analyses bolster our tentative conclusion that they are conspecific with S. albertinae, we hypothesize that ancestral populations of the new species and Scinax sp. clade 72 might have inhabited the eastern slopes of the Andes (where the sister species S. wandae currently occurs) and then split toward their current range. An alternate hypothesis is that emergence of the Negro River was a vicariant event responsible for limiting gene flow between northern and southern populations. Regardless of which of these hypotheses proves most likely, testing them requires a robust phylogeographic approach as well as an extensive sampling over intervening regions among target taxa.
According to our phylogenetic analyses, Scinax albertinae clusters with S. cruentomma, S. wandae, S. strussmannae and several undescribed species in a well-defined, monophyletic clade. This clade, together with its sister clade containing S. fuscomarginatus, S. madeirae, S. villasboasi, S. parkeri (Gaige, 1929) and S. staufferi (Cope, 1865), constitutes a well-supported, major clade of Scinax (Fig. 2, Appendix 3). Delimitation, structure and content of this large clade correspond well to earlier results by Ferrão et al. (2016) and Ferrão et al. (2018b).
By combining characters from morphology, karyology, and reproductive biology, Faivovich (2002) evaluated the monophyly of Scinax species groups proposed by Fouquette and Delahoussaye (1977) and Duellman and Wiens (1992). He found that the S. staufferi species group as previously defined is polyphyletic, but at the same time he recovered a monophyletic S. staufferi clade containing S. cruentomma, S. fuscomarginatus, S. nasicus, S. squalirostis, S. staufferi and an undescribed species, Scinax sp. 2 (S. nasicus and S. squalirostis assume a different topology in Ferrão et al. [2018b] and in the present study). Within his S. staufferi clade, Faivovich (2002) recovered a minor clade composed of S. cruentomma, S. fuscomarginatus, and S. sp. 2, which is supported by two morphological synapomorphies: larynx strongly rotated in relation to the posteromedial processes, and subcylindrical cardiac process of the cricoid ring slightly wider than adjacent parts of the ring. These characters need to be assessed in other members of the major clade containing those species evidenced in the present study. Faivovich (2002) also noted that S. cruentomma and S. fuscomarginatus have a similar vocalization. Later, Carvalho et al. (2017) reported that the advertisement call of S. exiguus shares structural similarities with that of S. fuscomarginatus. The same structural pattern is observed herein in the advertisement calls of S. albertinae, S. cruentomma, S. madeirae, S. staufferi, S. strussmannae, S. wandae, S. trilineatus (Hoogmoed and Gorzula, 1979) and S. parkeri (Fig. 13). Indeed, the latter two species were recently synonymized with S. fuscomarginatus (Brusquetti et al., 2014). In light of these findings, reevaluation of the monophyly of the S. staufferi species group and its possible redefinition appear warranted.
ACKNOWLEDGMENTS
We dedicate this study to the memory of Dona Lúcia Toga, who always kindly supported research in the PPBio sampling module at the RDS Rio Negro. This study was funded by the Programa de Apoio aos Núcleos de Excelência (PRONEX) of the National Council for Scientific and Technological Development (CNPq) and Fundação de Amparo à Pesquisa do Estado do Amazonas (FAPEAM; Proj. 003/2009, Proc. 653/2009); the Museum of Comparative Zoology [MCZ] and David Rockfeller Center for Latin American Studies [DRCLAS], Harvard University; the Czech Republic Ministry of Culture (DKRVO 2019-2023/6.VI.d, National Museum Prague, 00023272); and Centro de Estudos Integrados da Biodiversidade Amazônica (CENBAM). We thank Ariane Silva (INPA-H), Andrés Rymel Acosta-Galvis (IAVH), Fernanda Werneck (INPA-H), José Rosado (MCZ), Michael Franzen (ZSM), and Santiago Ron (QCAZ) for granting access to specimens under their care; to BIOWEB Ecuador, Pontificia Universidad Católica del Ecuador (PUCE), Santiago Ron, and the QCAZ staff for making public data of specimens housed at QCAZ; to A. R. Acosta-Galvis, Igor Y. Fernandes, Rafael de Fraga and S. Ron for sharing photographs of Scinax; to Domingos J. Rodrigues, Marcos Penhacek and Samuel Anjos for sharing the vocalization of S. aff. cruentomma; and to Igor Y. Fernandes, Jussara Dayrell, and Rafaela Pereira for assistance with sampling. Miquéias Ferrão received a fellowship from CNPq (PDJ process 154325/2018-0), an Edward O. Wilson Biodiversity Postdoctoral Fellowship from the Harvard Museum of Comparative Zoology, and a fellowship from the David Rockefeller Center for Latin American Studies of Harvard University. Specimens were collected under IBAMA/ICMBio/RAN permits (02001.000508/2008-99; 1337-1). Protocols for animal collection and care followed the Conselho Federal de Biologia resolution 148/2012.
LITERATURE CITED
Appendices
Appendix 1.
Appendix 1.
Specimens of Scinax and closely related taxa included in molecular analyses, with respective voucher numbers, localities, and GenBank accession numbers. Sequences of the new species are highlighted in bold. Asterisks denote samples shown in Figure 2.
Appendix 1.
Continued.
Appendix 1.
Continued.
Appendix 1.
Continued.
Appendix 2. Specimens examined. Hyla affinis: BRAZIL: “fluminis Amazonum″ = Rio Amazonas (ZSM 2495/0 [holotype, photo]). Scinax baumgardneri: VENEZUELA: Amazonas: (MCZ-A 28563 [holotype]). Scinax blairi: COLOMBIA: Vaupes, near junction of Rio Guaviare and Rio Ariari (MCZ-A 81819 [paratype]). Scinax boesemani: SURINAME: Paramaribo: near Zanderij (RMNH 12601 [holotype, photo], MCZ-A 52833 [paratype]). BRAZIL: Roraima: Caracaraí, Viruá National Park (INPA-H 25972, 25974). Scinax chiquitanus: BRAZIL: Rondônia: Porto Velho (INPA-H 35554–58, 35560). Scinax cruentomma: ECUADOR: Napo: Santa Cecilia (KU 126587 [holotype, photo]); Orellana: Parque Nacional Yasuní (QCAZ-A 8184), Río Napo (QCAZ-A 43772, 44754). BRAZIL: Amazonas: Careiro da Várzea, Ramal do Purupuru (INPA-H 34697). Scinax funereus: ECUADOR: Orellana: Río Napo, Primavera (QCAZ-A 43799, photo), Tambococha (QCAZ-A 55280, 55283; photo); Napo: Limoncocha (MCZ-A 97672). Scinax fuscomarginatus: BRAZIL: Roraima: Boa Vista, Maracá Ecological Station (INPA-H 34634, 34646, 34661–62); Caracaraí, Viruá National Park (INPA-H 19371–72, 19376, 19378, 19383–84). Scinax garbei: BRAZIL: Roraima: Caracaraí, Viruá National Park (INPA-H 25964, 27496). ECUADOR: Napo: Limoncocha (MCZ-A 97672). Scinax madeirae: BRAZIL: Rondônia: Alta Floresta, Corumbiaria Park (INPA-H 7050–51); Porto Velho (MCZ-A 64371 [paratype]). Scinax nebulosus: BRAZIL: Pará: Alter do Chão (INPA-H 34647, 34653); Rondônia: Costa Marques, Real Forte Príıncipe da Beira (INPA-H 34641); Roraima: Caracaraí, Parque Nacional do Viruá (INPA-H 27535–37). Scinax onca: BRAZIL: Amazonas: Beruri (INPA-H 34584 [holotype], 20582, 20585–86 34581, 34583, 34587 [praatypes]); Rondônia: Porto Velho (INPA-H 34588–95 [paratypes]). Scinax proboscideus: BRAZIL: Amazonas: Manaus, Colosso Reserve at PDBFF (INPA-H 10304); Presidente Figueiredo, Vila Pitinga (INPA-H 1870); Pará: Oriximiná (INPA-H 304). Scinax ruberoculatus: BRAZIL: Amazonas: Careiro da Várzea, BR-319, km 100 (INPA-H 34598, 34600–01, 34604, 34614–15, 34622, 34624, 34627, 34629 [paratypes]), km 168 (INPA-H 34602 [paratype]); Borba, BR-319, km 220 (INPA-H 34610, 34620 [paratypes]); Beruri, BR-319, km 220 (INPA-H 34608 [paratype]), km 360 (INPA-H 34599, 34607, 34609, 34611–12, 34617–18, 34621, 34625–26, 34628, 34630 [paratypes]); Manicoré, BR-319, km 400 (INPA-H 34603, 34606, 34616, 34623 [paratypes]); Tapauá, BR-319 km 450, Nascentes do Lago Jari National Park (INPA-H 34665 [holotype], 34605, 34613, 34619 [paratypes]). Scinax squalirostris: URUGUAY: 15 km northeast of San Carlos, Alvarez Farm (MCZ-A 25761 [holotype of Hyla evelynae]). Scinax strussmannae: BRAZIL: Amazonas: Tapauá, Nascentes do Lago Jari National Park (INPA-H 34688 [holotype]; 34689–92, 34700 [paratypes]). Scinax wandae: COLOMBIA: Meta: La Macarena (IAvH-Am 15107, 15256–59, 15263–64, 15276); Casarane: Sabanalarga (ANDES-A 1068, 1072, 1077, 1234).
Appendix 3.
Bayesian phylogenetic inference of Scinax diversification based on a fragment of the 16S rDNA mitochondrial gene, including sequences of the three most closely related genera Julianus, Ololygon, and Sphaenorhynchus. The red rectangle encloses the major clade where the new species is nested; that clade is fully expanded in Figure 2. Two support values are shown along each branch: bootstrap of a Maximum Likelihood analysis (before slash) and Bayesian posterior probability (after slash); values below 75 and 0.95 are omitted, above 95 and 0.95 are denoted by an asterisk.
