Amphibians and reptiles from a protected area in western Brazilian Amazonia (Reserva Extrativista do Baixo Juruá)

. Sampling gaps across the logistically challenging and extremely biodiverse Amazonia largely hamper our understanding of broad‑scale amphibian and reptile diversity patterns in this ecosystem. The Juruá River basin, a southwestern tributary of the Amazon River, is one of these undersampled areas, with only punctual information documented for these vertebrates that are spatially or temporally biased. This is especially the case for the lower‑middle courses of the Juruá River, which also has comparatively less protected areas than its upper course. In order to fill some biodiversity knowledge gaps associated with amphibians and reptiles in this river basin, we combined results of our field expeditions carried out in 1992, 2005‑2007, and 2018 to the Reserva Extrativista (Extractive Reserve) do Baixo Juruá, a Brazilian protected area in the right bank of the lower Juruá River. Amphibians and reptiles were sampled using four complementary methods: active surveys, pitfall traps, funnel traps, and trammel nets. We identified species or updated their taxonomic status with a reanalysis of the external morphology of the preserved material in the light of novel taxonomic literature (more than 1,500 specimens) and employment of DNA barcoding analyses for some newly collected specimens with contentious taxonomic status. Our combined sampling evidenced 149 species of amphibians and reptiles occurring in this protected area (72 amphibians, 68 squamates, six chelonians, and three crocodilians). Recorded species highlight the value of the lower Juruá River region as harboring quite diverse assemblages for these vertebrates, with species typical of the western and southwestern Amazonia sub‑regions. Remarkable species records are presented, as well as accounts on species with lower taxonomic resolution. Furthermore, we discuss the biogeographic affinities of recorded assemblages based on the species geographic range and preferred habitats, and the value of this protected area to preserve the regional biological diversity.


INTRODUCTION
Amphibians and reptiles are among the most taxonomically and ecologically diverse vertebrates of the South American biota (Duellman, 1979). In the lowland tropical rainforest of South America (Amazonia), the largest and most diverse tropical forest in the world, more than 350 and 370 nominal species of amphibians and reptiles are respectively known (Prudente, 2017;Ávila-Pires, 2018Ávila-Pires, , 2020Hoogmoed, 2019). Despite that, our understanding of the Amazonian amphibian and reptile diversity goes through a transition period. The relatively recent implementation of molecular ISSN On-Line: 1807-0205 ISSN Printed: 0031-1049 ISNI: 0000-0004-0384-1825 data in integrative approaches has significantly increased our understanding regarding how these vertebrates have diversified in this ecosystem (e.g., Réjaud et al., 2020;Carvalho et al., 2021), and in their taxonomic boundaries and geographic distributions (e.g., Ferrão et al., 2016;Peloso et al., 2016;Ribeiro-Júnior et al., 2021;Vacher et al., 2020). At the same time, molecular elucidation also highlighted how much we still do not know about this diversity, at different scales of knowledge Vacher et al., 2020). This lack of knowledge is largely influenced by the fact that Amazonia harbors historically undersampled localities , and our increased diversity resolution is restricted to more accessible areas, such as the margins of large rivers and highways (Oliveira et al., 2016). As the biological diversity in some regions can be quite substantial, scientific novelties resulting from each new sampling event are common even at historically best explored localities (e.g., Fonseca et al., 2019, Jorge et al., 2020.
Western Amazonia is one of its sub-regions with the lowest resolution in terms of amphibian and reptile taxonomy and geographic distributions, mainly due to inaccessibility (França & Venâncio, 2010;Pantoja & Fraga, 2012). Knowledge associated with the westernmost tributaries of the Amazon River often results from few longterm inventories that remained as the principal data source over decades (e.g., Heyer, 1976;Gascon, 1996;Ávila-Pires et al., 2009). One of the most representative examples is the river basin of the southern tributary Juruá River. While the upper course of this river has historically been better explored and is well known for harboring a high species richness of amphibians and reptiles (e.g., see Ávila-Pires et al., 2009;Souza, 2009;França & Venâncio, 2010;Bernarde et al., 2013;Fonseca et al., 2019), its middle and lower courses have largely contrasting less associated information (Pantoja & Fraga, 2012). Despite such undersampling, these river sections are frequently highlighted as having rarely recorded amphibian and reptile species (e.g., Carvalho et al., 2017Carvalho et al., , 2020. However, knowledge advance in this case is still highly restricted to punctual observations, and we lack studies summarizing the information on their assemblages. Extensive geographic areas closer to the western limits of the Brazilian Amazonia are mainly protected by indigenous territories, though there are relatively fewer federal or state protected areas when compared to the eastern Amazonia (Silva-Jr. & Sites-Jr., 1995). The effective conservation of this region considerably benefits from its inaccessibility and unfavorable agroclimatic conditions, when compared to the highly threatened eastern Amazonia (Fearnside, 2008;Sombroek, 2001;Chomitz & Thomas, 2003). However, anthropogenic changes such as deforestation are surprisingly rapidly advancing across the entire Amazonia (Silva-Jr. et al., 2021), and Brazilian government has been recently marked by anti-environmental policies (Ferrante & Fearnside, 2019). Under this scenario, establishing baselines on biotic diversity across western Amazonia is urgent not only so we can better characterize large scale biodiversity patterns and trends within the Amazonia, but also so that the potential an-thropic impact in this region can be assessed to establish effective conservation strategies (Laurance et al., 2011). In the Juruá River basin, a higher number of protected areas are currently delimited in its upper course (9 areas) and progressively decreases towards its middle (3 areas) and lower course (2 areas) (ICMBio, 2009). Only one of the protected areas at the lower Juruá River region reaches the margin of the Juruá River: the Reserva Extrativista do Baixo Juruá (Extractive Reserve of the lower Juruá, hereafter RBJ), focus of our study. As is common for Amazonian protected areas (Correia et al., 2016), scientific knowledge associated with RBJ is still scarce.
Considering that the effective conservation of protected areas is highly dependent on our knowledge of their native biological diversity, here, we present and discuss the diversity of amphibian and reptile assemblages of the RBJ. Our study is based on a robust approach combining data collected in three distinct field expeditions to this protected area across three decades, allied with reanalyses of museum specimens from the region. We identified the recorded species combining morphological and molecular evidence, and discussed some remarkable records, biogeographic affinities of the assemblages, and conservation implications.

Study area
The RBJ is a federal protected area created in 2001 and located on the right bank of the lower course of the Juruá River, state of Amazonas, northwestern Brazil (Fig. 1). This sustainable use protected area (i.e., permit the direct use of natural resources) is contained within the limits of the municipalities of Juruá and Uarini and has 188,000 hectares (ICMBio, 2009). It is limited to the west by the Juruá River, to the south by the Andirá River and Igarapé do Branco (tributaries of the Juruá River), to the north by the Igarapé São Benedito (tributary of the Juruá River), and to the east by the Copacá River (tributary of the Uarini River) (ICMBio, 2009). The Juruá River flows to the north, also limiting the indigenous territory Kumaru do Lago Ualá (ICMBio, 2009), and the southern limits of the RBJ adjoin another protected area, the Floresta Nacional de Tefé (ICMBio, 2009) (Fig. 1).
This RBJ is home to ca. 16 traditional communities, but their areas correspond to low percentages of this terrain (ICMBio, 2009). The reserve is dominated by extensive and preserved landscapes of dense or open ombrophilous forests (Fig. 2), most of which are not subject to periodic flooding by the annual flood pulse of large rivers (ICMBio, 2009). As this non-flooded terrain is part of an Amazonian region that was only recently (Late Pleistocene; < 45,000 years ago) exposed to this flooding pulse regime (Pupim et al., 2019), they can be considered as paleovárzea. On the margins of the large rivers, the seasonal flood pulse creates many distinct environments, including diverse várzea forests, pioneer species monodominant forests, grasslands, patches of aquatic macro-phytes, and mud or sandy exposed banks (ICMBio, 2009) (Fig. 2). As is common in Amazonia, such environmental heterogeneity created by the flooding gradient largely influences the faunistic spatial occupation and diversity patterns (Moraes et al., 2016(Moraes et al., , 2022aRamalho et al., 2018).
The Juruá River is one of the major southern tributaries of the Amazon River. Their headwaters are in the Andean region and the river crosses the historically dynamic sedimentary basin of Western Amazonia (Sioli, 1968). As a consequence, this river carries high levels of suspended sediments (a white-water river; Sioli, 1984), and the low slope promotes a meandering morphology with highly dynamic margins subject to recent erosion and deposition processes (Kalliola et al., 1991). Considering the lower course of the Juruá River, a dissimilar geomorphological influence is clearly noticed between its margins. At the region of the confluence with the Andirá River, the Juruá River acquires an uncentralized position in relation to its alluvial plain by following geological breaks (Fernandes et al., 1977). As a result, the lower Juruá River has a higher proportion of higher terrains part of the Formação Solimões (Pleistocene deposition) on its right bank, where the RBJ is located, while the left bank has a broad and asymmetric extension of more recent alluvial plains (Holocene deposition) (Fernandes et al., 1977). Such a dissimilarity directly influences the proportions of vegetation types found in each bank (prevalence of várzea in lower terrains and paleovárzea in higher terrains), as well as in human occupation in this region, which is higher on the right bank (ICMBio, 2009). Altitudes within the RBJ range from 40 to 100 m above sea level, and the region's climate is characterized as tropical (Af ), being hot (mean annual temperature of 25℃) and humid (total annual rainfall of ca. 3,000 mm) (Alvares et al., 2013).

Data collection and sampling effort
We generated a database based on data collected in our field expeditions and reanalyses of museum specimens deposited at the Collection of Amphibians and Reptiles (INPA-H) of the Instituto Nacional de Pesquisas da Amazônia (INPA), Manaus, state of Amazonas, Brazil. Three main field expeditions conducted by the authors (hereafter indicated with the initials of their names) formed the baseline of our data set. This periodic sampling covered the entire hydrological cycle of the flooding pulse in a timeframe of 31 years, resulting in the total sampling effort of 95 days. The first expedition was led by CG in 1991-1992 (i.e., before the creation of the RBJ) through the entire Juruá River course and focused on amphibian sampling (see results in Gascon, 1996;Lougheed et al., 1999;Gascon et al., 2000). The study area was sampled in these expeditions from 02-19 May 1992 (18 sampling days) in a single locality (Table 1; see detailed information of this sampling in Patton et al., 2000). The second inventory in this region was led by VTC thirteen years later (2005)(2006)(2007), as part of the actions to generate a management plan for the recently created RBJ (ICMBio, 2009). During these years, both amphibians and reptiles were sampled using standardized efforts in eight distinct localities in the RBJ from 17 November-02 December 2005, 19 January-02 February 2006, 12 May-01 June 2006, and 04-13 May 2007. Lastly, eleven years later, remaining authors (LJCLM, RNR, FPW and AFSO) conducted a third amphibian and reptile inventory at the RBJ, sampling three distinct localities from 14-28 July 2018 (15 sampling days). In 2018, we also sampled through active surveys a locality of flooded habitats at the left margin of the Juruá River, i.e., outside the limits of the RBJ (Appendix 1). However, in order to maximize Figure 1. Geographic location of the Reserva Extrativista (RESEX) do Baixo Juruá in relation to South America (A) and the Juruá River basin, in Brazilian Western Amazonia (B); upon an elevation background). In (B), protected areas and indigenous territories that cover this river basin are delimited in white and pink, respectively. The focal area is zoomed at (C), showing the eleven localities sampled for amphibians and reptiles (numbered white symbols; see Table 1), and the Juruá municipality. Background in (C) is from Google Earth (Map data ©2015 Google/Landsat/Copernicus).
the assessment of diversity, we aggregate these few records in this study as part of the locality 8, which also represents flooded habitats but is within the limits of the RBJ (Fig. 1).
Individuals were recorded in those inventories using complementary sampling methods established for herpetofauna in Amazonia: (1) diurnal and nocturnal active surveys (Heyer et al., 1994), where the individuals are detected in all the microhabitats visually accessible or, in the case of anurans, through their vocalizations (all the inventories; total sampling effort of 245 person-days); (2) pitfall traps (Heyer et al., 1994) (Fig. 2), composed by stations of 7-12 100-liter buckets buried linearly with drift fences that were installed only in non-flooded forests (2005-2007 and 2018 inventories; 886 trap-days; details of sampling effort at Table 1). During sampling at locality 11 in 2007, due to logistical issues, only buckets were buried (i.e., without drift fences) in a "Y" setup ( Fig. 2); (3) funnel traps (Enge, 2001), installed in pairs between the buckets of pitfall traps, ranging from 12-22 funnel traps per pitfall station (2005)(2006)(2007) inventories; 1,184 trap-days) (Fig. 2). During sampling at the locality 9 in 2006, only funnel traps were installed in a drift fence station; and (4) trammel nets with distinct opening sizes to sample aquatic animals (2005-2007 inventories; non-standardized). Specimens were also recorded through non-standardized occasional encounters made by us or by members of local communities and other teams in the multidisciplinary expeditions.
We performed rarefaction (interpolation) and extrapolation curves with 95% confidence intervals (Chao et al., 2014) to compare the observed and expected species richness of amphibians and reptiles in the RBJ, as well as to investigate the effectiveness of our sampling effort in assessing the local diversity of those groups. The extrapolated rarefaction curves were generated using the R statistical software package "iNEXT" (Chao et al., 2014). We also estimated species richness for amphibians and reptiles in the RBJ based on our sampling using the Chao and Jackknife estimation indices implemented in the "specpool" function of the R package "vegan" (Oksanen et al., 2022). Due to the lack of standardization among our sampling events and the uncertainties associated with historical data, we only consider data associated with voucher specimens for these estimation analyses.
We collected specimens and/or tissue samples following appropriate ethical and legal guidelines, under the permits 67798-1 and 02005.001056/06-05 provided by the Brazilian Federal Agencies Instituto Chico Mendes de Conservação da Biodiversidade (ICMBio) and Instituto Brasileiro do Meio Ambiente e dos Recursos Naturais Renováveis (IBAMA). Voucher specimens and tissue samples were respectively deposited in the INPA-H and the Genetic Resources Collection (INPA-HT) of the same institute (INPA). During 2018-2020, we reanalyzed all the voucher specimens from the target region deposited at INPA-H (1,560 specimens, see Appendix 1), providing support for the species list of amphibians and reptiles inhabiting the study area. Taxonomic identifications were performed under combined approaches of external morphology analyses (for newly collected and historical material) and DNA barcoding (only for some specimens collected in 2018). We followed the taxonomic arrangements of Frost (2022) for amphibians and Uetz et al. (2022) for reptiles.

Species identification: DNA barcoding
In order to expand the investigation of taxonomic status, we applied a DNA-based identification approach for some recently collected specimens, mainly focused on recognized species complexes where further resolution is needed. We focused on genetic markers widely recognized in barcoding studies with these groups and sequenced fragments of mitochondrial DNA (mtDNA) loci: the RNA 16S subunit ribosomal RNA (16S) and the protein-coding gene (CDS) cytochrome oxidase I (COI) for amphibians, and the CDS cytochrome b (CYTB) and NADH dehydrogenase subunit 4 (ND4) for squamates. Genomic DNA was extracted from liver tissue samples using standard protocols of a commercial kit (Wizard® Genomic DNA Purification Kit, Promega, Wisconsin, USA), and the target loci were amplified via Polymerase Chain Reaction (PCR), using the primers and following the protocols described in Palumbi et al. (1991)  After editing the sequences using Geneious, version 8 (Kearse et al., 2012), we submit them to a BLAST® search under the Standard Nucleotide BLAST function (blastn) (Johnson et al., 2008). For each analyzed sample, we downloaded the 100 most similar sequences from the GenBank online repository (GB; Clark et al., 2016), and aligned each dataset using the MAFFT online server with default parameters, except by the use of the E-INS-i strategy for the RNA and G-INS-i strategy for the CDS (Katoh & Standley, 2013). Uncorrected pairwise genetic distances were estimated for each alignment and we inferred the taxonomic status of samples considering the minimum distance among sequences. We consider the possibility of candidate unnamed species when genetic divergence was above 3% (16S), 5% (COI), 23% (Gekkota CYTB), and 10% (ND4), based on thresholds recognized in the literature for these loci and taxonomic groups (Fouquet et al., 2007;Bergmann & Russel, 2007;Kok et al., 2018;Carvalho et al., 2021). Newly generated sequences were deposited on GB under the accession numbers MT472181-MT472182 , MZ018807-MZ018827 , MW916091 (Ribeiro-Júnior et al., 2021), and OP296595-OP296615, OP314270-OP314276 and OP277918 (this study).

Biogeographic affinities
To investigate the relative biogeographic influence on the assemblages composition of amphibians and reptiles in the focal area, we firstly classified the recorded species according to the main habitats of occurrence, based on our sampling evidence: non-flooded forests (NFF), seasonally flooded habitats -margin of large rivers (FLH), anthropic (ANT), and aquatic (AQU). Aquatic interface was subdivided in lakes (AQU-L) and streams (AQU-S). Additionally, we also classified the species according to their known geographic distributions verified in literature, using as categories some subdivisions of Amazonia based on the limits of its main geological compartments and major rivers (Gibbs & Barron, 1993;Aleixo & Rossetti, 2007;Hoorn et al., 2010). Such Amazonian geographic subdivisions were selected because they are widely known as sharing biotic similarities (Ávila-Pires, 1995;Silva-Jr. & Sites-Jr., 1995;Réjaud et al., 2020). Therefore, species were classified in the following categories: widely distributed in Amazonia (WD), for species with geographic ranges encompassing various of the major subdivisions delimited; western Amazonia (WA) and eastern Amazonia (EA), for species typical of these macro-regions, influenced by the sedimentary basin of the Amazon River and crystalline shields, respectively; central Amazonia (CA) for species with distribution encompassing the internal boundary between WA and EA, but not their extremes; central southern Amazonia (SA) and southwestern Amazonia (SW), for species typical of the southern bank of the Amazon River, and restricted to the westernmost portion of this river bank, respectively. Finally, we also considered species that are potentially endemic to the study area (PE) or occur at punctual restricted localities (PR) within major subdivisions delimited. We calculated relative percentages by dividing the number of species and specimens from each category by the total recorded for both amphibians and reptiles.

RESULTS
Combining the results of our three inventories, we recorded 149 species of amphibians and reptiles occurring at the RBJ (Table 2). Regarding amphibians, after the taxonomic update, we found that 37 species were recorded in the first inventory (1992). This number increased to 60 in the 2005-2007 inventories, and finally to 72 after the 2018 inventory. Amphibians were represented by a single species of the Order Caudata, two of the Order Gymnophiona, and 69 of the Order Anura (13 families). Among the anurans, the most species-rich family was Hylidae (28 species), followed by Leptodactylidae (14 species), Strabomantidae (six species), Microhylidae and Bufonidae (five species each), Aromobatidae (three species), and Dendrobatidae (two species). The remaining anuran families recorded were represented by single species: Allophrynidae, Ceratophryidae, Eleutherodactylidae, Phyllomedusidae, Pipidae, and Ranidae.
Regarding reptiles, the increase in species richness reported for the RBJ went from 69 in the 2005-2007 inventories to 77 after the 2018 inventory. Squamates were represented by 29 lizard species and 39 snake species. Gymnophthalmidae was the most species-rich lizard family (six species), followed by Dactyloidae and Teiidae (five Table 2   species each), and Alopoglossidae, Sphaerodactylidae, and Tropiduridae (three species each). Among snakes, the most species-rich family was Dipsadidae (23 species), followed by Colubridae (six species), Viperidae (four species), Elapidae (three species), and Boidae (two species). The remaining squamate families recorded were represented by single species: Gekkonidae, Iguanidae, Phyllodactylidae, Scincidae (lizards), and Aniliidae (snakes). The order Testudines was represented by six species, with the family Podocnemididae evidenced as the most species-rich (three species), followed by Chelidae (two species) and Testudinidae (one species). In addition, three species of the Order Crocodylia were recorded, all members of the family Alligatoridae. The species rarefaction curves showed that our sampling was well representative considering the expected amphibian and reptile richness for the RBJ, as for both groups the observed data almost reached the asymp-totes of curves (Fig. 3). However, as expected, both rarefaction curves also indicate that the RBJ still has potential to harbor additional species not recorded in our samplings, notably for reptiles. According to the applied estimation indices, the expected amphibian richness in the RBJ ranges from 80-87 species based on our sampling (representing an increment of 11-21% in the observed richness), while for reptiles this estimate ranges from 78-94 species (representing an an increment of 16-40% in the observed richness).
As for biogeographic affinities, disregarding the widely distributed species in the lowland Amazonia (53% of amphibians and 78% of reptiles), the recorded assemblage was mostly composed of typical species from western, southwestern, and southern lowland Amazonia (33% of amphibians and 22% of reptiles) ( Table 2). We detected changes in species richness and composition among habitat types. Assemblages from non-flooded  forests (paleovárzea) were the most species-rich, with 107 recorded species (73 unique species). Other 65 species compose the assemblages from the flooded habitats (várzea) (24 unique species) ( Table 2). We also recorded 25 species using anthropic habitats, while seven species are typical from the aquatic interface (Table 2).
Under the DNA barcoding approach, we also detected the occurrence of two divergent mtDNA lineages under the species complexes Rhinella margaritifera (Laurenti, 1768) (Fig. 4C, D), Boana steinbachi (Boulenger, 1905) (Fig. 6B), and Boana geographica (Spix, 1824) (Fig. 6G). With the genetic distances at the 16S fragment ranging to 4%, 3.2%, and 4.5%, respectively, those divergent lineages most likely correspond to distinct sympatric species. The first lineage of the R. margaritifera species complex (R. aff. margaritifera) matched to sequences from Atlantic Forest (AY680262 and KU495504, 99.2%), while the second one matched to sequences of the recently described Rhinella exostosica Ferrão et al., 2020 (KF992145, 99.6%). These distinct lineages are indeed morphologically divergent, with R. aff. margaritifera corresponding to a candidate species with head longer than wide, absence of prominent bony knobs at the angle of jaws, and females possessing considerably extended supratympanic crests (Fig. 4C), whereas the specimens of R. exostosica have the head as wide as long, prominent bony knobs at the angle of jaws, and females with less extended supratympanic crests (Fig. 4D). Sequences of another bufonid from our sampling matched to R. castaneotica sensu stricto from eastern Amazonia (MF479716 and MF479717, 98.8%) (Fig. 4B), but given the overall low resolution in the taxonomy of this group (Santos et al., 2015;Ferrão et al., 2022), this status should be considered preliminary, and the chances of this population belong to a new species are high. Specimens of R. castaneotica from the RBJ share a small body size (mean snout-vent length; SVL = 46.7 mm) and lacks supratympanic crests in both sexes, therefore differing from the two sympatrically recorded larger species of the R. margaritifera species group (mean SVL = 57.9 and 54.3 mm; supratympanic crests present).
Within the Boana steinbachi species complex (Fig. 6B), the two divergent genetic lineages detected at the RBJ are not conspecific with the nominal Boana alfaroi (Caminer & Ron, 2014) or Boana tetete (Caminer & Ron, 2014). In fact, one of the lineages is more closely related to the Boana "clade J" sensu Caminer & Ron (2014) (JN790138, 96%), which corresponds to B. steinbachi Rainha et al., 2021). The other lineage is genetically divergent from any nominal or candidate species and most likely represents an undescribed species (Fig. 6B). Analyzing the external morphology of specimens, we are unable to detect any distinctly evident qualitative or quantitative character segregating        these two evolutionary units, indicating a likely case of cryptic speciation, which is common in this species complex (Caminer & Ron, 2014;Fouquet et al., 2021). Considering the B. geographica species complex, we found that one lineage matched to the nominal B. geographica (KU168897-98, 100%) (Fig. 5G) and the second matched to the recently revalidated Boana appendiculata (Boulenger, 1882) (MG840867, 99.8%) (Caminer & Ron, 2020). Morphological analysis of the preserved material referred to the nominal B. geographica fully agreed with the species diagnosis (reviewed by Fouquet et al., 2016), differing from the B. appendiculata, for example, by having dark spots on belly (vs. immaculate belly in B. appendiculata), and orange-reddish hand and foot webbing (vs. cream colored in B. appendiculata). Lastly, two recorded specimens from the snake genus Atractus Wagler, 1828 represent a molecular lineage highly divergent from their congeners. However, this lineage is cryptic to the syntopic species A. snethlageae considering external morphology, and we opt to designate a conservative taxonomic status (Atractus aff. snethlageae) while pending the acquisition of additional data.
Allobates sp. is a species with smaller body size (maximum SVL = 17.8 mm) compared to the two sympatric large-sized Allobates Zimmermann & Zimmermann, 1988 (A. femoralis - Fig. 4A -and Allobates vanzolinius [Morales, 2002], with minimum SVL = 19.6 mm), also differing by its cryptic coloration (vs. brightly colored in A. femoralis), and cream-colored throat in males, in some specimens punctuated with small black dots (vs. uniformly dark throat in males of A. vanzolinius). Some of these specimens could correspond to Allobates gasconi (Morales, 2002), as INPA-H 004986 is a paratype of this species. However, we opt to maintain a conservative taxonomic status, due to the combination of overall uncertainty in attributing taxonomic status of cryptically colored Allobates, historical composite type series , and the fact that Juruá River basin harbor many morphologically similar small-bodied species. We also cannot safely discard the occurrence of more than one species contained under this taxon.

Amazophrynella sp.
Preserved specimens of Amazophrynella sp. have a cream-colored belly punctuated with small black dots, and males have mean SVL = 15.7 mm. Such characteristics are shared by some nominal species from western Amazonia , leading to uncertainty on the identity of these populations. In fact, several populations of Amazophrynella Fouquet et al., 2012 across Amazonia remain uncertainly identified even after the latest taxonomic and biogeographic studies with the genus, which prove the occurrence of extensive morpho-logical conservatism among species Moraes et al., 2022b).

Bolitoglossa sp.
The taxonomic status of the Bolitoglossa Duméril, Bibron & Duméril, 1854 populations from the lower Juruá River basin has been reviewed in a broader Amazonian context, but remains unresolved (Brcko et al., 2013). Given the high morphological similarity among the lowland Amazonian species of Bolitoglossa and the recent indication of at least four genetically distinct lineages occurring at the Juruá River basin , we opt to designate a conservative taxonomic status for these specimens, only recorded in the 1992 expedition. We also cannot discard the occurrence of more than one species contained under this taxon.

Pristimantis sp.
A small-sized species of speciose genus Pristimantis Jiménez de la Espada, 1870, with external morphology overlapping with the known variation of many species (relatively short hind limbs, tuberculate dorsum with a "W" scapular pattern, and an immaculate areolate venter). Historically, populations sharing this general morphology have been attributed across Amazonia to the name Pristimantis ockendeni (Boulenger, 1912). However, it is already evidenced that many undescribed species are masqueraded under this name (Elmer & Cannatela, 2008). In this case, assignments to any taxonomic status would be putative without additional data because of the same reasons for the case of Allobates sp. Therefore, we also cannot discard the presence of more than one species contained under this taxon.

Pristimantis cf. diadematus
A morphologically diagnosable subgroup within the Pristimantis unistrigatus (Günther, 1859) species group shares very large digital tips, relatively short hind limbs, tuberculate dorsum, a "W" scapular pattern, and prominent leg barring (Hedges & Schlüter, 1992). However, this subgroup (Pristimantis diadematus series) was evidenced to be non-monophyletic (Hedges et al., 2008). We recorded at the RBJ one species attributed to this morphological subgroup, with its characteristic mostly fitting in the diagnosis of Pristimantis eurydactylus Hedges & Schlüter, 1992, notably by possessing a visible tympanum and a smooth dorsum. However, P. eurydactylus is little known and much of its associated information comes only from its original description (Hedges & Schlüter, 1992). In addition, its diagnostic characteristics are quite subject to variation and is difficult to elucidate the taxonomic and geographic boundaries between P. eurydactylus and Pristimantis diadematus Jiménez de la Espada, 1875, which is more widely distributed in the Amazonian Andean foothills and adjacent lowlands (Jiménez de la Espada, 1875; Frost, 2022). Without a finer taxonomic resolution and new field data, it is impossible to assign a more precise taxonomic status to this RBJ's population. In addition to the aforementioned characters, P. cf. diadematus has a belly marked by contrasting small dark dots and greenish flanks in life.
Leptodactylus cf. mystaceus (Fig. 8C) The species-pair Leptodactylus mystaceus (Spix, 1824) and Leptodactylus didymus Heyer, García-Lopez, & Cardoso, 1996 is totally cryptic with regard to external morphology, only being phenotypically distinguished by advertisement calls (pulsed in L. mystaceus and nonpulsed in L. didymus) (Heyer et al., 1996). Leptodactylus didymus is known to occur at southwestern Amazonia, with punctual records at the lower Juruá River region (Heyer et al., 1996), while L. mystaceus is more widely distributed in Amazonia (Heyer et al., 1996;Silva et al., 2020). However, the widest knowledge gap on the distribution of this species pair is spatially congruent to where they are potentially sympatric, including our study area (Heyer et al., 1996;Silva et al., 2020). As external morphology was the only available information for the assignment of taxonomic status in this case, we opt to be conservative until acoustic vouchers are obtained in the focal area.

Remarkable records
Strabomantis sulcatus (Cope, 1874) (Fig. 5A) This species is the most widespread in the genus Strabomantis Peters, 1863, which is more diverse in upland areas of the Andean mountain range (Lynch, 1997;Hedges et al., 2008). This record represents an extension of its distribution range, ca. 400 km east of the nearest known locality of occurrence (see Lynch, 1997). As S. sulcatus is the only species of the genus widely distributed in lowland Amazonia, this record also expands the known distribution for the genus Strabomantis (Hedges et al., 2008). Based on the fact that Strabomantis species are rarely reported for Brazilian Amazonia, are quite similar considering external morphology, and we still have little knowledge on the genetic variation within S. sulcatus, other species could likely be masqueraded under this name across lowland Amazonia.

Phyzelaphryne nimio Simões et al., 2018 (Fig. 5C)
Several populations of small eleutherodactylids in lowland Amazonia are historically attributed to a single taxon of the genus Phyzelaphryne Heyer, 1997(Phyzelaphryne miriamae Heyer, 1977. Recently, the real diversity within the widespread P. miriamae has been clarified, indicating that at least two molecularly divergent lineages representing candidate new species were masqueraded under this name (Simões et al., 2018). Considering this taxonomic history, we initially thought that specimens from the RBJ could be conspecific with the "candidate new species 1" sensu Simões et al. (2018), which occur at the middle Juruá River basin, or representatives of a never sampled new species. Surprisingly, our molecular and morphological evidence supports the identity of this population as conspecific with the recently described P. nimio, known only from the Japurá River basin (Simões et al., 2018). This record represents a southeast extension of ca. 280 km for the distributional range of P. nimio, as well as the first evidence of its occurrence at the southern bank of the Amazon River.

Rhinella exostosica Ferrão et al., 2020 (Fig. 4D)
Recent descriptions of Amazonian species of the taxonomically challenging R. margaritifera species group (Santos et al., 2015;Ávila et al., 2020;Ferrão et al., 2020) significantly improved the taxonomic resolution of different populations across this ecosystem. Due to these advances, we were able to confirm the presence of R. exostosica in the RBJ. This species was recently described to the extreme southwestern Amazonia, and our record represents the first in the state of Amazonas, extending its distribution by ca. 650 km northeast. Such a finding rejects the hypothesis that R. exostosica distribution is limited to the west by the Madeira River , at least considering a broad spatial scale.

Dendropsophus kamagarini Rivadeneira et al., 2018 (Fig. 6C)
The taxon Dendropsophus parviceps (Boulenger, 1882), formerly considered to be widespread in western Amazonia, was recently taxonomically revised under integrative approaches (Rivadeneira et al., 2018). This revision revealed three molecularly and morphologically diagnosable lineages contained under this name (Rivadeneira et al., 2018). However, due to the undersampling in this study, an overall low resolution on the identity of "D. parviceps" populations occurs in Brazilian Amazonia. Our molecular and morphological analyses support the presence in the RBJ of the "southwestern clade" of the former widespread D. parviceps, now named D. kamagarini (Rivadeneira et al., 2018).

Dendropsophus mapinguari Peloso et al., 2016
A member of the Dendropsophus leucophyllatus (Beireis, 1783) group recently described, D. mapinguari is until now only known to three localities in central Amazonia, Brazil (Peloso et al., 2016). In the original description, authors suggest that this narrow distribution could be underestimated, as they are segregated by large rivers already recognized as geographic barriers for some Amazonian frogs (Fouquet et al., 2014), and D. mapinguari frequently exploit the flooded habitats, which favors individual dispersion and gene flow (Harvey et al., 2017). In fact, D. mapinguari was recently reported for a new locality in the Negro River basin (Igor Fernandes, pers. comm.), and this record extends the known distribution of this species ca. 600 km west of this (nearest) locality.

Dendropsophus miyatai (Vigle & Goberdhan-Vigle, 1990)
This small-sized and elusive species is originally described as distributed in Ecuadorian Amazonia (Cisneros-Heredia, 2005). Since then, populations within western lowland Amazonia in Peru, Colombia, and Brazil have been attributed to this name. However, these localities of occurrence are in some cases patchily distributed, and the full range of this species is overall irresolute (Cisneros-Heredia, 2005). This is partially explained by the fact that the species mostly exploit seasonally flooded habitats, which are historically neglected in standardized inventories (Moraes et al., 2022a), and by misidentification with other Dendropsophus Fitzinger, 1843. In the light of available evidence, D. miyatai is widely distributed in the margins of the main western Amazonian rivers, extending through the east following the course of the Amazon River.

Alopoglossus indigenorum Ribeiro-Júnior et al., 2021 (Fig. 10C)
Both the analyses of molecular and morphological (meristic) variation of a recently collected specimen of the Alopoglossus angulatus (Linnaeus, 1758) species complex from the RBJ showed that it represented an undescribed taxon. This species was then recently described as A. indigenorum, containing the specimen from the RBJ as part of its type series (Ribeiro-Júnior et al., 2021). The species is distributed in the extreme western Brazilian Amazonia, at south of the Amazon River.
Atractus poeppigi (Jan, 1862) (Fig. 12F) The combination of cryptozoic habits of species and shortfalls of biological knowledge on Amazonia usually provide challenging cases in defining distributional limits for the snake genus Atractus Wagler, 1928 . For instance, A. poeppigi (already considered a synonym of Atractus elaps Günther, 1858 [Dixon et al., 1976]) is a representative of an Atractus species with an irresolute distribution, composed by patchily distributed localities of occurrence. With occurrence records ranging from the Andean foothills to the central lowland Amazonia (Nogueira et al., 2019), A. poeppigi can be considered widely distributed, and records in the RBJ (three individuals) help to fill some of the knowledge gaps in its distribution.

DISCUSSION
Our study reinforces the importance of maintaining effective protected areas at the lower Juruá River and shows the value of the RBJ as a conservation unit in the context of the Amazonian amphibian and reptile diversity, also helping to promote relevant subsidies for its management. This assumption was exemplified by the presence of many species in the RBJ usually rarely recorded across Amazonia, typical of the western and southwestern Amazonia sub-regions, or potentially endemic to the study area.
Our first sampling at the RBJ (1992) provided initial insights regarding the local diversity of amphibians and reptiles, while most recent expeditions (2005)(2006)(2007)2018) complemented this knowledge and supported the creation, management and monitoring of this protected area. Even after the first two expeditions carried out to this region, 18 new species of amphibians and reptiles were reported for the RBJ in 2018. Therefore, an increase in the species richness of these vertebrates known for the RBJ as a result of new inventories and refinement in their taxonomic resolution in Amazonia is very likely. In fact, additional inventories on the RBJ are highly encouraged in order to bring taxonomic resolution to amphibian species with uncertain status in the genera Allobates, Pristimantis, Bolitoglossa, Leptodactylus, and Amazophrynella, as well as to investigate the presence and current population status of species that were only reported in the 1992 sampling, as A. vanzolinius, D. miyatai, D. mapinguari, and Scinax cruentomma (Duellman, 1972).
By sampling a western Amazonian locality over a long period through active surveys, Duellman (1978) concluded that the vast majority of local species (75%) were recorded in the first 100 person-days of fieldwork. Based on this result and on our estimated rarefaction curves, we believe that the combination of our 245 person-days of active survey sampling plus other three sampling methods (including 2,070 trap-days) has sufficiently assessed the typical assemblages of amphibians and reptiles occuring in the RBJ. Rare species were found through active surveys in the locality sampled by Duellman (1978) only after more than 1,000 person-days of fieldwork. Therefore, and as evidenced in our richness estimation analyses, the list of species presented here certainly should expand with the advancement of knowledge. This might be the case especially considering reptile diversity, which were standardized sampled only in two of our expeditions, and have many species barely detectable due to secretive habits and natural lower densities (Fraga et al., 2014). In this sense, many snake species inhabiting broad geographic ranges in Amazonia, such as Chironius scurrulus (Wagler, 1824), Atractus latifrons (Günther, 1868), Hydrops martii (Wagler, 1824), Hydrops triangularis (Wagler, 1824), Siphlophis compressus (Daudin, 1803), and Xenopholis scalaris (Wucherer, 1861), are highly expected to occur in the RBJ (see Nogueira et al., 2019).
Data obtained on our expeditions were relevant to improve the knowledge on the effects of the Juruá River as geographical barriers to amphibians (see Gascon, 1996;Lougheed et al., 1999;Gascon et al., 2000), but also on the taxonomy of Amazonian amphibians and reptiles. For example, the microhylid Chiasmocleis avilapiresae (Peloso & Sturaro, 2008), the aromobatids A. vanzolinius and A. gasconi (Morales, 2002), and the gymnophthalmid A. indigenorum (Ribeiro-Júnior et al., 2021) were described using this material. Furthermore, specimens collected at the RBJ also provide relevant geographic novelties, supporting distribution extensions for the allophrynid Allophryne resplendens Castroviejo-Fisher et al., 2012, the leptodactylid A. simonstuarti , the dipsadid Erythrolamprus pygmaeus (Cope, 1868) (Kawashita-Ribeiro et al., 2011), and the viperid Botrocophias hyoprora (Amaral, 1935) (Carvalho et al., 2013). Conversely, some of the records reported to the RBJ in the literature are uncertain and noteworthy. In a revision dealing with Amazonian caecilians diversity, Maciel & Hoogmoed (2011) reported a specimen of the siphonopid Siphonops annulatus (Mikan, 1822) to the RBJ (INPA-H 004401). However, this specimen was actually collected in the upper Juruá River, and to our knowledge, no record of this species is unambiguously attributed to the RBJ. Similarly, the specimen INPA-H 002417, defined as a paratype of the hylid T. cunauaru (Gordo et al., 2013), was collected in the middle Juruá River, but mistakenly reported to the RBJ. We eventually confirmed the presence of T. cunauaru in the RBJ in our inventories, but S. annulatus' presence remains uncertain. Such a pattern of geographic errors associated with the material collected in the 1991-1992 expedition across the Juruá River indicates that such mistakes may be occurring more widely in the literature using this material.
Although Amazonia represents a region of historical interest for inventories of amphibians and reptiles, the comparability across studies over time is hampered by the highly distinct sampling efforts caused by logistical challenges and pervasive biodiversity shortfalls, which introduce many biases in the reported results (Bush & Lovejoy, 2007;Hortal et al., 2015;Oliveira et al., 2016). However, some biogeographic patterns are repeatedly recovered and appear to be robust. For instance, a westeast gradient of decrease in species richness is consistently reported considering broad taxonomic scales (e.g., Rangel et al., 2018;Oberdorff et al., 2019;Ritter et al., 2019). The west-east diversity gradient is also noticed on a regional scale in the Juruá River basin. In addition to the higher diversity reported in the upper Juruá River basin, there are several species substitutions reported along the course of this river (Gascon et al., 2000;Azevedo-Ramos & Galatti, 2002). Such a spatial variation in species composition and richness along the course of the Juruá River was also noticed for plants (Tuomisto et al., 2016), and other vertebrates as small mammals . This diversity gradient is likely a result of the geological, geomorphological, and pluviometric gradients crossed by this river basin Tuomisto et al., 2016;Pupim et al., 2019). In fact, a considerable number of amphibian species are geographically constrained to the upper Juruá River basin (Gascon et al., 2000;Fonseca et al., 2019), particularly in the montane region of the Serra do Divisor, which emerges as one of the most species-rich localities for these vertebrates in the entire Neotropics (Souza, 2009). Even so, our study reveals that much of the diversity previously reported in the Juruá River basin is underestimated by limited sampling effort, since the number of species found in the RBJ is among the highest reported for single Amazonian localities (see compilation by Freitas et al., 2020).
Amphibian and reptile species substitutions across Amazonian assemblages are also especially evident considering: (1) the main geological compartments, i.e., sedimentary basin of western Amazonia vs. Quaternary crystalline shields of eastern Amazonia (e.g., Ávila-Pires, 1995;Silva-Jr. & Sites-Jr., 1995;Rojas et al., 2018;Réjaud et al., 2020;Ribeiro-Júnior et al., 2021); and (2) the Amazon River as a geographic barrier, i.e., its northern vs. southern banks (e.g., Silva-Jr. & Sites-Jr., 1995;Godinho & da Silva, 2018;Rojas et al., 2018). Considering these main biogeographic units, the RBJ is inserted at western Amazonia sedimentary basin and at the southern bank of the Amazon River. This influence is particularly evident considering the high number of species recorded which are typical of these biogeographic units, while no species typical from the eastern Amazonia was recorded. Meanwhile, a degree of faunistic compositional homogeneity in the entire southwestern Amazonia sub-region is expected because of the high course dynamism of the Juruá River, and this pattern has already been evidenced for amphibians and small mammals (Gascon, 1996;Gascon et al., 2000). However, some species typical of western Amazonia recorded in the RBJ should not be considered as homogeneously distributed in the eastern direction, i.e., limited by the Madeira River, historically attributed as a boundary of a southwestern Amazonia area of endemism for birds (Cracraft, 1985). In fact, some of the species recorded in the RBJ have their geographic distribution expanding towards the central Juruá-Purus interfluve (Prudente et al., 2010(Prudente et al., , 2013, but apparently no longer occurring further East (e.g., the amphibians S. sulcatus, A. simonstuarti, A. resplendens, P. nimio, Edalorhina perezi Jiménez de la Espada, 1870, and the lizard A. indigenorum). This is supported by absence of these species reported so far in the Purus-Madeira interfluve, which has been historically well-sampled for both amphibians and reptiles (see Dias-Terceiro et al., 2015;Ferrão et al., 2016;Peixoto et al., 2019;Nogueira et al., 2019). This distributional pattern suggests that the RBJ's longitudinal zone may represent the maximum limit for their geographic occurrence. This would explain their rarity in our samplings, and they may also constitute isolated populations (see Carvalho et al., 2020). In addition to the Purus River as a geographic barrier, the distribution and dispersal of these species typical from the western Amazonia may be controlled by subtle gradients of climate, soils and vegetation types (see Tuomisto et al., 2019).
Regarding conservation status, according to the latest national list of threatened species (ICMBio, 2018), the amphibian and reptile species reported here for the RBJ are mostly in the Least Concern category (LC; 79%), which is a common event for Amazonian species since many of them are attributed to large extents of occurrence (Neckel- . However, a consid-erable percentage of the recorded species has not yet been properly assessed as to its conservation status (NE; 12%); the strabomantid Pristimantis reichlei Padial & de la Riva, 2009 is considered as Data Deficient (DD); and the three podocnemidids Podocnemis expansa (Schweigger, 1812), Podocnemis sextuberculata (Cornalia, 1849), and Podocnemis unifilis Troschel 1848 are considered to be Near Threatened (NT) of extinction. It is noteworthy that the scenario of ongoing and future anthropic changes in Amazonia is quite worrying (e.g., Latrubesse et al., 2017) and can directly or indirectly affect the biological diversity even in localities that are difficult to access such as the case of the RBJ.
Many traditional communities live in the RBJ, surviving mostly from extractivism, and the effective conservation of these environments can only be achieved by following the sustainable use of natural resources (ICMBio, 2009). In fact, even with a more permissive character, the value of sustainable use protected areas as the RBJ in preserving Amazonian habitats has been proven (Milien et al., 2021). We also highlight the need to monitor the expansion of the Juruá municipality, the largest urban nucleus in this region. This municipality is completely surrounded by the boundaries of the RBJ (ICMBio, 2009), and the permissive character of this conservation unit can lead to unsustainable advances that would irreversibly alter some of the RBJ habitats and associated biodiversity. It is also noteworthy the biased conservation caused by environmental differences between the banks of the lower Juruá River. While the non-flooded habitats are more protected on the right bank of the Juruá River by the RBJ and the indigenous territory Kumaru do Lago Ualá, an immense portion of seasonally flooded habitats on the left bank are less represented by protected areas. These habitats are essential contributors to the generation and maintenance of regional diversity, and we strongly suggest that this environmental difference should be taken into account in future assessments of biological conservation in this region.