Zoological Journal of the Linnean Society, 2020, XX, 1–32. With 13 figures. MARCEL A. CAMINER1,2, and SANTIAGO R. RON1*, Museo de Zoología, Facultad de Ciencias Exactas y Naturales, Pontificia Universidad Católica del Ecuador, Av. 12 de Octubre y Roca, Aptdo. 17-01-2184, Quito, Ecuador 2 Institute of Organismic and Molecular Evolution, Johannes Gutenberg University Mainz, Hanns-DieterHüsch-Weg 15, 55128 Mainz, Germany 1 Received 17 April 2019; revised 13 December 2019; accepted for publication 8 January 2020 The combination of genetic and phenotypic characters for species delimitation has allowed the discovery of many undescribed species of Neotropical amphibians. In this study, we used DNA sequences (genes 12S, 16S, ND1 and COI) and morphologic, bioacoustic and environmental characters of the Boana semilineata group to evaluate their phylogenetic relationships and assess their species limits. In addition, we included DNA sequences of several species of Boana to explore cryptic diversity in other groups. We found three Confirmed Candidate Species (CCS) within the B. semilineata group. Holotype examination of Hyla appendiculata shows that it is a valid species that corresponds to one of the CCS, which is here transferred to Boana. We describe the two remaining CCS. Our phylogeny highlights a number of secondary but meaningful observations that deserve further investigation: (1) populations of B. pellucens from northern Ecuador are more closely related to B. rufitela from Panama than to other Ecuadorian populations of B. pellucens; (2) we report, for the first time, the phylogenetic relationships of B. rubracyla showing that it is closely related to B. rufitela and B. pellucens; and (3) B. cinerascens and B. punctata form two species complexes consisting of several unnamed highly divergent lineages. Each of these lineages likely represents an undescribed species. ADDITIONAL KEYWORDS: Amazon – amphibian – candidate species – cryptic diversity – phylogeny – taxonomy. INTRODUCTION The genus Boana Gray, 1825 is distributed in the Neotropical region from Nicaragua to Argentina and Trinidad and Tobago. Boana is the second most species-rich genus of the family Hylidae, containing 93 recognized species (Frost, 2019) divided into seven species groups (Faivovich et al., 2005): B. alopunctata, B. benitezi, B. faber, B. pellucens, B. pulchella, B. punctata and B. semilineata. A putative synapomorphy for Boana is the presence of a prepollical spine (Faivovich et al., 2005; but see: Peloso et al., 2018). Recent integrative systematic reviews of small clades of Amazonian Boana (e.g. Caminer & Ron, 2014; *Corresponding author. E-mail: santiago.r.ron@gmail.com [ Ve r s i o n o f r e c o r d , p u b l i s h e d o n l i n e 2 1 M a r c h 2020; http://zoobank.org/ urn:lsid:zoobank. org:pub:AD80E63E-EF0E-42FB-8AE3-47B705CA4454] Fouquet et al., 2016; Peloso et al., 2018) have shown that undescribed cryptic species can even outnumber described species. These results highlight the need for systematic reviews of Boana based on genetic data. One clade that has been the subject of recent systematic reviews based on molecular phylogenies is the B. semilineata species group (Fouquet et al., 2016; Peloso et al., 2018). It occurs in the Atlantic Forests, the Amazon, the Chocó region of Colombia and Ecuador, and in eastern Panama (Fouquet et al., 2016). The group contains nine nominal species: B. boans (Linnaeus, 1758), B. diabolica (Fouquet et al., 2016), B. geographica (Spix, 1824), B. hutchinsi (Pyburn & Hall, 1984), B. icamiaba Peloso et al., 2018, B. pombali (Caramaschi et al., 2004), B. secedens (Lutz, 1963), B. semilineata (Spix, 1824) and B. wavrini (Parker, 1936). One putative morphological synapomorphy for the group is the presence of a reticulated lower lid (Faivovich et al., 2005). A species with a challenging © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 1 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 Systematics of the Boana semilineata species group (Anura: Hylidae), with a description of two new species from Amazonian Ecuador 2 M. A. CAMINER and S. R. RON 2016). The only species left with a trans-Andean distribution in lowland rainforests is B. boans. Genetic and morphological analyses of populations on both sides of the Andes are needed to determine whether they represent two or one species. Here, we present a new phylogeny of Boana, including all Ecuadorian species of the genus, some of which had unknown phylogenetic relationships. The main aim of our study is to delimit species boundaries among populations of the B. semilineata species group by combining genetic, morphologic, environmental and bioacoustic characters. In addition, we explore the existence of candidate species in other species groups of Boana by adding genetic characters from additional populations. Finally, we determine whether populations from opposite sides of the Andes of B. boans belong to a single or more species. We predict that populations on opposite sides of the Andes have high genetic divergence as a result of the isolating effect of the Andes. MATERIAL AND METHODS DNA extraction, sequencing and analyses We selected available mitochondrial sequences in GenBank (until October 2018) attributed to 63 of the 93 species formally described for the genus Boana (Frost, 2019). We also included sequences from 118 individuals of Boana from Ecuador (Supporting Information, Table S1). Total genomic DNA was extracted from muscle or liver tissue preserved in 95% ethanol or tissue storage buffer using standard phenol-chloroform extraction protocols (Sambrook et al., 1989). We amplified the mitochondrial genes 12S rRNA, 16S rRNA, NADH dehydrogenase 1 (ND1) and cytochrome c oxidase I (COI). PCR amplification was carried out using standard protocols and primers are presented in Supporting Information, Table S2. PCR products were cleaned using Exo I/SAP digest and sequenced in both directions by Macrogen Co. Ltd (South Korea). Sequences were edited and aligned with GENEIOUS v.5.4.4 software (GeneMatters Corp.) using the GENEIOUS alignment algorithm. Leading and trailing ends were trimmed manually to remove low-quality sequences. The matrix was imported into MESQUITE (v.3.04; Maddison & Maddison, 2015) and the alignment was checked and manually revised. The final combined DNA matrix contained sequence data for: 159 samples for 12S (923 bp, 5% of missing data), 290 for 16S (1767 bp, 56.9%), 136 for ND1 (1072 bp, 4.8%), 138 for COI (670 bp, 1.4%) and 44 for cytochrome b (385 bp, 2.6%). Because it is likely that variation in each of our sampled genes (or codon positions within proteincoding genes) were shaped by different evolutionary © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 taxonomy in this group is Boana geographica. Its distribution range has ~10.4 million km 2, the 13 th largest among the ~8000 described amphibian species (Pimm et al., 2014). Duellman (1973) reported significant geographic variation in life coloration and advertisement call, suggesting the existence of cryptic species. Fouquet et al. (2007) and Jansen et al. (2011) reported deeply divergent lineages within B. geographica, providing additional evidence of the existence of cryptic species. More recently, Fouquet et al. (2016) and Peloso et al. (2018) carried out systematic reviews based on morphological, genetic and bioacoustic evidence that resulted in the identification of seven unnamed candidate species and the description of B. diabolica and B. icamiaba. The unnamed candidate species belong to a clade that includes B. diabolica, B. geographica, B. hutchinsi and B. semilineata. A putative synapomorphy for this clade is the absence of an enlarged prepollex forming a projecting spine (Peloso et al., 2018). The reviews of Fouquet et al. (2016) and Peloso et al. (2018) are significant advances to solve the taxonomy of B. geographica. However, those studies only include one genetic sample of ‘B. geographica’ from Ecuador. We have found morphological differences between sympatric individuals of B. geographica in the Ecuadorian Amazon that suggest the existence of at least two cryptic species there (Ron & Read, 2018). Therefore, a thorough review of Ecuadorian populations of the B. semilineata group is needed. An enigmatic member of the Boana semilineata species group is B. boans. This large hylid is unique among neotropical amphibians, because of its geographic distribution. Unlike other amphibians inhabiting lowland Amazonian rainforests, B. boans also occurs in lowland Chocoan rainforests, on the other side of the Andes (La Marca et al., 2010; Ron et al., 2019). This distribution is unexpected, because the Andes appear to be an impassable barrier for lowland amphibians. Only three other amphibians were thought to be shared between lowland rainforests on opposite sides of the Andes: Rhinella marina (Linnaeus, 1758), R. margaritifera (Laurenti, 1768) and Trachycephalus typhonius (Laurenti, 1768). Although a few anuran species are shared across the Merida Cordillera in Venezuela [e.g. Boana pugnax (Schmidt, 1857), B. xerophylla (Duméril & Bibron, 1841); Escalona et al., 2019], their populations likely connect through adjacent dry lowlands along the Andes. Systematic reviews of Rhinella marina, R. margaritifera and Trachycephalus typhonius have shown that their shared distribution across the Andes was an artefact of incorrect delimitation of species boundaries because Chocoan populations turned out to be distinct species from those in the Amazon (dos Santos et al., 2015; Acevedo et al., 2016; Ron et al., SYSTEMATICS OF BOANA SEMILINEATA GROUP Morphological analyses Our morphological analyses focus on Ecuadorian populations of the Boana semilineata group. The specimens examined for the description of the new species (listed in the type-series and Appendix), were fixed in 10% formalin, preserved in 70% ethanol and deposited in the collections of the Museo de Zoología at Pontificia Universidad Católica del Ecuador (QCAZ) and the Herpetology Division at the Ecuadorian Instituto Nacional de Biodiversidad, INABIO (DHMECN, Quito, Ecuador). We also examined the type material of Hyla appendiculata Boulenger, 1882 (lectotype BMNH 1947.2.13.1) deposited at the Natural History Museum (NHM, London, UK) and specimens identified as Boana from the herpetological collection of Universidad de los Andes (ANDES, Bogotá, Colombia). To evaluate morphometric variation, we measured 133 specimens of ‘Boana geographica’ from Ecuador with digital callipers (± 0.01 mm) according to the methodology described in Duellman (1970, 1973). Webbing formulae of hands and feet follow Savage & Heyer (1967) as modified by Myers & Duellman (1982). Acronyms for morphological measurements are: SVL (snout–vent length), HL (head length), HW (head width), ED (eye diameter), TD (tympanum diameter), TL (tibia length), FEL (femur length), FL (foot length) and CL (calcar length). Sex was determined by gonadal examination or by the presence of vocal sac folds and nuptial excrescences in males. We also examined colour and patterning differences between the new species and their closest relatives. In the Diagnosis sections, coloration always refers to preserved specimens unless otherwise stated. Eleven qualitative morphological characters were evaluated: (1) dorsal skin texture (i, smooth; ii, shagreen; iii, finely granular); (2) dorsal coloration (X-shaped mark on scapular region: i, present; ii, absent); (3) marks on flanks and hidden surfaces of thighs (i, dark vertical bars: ii, dark vertical stripes; iii, uniform black; iv, speckled pattern); (4) ventral skin texture (i, coarsely areolate; ii, areolate; iii, smooth); (5) ventral coloration; (6) in life, webbing coloration; (7) palpebrum reticulated (i, present; ii, absent); (8) nuptial excrescences in males (i, present; ii, absent); (9) projecting prepollices (i, present; ii, absent); (10) tympanum (i, present; ii, absent); and (11) calcar size and shape (i, small and triangular; ii, small and tubercular; iii, absent). Coloration in life was obtained from digital photographs. Bioacoustic analyses We analysed advertisement calls from five individuals of ‘Boana geographica’ recorded at Yasuní (Provincia Orellana), Sardinayacu (Provincia Morona Santiago) and Miazi-Nuevo Paraíso (Provincia Zamora Chinchipe) in Ecuador. Recordings were made with a Sennheiser K6-ME67 directional microphone and Olympus LS-10 and WS-802 digital recorders. Calls © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 processes, we partitioned the data according to gene and codon position to analyse each partition under separate models of evolution. We used software PartitionFinder v.1.1.1 (Lanfear et al., 2012) to simultaneously estimate both the best-fit model for each partition and the best-partition strategy for our data. The phylogenetic relationships among the species were inferred from the concatenated matrix of the mitochondrial genes 12S rRNA, 16S rRNA, ND1, COI and cytochrome b (Cytb). For the outgroup, we added samples of Dendropsophus bifurcus (Andersson, 1945) and D. sarayacuensis (Shreve, 1935). Phylogenetic analyses were conducted using maximum likelihood (ML) and Bayesian inference (BM). Bayesian phylogenetic analyses were carried out in MrBayes v.3.2 (Ronquist et al., 2012). Each Bayesian analysis consisted of four parallel runs of the Metropolis coupled Markov chain Monte Carlo for 60 × 106 generations. Each run had four chains with a temperature of 0.1. For each analysis, the chain was sampled every 1000 generations. The first 25% of sampled trees were discarded as burn-in and the remaining trees were used to estimate the Bayesian tree, posterior probabilities and other model parameters. Convergence of the Markov chain Monte Carlo simulations was assessed by checking if the average standard deviation of split frequencies were < 0.05 between runs. Search convergence and stationarity was visualized and effective using TRACER v.1.5 (Rambaut & Drummond, 2007). TRACER was also used to obtain estimate sample-size values (ESS) for parameter estimates. We assumed that ESS values > 200 were indicative of an adequate search. Maximum likelihood analyses were conducted using GARLI v.2.0 (Genetic Algorithm for Rapid Likelihood Inference; Zwickl, 2006) under default settings. Analysis was terminated after 500 000 generations without topology improvement. We ran a total of 20 independent searches and used random starting trees (streefname = random) to reduce the probability of inferring a suboptimal likelihood solution. Likelihood values of the 20 searches differed by less than 1.0 likelihood units between each other, indicating that all searches converged on similar optimal trees. Nodal support was obtained using 100 bootstrap replicates. MESQUITE (Maddison & Maddison, 2015) was used to generate a majority-rule consensus tree from the bootstrap replicates. Nodes corresponding to partitions present in less than 50% of the bootstrap replicates were collapsed. 3 4 M. A. CAMINER and S. R. RON Environmental analyses We considered environmental differentiation as additional evidence to define species boundaries (e.g. Rissler & Apodaca, 2007; Aguilar-Puntriano et al., 2016). We obtained bioclimatic values at localities of the B. semilineata species group using software QGIS v.3.4 (QGIS Development Team, 2018). Rasters of bioclimatic variables, with 1 km resolution, were downloaded from WorldClim v.2.0 (http://www.worldclim.org; Fick & Hijmans, 2017). The variables were: (1) Annual Mean Temperature, (2) Mean Diurnal Range, (3) Isothermality, (4) Temperature Seasonality (standard deviation *100), (5) Max. Temperature of Warmest Month, (6) Min. Temperature of Coldest Month, (7) Temperature Annual Range, (8) Mean Temperature Table 1. Character loadings, eigenvalues, and percentage of explained variance for Principal Components (PC) I–II. The analysis was based on eight acoustic variables from the advertisement calls of Boana appendiculata, B. nigra and B. ventrimaculata. Bold figures indicate highest loadings Variable PC I PC II Call duration Number of notes per call Rise time of the call Pulses per note Pulse rate Note duration Dominant frequency call Frequency bandwidth Eigenvalue % 0.86 0.94 0.65 –0.2 0.24 –0.40 0.77 0.84 3.64 45.58 0.40 –0.17 0.71 0.90 –0.84 0.89 –0.26 0.36 3.23 40.46 of Wettest Quarter, (9) Mean Temperature of Driest Quarter, (10) Mean Temperature of Warmest Quarter, (11) Mean Temperature of Coldest Quarter, (12) Annual Precipitation, (13) Precipitation of Wettest Month, (14) Precipitation of Driest Month, (15) Precipitation Seasonality (Coefficient of Variation), (16) Precipitation of Wettest Quarter, (17) Precipitation of Driest Quarter, (18) Precipitation of Warmest Quarter and (19) Precipitation of Coldest Quarter. To avoid spatial autocorrelation, we reduced the number of localities using a buffer with a 5-km radius around each point. If two or more localities of the same species were at a distance < 10 km from each other, we randomly chose one of them. We obtained a total of 88 localities from individuals sequenced or morphologically identified as belonging to the B. semilineata species group. Since species identification in the literature could be incorrect, we only included geographic records based on specimens examined by us. We used Principal Component Analysis (PCA) to examine environmental differentiation between candidate species using software JMP v.9.0.1 (SAS Institute, 2010). Species limits Species limits were evaluated under the framework of integrative taxonomy (Dayrat, 2005; Vieites et al., 2009). We identified undescribed species by using the categories defined by Vieites et al. (2009). During the first stage, we partitioned genetic lineages into candidate species using two criteria. First, we used a Poisson Tree Processes (PTP) model (Zhang et al., 2013) to infer the most likely number of species within the Boana geographica–B. semilineata complex in our dataset, as implemented in the bPTP server (http:// species.h-its.org/ptp/). The input tree was obtained © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 of B. semilineata recorded by Axel Kwet from Brusque (Santa Catarina, Brazil) and B. geographica described by Duellman & Pyles (1983) from Santa Cecilia (Sucumbíos, Ecuador) were also analysed (downloaded from Fonoteca Zoológica of MNCN, http://www.fonozoo. com and the Macaulay Library at the Cornell Lab of Ornithology, http://macaulaylibrary.org). Spectrograms were generated with RAVEN v.1.5 software (2010) using a Fast Fourier Transformation (FFT) from a sample of 1024 points and a frequency resolution of 44.1 Hz. Recordings are deposited at the QCAZ audio archive (available at the Anfibios del Ecuador website: https://bioweb.bio/faunaweb/amphibiaweb/). Ten acoustic parameters (modified from: Cocroft & Ryan, 1995) were measured to describe the structure of each call: (1) call duration = time from beginning to end of the call, measured in the oscillogram; (2) notes per call = number of notes per call; (3) note duration = average time from beginning to end of a note; (4) pulses per call = number of pulses per call; (5) pulses per note = number of pulses per note; (6) pulse rate = number of pulses repeated in a defined period of time within a note; (7) interval between notes = average time from ending of a note and beginning of next note; (8) call rise time = time from beginning of call to point of maximum amplitude; (9) dominant frequency = frequency in call containing the greatest amount of energy, measured along the entire call; and (10) frequency bandwidth = difference between the upper and lower frequencies measured along the entire call. Several calls of each individual were measured to calculate an average. The species comparisons were based on a single value (average) per male. Eight call variables (listed in Table 1) were used to run a Principal Components Analysis (PCA) to assess acoustic differentiation between calls from candidate species using software JMP v.9.0.1 (SAS Institute, 2010). SYSTEMATICS OF BOANA SEMILINEATA GROUP without character data to compare, were assigned to the Unconfirmed Candidate Species category. As a general reference we also obtained uncorrected mean genetic distances using MEGA v.6.06 (Tamura et al., 2011) for gene 16S, given that this is a reference standard to identify candidate species (e.g. Fouquet et al., 2007; Jansen et al., 2011; Caminer et al., 2017). We only used sequences that overlapped by more than 330 bp to calculate the distances. RESULTS Phylogenetic relationships The complete DNA sequence data matrix has 4925 bp for 304 samples. PartitionFinder chose four partitions as the best strategy (best model in parentheses): 12S, 16S, COI 3 rd , ND1 3 rd and Cytb 3 rd position (GTR + I + G), COI 1st, ND1 1st and Cytb 1st position (SYM + I + G), COI 2nd position (F81), ND1 2 nd and Cytb 2 nd position (HKY + I + G). The tree topologies for the maximum likelihood and Bayesian phylogenies are similar except for weakly supported nodes (posterior probability, pp < 0.5 and bootstrap < 52). The phylogeny shows strong support for the B. semilineata species group (Fig. 1A, B).Within it, a wellsupported clade includes all species except B. pombali and B. icamiaba. The next most-inclusive clade is composed of all species except B. boans. We also found strong support for the B. geographica–B. semilineata clade (sensu Peloso et al., 2018). Hereafter we refer to it as the B. geographica–B. semilineata species complex. This clade also includes ‘B. hutchinsi’, B. diabolica and several undescribed species. Their status is assessed in the ‘Identification of candidate species’ section. Surprisingly, populations of B. boans on both sides of the Andes are separated by short branches and low genetic distances. Uncorrected-p distances for gene 16S range from 0.7 to 0.9% (Chocó vs. Amazonian Ecuador). Chocoan populations form a well-supported clade. Samples from Amazonian Ecuador also form a clade, sister to Chocoan samples. Samples from Amazonian Brazil, Colombia, Peru and the Guianas also show short branches and weak phylogenetic structure. Regarding other species groups of Boana, our phylogeny (Fig. 1C) is consistent with the strongly supported species groups found by Duellman et al. (2016). A notable difference occurs in the Boana calcarata species complex, which shows a clade containing B. almendarizae (Caminer & Ron, 2014), B. calcarata (Troschel, 1848), B. fasciata (Günther, 1858) and B. maculateralis (Caminer & Ron, 2014). However, the Duellman et al. (2016) tree shows B. fasciata and B. maculateralis more closely related to © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 from the 16S gene matrix alignment using maximum likelihood. We ran 100 000 MCMC generations, with a thinning value of 100 and a burn-in of 10%. Second, we partitioned genetic lineages into candidate species using the Automatic Barcode Gap Discovery (ABGD) method (available at https://bioinfo. mnhn.fr/abi/public/abgd; Puillandre et al., 2012). This distance-based approach identifies a barcode gap in the frequency distribution of pairwise genetic distances among samples. Based on that gap, it proposes candidate species among samples. ABGD has shown good performance to identify candidate species compared to other DNA barcode methods (Paz & Crawford, 2012). The analyses were based on the 16S gene matrix, Kimura two-parameter distances, a prior for the maximum value of intraspecific divergence from 0.001 to 0.05, ten recursive steps and a gap width of 0.5. A consensus between the candidate species found by bPTP and ABGD was obtained by applying a conservative criterion to minimize type I errors (e.g. one species incorrectly split into two or more). In case of disagreements between bPTP and ABGD, we chose the alternative, implying the lower number of candidate species. On a second stage, we assigned the candidate species to two mutually exclusive categories: (1) Unconfirmed Candidate Species (UCS) and (2) Confirmed Candidate Species (CCS; Vieites et al. 2009). Confirmed Candidate Species were those showing species-level differences in adult morphological, bioacoustic or environmental characters when compared to its most closely related species. We assumed that species-level differences in morphology were: (1) fixed differences in qualitative morphological characters between sister species or (2) quantitative morphological differences between sister species. For both categories we evaluated morphological characters of known diagnostic value in Boana. For comparisons of bioacoustic characters we applied Kohler et al. (2017) criteria to define species limits. Specifically, we assumed that qualitative call differences (Kohler et al., 2017: 46) or quantitative differences in note duration, dominant frequency or pulse rate, which are useful to assess species limits (Kohler et al., 2017: 51), were indicative of species-level differences. We also compared environmental envelopes between sister species (see ‘Environmental analyses’ section) as additional evidence of species distinctiveness. Finally, the existence of sympatric regions between candidate species was also considered evidence of lineage independence. This is because deep genetic divergence in sympatry is indicative of lack of gene flow and the existence of non-geographic reproductive barriers (and, therefore, lineage independence). Candidate species not showing those differences (morphological, bioacoustical or environmental), or 5 6 M. A. CAMINER and S. R. RON B. tetete and B. alfaroi. Our topology for those species is the same as that reported by Caminer & Ron (2014). The Boana pellucens species group has strong support and includes samples of B. rubracyla (Cochran & Goin, 1970), which are sister to the clade B. rufitela + B. pellucens (pp = 1 and bootstrap = 100). Populations from Durango and Lita, Ecuador, of ‘B. pellucens’ are more closely related to B. rufitela © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 Figure 1. A, Maximum likelihood phylogram depicting relationships within Boana. The phylogram was derived from analysis of 4916 bp of mitochondrial DNA (gene fragments 12S, 16S, ND1, COI and Cytb). Museum catalogue numbers and localities are shown for each sample. Bootstrap values are above the branches and Bayesian posterior probabilities are below; asterisks indicate values > 99 (bootstrap) or > 0.99 (posterior probability); missing values indicate values < 50 or 0.5. Outgroup species are not shown. Abbreviations are: AR = Argentina, BO = Bolivia, BR = Brazil, CO = Colombia, COS = Costa Rica, EC = Ecuador, FG = French Guiana, GU = Guyana, PA = Panama, PE = Peru, SU = Suriname, VE = Venezuela. Vertical bars in the B. semilineata group indicate species partitions inferred using genetic (bPTP and ABGD), morphologic, and bioacoustic criteria. Horizontal grey boxes in 1c indicate uncorrected genetic distances > 2% for the mitochondrial gene 16S. See text for details. SYSTEMATICS OF BOANA SEMILINEATA GROUP 7 than to other populations of B. pellucens (pp = 1 and bootstrap = 99; Fig. 1C). Identification of candidate species The bPTP analysis recovered 17 species (Figs 1A, B; Supporting Information, Fig. S1). The ABGD method recovered 14 species across a wide range of priors for maximal intraspecific thresholds (0.024–0.088; Figs 1A, 1B and Supporting Information, Fig. S2). The results of both methods are highly similar except for ABGD lumping into a single species, four species detected in bPTP (part of the B. semilineata species complex in Fig. 1B). Based on our conservative criterion, we chose the 14 species partition for subsequent analyses. Of the 14 species, six correspond to B. boans, B. diabolica, B. geographica, B. icamiaba, B. semilineata and B. wavrini; its assignment has been documented before (Fouquet et al., 2016; Peloso et al., 2018). Boana pombali is of uncertain assignment, because our bPTP and © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 Figure 1. B, Continued. 8 M. A. CAMINER and S. R. RON ABGD analyses identified two candidate species within populations identified as ‘B. pombali’. Resolution of the status of those populations falls outside the scope of our review and we leave it unresolved. Therefore, six candidate species could be undescribed (clades A, B, C, D, E and F in Fig. 1A, B). Morphological, acoustic and environmental data are available for clades A, B and D, which allows us to assign them to the © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 Figure 1. C, Continued. SYSTEMATICS OF BOANA SEMILINEATA GROUP Morphometric comparisons The following comparisons pertain to male SVL. There are significant differences between the three species. Boana ventrimaculata sp. nov. (clade D) is larger than B. appendiculata (clade A) and B. nigra sp. nov. (clade B) (all P values for t tests < 0.001), while B. nigra sp. nov. is smaller than B. appendiculata (t = 2.92, d.f. = 56.4, P = 0.005). Relative tympanum size of males and females of B. ventrimaculata sp. nov. is larger than that of B. appendiculata and B. nigra sp. nov. (all P values for t tests < 0.001), while B. nigra sp. nov. presents a smaller tympanum compared to B. appendiculata (all P values for t tests < 0.03; comparisons were made on the residuals of the regression between the tympanum diameter and SVL). In the DFA classification, most specimens are assigned to the correct species. In B. appendiculata, 29 out of 35 males are correctly classified (four are misclassified as B. nigra sp. nov. and two as B. ventrimaculata sp. nov.); in B. nigra sp. nov., three out of 26 males are incorrectly classified (all as B. appendiculata); in B. ventrimaculata sp. nov., two out of 20 males are incorrectly classified (as B. appendiculata) Among females, in B. appendiculata, 27 out of 30 specimens are correctly classified (three are misclassified as B. nigra sp. nov.); in B. nigra sp. nov., all five females are correctly classified; in B. ventrimaculata, only one out of 16 females is incorrectly classified (as B. nigra sp. nov.). Overall, the DFA show morphometric differentiation among species, specially between B. ventrimaculata sp. nov. and B. nigra sp. nov. Bioacoustic comparisons Our PCA of advertisement calls from six males result in two PCs with eigenvalues > 1.0. These two PCs Figure 2. Phylogenetic relationships within the Boana semilineata group showing the Confirmed Candidate Species (CCS) and Unconfirmed Candidate Species (UCS). © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 Confirmed Candidate Species category (Fig. 2). Clade A is sister to clade B; they differ from each other in environmental envelope and qualitative morphological characters (see species descriptions for details). Clade A corresponds to Hyla appendiculata Boulenger, 1882, a binomen mistakenly considered a junior synonym of B. geographica (see ‘Taxonomic review’ section). Clade B is a new species that we describe below as Boana nigra sp. nov. Clade D is sister to a clade that includes B. semilineata and Clade C. Calls of clade D and B. semilineata show quantitative differences in note duration, dominant frequency and pulse rate (ranges do not overlap between species; Table 2). Although our sample size was low (one male in B. ventrimaculata sp. nov. and two males in B. semilineata), note duration, dominant frequency and pulse rate are static variables and have low intraspecific variation (Kohler et al., 2017). Clades C, E and F are left as Unconfirmed Candidate Species due to lack of data. The following sections show the comparisons between B. appendiculata, B. nigra sp. nov. and B. ventrimaculata sp. nov. 9 © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 2.07 ± 1.6 (1.22–3.92) 5.16 ± 1.04 (4–6) 0.17 ± 0.02 (0.15–0.17) 94 ± 26.8 (75–113) 20.5 ± 1.6 (18–22) 120.4 ± 18.2 (107.6–133.3) 0.16 ± 0.1 (0.1–0.27) 0.44 ± 0.13 (0.35–0.53) 1387.6 ± 65.7 (1341.2–1434) 1841.9 ± 46.8 (1808.8–1875) 0.45 ± 0.04 (0.43–0.51) 1 0.45 ± 0.04 (0.43–0.51) 33.9 ± 2.6 (31–37) 33.9 ± 2.6 (31–37) 75.8 ± 9.2 (62.5–78.7) NA 0.34 ± 0.03 (0.32–0.39) 974.2 ± 183.3 (795.2–1205.6) 1316.2 ± 258.1 (1076.6–1645.2) Type II 890.1 ± 65.8 (818.3–947.5) 0.25 ± 0.11 (0.15–0.22) 1044.2 ± 41.4 (996.6–1063.8) 87.9 ± 17.4 (76.3–108) NA 28.3 ± 8.1 (21–37) 0.33 ± 0.11 (0.25–0.46) 1 0.33 ± 0.11 (0.25–0.46) 28.3 ± 8.1 (21–37) B. nigra (N = 1) 680.5 ± 93.4 (559.9–818.3) 0.08 ± 0.04 (0.02–0.11) 561.4 ± 79.1 (475.2–685.4) 57.5 ± 3.3 (53.5–61.9) NA 13.8 ± 1.9 (11–16) 0.24 ± 0.03 (0.18–0.28) 1 0.24 ± 0.03 (0.18–0.28) 13.8 ± 1.9 (11–16) B. ventrimaculata (N = 1) 1068.1 ± 146.2 (964.6–1171.4) 0.14 ± 0.01 (0.13–0.15) 1240.3 ± 194.8 (1102.5–1378.1) 258.8 ± 30.1 (237.6–280.1) 0.04 ± 0.002 (0.04–0.05) 11.7 ± 0.6 (9–15) 0.14 ± 0.01 (0.13–0.15) 2.1 ± 0.14 (1–3) 0.04 ± 0.002 (0.04–0.05) 23.6 ± 0.28 (14–32) B. semilineata (N = 2) Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 Call duration (s) Notes per call Note duration (s) Pulses number per call Pulses number per note Pulse rate Interval between notes (s) Call rise time (s) Dominant frequency (Hz) Frequency bandwidth (Hz) Type I B. appendiculata (N = 5) Table 2. Descriptive statistics for calls parameters of members of the Boana semilineata group. Values for B. nigra and B. ventrimaculata were obtained from five calls from a single male. Call vouchers: B. appendiculata (KU 143144–45, QCAZ 67113–34), B. nigra (DHMECN 12130) and B. ventrimaculata (QCAZ 49233). The N values indicate the number of males analysed. If available, several calls of each individual were measured to calculate an average. When N > 1, descriptive statistics are based on those averages only (one value per male); when N = 1, we report descriptive statistics for several calls of the same male. Mean ± SD is given with range in parentheses. Calls of B. nigra, B. ventrimaculata and B. semilineata are of Type I only 10 M. A. CAMINER and S. R. RON SYSTEMATICS OF BOANA SEMILINEATA GROUP Environmental comparisons Two components with eigenvalues > 1.0 are extracted from the PCA for 88 localities. The two components account for 81.2% of the total variation. The first component (PC I) explains 41.9% of the total variation and the second component (PC II) explains 39.3%. There is partial segregation between B. nigra sp. nov. and the other species along PC I but specially PC II, a component Figure 3. Axes I and II from Principal Components Analysis. A and B: based of six size–corrected morphological variables for 81 males (A) and 52 females (B) of Boana. C: based on eight advertisement calls variables of B. appendiculata (four males), B. nigra (1), B. ventrimaculata (1), and B. semilineata (2). The only calls type II of B. appendiculata are shown in hollow green triangles. D: based on 19 bioclimatic variables for 88 localities. See Tables 1, 3, and 4 for character loadings on each component. © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 account for 86.05% of the total variation. PC I (45.6% of the variance) has high loadings on number of notes and call duration; PC II (40.5% of the variance) has high loadings on number of pulses per note and note duration (Table 1). The acoustic space (as represented by PC I and PC II) shows segregation among the advertisement calls of the three species (Fig. 3C). We also included calls of B. semilineata, which separate from the other species along PC II. 11 12 M. A. CAMINER and S. R. RON Variable Foot length Head length Head width Tympanum diameter Tibia length Femur length Eigenvalue % PCA Males PCA Females PC I PC II PC I PC II 0.31 –0.10 0.47 0.16 0.85 0.90 1.91 31.94 –0.74 0.43 0.37 0.82 0.02 0.01 1.56 26.06 0.39 0.31 0.68 0.13 0.89 0.83 2.14 37.41 0.41 –0.57 0.35 0.69 –0.05 –0.32 1.21 20.26 Table 4. Character loadings, eigenvalues and percentage of explained variance for Principal Components (PC) I– II of environmental variables. The analysis was based on 88 localities of Boana appendiculata, B. nigra and B. ventrimaculata. Bold figures indicate highest loadings Variable PC I PC II BIO1 = Annual Mean Temperature BIO2 = Mean Diurnal Range BIO3 = Isothermality (BIO2/BIO7) (* 100) BIO4 = Temperature Seasonality (standard deviation *100) BIO5 = Max Temperature of Warmest Month BIO6 = Min Temperature of Coldest Month BIO7 = Temperature Annual Range (BIO5–BIO6) BIO8 = Mean Temperature of Wettest Quarter BIO9 = Mean Temperature of Driest Quarter BIO10 = Mean Temperature of Warmest Quarter BIO11 = Mean Temperature of Coldest Quarter BIO12 = Annual Precipitation BIO13 = Precipitation of Wettest Month BIO14 = Precipitation of Driest Month BIO15 = Precipitation Seasonality (Coefficient of Variation) BIO16 = Precipitation of Wettest Quarter BIO17 = Precipitation of Driest Quarter BIO18 = Precipitation of Warmest Quarter BIO19 = Precipitation of Coldest Quarter Eigenvalue % 0.34 –0.11 –0.2 0.24 0.04 –0.24 –0.12 0.09 0.34 0.33 –0.01 0.07 0.08 –0.2 0.34 0.03 0.34 0.34 0.06 0.04 0.34 0.04 –0.05 –0.07 –0.08 0.04 0.35 0.33 0.34 –0.19 –0.05 –0.09 –0.09 –0.08 7.96 41.94 0.34 0.33 0.33 0.33 7.46 39.26 with high loadings in precipitation-related variables (all P values for t tests < 0.03; Fig. 3D; Table 4). Taxonomic review of the Boana geographica– Boana semilineata species complex Fouquet et al. (2016) assigned the binomen ‘Hypsiboas geographicus’ to populations from Tefé, Purus, Serra do Divisor (Brazil) and Concepción (Bolivia), which form a single clade. We agree with the assignment of Boana geographica, given its type locality ‘flumen Teffé’. Fouquet et al. (2016) assigned Boana semilineata to two specimens from Rio de Janeiro (e.g. specimen CFBH 5424 in Fig. 1B). We follow that assignment and highlight that our bPTP and ABGD analyses indicate that the sister clade of B. ventrimaculata is a species complex. The assignment of other currently recognized binomials is shown in Figure 1 (A, B). Available junior synonyms in the Boana geographica–Boana semilineata species complex Junior synonyms are: Hyla appendiculata Boulenger, 1882; Hylella punctatissima Reinhardt & Lutken, 1862; Centrotelma cryptomelan Cope, 1867; and Cophomantis punctillata Peters, 1870. Except for Hyla appendiculata, all type localities are in the Atlantic Forest of Brazil and are at a distance of > 2000 km from the distribution range of our Confirmed Candidate Species. Therefore, those binomials are unlikely to be applicable to them. The type locality of Hyla appendiculata, erroneously synonymized under B. geographica, overlaps with the distribution range of two Confirmed Candidate Species: clades A and D (Fig. 6). Examination of the lectotype and published descriptions from the literature allows us to assign the name Hyla appendiculata to clade A (Fig. 1). We document this assignment in the following section. Taxonomic status of Hyla appendiculata The lectotype of Hyla appendiculata is an adult female with SVL = 63.70 mm (BMNH 1947.2.13.1 formerly 80.12.5.168; Fig. 4) from Canelos, Provincia Pa s t a z a , E c u a d o r. H y l a a p p e n d i c u l a t a w a s synonymized under Hyla geographica geographica by Parker (1935; for a detailed discussion see: Duellman, 1973). The lectotype closely resembles individuals of B. geographica s.s. and clade A, which also have brown-paired vertical stripes on the flanks and hidden surfaces of the thighs (Fig. 5). The type locality lies within the distribution range of clade A (Fig. 6) and at a distance of 780 km from the geographic range of B. geographica s.s. Based on its resemblance to the holotype and its geographic distribution, we conclude that clade A is conspecific with Hyla appendiculata, which we remove from synonymy of B. geographica. This taxonomic decision results in the new binomial Boana appendiculata. © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 Table 3. Character loadings, eigenvalues, and percentage of explained variance for Principal Components (PC) I–II. The analysis was based on six morphometric variables of adult Boana appendiculata, B. nigra and B. ventrimaculata. Bold figures indicate highest loadings SYSTEMATICS OF BOANA SEMILINEATA GROUP 13 Figure 5. Adult preserved specimens of Boana appendiculata showing variation in dorsal and ventral coloration. From left to right, first and second rows: QCAZ 64085, 25955, 28227, 30921 (males), 31035 (subadult female), 36932, 36934, 39416, 43850 (males); third and fourth rows: QCAZ 44160, 44168, 44174, 52610, 52734 (males), 37873, 5076, 25797 (females). See Appendix for locality data. All specimens are shown at the same scale. Identity of Boana hutchinsi According to Peloso et al. (2018), specimens of clade D (clade S5 in that publication) represent B. hutchinsi. This assignment was based on comparisons between specimens from Vaupés (Colombia), Amazonas (Brazil) and the type material of B. hutchinsi (holotype UTA-A 24819 and paratypes 24830 and 24835). Nevertheless, a thorough review of the description of B. hutchinsi indicates that it © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 Figure 4. Dorsal and ventral views of the lectotype of Hyla appendiculata (BMNH 1947.2.13.1; adult female). 14 M. A. CAMINER and S. R. RON is not conspecific with clade D. We base this conclusion on the original description of type material (Pyburn & Hall, 1984), which differs from that of clade D: webbing between second, third and fourth fingers black in adults (yellowish white, brown or cream in clade D), dorsum lavender-grey in recently metamorphosed individuals (brown with black flecks in clade D) and advertisement calls having four to six notes with a mean duration of 0.68 s (one note with mean duration of 0.24 s in clade D). The available evidence indicates that B. hutchinsi have not been included in any molecular phylogeny. Systematic accounts Boana appendiculata (Boulenger, 1882) comb. nov. (Figs 4, 5, 7B; Table 5) Hyla appendiculata Boulenger, 1882. Lectotype (Fig. 4): BMNH 1947.2.13.1 (designated by Parker, 1933), an adult female from ‘Canelos’, Ecuador. Common names: Proposed standard English name: Canelos treefrog. Proposed standard Spanish name: rana arbórea de Canelos Diagnosis (Figs 5, 7B): Boana appendiculata is characterized by: (1) mean SVL 45 mm in males (range 38.5–53.1; N = 96), 63.8 mm in females (range 50.2–72.6; N = 40; Table 5); (2) thighs long (TL/SVL 0.45–0.56); (3) dorsal skin finely granular; (4) dorsum coloration dull tan, brown, dark brown or cream usually with irregular dark brown markings frequently including an X-shaped mark in scapular region; (5) flanks and hidden surfaces of thighs creamy or pale grey with black vertical stripes (either single or paired); (6) areolate skin on the posterior surface of throat, belly and ventral surfaces of thighs; (7) venter vary from creamy white to yellowish white or orange, with or without dark brown spots; (8) in life, webbing yellow, orange or brown; (9) palpebrum reticulated; (10) nuptial excrescences present in breeding males; (11) projecting prepollices absent in both sexes; (12) tympanum large with distinct tympanic ring; (13) small, triangular calcar on heel, often coloured © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 Figure 6. Distribution of Boana geographica, B. appendiculata, B. nigra, B. ventrimaculata, and B. sp. (clade F in Fig. 1B). Localities are based on adult and sequenced specimens deposited at Museo de Zoología of Pontificia Universidad Católica del Ecuador and from Fouquet et al. (2016). F M F M F Mean ± SD is given with range in parentheses. Abbreviations are: SVL = Snout–vent length; HL = Head length; HW = Head width; ED = Eye diameter; TD = Tympanum diameter; TL = Tibia length; FEL = Femur length; FL = Foot length; CL = Calcar length; F = females; M = males. All measurements are in mm. 1.5 ± 0.4 (0.7–2.27) 2.1 ± 0.4 (1.1–2.7) 1.1 ± 0.5 (0.9–1.6) 2.3 ± 0.4 (1.9–2.8) 1.7 ± 0.4 (1–2.4) 2.5 ± 0.5 (1.6–3.7) 16.3 ± 1.8 (13–19.8) 24 ± 2.1 (18.4–28) 16.4 ± 0.9 (14.9–18.6) 25.2 ± 2.8 (21.2–28.8) 19.2 ± 1.9 (15.8–23.6) 31 ± 2.3 (27.8–34.4) B. appendiculata (N = 35) B. appendiculata (N = 30) B. nigra (N = 26) B. nigra (N = 5) B. ventrimaculata (N = 20) B. ventrimaculata (N = 17) M 44.3 ± 3.5 (38.5–51.9) 64.4 ± 4.8 (53.5–72.6) 42.2 ± 2.0 (38.5–46.3) 64.8 ± 7.9 (56.4–76.7) 52.7 ± 4.6 (44.5–60.7) 81.6 ± 5.4 (72.9–90.5) 11.5 ± 1.5 (8–14.3) 15.6 ± 1.9 (12.1–18.6) 10.8 ± 1.3 (8.8–13.9) 15.2 ± 3.1 (11.2–19.5) 12.9 ± 1.9 (9.8–17.3) 18.9 ± 2.6 (14.6–22.9) 16.6 ± 1.5 (11.9–18.4) 22.7 ± 1.5 (19.1–25) 14.5 ± 1.2 (10.2–16.4) 23.5 ± 2.8 (21–27.9) 18.8 ± 1.6 (16.2–22.2) 29.9 ± 2.3 (26.6–34.9) 4.6 ± 0.5 (3.7–5.6) 5 ± 0.5 (4–5.9) 4.1 ± 0.5 (3.4–5.5) 5.3 ± 0.5 (4.6–5.8) 5.3 ± 0.6 (3.9–6.3) 6.7 ± 0.8 (5.3–8.4) 3.9 ± 0.4 (3–4.8) 5.1 ± 0.5 (4–5.8) 3.3 ± 0.3 (2.8–3.9) 4.6 ± 0.3 (4.3–5.1) 4.8 ± 0.5 (3.7–5.6) 6.4 ± 0.8 (5.1–7.7) 21.8 ± 2.1 (18.1–25.8) 32.9 ± 2.7 (26.9–37.8) 20.7 ± 1.2 (18.5–23.5) 33 ± 4.1 (29.3–38.2) 26.8 ± 2 (22.8–29.8) 41.9 ± 2.9 (36.3–47.8) 22.8 ± 2 (19.1–26.2) 33.9 ± 2.2 (26.9–37.7) 21.9 ± 0.8 (20.4–23.6) 34.5 ± 4.4 (31.1–40.2) 27.3 ± 1.9 (23.8–30.9) 42.4 ± 2.8 (36.9–47.9) CL FL FEL TL TD ED HW HL SVL Species 15 with a cream spot; (14) advertisement call consists of two types of calls: Type I with one note and Type II with four to six notes; (15) pulses per note varying from 31 to 37 (Type I) and 18 to 22 (Type II); and (16) call dominant frequency varying from 1076–1645 Hz (Type I) and 818–947 Hz (Type II). Comparisons with other species: Boana appendiculata is most similar to B. semilineata, B. diabolica, B. geographica and B. hutchinsi. It differs from the last three species by having yellow to orange or brown webbing in life (webbing black in B. hutchinsi, red in B. diabolica and B. geographica; Fouquet et al., 2016). Although webbing between fingers III–IV and toes IV–V is grey dorsally in B. geographica, colour between the other fingers differs as indicated above. Bright webbing colours fade to cream in preservative and become indistinguishable among species. Boana appendiculata can be further distinguished by the presence of black vertical stripes on the flanks (speckled pattern in B. diabolica, B. hutchinsi and B. semilineata; Fouquet et al., 2016). Boana appendiculata differs from B. ventrimaculata sp. nov. in having black vertical stripes on the flanks (a speckled pattern or dark blotches in B. ventrimaculata sp. nov.) and by its smaller size (differences are significant in males: Student’s t = 7.16, d.f. = 31.9, P < 0.001). Boana appendiculata is similar to B. nigra sp. nov. but it differs in having a combination of yellow or orange ventral coloration in life (cream to pale orange in B. nigra sp. nov.) and black vertical stripes on the flanks and hidden surfaces of thighs (uniform black in males of B. nigra sp. nov.; Fig. 7A, 10). However, small adult specimens of B. appendiculata (< 42 mm) may present a similar colour pattern to B. nigra sp. nov. (cream ventral coloration with uniform black flanks). Both species also differ in environmental envelope (Fig. 3D; scores along PC I and PC II are significant different: PC I, t = 8.51, P < 0.001; PC II, t = 2.55, P = 0.034). Boana appendiculata occurs in warmer localities than B. nigra sp. nov. as shown by annual mean temperature (bio01): B. appendiculata mean = 24.28 ËšC (SD = 1.53, N = 59) vs. B. nigra mean = 19.48 ËšC (SD = 1.78, N = 8). Despite high genetic distances (4.2–5.2% for gene 16S; Table 6), both species are sympatric in at least one locality in Pastaza Province. This suggests the existence of non-geographic reproductive barriers and, therefore, lineage independence. Boana appendiculata is easily distinguished from B. boans, B. icamiaba, B. pombali and B. wavrini by its prepollex not modified into a projecting spine and by the presence of pigmented nuptial excrescences. Boana appendiculata further differs from B. boans in having well-defined vertical stripes on the flanks and hidden surfaces of thighs (diffuse vertical bars and barely distinguishable in B. boans), and webbing © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 Table 5. Descriptive statistics for morphometric of Boana appendiculata, B. nigra and B. ventrimaculata used for Principal Component Analysis SYSTEMATICS OF BOANA SEMILINEATA GROUP 16 M. A. CAMINER and S. R. RON extends to about the middle of the outer finger (fully webbed hands in B. boans). Variation in preservative: Morphometric data for adult specimens are summarized in Table 5, while variation in dorsal and ventral coloration of preserved specimens is shown in Figure 5. The most noticeable variation occurs in the dorsal coloration and in the shape of the calcars. Calcar size varies from inconspicuous (e.g. QCAZ 44174) to prominent (e.g. QCAZ 43850). Sexual dimorphism was observed in SVL, with 38.5–51.9 mm (44.3 ± 3.5 mm, N = 35) in males and 53.5–72.6 mm (64.4 ± 4.8 mm, N = 30) in females. Females are significantly larger than males (t = –20.13, d.f. = 52, P < 0.001). Head shape varies between subacuminate and rounded in dorsal view; in lateral view it varies between truncate and short, rounded. The head is slightly wider than the body in most of the specimens. Breeding males have keratinized nuptial excrescences on the inner surfaces of the thumbs. Webbing on the hand varies slightly among specimens: fingers I basal II (1+–1½)—(2–3–) III (1½–2½)—(1–2) IV. Variation of webbing between toes is I (1–2–)—(1–2–) II (1–1½)— (1–2+) III (1–1+)—(1–2+) IV (1–2+)—(1––1+) V. Vomerine teeth (N = 15): 10–14 (right = 11.5) and 10–15 (left = 12.1). Background dorsal coloration varies from dull tan (e.g. QCAZ 36932), brown (e.g. QCAZ 25955), dark brown (e.g. QCAZ 64085) to cream (e.g. QCAZ 44167) with irregular faint to well-defined dark brown marks in diverse patterns. Most frequently there is an X-shaped mark on the scapular region (e.g. QCAZ 25797, 44168, 52734). A dark brown transversal band (e.g. QCAZ 36934, 52610) or sometimes two narrow transversal stripes interconnected (e.g. QCAZ 44168) are usually present on the sacral region. In some individuals © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 Figure 7. Dorsal and ventral views of adults of: A, Boana nigra, QCAZ 58786 (holotype; SVL = 42.53 mm), 58783 (SVL = 45.95 mm), 59114 (SVL = 45 mm), adult males, 58798 (SVL = 68.03 mm), adult female; B, B. appendiculata, QCAZ 64085 (SVL = 45.67 mm), adult male, 52845 (SVL = 66.82 mm), adult female; C, B. ventrimaculata, QCAZ 51241 (holotype; SVL = 57.57 mm), adult male, 39137 (SVL = 79.98 mm), adult female. N=2 0 0.086 0.077 0.068 (0.061–0.074) 0.062 (0.057–0.069) 0.082 (0.079–0.088) 0.061 (0.059–0.067) 0.078 (0.074–0.086) 0.059 (0.057–0.064) 0.044 (0.032–0.059) 0.049 (0.049–0.052) 0 0.045 (0.035–0.059) 0.049 (0.044–0.052) 0.035 N=9 N=9 0.09 (0.088–0.093) 0.079 (0.064–0.089) 0.086 (0.079–0.091) 0.07 (0.064–0.077) 0.046 (0.029–0.066) 0.008 (0–0.017) N = 41 N = 41 N = 48 0.080 (0.069–0.093) 0.068 (0.054–0.082) 0.078 (0.064–0.094) 0.060 (0.052–0.076) 0.022 (0–0.054) N=4 N=4 N = 11 0 0.046 (0.042–0.052) 0.062 (0.059–0.069) 0.059 (0.057–0.061) N = 43 N = 17 N = 12 N = 25 N = 17 N = 12 N = 25 N = 24 N = 19 N = 32 N = 19 N = 14 N = 27 N = 27 0.004 (0–0.012) 0.047 (0.039–0.057) B. diabolica B. geographica B. appendiculata Clade A B. nigra Clade B B. semilineata species complex B. ventrimaculata Clade D Clade E Clade F 0.001 (0–0.005) 0.045 (0.04–0.052) 0.06 (0.052–0.071) N = 30 N = 35 0.017 (0–0.03) N = 56 N = 51 N = 64 Clade F Clade E B. ventrimaculata Clade D B. semilineata species complex B. nigra Clade B B. appendiculata Clade A B. geographica B. diabolica 17 there are scattered minute black flecks on the dorsal surfaces (e.g. QCAZ 25797, 36932, 39416). A dark mid-dorsal line may extend from the tip of the snout to the mid-sacrum, but in some specimens, the line is restricted to the head. Individuals may display asymmetrically distributed small to large cream or dark blotches on the dorsum (e.g. QCAZ 5076, 28227, 44160, 64085). The coloration of flanks and hidden surfaces of thighs varies from creamy to pale grey, with vertical dark stripes to black bars (e.g. QCAZ 30921, 36934, 39416); bars on the flanks, and less frequently on the thighs, are sometimes arranged in pairs. Flanks are uniform black in juveniles, some subadults and a few small adults (< 42 mm). In some individuals, the bars can also be present on the hidden surfaces of the shanks (e.g. QCAZ 25797). Dark brown transversal bands are present on the dorsal surfaces of the limbs (one or two on the upper arm and forearm, and three to five on the thigh, shank and foot). The heels are usually cream, contrasting with the darker brown coloration of the dorsum (e.g. QCAZ 5076, 64085), but in some individuals, heels have the same colour as the rest of the body (e.g. QCAZ 36932, 44160). Ventral surfaces of preserved specimens vary from creamy white (e.g. QCAZ 30921) to yellowish white (e.g. QCAZ 39416). In some individuals there are dark brown spots on the belly (e.g. QCAZ 5076). Coloration of webbing and discs varies from yellowish white to cream, brown or grey. Coloration of bones is white or green (e.g. QCAZ 64085). Coloration in life (Fig. 7B): Dorsal surfaces vary from brown (e.g. QCAZ 51182) to dark brown (e.g. QCAZ 23855), reddish brown (e.g. QCAZ 26267) or yellowish tan (e.g. QCAZ 52845); transversal bands on dorsal surfaces on limbs and marks on dorsum vary from light brown to dark brown, yellow or white (e.g. QCAZ 52733, 61767, 64101); in some individuals there are scattered minute black flecks on dorsal surfaces (e.g. QCAZ 36934); flanks and hidden surfaces of thighs are cream or yellowish with dark brown vertical stripes or bars (e.g. QCAZ 51179, 51182, 64085), uniform black in some small adult individuals; belly and ventral surfaces of limbs vary from cream to yellowish or orange (e.g. QCAZ 37879, 41009, 64099); throat cream white or yellowish; webbing varies from yellow to orange or brown; fingers and toes are dorsally grey or with same coloration of webbing; iris brown or bronze (e.g. QCAZ 52845, 55356); palpebrum finely reticulated with golden yellow. In recently metamorphosed individuals, flanks, hidden surfaces of thighs and webbing are black; venter is grey; dorsal surfaces of the limbs and sides of the head are dark brown; dorsum is brown with scattered black flecks (Fig. 8A). © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 Table 6. Pairwise genetic distances (uncorrected p) of 16S DNA sequences (409 pb) among members of the Boana semilineata group. Mean is given with range in parentheses (below diagonal). Number of individuals compared is shown above diagonal. Diagonal shows intra-clade genetic distances SYSTEMATICS OF BOANA SEMILINEATA GROUP 18 M. A. CAMINER and S. R. RON Vocalization: Based on one male, not collected, from Yasuní on the road to Pompeya at km 73 (recorded by Morley Read); two males from Santa Cecilia (KU 143144–45; recorded by W. E. Duellman on 10 October 1971; 20:50–21:00 h; 23 °C) and two males from MiaziNuevo Paraíso (QCAZ 67113–14; collected by D. Paucar on 4 March 2017; 21:00 h). The call that lacks voucher specimen was assigned to Boana appendiculata based on the similarity of its structure compared to vouchered recordings of B. appendiculata. The call varies from a long groan (Type I, mean duration of 0.45 s and 31–37 pulses/note) to a series of short chuckle-like notes (Type II, mean 2.07 s and four to six notes/call) with fewer pulses per note (18–22 pulses/note). The long note (Fig. 9A, B) has a mean dominant frequency of 973.2 Hz, rise time of 0.34 s and frequency bandwidth 1316.2 Hz. The serial of short notes (Fig. 9C, D) has a mean dominant frequency of 1387.6 Hz, rise time of 0.44 s, frequency bandwidth of 1841.9 Hz and interval between notes of 0.16 s. Other call parameters are listed in Table 2. Distribution and ecology: Boana appendiculata occurs in the Amazon basin of Ecuador (Morona Santiago, Napo, Orellana, Pastaza, Sucumbíos and Zamora Chinchipe provinces), Brazil (Amazonas, Pará and RondoÌ‚nia states) and Colombia (Fig. 6). Localities with known elevation range from 14 to 1050 m above sea level. The elevation at Shell (1050 m) is the highest, while Caxiuanã (14 m) is the lowest. Specimens of B. appendiculata occur in terra firme forest, swamps, semi-flooded and flooded forests and artificial open areas. They are generally found at night in primary and secondary forest, perching on vegetation 20 to 600 cm above the ground, next to lakes, flooded areas and temporary ponds in clearings. Few individuals were found in flooded areas with pastures or forest away from water bodies. Their occurrence in secondary forests and artificial open areas suggest at least some tolerance of anthropogenic habitat disturbance. Vegetation types for Ecuadorian localities are: (1) Amazonian evergreen foothill forest, characterized by a mixture of Amazonian and Andean vegetation with a canopy of 30 m; (2) Amazonian lowland evergreen forest, characterized by high plant alphadiversity and a canopy height of 30 m with emergent trees that reach 40 m; (3) floodplain lowland forest of white-waters, characterized by periodical flooding with white-waters from large rivers, with the vegetation reaching 35 m in height and several horizontal strata of vegetation; and (4) lowland forest of palms and black-waters characterized by a canopy height of 30 m with dense understory and dominance of the palm Mauritia flexuosa L.f. (Sierra et al., 1999). Vegetation types at localities in Brazil include Purus-Madeira moist forests, Purus Varzea and Xingu-Tocantins-Araguaia moist forests (according to the World Wildlife Fund, 2017 classification scheme; available at http://www.eoearth.org/view/article/151948). Remarks: A 16S sequence of ‘Boana cf. geographica’ from Valle del Guamuez, Colombia, published by Meza-Joya et al. (2019; GenBank accession number MF583739) is identical to the homologous sequence of specimen QCAZ 61767 of B. appendiculata from Limoncocha, Ecuador (539 bp). This similarity indicates the presence of B. appendiculata in Colombia. The Meza-Joya et al. (2019) publication became available while our manuscript was under review. Therefore, we did not include its sequence into our phylogeny. Boana nigra sp. nov. (Figs 7A, 10; Table 5) Lsid: urn:lsid:zoobank.org:act:FE5D2127-D2BD4F4D-A4E5-4CE742A32F21 Holotype (Figs 7A, 10): QCAZ 58786 (field no. SC 48236), adult male from Ecuador, Provincia Morona Santiago, Parque Nacional Sangay, Sardinayacu (2.1028° S, 78.1521° W), 1348 m a.s.l., collected by © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 Figure 8. Dorsolateral views of juveniles of: A, B. appendiculata, QCAZ 39339 (SVL = 28.04 mm); B, Boana nigra, QCAZ 61883 (SVL = 25.04 mm); C, B. ventrimaculata, QCAZ 38663 (SVL = 27.06 mm). SYSTEMATICS OF BOANA SEMILINEATA GROUP 19 Daniel Rivadeneira, David Velalcázar, Javier Pinto, Francy Mora, Darwin Núñez, Juan Carlos Sánchez and Andrea Correa on 29 January 2015. Paratopotypes: QCAZ 58783–85, 58787–89, 58791– 97, adult males, 58790, 58798, adult females, same collectors as the holotype, 15–29 January 2015; © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 Figure 9. Calls of: A–D, Boana appendiculata from Yasuní, Province Orellana; E, F, B. nigra from Sardinayacu, Province Morona Santiago; G, H, B. ventrimaculata (QCAZ 49233) from Yasuní, Province Napo. A, C, E, G are oscillograms and B, D, F, H spectrograms. 20 M. A. CAMINER and S. R. RON Paratypes: ECUADOR: PROVINCIA MORONA SANTIAGO: Bosque Protector Abanico (2.2583° S, 78.1983° W), 1560 m, QCAZ 48984, adult female, collected by D. Morocho on 1 June 2007; bridge N of Macas (2.3017° S, 78.1137° W), 944 m, QCAZ 47044, collected by S. Poe, F. Ayala, L. Gray, J. Davis and I. Latella, on 15 December 2009; Parque Nacional Sangay (1.786° S, 77.979° W), 982 m, QCAZ 54605–6, 51776, 51780, tissues, collected by L. MacLean and A. Almendáriz, on 4 December 1991 and 29 August 2011; PROVINCIA PASTAZA: Centro Ecológico Zanja Arajuno, km 6 on the road San Ramón-El Triunfo (1.373° S, 77.860° W), 910 m, QCAZ 36929, 36931, 40959, 45282, adult males, 45283, female, 37903, 37988, tadpoles, 40265, 40958, subadult females, collected by S. R. Ron, I. Tapia, L. Coloma, A. Carvajal, A. Merino, E. Arbeláez, A. Tapia, D. Salazar, D. Acosta and C. Korfel, on 29 April 2004 and 2 February 2009. Common names: Proposed standard English name: black-flanked treefrog. Proposed standard Spanish name: rana arbórea de flancos negros. Diagnosis (Figs 7A, 10): Boana nigra is characterized by: (1) mean SVL 42.2 mm in males (range 38.5–46.3; N = 26), 64.8 mm in females (range 56.4–76.7; N = 5; Table 5); (2) thighs long (TL/SVL 0.44–0.53); (3) dorsal skin finely granular; (4) dorsum coloration brown, dark brown, cream or grey usually with an X-shaped mark on scapular region or rounded marks over the entire dorsum; (5) flanks and hidden surfaces of thighs vary from uniform black to dark vertical bars; (6) areolate skin on the posterior surface of throat, belly and ventral surfaces of thighs; (7) venter creamy white to pale orange, plain or spotted with dark brown; (8) in life, webbing pale orange, orange or black; (9) palpebrum reticulated; (10) nuptial excrescences present in breeding males; (11) projecting prepollices absent in both sexes; (12) tympanum large with distinct tympanic ring; (13) small triangular calcar on heel, often coloured with a cream spot; (14) advertisement call with one note; (15) pulses per note varying from 21 to 37; and (16) call dominant frequency varying from 996 to 1063 Hz. Comparisons with other species: Boana nigra can be distinguished from B. ventrimaculata sp. nov., B. hutchinsi, B. diabolica and B. semilineata by having uniform black or dark vertical stripes (in females) on flanks (speckled pattern or blotches in all the four species; Fouquet et al., 2016). In addition, the hidden surfaces of thighs and webbing in life are red in B. diabolica (uniform black in B. nigra, sometimes with dark vertical bars on the hidden surfaces of thighs). The individuals of B. nigra with dark vertical bars pattern on flanks and hidden surfaces of the thighs can be confused with B. geographica and B. appendiculata. However, B. nigra differs in having orange or black webbing in life (red webbing except between fingers III–IV and toes IV–V in B. geographica) and a cream to pale orange ventral coloration in life (yellow or orange in B. appendiculata). Some small adult B. appendiculata (< 42 mm) can also have a cream venter with uniform black flanks. Ventral and webbing bright coloration fade to cream in preservative and become indistinguishable among the three species. Boana nigra and B. appendiculata also differ in environmental envelope (Fig. 3D; for details see ‘Comparisons with other species’ in the description of B. appendiculata). Despite being separated by high genetic distances (Table 6), they occur in sympatry, which indicates the existence of non-geographic reproductive barriers. Boana nigra is easily distinguished from B. boans, B. icamiaba, B. pombali and B. wavrini by its prepollex not modified into a projecting spine and by the presence of pigmented nuptial excrescences. Boana boans further differs from B. nigra in having diffuse vertical bars on the flanks and barely distinguishable on the hidden surfaces of the thighs (well-defined vertical bars or uniform black in B. nigra) and by its fully webbed hands (webbing extends to about the middle of the outer finger in B. nigra). Description of the holotype: Adult male, 42.5 mm SVL, foot length 16.5 mm, head length 9.9 mm, head width 14.5 mm, eye diameter 4.5 mm, tympanum diameter 3.7 mm, tibia length 22.4 mm, femur length 22.2 mm, calcar length 0.9 mm, arm length 7.7 mm, eye–nostril distance 4.3 mm, head wider than long and wider than body; snout sloping in lateral view, subacuminate in dorsal view; distance from nostril to eye shorter than diameter of eye; canthus rostralis rounded; loreal region concave; internarial region subtly depressed; nostrils not protuberant, directed laterally; interorbital area slightly convex; eye large, protuberant; lower eyelid finely reticulated; diameter of eye 1.2 times diameter of tympanic annulus; supratympanic fold inconspicuous; tympanic membrane roundish, slightly longer than high; tympanic annulus evident, rounded, separated from eye by c. 0.5 times its diameter. Arm slender, axillary membrane absent; relative length of fingers I < II < IV < III; finger discs broadly expanded, round, © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 QCAZ 59110–12, 59114, 59116–17, adult males, 59113, metamorph, 59115, 59118, tadpoles, 61881–87, juveniles, collected by S. R. Ron, D. Paucar, P. Venegas, P. Baldeón, M. Caminer and E. Nusirquia, 27 February 2015; QCAZ 58337–8, tissues, collected by J. Brito on 29 October 2014; DHMECN 12130, adult male, collected by D. R. Batallas and P. Bejarano on 16 June 2014. SYSTEMATICS OF BOANA SEMILINEATA GROUP Colour of holotype in preservative (Fig. 10): Dorsum light grey with extensive dark brown rounded marks, sometimes interconnected; faint brown transversal bands on the dorsal surfaces of shanks; flanks, hidden surfaces of limbs and dorsal surfaces of thighs, upper arms, hands and feet black; venter cream with dark, diffuse, minute dots on the belly; ventral surfaces of thighs, forearms and hands cream; ventral surfaces of shanks and upper arms black; discs and webbing grey. Colour of holotype in life (Fig. 7A): Dorsum brown with dark brown rounded marks over the entire dorsum; dorsal and hidden surfaces of limbs and flanks black; faint brown transversal bands on the dorsal surface of hindlimbs; venter and ventral surfaces of thighs and hands cream; belly cream with brown diffused dots; ventral surfaces of upper limbs, shanks and feet black; discs and webbing black; iris bronze. Etymology: The specific epithet nigra is a Latin adjective in the nominative case meaning black, and refers to the black coloration on the flanks of these frogs. Variation in preservative: Morphometric data for adult specimens are summarized in Table 5, and variation in dorsal and ventral coloration of preserved specimens is shown in Figure 10. Sexual dimorphism is observed in SVL, with 38.5–46.3 mm (42.2 ± 2 mm, N = 26) in males and 56.4–76.7 mm (64.8 ± 7.9 mm, N = 5) in females. Females are significantly larger than males (t = –6.40, d.f. = 4.1, P = 0.002). Head shape varies between subacuminate and rounded in dorsal view; in lateral view it varies between truncate and short, rounded. The head is wider than long and slightly wider than the body in most of the specimens. Breeding males have keratinized nuptial excrescences on the inner surfaces of the thumbs. Calcar size ranges from inconspicuous (e.g. QCAZ 58783) to prominent (e.g. QCAZ 45283). Webbing on the hand varies slightly among specimens: fingers I basal II (1 +–2 –)—(2½–3 –) III (2 ––2½)—(2–2 +) IV. Variation of webbing between toes is I (1–2–)—(1–2–) II (1–1+)—(1–2+) III (1–1½)—(1+–2+) IV (2––2+)—(1–1+) V. Vomerine teeth (N = 15): 8–12 (right = 9.14) and 7–12 (left = 9.71). Most variation involves dorsal coloration and pattern. Background dorsal coloration varies from brown (e.g. QCAZ 45282), dark brown (e.g. QCAZ 48984), cream (e.g. QCAZ 45283) to grey (e.g. QCAZ 47044) with a dark brown, X-shaped mark on the scapular region or dark brown rounded marks over the entire dorsum (e.g. QCAZ 59114); sometimes rounded marks are interconnected (e.g. QCAZ 59117). Scattered, minute, black or white flecks are present on the dorsal surfaces (e.g. QCAZ 40958). A dark mid-dorsal line extends from the tip of the snout to the mid-sacrum (e.g. QCAZ 45283). Dark brown or grey transversal bands are present on the dorsal surfaces of the limbs (one or two on the upper arm and forearm and three to five on the thigh, shank and foot). The coloration of flanks and hidden surfaces of thighs and shanks are uniform black in males (e.g. QCAZ 58783) or grey with dark vertical bars in females (e.g. QCAZ 48984). However, in some cases males can present dark vertical bars on the flanks (e.g. QCAZ 6478). The heels have usually the same colour as the rest of the body (e.g. QCAZ 45282), but in some individuals heels are cream, contrasting with the background dorsal coloration (e.g. QCAZ 40958). Ventral surfaces of preserved specimens vary from creamy white (e.g. QCAZ 40958) to yellowish white (e.g. QCAZ 48984). In some individuals, there are dark brown spots on the belly (e.g. QCAZ 58783). Coloration of webbing and discs vary from yellowish white to cream, black or grey. Coloration in life (Fig. 7A): Dorsal surfaces vary from brown (e.g. QCAZ 58783) to dark brown (e.g. QCAZ 59110) or reddish brown (e.g. QCAZ 59111) with brown marks on the dorsum; brown transversal bands on dorsal surfaces on limbs; brown rounded spots and blotches over the entire dorsum (e.g. QCAZ 59110, 59111). In some individuals, there are scattered minute black or white flecks on dorsal surfaces (e.g. © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 as wide as long; subarticular tubercles prominent, round to ovoid, single; supernumerary tubercles absent; palmar tubercle inconspicuous; large and rugose dark P-shaped nuptial pad on dorsal surface of finger I; prepollex not modified into a projecting spine; webbing formula of fingers I basal II1½—3–III2½— 2IV. Toes bearing discs broadly expanded, rounded and slightly smaller than those of fingers; relative length of toes I < II < V < III < IV; inner metatarsal tubercle large, elliptical; outer metatarsal tubercle absent; subarticular tubercles single, round, large and protuberant; supernumerary tubercles small, round; webbing formula of toes I1+—2–II1—2–III1+—2+IV2+— 1+V. Vocal sac single and subgular. Skin on dorsum, head and dorsal surfaces of limbs finely granular; throat smooth anteriorly without mental gland; posterior surface of throat, belly and ventral surfaces of thighs areolate, those of shanks smooth. Cloacal opening directed posteriorly at upper level of thighs, round tubercles below and on sides of vent. Tongue slightly cordiform, widely attached to mouth floor; vomerine odontophores situated parallel to choanae, in two oblique series narrowly separated, each bearing 9–11 vomerine teeth; choanae trapezoidal, oblique. 21 22 M. A. CAMINER and S. R. RON © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 Figure 10. Adult preserved specimens of Boana nigra showing variation in dorsal and ventral coloration. From left to right, first and second rows: QCAZ 58786 (holotype), 58783, 45282 (males), 40958, 2657 (subadult females); third and fourth rows: QCAZ 6478, 58793, 59114 (males), 48984 (female); fifth and sixth rows: QCAZ 45283, 47044, 58790, 58798 (females). See the type-series list for locality data. All specimens are shown at the same scale. SYSTEMATICS OF BOANA SEMILINEATA GROUP Vocalization: Based on one male (DHMECN 12130) recorded at Sardinayacu (Provincia Morona Santiago) by D. R. Batallas on 16 June 2014 at 22:30 h, air temperature 14.8 °C. The call (Fig. 9E, F; Table 2) consists of one groan-like note with a mean duration of 0.33 s and dominant frequency of 1044.2 Hz. Other parameters are available in Table 2. The individual was perched on vegetation 40 cm above the ground near a lake. Distribution and ecology: Boana nigra occurs on the eastern Andean slopes of Ecuador (provinces of Morona Santiago, Napo, Pastaza, Sucumbíos and Zamora Chinchipe) (Fig. 6). Localities with known elevation range from 910 to 1847 m a.s.l. The elevation at Sardinayacu (1847 m) is the highest known for B. nigra, while Centro Ecológico Zanja Arajuno (910 m) is the lowest. It is generally found next to streams, swamps, ponds and lakes. Individuals have been recorded at night perching on vegetation 30 to 400 cm above ground or water. The specimen from 9 de Octubre was found next to a highway on the leaf litter. Vegetation types at known localities include: (1) Amazonian evergreen foothill forest, characterized by a mixture of Amazonian and Andean vegetation with a canopy of 30 m; (2) montane cloud forest of the eastern Andes, characterized by trees covered by mosses and abundant epiphytes; (3) evergreen lower montane forest of the east of the northern and central Andes, characterized by a canopy height of 25 to 30 m, with abundant epiphytes, and by the absence of species of trees characteristic of the lowlands (e.g. species from the family Bombacaceae and Myristicaceae); and (4) lowland floodplain forest of white-waters, characterized by periodical flooding with white-waters from large rivers, with the vegetation reaching 35 m in height, and several horizontal strata of vegetation (Sierra et al., 1999). Boana ventrimaculata sp. nov. (Figs 7C, 11; Table 5) Lsid: urn:lsid:zoobank.org:act:34C45CDE-5F974AE8-A655-E2C1DA6E9787 Holotype (Figs 7C, 11): QCAZ 51241 (field no. SC37210), adult male from Ecuador, Provincia Orellana, Estación Científica Yasuní PUCE (0.673° S, 77.407° W), 245 m a.s.l., collected by Italo Tapia, Teresa Camacho, Francy Mora and Santiago R. Ron, on 1 June 2011. Paratopotypes: QCAZ 51236, adult male, collected with the holotype; QCAZ 17517, juvenile, collected by O. Pérez and M. Bustamante on 17 February 2000. Paratypes: ECUADOR: PROVINCIA MORONA SANTIAGO: Pankints (2.9019° S, 77.894° W), 332 m, QCAZ 46429, adult female, collected by J. Brito on 26 December 2010; PROVINCIA NAPO: Huino (0.6449° S, 77.149° W), 273 m, QCAZ 30928, adult male, collected by D. Alvarado and G. Buitrón on 3 February 2003; Misahuallí (1.0343° S, 77.6685° W), 401 m, QCAZ 20000, adult female, collected by F. Ayala and M. Díaz on 26 March 2002; PROVINCIA SUCUMBÍOS: Río Aguarico, San Pablo (0.255° S, 76.426° W), 220 m, QCAZ 6314–15, adult females, collected on 2 August 1994; PROVINCIA PASTAZA: Campamento K10, Campo Villano (1.476° S, 77.5347° W), 475 m, QCAZ 38772, adult male, 39247, 56429, 56307, adult females, 38663, juvenile, collected by Y. Mera, D. Paucar, F. Ayala, E. Carrillo, G. Díaz, J. Brito and D. Velalcázar on 14 February 2008 and 18 November 2013; Comunidad Tarangaro, Campo Villano (1.3951° S, 77.3838° W), 338–1035 m, QCAZ 39130, 54233, 54238, adult males, 39137, adult female, collected by Y. Sagredo, A. Barahona, D. Paucar, G. Díaz, Y. Mera and F. Ayala on 23 August 2008 and 20 June 2012; Juyuintza, 4.5 km SE (2.1132° S, 76.1646° W), 164 m, QCAZ 53533, adult male, collected by M. Ortega and F. Timias on 26 June 2012; Lorocachi (1.6386° S, 75.9711° W), 207 m, QCAZ 56114, adult male, collected by S. R. Ron, D. Paucar, F. Ayala, D. Velalcázar, M. J. Navarrete, G. Galarza, J. Pinto, V. Chasiluisa, S. Arroyo and D. Zurita on 23 June 2013; PROVINCIA ORELLANA: Nuevo Rocafuerte, Río Napo (0.9165° S, 75.423° W), 194 m, QCAZ 44845–46, adult males, collected by S. R. Ron, E. Toral and I. Tapia on 12 July 2009; Parque Sumaco Balsayacu (0.3724° S, 77.3411° W), 600 m, QCAZ 23059, adult male, collected by P. Salvador on 16 January 2002; Estación de Biodiversidad Tiputini USFQ (0.633° S, 76.1473° W), 231 m, QCAZ 20398, adult male, collected by M. D. Proaño on 2 July 2002; Estación Científica Yasuní PUCE, km 8 on the road Tivacuno (0.6815° S, 76.3897° W), 251 m, QCAZ 18280, adult female, collected by © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 QCAZ 36929, 58783). Flanks and hidden surfaces of thighs are uniform black (in males) or grey with dark vertical bars (in females); belly and ventral surfaces of limbs vary from cream to pale orange (e.g. QCAZ 59111, 59114); throat cream white or pale orange; webbing vary from pale orange to orange or black; fingers and toes dorsally vary from black to grey or have the same coloration of webbings; iris brown (e.g. QCAZ 58797) or bronze (e.g. QCAZ 36929); palpebrum finely reticulated with golden yellow. In recently metamorphosed individuals, flanks, hidden surfaces of thighs, dorsal surfaces of the limbs, sides of the head and webbing are black; venter is grey sometimes with scattered dark brown spots; dorsum is cream with black spots and irregular blotches (Fig. 8B). 23 24 M. A. CAMINER and S. R. RON Common names: Proposed standard English name: Yasuní treefrog. Proposed standard Spanish name: rana arbórea del Yasuní. Diagnosis (Figs 7C, 11): Boana ventrimaculata is characterized by: (1) mean SVL 52.7 mm in males (range 44.5–60.7; n = 20), 81.6 mm in females (range 72.9–90.5; n = 17; Table 5); (2) thighs long (TL/SVL 0.45–0.56); (3) dorsal skin finely granular; (4) dorsum brown, dark brown, cream or dull tan usually with an X-shaped mark on scapular region; (5) flanks grey with white speckles, and hidden surfaces of thighs grey with black vertical bars; (6) areolate skin on the posterior surface of throat, belly and ventral surfaces of thighs; (7) venter creamy white to yellowish white, plain or with brown blotches; (8) in life, webbing brown or orange; (9) palpebrum reticulated; (10) nuptial excrescences present in breeding males; (11) projecting prepollices absent in both sexes; (12) tympanum large with distinct tympanic ring; (13) small, triangular calcar on heel, often with a cream spot; (14) advertisement call with one note; (15) 11 to 16 pulses per note; and (16) call dominant frequency varying from 475 to 685 Hz. Comparisons with other species: Boana ventrimaculata differs from B. geographica, B. appendiculata and B. nigra by having a speckled pattern or blotches on the flanks (black flanks or dark vertical bars or stripes in the three species) and brown blotches on the ventral surfaces (immaculate or spotted with dark brown in the three species). Boana ventrimaculata is larger (male mean SVL = 52.72 mm; female mean SVL = 81.62 mm) than B. diabolica (male mean SVL = 43.50 mm, female mean SVL = 56.30 mm; Fouquet et al., 2016), B. geographica (male mean SVL = 43.00 mm, female mean SVL = 48.60 mm; Fouquet et al., 2016), B. appendiculata (t = 7.16, d.f. = 31.9, P < 0.001) and B. nigra (t = 9.61, d.f. = 24.9, P < 0.001) (Table 5). Among species of the B. geographica–B. semilineata species complex, B. ventrimaculata is morphologically more similar to B. diabolica and B. hutchinsi. However, it can be distinguished by having a speckled pattern on the hidden surfaces of thighs and brown to orange webbing in life (hidden surfaces of thighs and webbing are immaculate red in B. diabolica; black in B. hutchinsi; Pyburn & Hall, 1984; Fouquet et al., 2016). Boana ventrimaculata can be further distinguished from B. hutchinsi by having a brown dorsum with black flecks in recently metamorphosed individuals (lavender-grey in B. hutchinsi), and a shorter duration and lower number of notes in the advertisement call (one groan-like note with a mean call duration of 0.24 s in B. ventrimaculata; four to six chuckle-like notes with a mean call duration of 0.68 s in B. hutchinsi; Pyburn & Hall, 1984). Boana hutchinsi has not been included in molecular phylogenies and, therefore, current diagnosis are based exclusively on morphological and bioacoustic characters. Specimens of B. ventrimaculata can be easily confused with B. semilineata due to their similar colour pattern (brown webbing, black flanks with white speckles and brown spots on the abdomen; Fouquet et al., 2016; Peloso et al., 2018), but these species can be clearly distinguished by having markedly differentiated advertisement calls (one note with a mean duration of 0.24 s, and mean dominant frequency of 561 Hz in B. ventrimaculata vs. one to three notes with a mean duration per note of 0.04 s, and mean dominant frequency of 1240 Hz in B. semilineata). The type localities of both species are separated by ~4500 km making it extremely unlikely that they represent the same species. Boana ventrimaculata is easily distinguished from B. boans, B. icamiaba, B. pombali and B. wavrini by its prepollex not modified into a projecting spine and by the presence of pigmented nuptial excrescences. Boana boans further differs from B. ventrimaculata in having diffuse vertical bars on the flanks (speckled pattern in B. ventrimaculata) and by its fully webbed hands (webbing extends to about the middle of the outer finger in B. ventrimaculata). Description of the holotype: Adult male, 57.6 mm SVL, foot length 21.5 mm, head length 12.4 mm, head width © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 D. Paucar on 20 December 2001; Yasuní (0.6771° S, 76.4011° W), 230 m, QCAZ 13962, adult male, 30271, adult female, collected by D. Prado on 18 November 1999; Yasuní, Apaika (0.8111° S, 76.704° W), 200 m, QCAZ 27788–89, adult males, collected by F. Nogales on 11 October 2000; Parque Nacional Yasuní, km 38 on the road Pompeya Sur-Iro (0.6535° S, 76.4535° W), 238 m, QCAZ 17251, adult male, collected by S. de la Torre on 25 February 1994; Parque Nacional Yasuní, km 45 on the road Pompeya Sur-Iro (0.6940° S, 76.4562° W), 247 m, QCAZ 8181, adult female, collected by M. Read on 29 January 1995; km 74 on the road Pompeya SurIro (0.8358° S, 76.3533° W), 257 m, QCAZ 5105, adult female, collected by S. de la Torre on 20 May 1994; Parque Nacional Yasuní, Ginta-Bloque 16 (0.9981° S, 76.1963° W), 243 m, QCAZ 20465, adult male, collected by M. Díaz on 27 January 2001; Parque Nacional Yasuní, on the road Pompeya Sur-Iro km 21 (0.6504° S, 76.5201° W), 306 m, QCAZ 13301, adult female, collected by F. Sornoza on 21 August 1996; km 38 on the road Pompeya Sur-Iro (0.6539° S, 76.4518° W), 230 m, QCAZ 49129, adult female, collected by M. Read on 11 January 1996; km 107 on the road Pompeya SurIro (0.9811° S, 76.2476° W), 249 m, QCAZ 6821, 49233, adult males, 8325, 30972, adult females, collected by S. Harris and M. Read on 12 November 1994 and 4 November 1996. SYSTEMATICS OF BOANA SEMILINEATA GROUP Colour of holotype in preservative (Fig. 11): Dorsum brown with a dark brown X-shaped mark on the scapular region and a dark brown transversal band on the sacrum; dark brown triangle with its base in the interorbital area and its apex between the nostrils; dark brown transversal bands on the dorsal surfaces of the limbs; flanks grey with dark brown blotches; hidden surfaces of the thighs grey with brown vertical bars; venter cream with brown rounded blotches on the belly; discs and webbing grey. Colour of holotype in life (Fig. 7C): Dorsum background brown with dark brown marks; dark brown transversal bands on the dorsal surface of limbs; flanks grey with brown blotches; hidden surfaces of thighs and shanks grey with brown vertical bars; venter and ventral surfaces of thighs yellowish with brown rounded blotches on the belly; throat cream; webbing orange; iris brown; bones green. Etymology: The specific name is derived from the Latin words ventriculus = belly, and macula = spot or stain, in reference to the brown blotches on the chest and belly of these frogs. Variation in preservative: Morphometric data for adult specimens are summarized in Table 5. Variation in dorsal and ventral coloration of preserved specimens is shown in Fig. 11. Sexual dimorphism is observed in SVL, with 44.5–60.7 mm (52.7 ± 4.6 mm, N = 20) in males and 72.9–90.5 mm (81.6 ± 5.4 mm, N = 17) in females. Females are significantly larger than males (t = –17.39, d.f. = 31.5, P < 0.001). Head shape varies between subacuminate and rounded in dorsal view; in lateral view it varies between truncate and short, rounded. The head is slightly wider than the body, but in some cases equal (e.g. QCAZ 39137). Breeding males have keratinized nuptial excrescences on the inner surfaces of the thumbs. Calcar size ranges from inconspicuous (e.g. QCAZ 30928) to prominent (e.g. QCAZ 51241). Webbing on the hand varies slightly among specimens: fingers I basal II (1––1+)—(2––2+) III (1+–2–)—(1–1½) IV. Variation of webbing between toes is I (1––1+)—(1–1+) II (1–1–)—(1–2) III (1––1+)—(1–2–) IV (1–2–)—(1––1+) V. Vomerine teeth (N = 10): 14–21 (right = 16.9) and 12–22 (left = 18). Variation in dorsal coloration of preserved specimens is extensive. Dorsal coloration varies from brown (e.g. QCAZ 30928) to dark brown (e.g. QCAZ 56307), cream (e.g. QCAZ 49233) or dull tan (e.g. QCAZ 39137) with irregular faintly or well-defined dark brown marks in diverse patterns, but most frequently consist of an X-shaped mark on the scapular region (e.g. QCAZ 13962). Individuals usually have a dark brown transversal band (e.g. QCAZ 39247) or two narrower transversal bars usually interconnected on the sacral region (e.g. QCAZ 13962). In some specimens there are scattered minute black flecks and white rounded spots on the dorsal surfaces (e.g. QCAZ 38772, 56114). Individuals may display asymmetrically distributed large cream blotches on the dorsum (e.g. QCAZ 44845). The coloration of flanks is grey with white speckles or sometimes with brown blotches. Vertical bars are present on the hidden surfaces of the thighs and sometimes on the shanks (e.g. QCAZ 39137). Dark transversal bands are present on the dorsal surfaces of the limbs (one or two on the upper arm and forearm and three to five on the thigh, shank and foot). The © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 19.3 mm, eye diameter 6.2 mm, tympanum diameter 4.8 mm, tibia length 27.8 mm, femur length 26.4 mm, calcar length 2.3 mm, arm length 9.9 mm, eye–nostril distance 7.1 mm, head wider than long and wider than body; snout truncate in lateral view, subacuminate in dorsal view; distance from nostril to eye larger than diameter of eye; canthus rostralis rounded; loreal region concave; internarial region subtly depressed; nostrils not protuberant, directed laterally; interorbital area slightly convex; eye large, protuberant; lower eyelid finely reticulated; diameter of eye 1.3 times diameter of tympanic annulus; supratympanic fold inconspicuous; tympanic membrane roundish, slightly longer than high; tympanic annulus evident, rounded, separated from eye by c. 0.4 times its diameter. Arm slender, axillary membrane absent; relative length of fingers I < II < IV < III; finger discs broadly expanded, round, as wide as long; subarticular tubercles prominent, round to ovoid, single; supernumerary tubercles absent; palmar tubercle inconspicuous; large and rugose dark P-shaped nuptial pad on dorsal surface of finger I; prepollex not modified into a projecting spine; webbing formula of fingers I basal II1—2–III2– —1+IV. Toes bearing discs broadly expanded, rounded and slightly smaller than those of fingers; relative length of toes I < II < V < III < IV; inner metatarsal tubercle large, elliptical; outer metatarsal tubercle absent; subarticular tubercles single, round, large and protuberant; supernumerary tubercles small, round; webbing formula of toes I1—1II1–—1III1—1IV1— 1–V. Vocal sac single and subgular. Skin on dorsum, head and dorsal surfaces of limbs finely granular; throat smooth anteriorly without mental gland; posterior surface of throat, belly and ventral surfaces of thighs areolate, those of shanks smooth. Cloacal opening directed posteriorly at upper level of thighs, round tubercles below and on the sides of vent. Tongue slightly cordiform, widely attached to mouth floor; odontophores situated parallel to choanae, in two oblique series closely set, each bearing 19–21 vomerine teeth; choanae trapezoidal, oblique. 25 26 M. A. CAMINER and S. R. RON heels have usually the same colour as the rest of the body (e.g. QCAZ 8181, 13962), but in some individuals, heels are cream contrasting with the background dorsal coloration (e.g. QCAZ 49233). Ventral surfaces of preserved specimens vary from creamy white (e.g. QCAZ 30928) to yellowish white (e.g. QCAZ 8181) with a belly usually yellowish. Lips, throat, chest, belly and ventral surfaces of the limbs usually with brown blotches (e.g. QCAZ 39247, 49233, 56114). Coloration of webbing and discs vary from yellowish white to brown or cream. Coloration of bones is white or green. Coloration in life (Fig. 7C): Dorsal surfaces vary from brown (e.g. QCAZ 56114) to dark brown (e.g. QCAZ 56307) or reddish brown (e.g. QCAZ 38772); brown transversal bands on dorsal surfaces of limbs; brown marks and white or yellow blotches on dorsum (e.g. QCAZ 56114, 56429); in some individuals, there are scattered minute black dots on dorsal surfaces (e.g. QCAZ 39130); flanks are grey or bluish with white speckles; hidden surfaces of thighs and shanks are bluish with dark vertical bars; venter and ventral surfaces of thighs vary from yellowish to cream white usually with brown blotches on the belly (e.g. QCAZ 39130, 56429); throat cream white or yellowish with or without brown blotches; webbings vary from brown to orange; fingers and toes dorsally vary from brown to grey; iris copper (e.g. QCAZ 51236); palpebrum finely reticulated with golden yellow. © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 Figure 11. Adult preserved specimens of Boana ventrimaculata showing variation in dorsal and ventral coloration. From left to right, first and second rows: QCAZ 51241 (holotype), 13962, 30928, 38772, 56114, 49233 (males); third and fourth rows: QCAZ 44845 (male), 56307, 39137, 39247, 8181 (females). See listed in the type-series for locality data. All specimens are shown at the same scale. SYSTEMATICS OF BOANA SEMILINEATA GROUP Vocalization: One male (QCAZ 49233) was recorded at Yasuní, on the road Pompeya Sur-Iro, km 108 (Provincia Orellana) by Morley Read on 4 November 1996 at 20:00 h. The call (Fig. 9G, H; Table 2) consists of one groan-like note with a mean duration of 0.24 s and dominant frequency of 561.4 Hz. Other call parameters are shown in Table 2. The individual was perched on vegetation 1 m above a stream. Distribution and ecology: Boana ventrimaculata inhabits the Amazon basin of Ecuador (provinces of Morona Santiago, Napo, Orellana, Pastaza and Sucumbíos) and Brazil (Estado do Amazonas) (Fig. 6). Localities with known elevation range vary between 64 and 1035 m of elevation. Among localities with known elevation, Comunidad Tarangaro (1035 m) is the highest and Rio Içá (64 m) is the lowest. Most specimens of B. ventrimaculata were found in terra firme forest and semi-flooded forests. Frogs have been found at night in primary and secondary forest, perching on vegetation 20 to 500 cm above the ground, near streams or along trails away from water bodies. Vegetation types for Ecuadorian localities are: (1) Amazonian evergreen foothill forest, characterized by a mixture of Amazonian and Andean vegetation with a canopy of 30 m; (2) Amazonian lowland evergreen forest, characterized by high plant alpha-diversity and a canopy height of 30 m with emergent trees that reach 40 m; (3) lowland floodplain forest of white-waters, characterized by periodical flooding with white-waters from large rivers, with the vegetation reaching 35 m in height and several horizontal strata of vegetation; and (4) lowland forest of Mauritia flexuosa palms and blackwaters characterized by a canopy height of 30 m with dense understory (Sierra et al., 1999). Localities in Brazil fall within the Purus Varzea vegetation type (according to the World Wildlife Fund, 2017 classification scheme; available at http://www.eoearth.org/view/article/151948). DISCUSSION Similar to recent reviews of Neotropical hylids based on genetic characters (e.g. Ron et al., 2012, 2018; Caminer & Ron 2014; Ferrão et al., 2016; Caminer et al., 2017; Rivadeneira et al., 2018), our results and those of Fouquet et al. (2016) and Peloso et al. (2018) show that Boana geographica, which until recently was considered a single widespread species, is a species complex. Four new species have been described and one resurrected since 2016 [herein and in Fouquet et al. (2016) and Peloso et al. (2018)]. Nevertheless, there are at least six seemingly undescribed species that would require further systematic work (in addition of B. hutchinsi that remains to be included in a molecular phylogeny), especially in Amazonian Brazil. The B. semilineata group exemplifies the vast work on systematics needed to solve the Linnean shortfall in Neotropical amphibians. The phylogeny and geographic distribution of species of the Boana semilineata group suggest an origin in the Amazon basin. This scenario is suggested by the Amazonian distribution of most species of the group, including those diverging more basally: B. boans, B. icamibia and B. wavrini. The only exception is B. pombali, which occurs in the Atlantic forest. An Amazonian origin would suggest that species occurring in the Atlantic forest and the Guianan region dispersed from the Amazon region. Boana nigra is the only species of the group occurring almost exclusively above 1000 m of elevation (range from 910 to 1847 m in B. nigra; < 1100 m in all other species). This implies a recent dispersal from lowland tropical rainforests to Andean montane forests. Boana nigra has a parapatric distribution with its sister species, B. appendiculata. Other example of recent dispersion from the Amazon lowlands to the Andean slopes have been shown in the Boana calcarata–fasciata species complex (Caminer & Ron, 2014), where the sister species B. almendarizae and B. fasciata are mainly distributed in Andean montane forests (range from 500 to 1950 m), while the rest of the group occurs in adjacent Amazonian lowlands (< 650 m). These types of distribution pattern could result from parapatric speciation along the elevation gradient. We combine molecular, morphological, bioacoustic and environmental data to delimit species within the B. semilineata group. Although we find differences in advertisement calls and morphology (large males and females are easy to differentiate), there is remarkable similarity between small individuals of B. appendiculata and B. nigra. Subadults, and small adults of both species, show uniform black coloration on flanks and hidden surfaces of the thighs that make their identification challenging. Without genetic data it is difficult to recognize some specimens (e.g. QCAZ 52610, 52734) as distinct. The scant ecological data suggest differences in habitat preference between the sympatric species B. appendiculata and B. ventrimaculata. Boana ventrimaculata has been recorded in forests near lotic waters. Boana appendiculata, in contrast, usually occurs near lentic waters in open areas. Additional data on habitat preferences are needed to test these presumed differences. Our results allow us to determine the status of trans-Andean Chocoan populations of B. boans. © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 In recently metamorphosed individuals, flanks, hidden surfaces of thighs and webbing are black; venter is grey; dorsal surfaces of the limbs are dark brown; dorsum is brown with black flecks over the entire surface. (Fig. 8C). 27 28 M. A. CAMINER and S. R. RON ACKNOWLEDGEMENTS Field and laboratory work were funded by grants from the Secretaría Nacional de Educación Superior, Ciencia, Tecnología e Innovación del Ecuador SENESCYT (Arca de Noé initiative; SRR and Omar Torres principal investigators) and PUCE-DGA (SRR principal investigator). Charlie Barnes, D. Flores, G. Nicholls, A. Manzano and S. Espinoza carried out extractions and amplification of DNA. Diego PáezMoscoso helped organizing sequences from GenBank. William E. Duellman, Diego R. Batallas and The Cornell Lab of Ornithology provided call recordings. Andrea Varela provided photographs of the lectotype of Hyla appendiculata. Andrew Crawford provided access to the amphibian collection of Universidad de los Andes, Colombia. Mario Yánez-Muñoz from INABIO loaned a tissue of Boana rubracyla. Ministerio de Ambiente del Ecuador provided collect and research permits (001-11 IC-FAU-DNB/MA; 005-12-IC-FAU-DNB/MA; 003-15-IC-FAU-DNB/ MA; 002-16-IC-FAU-DNB/MA; 003-17-IC-FAU-DNB/ MA; 005-2009-INVESTIGACIÓN-B-DPMS/MAE; 006-2012-FAU-MAE-DPO-PNY; 010-2013-FAU-MAEDPAO-PNY; 008-IC-FAU-DPPZ/MA; 016-2015-FAUMAE-DPAO; MAE-DNB-ARRGG-CM-2014-0002). For collection of specimens of the new species we are indebted to D. Rivadeneira, D. Velalcázar, J. Pinto, F. Mora, D. Núñez, J. C. Sánchez, A. Correa, D. Paucar, P. Venegas, P. Baldeón, E. Nusirquia, J. Brito, D. R. Batallas, P. Bejarano, D. Morocho, S. Poe, F. Ayala, L. Gray, J. Davis, I. Latella, L. MacLean, A. Almendáriz, I. Tapia, L. Coloma, A. Carvajal, A. Merino, E. Arbeláez, A. Tapia, D. Salazar, D. Acosta, C. Korfel, T. Camacho, O. Pérez, M. Bustamante, D. Alvarado, G. Buitrón, M. Díaz, V. Ruales, Y. Mera, E. Carrillo, G. Díaz, Y. Sagredo, A. Barahona, M. Ortega, F. Timias, M. J. Navarrete, G. Galarza, V. Chasiluisa, S. Arroyo, D. Zurita, E. Toral, P. Salvador, M. D. Proaño, M. Read, D. Prado, S. de la Torre, F. Nogales and S. Harris. REFERENCES Acevedo AA, Lampo M, Cipriani R. 2016. The cane or marine toad Rhinella marina (Anura: Bufonidae): two genetically and morphologically distinct species. Zootaxa 4103: 574–586. Aguilar C, Wood PL Jr, Belk MC, Duff MH, Sites JW Jr. 2016. Different roads lead to Rome: integrative taxonomic approaches lead to the discovery of two new lizard lineages in the Liolaemus montanus group (Squamata: Liolaemidae). Biological Journal of the Linnean Society 120: 448–467. Caminer MA, Ron SR. 2014. Systematics of treefrogs of the Hypsiboas calcaratus and Hypsiboas fasciatus species © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 Unexpectedly, we find low genetic distances between samples from the Chocó region and the Amazonian region of Ecuador, suggesting that they represent a single species. This is a surprising finding, because the presence of the Andean barrier between them should have prevented gene flow for an extended period. All other amphibian species from lowland rainforests occur either in the Chocó or in Amazonia, but not in both. The few species that were believed to have a trans-Andean distribution turned out to represent separate species, one on each side of the mountain range. Rhinella marina, for example, was split in R. horribilis for populations west of the Andes and R. marina for population on the east (Acevedo et al., 2016). Boana boans and B. xerophylla (Escalona et al., 2019) appear to be the only two species in which a trans-Andean distribution has been confirmed with genetic evidence. However, B. boans is the only species occurring in Chocoan rainforests. At a broader scale, our phylogenetic analysis indicates the existence of a high number of potential new species in the genus Boana in Ecuador and Colombia. Ecuadorian populations of ‘B. pellucens’ from Durango and Lita (Provincia Esmeraldas) are most closely related to B. rufitela from Panama than to other Ecuadorian populations of B. pellucens. High genetic distances between both clades of Ecuadorian populations (5%) also hint that Ecuadorian populations encompass two species. Our results suggest that populations from Esmeraldas are B. rufitela or an undescribed species. Until this issue is solved, we consider B. rufitela as occurring in Ecuador. The non-monophyly of the B. punctata group is congruent with results by Pyron & Wiens (2011) and Duellman et al. (2016). Paraphyly is due to the position of B. picturata, which has weak support. Although other species have been assigned to this group (i.e. B. alemani, B. atlantica, B. hobbsi, B. jimenezi and B. liliae) on the basis of morphological similarity (Faivovich et al., 2005; Kok, 2006; Señariz & Ayarzagüena, 2006), their position based on genetic characters needs to be determined. Our phylogenetic analyses indicate that Boana cinerascens and B. punctata form two species complexes consisting of four to six divergent lineages each (uncorrected p-distances among lineages range from 4.1 to 11% in B. cinerascens and 2.9 to 6.6% in B. punctata). Some of these lineages were reported by Faivovich et al. (2004, 2005), Fouquet et al. (2007), Jansen et al. (2011) and Guarnizo et al. (2015). Elucidation of the status of candidate species outside the B. semilineata species group is beyond the scope of this publication. Therefore, we leave them as mtDNA highly divergent lineages that could represent undescribed species (Fig. 1). SYSTEMATICS OF BOANA SEMILINEATA GROUP Fouquet A, Martinez Q, Zeidler L, Courtois EA, Gaucher P, Blanc M, Dias Lima J, Marques Souza S, Rodrigues M, Kok PJR. 2016. Cryptic diversity in the Hypsiboas semilineatus species group (Amphibia, Anura) with the description of a new species from the eastern Guiana Shield. Zootaxa 4084: 79–104. Frost DR. 2019. Amphibian species of the world: an online reference, version 6.0. New York: American Museum of Natural History. Electronic Database accessible. Available at: http://research.amnh.org/herpetology/amphibia/index. html (accessed 14 May 2019). Guarnizo CE, Paz A, Muñoz-Ortiz A, Flechas SV, MéndezNarváez J, Crawford AJ. 2015. DNA barcoding survey of anurans across the Eastern Cordillera of Colombia and the impact of the Andes on cryptic diversity. PLoS One 10: e0127312. Jansen M, Bloch R, Schulze A, Pfenninger M. 2011. Integrative inventory of Bolivia’s lowland anurans reveals hidden diversity. Zoologica Scripta 40: 567–583. Kohler J, Jansen M, Rodriguez A, Kok PJR, Toledo LF, Emmrich M, Glaw F, Haddad CFB, Rodel MO, Vences M. 2017. The use of bioacoustics in anuran taxonomy: theory, terminology, methods and recommendations for best practice. Zootaxa 4251: 1–124. Kok P. 2006. A new species of Hypsiboas (Amphibia: Anura: Hylidae) from Kaiateur National Park, eastern edge of the Pakaraima mountains, Guyana. Bulletin de l’Institut Royal des Sciences Naturelles de Belgique 76: 191–200. La Marca E, Azevedo-Ramos C, Coloma LA, Solis F, Ibáñez R, Jaramillo C, Fuenmayor Q, Ron SR, Hardy J. 2010. Hypsiboas boans. The IUCN Red List of Threatened Species 2010: e.T55415A11304871. Available at: https://dx.doi.org/10.2305/IUCN.UK.2010-2.RLTS. T55415A11304871.en (accessed 15 October 2018). Lanfear R, Calcott B, Ho SYW, Guindon S. 2012. PartitionFinder: combined selection of partitioning schemes and substitution models for phylogenetic analyses. Molecular Biology and Evolution 29: 1695–1701. Linnaeus C. 1758. Systema naturae per regna tria naturae, secundum classes, ordines, genera, species, cum characteribus, differentiis, synonymis, locis, 10th edn, Vol. 1. Stockholm: L. Salvius. Lutz B. 1963. New species of Hyla from southeastern Brazil. Copeia 1963: 561–562. Maddison WP, Maddison DR. 2015. Mesquite: a modular system for evolutionary analysis, version 3.04. Available at: http://mesquiteproject.org/ (accessed 20 August 2015). Meza-Joya FL, Ramos-Pallares E, Hernández-Jaimes C. 2019. Hidden diversity in frogs within Boana calcarata– fasciata and Boana geographica species complexes from Colombia. Herpetology Notes 12: 391–400. Myers CW, Duellman WE. 1982. A new species of Hyla from Cerro Colorado, and other tree frog records and geographical notes from Western Panama. American Museum Novitates 2752: 1–32. Parker HW. 1933. A list of the frogs and toads of Trinidad. Tropical Agriculture 10: 8–12. © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 complex (Anura, Hylidae) with the description of four new species. ZooKeys 370: 1–68. Caminer MA, Milá B, Jansen M, Fouquet A, Venegas PJ, Chávez G, Lougheed SC, Ron SR. 2017. Systematics of the Dendropsophus leucophyllatus species complex (Anura: Hylidae): cryptic diversity and the description of two new species. PLoS One 12: e0171785. Caramaschi U, Pimenta BVS, Feio RN. 2004. Nova especie do grupo de Hyla geographica Spix, 1824 da floresta Atlântica, Brasil (Amphibia, Anura, Hylidae). Boletim do Museu Nacional. Nova Serie, Zoologia. Rio de Janeiro 518: 1–14. Cocroft RB, Ryan MJ. 1995. Patterns of advertisement call evolution in toads and chorus frogs. Animal Behaviour 49: 283–303. Dayrat B. 2005. Towards integrative taxonomy. Biological Journal of the Linnean Society 85: 407–415. Dos Santos SP, Ibáñez R, Ron SR. 2015. Systematics of the Rhinella margaritifera complex (Anura, Bufonidae) from western Ecuador and Panama with insights in the biogeography of Rhinella alata. ZooKeys 501: 109–145. Duellman WE. 1970. Hylid frogs of Middle America. Monograph of the Museum of Natural History University of Kansas 1: 1–753. Duellman WE. 1973. Frogs of the Hyla geographica group. Copeia 1973: 515–533. Duellman WE, Pyles RA. 1983. Acoustic resource partitioning in anuran communities. Copeia 1983: 639–649. Duellman WE, Marion AB, Hedges B. 2016. Phylogenetics, classification, and biogeography of the treefrogs (Amphibia: Anura: Arboranae). Zootaxa 4104: 1–109. Escalona MD, Acuña Juncá F, Giaretta AA, Crawford AJ, La Marca E. 2019. Contrasting genetic, acoustic, and morphological differentiation in two closely related gladiator frogs (Hylidae: Boana) across a common Neotropical landscape. Zootaxa 4609: 519–547. Faivovich J, García PCA, Ananias F, Lanari L, Basso NG, Wheeler WC. 2004. A molecular perspective on the phylogeny of the Hyla pulchella species group (Anura, Hylidae). Molecular Phylogenetics and Evolution 32: 938–950. Faivovich J, Haddad CFB, García PC, Frost DR, Campbell JA, Wheeler WC. 2005. Systematic review of the frog family Hylidae, with special reference to Hylinae: phylogenetic analysis and taxonomic revision. Bulletin of the American Museum of Natural History 294: 1–240. Ferrão M, Colatreli O, de Fraga R, Kaefer IL, Moravec J, Lima AP. 2016. High species richness of Scinax treefrogs (Hylidae) in a threatened Amazonian landscape revealed by an integrative approach. PLoS One 11: e0165679. Fick SE, Hijmans RJ. 2017. Worldclim 2: new 1-km spatial resolution climate surfaces for global land areas. International Journal of Climatology 37: 4302–4315. Fouquet A, Gilles A, Vences M, Marty C, Blanc M, Gemmell NJ. 2007. Underestimation of species richness in neotropical frogs revealed by mtDNA analyses. PLoS One 2: e1109. 29 30 M. A. CAMINER and S. R. RON Ron SR, Venegas PJ, Ortega-Andrade HM, GagliardiUrrutia G, Salerno P. 2016. Systematics of Ecnomiohyla tuberculosa with the description of a new species and comments on the taxonomy of Trachycephalus typhonius (Anura, Hylidae). ZooKeys 630: 115–154. Ron SR, Duellman WE, Caminer MA, Pazmiño D. 2018. Advertisement calls and DNA sequences reveal a new species of Scinax (Anura: Hylidae) on the Pacific lowlands of Ecuador. PLoS One 13: e0203169. Ron SR, Merino-Viteri A, Ortiz DA. 2019. Anfibios del Ecuador, version 2019.0. Available at: https://bioweb.bio/ faunaweb/amphibiaweb/ (accessed 2 February 2019). Ronquist F, Teslenko M, van der Mark P, Ayres DL, Darling A, Höhna S, Larget B, Liu L, Suchard MA, Huelsenbeck JP. 2012. MrBayes 3.2: efficient bayesian phylogenetic inference and model choice across a large model space. Systematic Biology 61: 539–542. Sambrook J, Fritsch EF, Maniatis T. 1989. Molecular cloning: a laboratory manual. New York: Cold Spring Harbor Laboratory Press. SAS Institute. 2010. User guide, version 9.01. Cary: SAS Institute. Available at: http://www.jmp.com/ (accessed 12 December 2010). Savage JM, Heyer WR. 1967. Variation and distribution in the tree-frog genus Phyllomedusa in Costa Rica, Central America. Studies on Neotropical Fauna and Environment 5: 111–131. Señariz JC, Ayarzagüena J. 2006. A new species of Hypsiboas (Amphibia; Anura; Hylidae) from the Venezuelan Guyana, with notes on Hypsiboas sibleszi (Rivero, 1972). Herpetologica 62: 308–318. Sierra R, Cerón C, Palacios W, Valencia R. 1999. Mapa de vegetación del Ecuador Continental 1:1’000.000. Quito: Proyecto INEFAN/GEF-BIRF, Wildlife Conservation Society y Ecociencia. Spix JB. 1824. Animalia nova sive species novae Testudinum et Ranarum quas in itinere per Brasiliam annis MDCCCXVII– MDCCCXX jussu et auspiciis Maximiliani Josephi I. Bavariae Regis. München: F. S. Hübschmann. Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar S. 2011. MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Molecular Biology and Evolution 28: 2731–2739. Vieites DR, Wollenberg KC, Andreone F, Köhler J, Glaw F, Vences M. 2009. Vast underestimation of Madagascar’s biodiversity evidenced by an integrative amphibian inventory. Proceedings of the National Academy of Sciences of the United States of America 106: 8267–8272. Zhang J, Kapli P, Pavlidis P, Stamatakis A. 2013. A general species delimitation method with applications to phylogenetic placements. Bioinformatics 29: 2869–2876. Zwickl DJ. 2006. Genetic algorithm approaches for the phylogenetic analysis of large biological sequence datasets under the maximum likelihood criterion. Unpublished Ph.D. Thesis, University of Texas, Austin, 125. © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 Parker HW. 1935. The frogs, lizards and snakes of British Guiana. Proceedings of the Zoological Society of London 1935: 505–530. Parker HW. 1936. A collection of reptiles and amphibians from the Upper Orinoco. Bulletin du Musée Royal d’Histoire Naturelle de Belgique 12: 1–4. Paz A, Crawford AJ. 2012. Molecular-based rapid inventories of sympatric diversity: a comparison of DNA barcode clustering methods applied to geography-based vs clade-based sampling of amphibians. Journal of Biosciences 37: 887–896. Peloso PLV, De Oliveira RM, Sturaro MJ, Rodrigues MT, Lima-Filho GR, Bitar YOC, Wheeler WC, Aleixo A. 2018. Phylogeny of map tree frogs, Boana semilineata species group, with a new Amazonian species (Anura: Hylidae). South American Journal of Herpetology 13: 150–169. Pimm SL, Jenkins CN, Abell R, Brooks TM, Gittleman JL, Joppa LN, Raven PH, Roberts CM, Sexton JO. 2014. The biodiversity of species and their rates of extinction, distribution, and protection. Science 344: 1246752. Puillandre N, Lambert A, Brouillet S, Achaz G. 2012. ABGD, automatic barcode gap discovery for primary species delimitation. Molecular Ecology 21: 1864–1877. Pyburn WF, Hall DH. 1984. A new stream inhabiting treefrog (Anura: Hylidae) from southeastern Colombia. Herpetologica 40: 366–372. Pyron RA, Wiens JJ. 2011. A large-scale phylogeny of Amphibia including over 2800 species, and a revised classification of extant frogs, salamanders, and caecilians. Molecular Phylogenetics and Evolution 61: 543–583. QGIS Development Team. 2018. QGIS geographic information system. Open Source Geospatial Foundation Project. Available at: http://qgis.osgeo.org (accessed 11 April 2019). Rambaut A, Drummond A. 2007. TRACER. MCMC trace analysis tool, version 1.4. University of Oxford. Available at: http://tree.bio.ed.ac.uk/software/tracer (accessed 10 September 2010). Rissler LJ, Apodaca JJ. 2007. Adding more ecology into species delimitation: ecological niche models and phylogeography help define cryptic species in the black salamander (Aneides flavipunctatus). Systematic Biology 56: 924–942. Rivadeneira CD, Venegas PJ, Ron SR. 2018. Species limits within the widespread Amazonian treefrog Dendropsophus parviceps with descriptions of two new species (Anura, Hylidae). ZooKeys 726: 25–77. Ron SR, Read M. 2018. Boana geographica. In: Ron SR, Merino-Viteri A, Ortiz DA, eds. Anfibios del Ecuador, version 2018.0. Museo de Zoología, Pontificia Universidad Católica del Ecuador. Database accessible. Available at: https:// bioweb.bio/faunaweb/amphibiaweb/FichaEspecie/Boana%20 geographica (accessed 11 October 2018). Ron SR, Venegas PJ, Toral E, Read M, Ortiz DA, Manzano AL. 2012. Systematics of the Osteocephalus buckleyi species complex (Anura, Hylidae) from Ecuador and Peru. ZooKeys 229: 1–52. SYSTEMATICS OF BOANA SEMILINEATA GROUP 31 SUPPORTING INFORMATION Table S1. Genbank accession numbers for DNA sequences used in the phylogenetic analysis. Table S2. Primers used for DNA amplification. Figure S1. Species delimitation tree based on the Poisson Tree Process (PTP) model, using a fragment of the 16S mitochondrial gene. The blue lines indicate branching processes among species, while red lines indicate taxa that should be considered as part of the same lineage. Figure S2. ABGD species delimitation. APPENDIX EXAMINED SPECIMENS Boana appendiculata. ECUADOR: PROVINCIA MORONA SANTIAGO: on the road Puerto Morona-San José de Morona (2.8802° S, 77.6788° W), 200 m (QCAZ 52610); PROVINCIA ZAMORA CHINCHIPE: Centro Shuar Yawi (4.4495° S, 78.6489° W), 945 m (QCAZ 31035–36); PROVINCIA NAPO: Ahuano (1.0545° S, 77.5483° W), 385 m (QCAZ 36723); Cotundo, Huasquila Lodge (0.8237° S, 77.8015° W), 804 m (QCAZ 57191); Huino (0.6449° S, 77.149° W), 273 m (QCAZ 30910, 30921); Reserva Ecológica Yachana (0.8457° S, 77.2287° W), 317 m (QCAZ 48846); Serena (1.0961° S, 77.9243° W), 551 m (QCAZ 25849–50); on the road Miazi-Nuevo Paraíso (4.3091° S, 78.645° W) 1006 m (QCAZ 67108– 24); PROVINCIA SUCUMBÍOS: Reserva Faunística Cuyabeno (0.0885° S, 76.142° W), 222 m (QCAZ 2157, 2061, 2123–24, 2208, 2213–14, 2216, 2220, 2222, 2798, 3357, 4930–31, 6210–16, 25955); Zábalo (0.3181° S, 75.7662° W), 220 m (QCAZ 27729, 27937); Comunidad Chonta Yacu (0.1115° S, 77.3743° W), 610 m (QCAZ 25797); Selva Lodge (0.4981° S, 76.3738° W), 250 m (QCAZ 12025); Rey de los Andes (0.2082° S, 76.2368° W), 270 m (QCAZ 28477); Puerto Libre (0.1912° N, 77.4882° W), 700 m (QCAZ 21406); Parque Pacayacu (0.0378° S, 76.5857° W), 262 m (QCAZ 23055); Puerto Bolívar (0.0886° S, 76.142° W), 240 m (QCAZ 28227); bridge of the Río Cuyabeno, on the road Tarapoa-Palma Roja (0.032° S, 76.3205° W), 240 m (QCAZ 6060–61); Limoncocha (0.4062° S, 76.6194° W), 261 m (QCAZ 3357–62, 4679, 8016, 8019, 43082, 43095, 43112, 43136, 43139, 43141, 61766–68, 63542); PROVINCIA PASTAZA: Shell (1.5064° S, 78.0607° W), 1050 m (QCAZ 36088–90, 36095, 36100, 36201–3, 36209–14, 36217, 36220); Sarayacu (1.7245 °S, 77.4755 °W), 383 m (QCAZ 52727, 52733–34); Río Tiguino, Bataburo Lodge (1.25° S, 76.6667° W), 238 m (QCAZ 20015, 20029–30, 20080–81, 39416); Pomona, Fundación Hola Vida (1.625° S, 77.9072° W), 846 m (QCAZ 25701, 25714, 25626, 25636, 37161–62, 37197); 5 km N of the Puyo, on the road to the Tena Hostería Safari (1.443° S, 77.9966° W), 954 m, (QCAZ 26255, 26267); Bobonaza (1.4945° S, 77.8696° W), 660 m, (QCAZ 40749); Canelos (1.5793° S, 77.7321° W), 650 m (QCAZ 48704–7, 52845, 56610, 56586–87); Lorocachi (1.6377° S, 75.9703° W), 210 m (QCAZ 55943, 55973); Centro Etnoturístico Indichuris (1.6509° S, 77.9213° W), 799 m (QCAZ 57791); Fátima (1.4113° S, 78° W), 1023 m (QCAZ 6244–45, 6248, 49320–21, 49324, 49333, 49341, 49353, 30323); Chunitayo (1.9625° S, 77.98° W), 1000 m (QCAZ 11063); Estación Científica Oglán (1.3235° S, 77.9912° W), 568 m (QCAZ 58102); Centro Ecológico Zanja Arajuno (1.3567° S, 77.8594° W), 903 m (QCAZ 36932–34, 62064, 62081); PROVINCIA ORELLANA: Estación Científica Yasuní PUCE (0.6771° S, 76.4011° W), 250 m (QCAZ 2385, 8419–20, 14819, 16619, 16801– 2, 16804–8, 16809, 16812, 18293, 19197–200, 19678, 20833, 22461–63, 22500, 22572–75, 22673–74, 22676, 22769, 23073, 36027, 41004, 41007–10); Añango, Parque Nacional Yasuní (0.53333° S, 76.45° W), 243 m (QCAZ 5528–30, 5086–87, 17253); on the road Pompeya-Iro, Parque Nacional Yasuní (0.6535° S, 76.4535° W), 238 m (QCAZ 5076–77, 5083–84, 5099, 5102, 5145, 5150, 5154, 5162, 5175, 5228–30, 5239, 5242–44, 5247, 5376– 78, 5383, 5418–20, 8204–9, 8211–12, 9678–79, 9700–1, 12382, 13311–12, 17252, 32719, 36028, 46499, 51170, 51179–82, 53399); Río Yasuní, Parque Nacional Yasuní (0.9047° S, 76.144° W), 228 m (QCAZ 64083–107); Edén, Río Napo (0.4983° S, 76.0711° W), 216 m (QCAZ 44159–75, 44226); Tambococha (0.9783° S, 75.4256° W), 177 m (QCAZ 55227, 55234, 55242, 55253, 55261, 55315, 55356, 55360–61, 55393); Huiririma, Río Napo (0.7116° S, 75.6239° W), 194 m (QCAZ 44578–83); Pompeya, Sacha Lodge (0.4705° S, 76.4597° W), 247 m (QCAZ 37873, 37879, 39339); Reserva Biológica Río Bigal (0.5525° S, 77.4199° W), 907 m (QCAZ 48491); Río Huataraco, 70 km E of Hollín (0.747° S, 77.354° W), 380 m (QCAZ 7117); Chiroisla, Río Napo (0.5756° S, 75.8998° W), 203 m (QCAZ 44434); Comunidad Sinchichicta, Río Napo (0.6793° S, 75.8454° W), 220 m (QCAZ 57927); La Primaver, Río Napo (0.4442° S, 76.868° W), 244 m (QCAZ 43850); Río Pusuno (0.7263° S, 76.7266° W), 266 m (QCAZ 14078); bridge over the Río Tiputini, near campamento ELF (0.7417° S, 76.8611° W), 295 m (QCAZ 7898, 7913); Pindo (0.7489° S, 76.7693° W), 304 m (QCAZ 49738); Coca (0.4691° © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 Additional Supporting Information may be found in the online version of this article at the publisher's web-site. 32 M. A. CAMINER and S. R. RON Nacional Yasuní (0.4716° S, 76.5832° W), 243 m (QCAZ 43039); PROVINCIA SUCUMBÍOS: Puerto Bolívar (0.0886° S, 76.142° W), 240 m (QCAZ 28184). Boana pellucens. ECUADOR: PROVINCIA EL ORO: route from Huaquillas to La Cuca Arenillas (3.5357° S, 80.0671° W), 53 m (QCAZ 23680); 8 km on the road La Avanzada-Playó (3.588° S, 79.919° W), 230 m (QCAZ 30594); near Valle Hermoso (3.5116° S, 79.8202° W), 282 m (QCAZ 37254); Carabota (3.3283° S, 75.581° W), 395 m (QCAZ 45317); PROVINCIA ESMERALDAS: Durango (1.0724° N, 78.6534° W), 120 m (QCAZ 13192, 40270); PROVINCIA PICHINCHA: La Unión del Toachi (0.3318° S, 78.9379° W), 901 m (QCAZ 24359); PROVINCIA SANTO DOMINGO DE LOS TSÁCHILAS: Alluriquín-La Florida (0.2777° S, 79.0181° W), 971 m (QCAZ 13875). B o a n a p i c t u r a t a . E C UA D O R : P R OV I N C I A ESMERALDAS: 3 km of Durango (1.0643° N, 78.6443° W), 158 m (QCAZ 15549); PROVINCIA PICHINCHA: Amanecer Campesino, Río Santiago (0.0485° S, 78.9641° W), 838 m (QCAZ 32452); PROVINCIA SANTO DOMINGO DE LOS TSÁCHILAS: Alluriquín-La Florida (0.2777° S, 79.0181° W), 971 m (QCAZ 13894). B o a n a p u n c t a t a . E C UA D O R : P R OV I N C I A MORONA SANTIAGO: Pankints (2.9019° S, 77.894° W), 332 m (QCAZ 46423); PROVINCIA NAPO: Estación Biológica Jatun Sacha (1.0645° S, 77.6128° W), 405 m (QCAZ 43790); PROVINCIA ORELLANA: La Primavera, Río Napo (0.4442° S, 76.7868° W), 244 m (QCAZ 43790); PROVINCIA SUCUMBÍOS: Reserva Biológica Limoncocha (0.4062° S, 76.6194° W), 261 m (QCAZ 43105). Boana rosenbergi. ECUADOR: PROVINCIA COTOPAXI: La Maná (0.9433° S, 79.2647° W), 163 m (QCAZ 26288); PROVINCIA ESMERALDAS: 5 km W of Durango (1.0724° N, 78.6534° W), 120 m (QCAZ 13194); Alto Tambo (0.9169° N, 78.5661° W), 253 m (QCAZ 17081); PROVINCIA LOS RÍOS: Patricia Pilar (0.5726° S, 79.369° W), 200 m (QCAZ 19770); PROVINCIA MANABÍ: Estero Ancho, 52 km W of El Carmen (0.0677° S, 79.7961° W), 329 m (QCAZ 23186); PROVINCIA PICHINCHA: Pedro Vicente Maldonado (0.0789° N, 79.6194° W), 678 m (QCAZ 35454, 45244). B o a n a r u b r a c y l a . E C UA D O R : P R OV I N C I A CARCHI: Tobar Donoso (1.1643° N, 78.5071° W), 257 m (QCAZ 26206–7). © 2020 The Linnean Society of London, Zoological Journal of the Linnean Society, 2020, XX, 1–32 Downloaded from https://academic.oup.com/zoolinnean/advance-article-abstract/doi/10.1093/zoolinnean/zlaa002/5810752 by University of Rochester user on 22 March 2020 S, 78.9863° W), 256 m (QCAZ 416); km 78 on the road Maxus (0.8346° S, 76.3198° W), 257 m (QCAZ 6822), Loreto-Avila Viejo (0.6833° S, 77.3166° W), 434 m (QCAZ 10740); COLOMBIA: DEPARTAMENTO AMAZONAS: km 9–10 on the road Leticia, Río Takana (4.136° S, 69.9412° W), 83 m (ANDES-A 881–82, 906– 07, 909–10, 1631–33). Boana boans. ECUADOR: PROVINCIA BOLÍVAR: La Pretoria (1.7621° S, 79.2889° W), 122 m (QCAZ 51309, 51319); Santa Rosa de Agua Clara (2.1856° S, 79.113° W), 424 m (QCAZ 32593–94); PROVINCIA ESMERALDAS: Durango (1.0427° N, 78.6245° W), 211 m (QCAZ 32049, 32062); PROVINCIA MORONA SANTIAGO: Limón Indanza (2.9266° S, 78.407° W), 1100 m (QCAZ 41810); Pankints (2.9019° S, 77.894° W), 332 m (QCAZ 46489); PROVINCIA NAPO: Ahuano (1.0545° S, 77.5483° W), 385 m (QCAZ 27029); Cotapino, near Río Pucuno (0.8011° S, 77.4318° W), 600 m (QCAZ 13682); Tena (0.9888° S, 77.8124° W), 508 m (QCAZ 16033); PROVINCIA ORELLANA: Estación Científica Yasuní PUCE (0.6771° S, 76.4011° W), 250 m (QCAZ 19248); PROVINCIA PASTAZA: Campo Villano (1.5048° S, 77.5106° W), 345 m (QCAZ 39038, 38455, 38783); PROVINCIA PICHINCHA: 10 km W Pedro Vicente Maldonado (0.0999° N, 79.1329° W), 422 m (QCAZ 45246). Boana cinerascens. ECUADOR: PROVINCIA MORONA SANTIAGO: Payapas Santiago (3.0465° S, 78.0129° W), 300 m (QCAZ 52536); PROVINCIA PASTAZA: Lorocachi (1.6407° S, 76° W), 185 m (QCAZ 56088); Canelos (1.592° S, 77.6932° W), 420 m (QCAZ 56580). Boana hutchinsi. COLOMBIA: DEPARTAMENTO VAUPÉS: Cano Kuduyari, Comunidad Pirasemo, (1.317° N, 70.3889° W), 199 m (ANDES-A 2508). Boana nigra. ECUADOR: PROVINCIA MORONA SANTIAGO: on the road Nueve de Octubre (2.2492° S, 78.2025° W), 1532 m (QCAZ 52557); Parque Nacional Sangay, Río Volcán (2.1035° S, 78.1571° W), 1350 m (QCAZ 2656–59); PROVINCIA NAPO: San Francisco (0.4206° S, 77.8458° W), 1730 m (QCAZ 7019, 7029–34); PROVINCIA SUCUMBÍOS: Reventador town (0.0462° S, 77.5291° W), 1400 m (QCAZ 6478–79). Boana nympha. ECUADOR: PROVINCIA MORONA SANTIAGO: Pankints (2.9019° S, 77.894° W), 332 m (QCAZ 46426); PROVINCIA ORELLANA: Parque