Abstract

We studied the taxonomic status of a population of Pristimantis from the southern Cordillera Azul, Departamento Huánuco, central Peru. A phylogenetic analysis based on the mitochondrial 16S rRNA gene revealed that it represents a lineage within the Pristimantis lacrimosus species group, being the closest relative of a species of uncertain taxonomic status from a lowland rainforest in central Peru (Panguana), and P. pulchridormientes from the Tingo Maria National Park. However, the focal lineage is divergent from all nominal species in the P. lacrimosus group for which respective data are available by >7.9% uncorrected pairwise distance in the 16S rRNA gene fragment. An integrative taxonomic approach, including morphological and bioacoustic analyses, provided multiple lines of evidence for the focal specimens belonging to an unnamed evolutionary lineage at the species level that we describe and name herein. The systematics of Peruvian populations associated with the P. lacrimosus group are discussed, particularly highlighting problematic taxa with uncertain taxonomic status and unknown relationships. We point to scientific challenges and actions needed to achieve a better taxonomic resolution of this species-rich clade of frogs.

Amphibia, bioacoustics, integrative taxonomy, molecular genetics, morphology, systematics

Within the species-rich Neotropical anuran genus Pristimantis, 17 species groups are currently recognized (Köhler et al. 2022; Ortega et al. 2022; Székely et al. 2023). Among these is the Pristimantis lacrimosus species group, which historically has been simply an assemblage within the purely phenetic P. unistrigatus group (Lynch and Duellman 1980) that has been demonstrated to be paraphyletic (see Hedges et al. 2008; Padial et al. 2014). Hedges et al. (2008) formally erected a P. lacrimosus species group and recognized 18 species within this group. In a comprehensive phylogenetic study, Padial et al. (2014) redefined the group according to monophyly criteria, and recognized 25 species as group members. Subsequently, Rivera-Correa and Daza (2016), by adding numerous Colombian samples, demonstrated that the group is composed of two distinct clades, one of which was later referred to as the P. boulengeri species group (Gonzáles-Durán et al. 2017). Finally, Ron et al. (2020) presented a comprehensive phylogeny of the group and related taxa, clarified the identity of the name-giving species P. lacrimosus, and removed and added species to the group, resulting in 36 recognized species. Due to subsequent species descriptions, exclusions, and additions to the group (Rivera-Correa and Daza 2020; Carrión-Olmedo and Ron 2021; Ttito and Catenazzi 2021), currently 40 Pristimantis species are placed in the P. lacrimosus group.

As demonstrated by the most recent progresses in understanding the phylogenetic relationships and taxonomic identity of various populations, the systematics of the P. lacrimosus group are far from being clarified. There are several documented cases of misidentifications and confusion among species, as the result of either the presence of intra-specific morphological variation, and/or inter-specific morphological crypsis (e.g., Paéz and Ron 2019; Ron et al. 2020; Carrión-Olmedo and Ron 2021). Evidently, with progressing research efforts, this group of frogs awaits the discovery of numerous new species as well as the clarification of so far unsolved taxonomic problems. This might be particularly true for Peru, where on the one hand new species of this group are regularly discovered (e.g., Morales 2007; Chávez and Catenazzi 2016; Shepack et al. 2016; Ttito and Catenazzi 2021), but on the other hand several taxonomic problems remain with respect to the identity of many populations (e.g., Morales and Icochea 2000). Duellman and Lehr (2009) listed eight species of the group to occur in Peru. Their species account was based on the assumption that the nominal species P. bromeliaceus, P. lacrimosus, and P. mendax occupy wide ranges along the eastern Andean slopes and the adjacent lowlands (see also Lehr et al. 2006). However, Ron et al. (2020) demonstrated that P. bromeliaceus and P. lacrimosus are actually restricted to Ecuador and northernmost Peru, rendering the majority of Peruvian samples previously identified as belonging to these two species actually unidentified and potentially belonging to unnamed species. Another uncertainty is connected with the name P. mendax, which to our knowledge has not been re-collected at its type locality and has not been genetically studied, leaving its identity partly obscure (see also Morales and Icochea 2000). As a consequence, many Peruvian populations from montane rainforests at the eastern Andean slopes, superficially similar to P. mendax, have been called P. mendax, P. aff. mendax or P. cf. mendax in provisional ways, as the data available prevented a proper evaluation of their taxonomic identity (e.g., Tasker and Twomey 2015; Angulo et al. 2016; Whitworth et al. 2016; Lehr and von May 2017; Lehr et al. 2017a, b; Moravec et al. 2020). The fact that species that were believed to have a wide distribution in Peru, actually occupy much more restricted ranges (P. lacrimosus and P. bromeliaceus; Ron et al. 2020), plus further uncertainties (identity of P. mendax), leaves a considerable lack of information concerning the taxonomic identity of numerous populations that occur in central and southern Peru. Furthermore, some Pristimantis species from Peru share morphological characters with the P. lacrimosus group (e.g., P. rhodostichus; see Chávez and Catenazzi 2016; Shepack et al. 2016), but they are unstudied genetically and thus their relationships remain unknown.

Summarizing the current knowledge, we can expect 16 species of the Pristimantis lacrimosus group to occur in Peru: P. achupalla, P. aureolineatus, P. acuminatus, P. deyi, P. galdi, P. lacrimosus, P. mendax, P. olivaceus, P. padiali, P. pardalinus, P. pluvialis, P. pseudoacuminatus, P. royi, P. schultei, P. tantanti, and P. pulchridormientes (Ron et al. 2020; Ttito and Catenazzi 2021). However, external morphology may argue for exclusion of P. deyi from the P. lacrimosus group (see Lehr et al. 2013), whereas P. rhodostichus could likely be a member of it (see Duellman and Pramuk 1999; Duellman and Lehr 2009; Chávez and Catenazzi 2016; Shepack et al. 2016).

The Pristimantis lacrimosus group has been demonstrated to exhibit great diversity in the montane forests of the Andean slopes (e.g., Rivera-Correa and Daza 2020; Ron et al. 2020; Carrión-Olmedo and Ron 2021), with an apparently high number of unnamed species in the group, as inferred by the presence of genetically highly divergent lineages. It therefore seems obvious that the presence of morphological variability and/or crypsis in combination with scarce molecular and bioacoustic information delayed taxonomic progress in many cases. Here, we contribute to the knowledge of this group by evaluating the relationships and taxonomic status of a population of rain frogs from the southern Cordillera Azul, central Peru, and describing it as a new species.

Fieldwork was conducted in November 2019 in different areas of the departments Huánuco and Ucayali in central Peru (see also Castillo-Urbina et al. 2021; Köhler et al. 2022, 2023). Specimens were observed and collected during opportunistic searching at night using torches and headlamps, often guided by the sounds emitted by calling males. Geographic position was recorded using handheld GPS receivers set to WGS84 datum.

Collected specimens were euthanized using an overdose of 5% lidocaine gel applied on ventral surfaces of the individuals (McDiarmid 1994). Tissue samples (thigh muscle and tongue pieces) were cut prior to fixation and stored in 99% ethanol, while specimens were fixed using 96% ethanol and subsequently stored in 70% ethanol. Specimens were deposited at the herpetological collection of the Museo de Historia Natural, Universidad Nacional Mayor de San Marcos (MUSM), Lima, Peru, and Hessisches Landesmuseum Darmstadt (HLMD), Germany. FGZC refers to Frank Glaw field numbers.

Morphometric measurements (in millimetres) were taken by the first author with a digital caliper to the nearest 0.1 mm. Measurements taken and used throughout the text are: SVL, snout–vent length; TL, tibia length; HW, head width (at level of angle of jaws); HL, head length (from posterior margin of lower jaw to tip of snout); IOD, interorbital distance; ED, horizontal eye diameter; E–N, eye–nostril distance (straight line distance between anterior corner of orbit and posterior margin of external nares); IND, inter-narial distance; TD, horizontal tympanum diameter; HandL, hand length (from proximal border of outer metacarpal tubercle to tip of third finger); FootL, foot length (from proximal border of inner metatarsal tubercle to tip of fourth toe). Fingers and toes are numbered preaxially to postaxially from I–IV and I–V, respectively. Lengths of fingers I and II were determined by adpressing the fingers against each other. For character state definitions, we followed Duellman and Lehr (2009). We use the term ‘hyperdistal tubercle’, as proposed by Ospina-Sarria and Duellman (2019) and Ron et al. (2020), for the tubercle underlying the articulation of the last phalange of each finger and toe. Description and diagnostic schemes follow Köhler et al. (2022). Coloration in life was described based on digital photographs.

Vocalizations in the field were recorded using an Olympus LS-05 digital recorder with built-in microphones at 44.1 KHz and 16-bit resolution and saved as uncompressed WAVE format. Recordings were re-sampled at 22.05 kHz and 32-bit resolution and computer-analyzed using the software Cool Edit Pro 2.0. We obtained frequency information through Fast Fourier Transformation (FFT; width 1024 points) at Hanning window function. Spectrograms were produced with Blackman window function at 256 bands resolution. Sensitive filtering was applied to remove background sounds, only to frequencies outside the prevalent bandwidths of calls. Temporal measurements are provided as range with mean ± standard deviation in parentheses. Analysis, description and terminology follow the recommendations of Köhler et al. (2017), using the call-centered terminological scheme.

Our genetic analysis aimed at identifying lineage divergence among focal lineages of Pristimantis. For representative taxon sampling, we largely followed an approach as described by Köhler et al. (2022). We used BLAST searches (Altschul et al. 1990) of newly generated 16S sequences of focal Pristimantis against the GenBank nucleotide archive and downloaded sequences identified to species rank with an identity >88%. Subsequently, we manually searched for sequences of nominal species that are currently placed in the P. lacrimosus species group (sensu Ron et al. 2020; Carrión-Olmedo and Ron 2021), plus representatives of its sister clade including P. appendiculatus and P. orcesi (Ron et al. 2020), and added a selection of sequences of relevant taxa (not the complete set of sequences available for the groups, which include many unidentified samples). For the P. lacrimosus species groups, we further added the sequences of two new samples of unidentified specimens collected by us at different localities in central Peru. Three sequences of Yunganastes species (Y. ashkapara, Y. mercedesae, Y. pluvicanorus), with Yunganastes being sister to Pristimantis (Padial et al. 2009; Canedo and Haddad 2012), were used to root the tree. For details of used samples see the Zenodo.org repository: DOI: 10.5281/zenodo.10048811.

We sequenced a DNA fragment of the mitochondrial 16S rRNA gene (16S) from tissue samples of newly collected specimens. DNA was extracted using a standard salt extraction protocol, Polymerase Chain Reaction (PCR) carried out with primers 16SAr-L (5’-CGCCTGTTTATCAAAAACAT-3’) and 16SBr-H (5’-CCGGT­CTGAACTCAGATCACGT-3’) (Palumbi et al. 1991), and the PCR products then directly sequenced on automated DNA sequencers by LGC Genomics (Berlin, Germany). All new DNA sequences were submitted to GenBank (accession numbers OR725032‒OR725036).

To estimate evolutionary lineage divergence within the P. lacrimosus species group we used a Maximum Likelihood (ML) phylogenetic approach. Sequences were exported to AliView v1.26 (Larsson 2014), edited and aligned using MAFFT v7.310 (Katoh and Standley 2013), with the L-INS-i parameter as the iterative refinement method (Katoh et al. 2005). Final alignment length of the 16S rRNA gene was 1637 bp, but many samples, including the focal lineage, were represented by a shorter 3’-terminal fragment of ca. 500 bp. We employed the W-IQ-TREE web server (Trifinopoulos et al. 2016) to infer a molecular phylogeny under ML as optimality criterion. We inferred the best evolution model using ModelFinder under the BIC criterion implemented in IQ-TREE, obtaining TIM2+F+I+G4 as the best fitting substitution model for the data set (-m TEST command; Kalyaanamoorthy et al. 2017). To evaluate branch support, we obtained ultrafast bootstrap values from 2000 pseudoreplicates and 10,000 iterations as the maximum number to stop (commands -bb 2000 and -nm 10000 in IQ-TREE) and Shimodaira–Hasegawa approximate likelihood ratio test (SH-aLRT) with 1000 replicates (-alrt 1000 command; Shimodaira and Hasegawa 1999; Guindon et al. 2010). Genetic divergences were quantified as uncorrected pairwise distances (p-distances) on pairwise alignments using MEGA v11 (Tamura et al. 2021) using default settings (pairwise deletion).

As apparent from most recent publications (e.g., Ron et al. 2020; Rivera-Correa and Daza 2020; Carrión-Olmedo and Ron 2021; Ttito and Catenazzi 2021), the Pristimantis lacrimosus species group forms a diverse and species-rich clade within Pristimantinae. With the description of P. loeslein, we added a further taxon to this clade of frogs by providing different lines of evidence for its distinctiveness (morphology, molecular genetics, and bioacoustics). However, taxonomic resolution within this species group is far from being complete. Prior to the application of molecular analyses and inclusion of topotypic samples, there have been numerous documented instances of misidentifications and confusion among species, as a distinction between intra-specific morphological variation on the one hand, and inter-specific morphological crypsis on the other hand is rather difficult (e.g., Morales and Icochea 2000; Ortega-Andrade et al. 2015; Ron et al. 2020; Carrión-Olmedo and Ron 2021). In light of the great species diversity in the P. lacrimosus group, particularly in central and southern Peru (unpublished data), uncovered by molecular analyses, it is paramount to assign existing names to living populations and clarify the taxonomic status of nominal species with uncertain identity. In most cases, this can only be achieved by sampling the type localities of respective taxa, with the goal of obtaining tissue samples for molecular analyses and ideally recordings of advertisement calls. With respect to Peru, certain nominal species deserve comments in this context.

The identity and relationships of Pristimantis mendax, with its type locality in southern Peru (Colonia Pistipata, SE of Huyro, 1820 m a.s.l., Departamento Cuzco) and paratypes from different low and high elevation localities in the Departamentos Ayacucho, Cuzco, and Huánuco (Duellman 1978) remains uncertain (see Introduction). By detailed morphological investigation, Morales and Icochea (2000) suggested that the type series of P. mendax is composed of at least three different species. However, their conclusion that paratypes from lowland Panguana, Departamento Huánuco, correspond to the two nominal species P. bromeliaceus and P. lacrimosus was later indirectly falsified by Ron et al. (2020), who for the first time studied molecular samples of both taxa, which apparently have a much more restricted distribution than previously suggested. One of the unidentified species from Panguana contained in the paratype series of P. mendax likely corresponds to the population included in our molecular analysis as Pristimantis sp. Panguana (Fig. 8), being the closest known relative of P. loeslein (Fig. 1; uncorrected p-distance 4.7%). Likely, this candidate species, with its taxonomic status yet to be clarified, has been identified as P. bromeliaceus by Morales and Icochea (2000), but is actually only distantly related to it (Fig. 1). Based on morphological examination of our Panguana material, we here also confirm that a second species of the P. lacrimosus group (voucher specimen FGZC 6300; Fig. 9) occurs at Panguana (likely the species identified as P. lacrimosus by Morales and Icochea 2000), that differs morphologically from P. mendax and shows greatest similarities with P. royi described by Morales (2007) from a locality only approximately 68 km southwest of Panguana (Fig. 7). To our knowledge, so far no sample reliably assignable to nominal P. mendax has been included in any molecular phylogenetic study. Our attempts to find P. mendax at the type locality in 2022 were unsuccessful.

Figure 8. 

Amplectant couple of Pristimantis sp. (FGZC 6314–6315) in situ at Panguana, Departamento Huánuco, Peru, 260 m a.s.l. This species is the closest known relative of P. loeslein sp. nov. (uncorrected p-distance 4.7%), with its identity yet to be clarified.

Figure 9. 

Living individual (FGZC 6300) of a second unidentified species of the Pristimantis lacrimosus group occurring at Panguana, Departamento Huánuco, Peru, 260 m a.s.l. (no sequence data available). Morphologically, this species is apparently most similar to P. royi, described from a nearby locality (Morales 2007).

Another taxonomic uncertainty concerns Pristimantis olivaceus, with its type locality situated at the lower Andean slopes of central Bolivia (Chapare region, Departamento Cochabamba), that has also been recorded from southern Peru (Pakitza, PN Manu, Departamento Madre de Dios) in the original description (Köhler et al. 1998). Although several southern Peruvian populations likely correspond to P. olivaceus (J.C. Chaparro, pers. comm.), the identity of many populations flagged as it or as P. cf. olivaceus (e.g., Villacampa-Ortega et al. 2017; Moravec et al. 2020; Carrión-Olmedo and Ron 2021) should be regarded with reservation. For example, the population from northern Bolivia (Palmira, Departamento Pando) included in our analysis as P. cf. olivaceus (Fig. 1) likely does not represent nominal P. olivaceus, and the same is true for the record from San Antonio, Departamento Pando, northern Bolivia, by Moravec et al. (2011), as is apparent from the photo provided by these authors (their fig. 2F), showing a species that differs from P. olivaceus by dorsal color pattern, iris color, and snout shape. Our attempts to obtain DNA sequences from topotypical samples of P. olivaceus unfortunately failed.

In addition to these difficulties, the presumed loss of the type specimens of P. royi (MUSM 11186–11187; Morales 2007), which are not present in the MUSM collection (absence of any information about their whereabouts), hampers the clarification of its identity and morphologically very similar populations in the same general area (e.g., Panguana; see above, Fig. 8).

These cases, together with the fact that other nominal species lacking phylogenetic information may actually be part of the P. lacrimosus group (e.g., P. rhodostichus), exemplify the challenging taxonomy of these frogs in Peru and highlight the need for future integrative taxonomic studies, including topotypic samples.

The closest nominal relative of the new species described herein, Pristimantis loeslein, was revealed by our analysis to be P. pulchridormientes occurring at Tingo Maria National Park (uncorrected p-distance 7.9%). Notably, these two species share a similar habitat, altitudinal distribution, and occur in close geographical proximity, with their respective type localities being only around 28 km airline distance apart (see Chávez and Catenazzi 2016). However, the Tingo Maria National Park and the southern Cordillera Azul are separated by the Huallaga river valley, which at Tingo Maria is at an elevation of approximately 640 m a.s.l., and thus about 1000 m lower in elevation than the habitats of both species. The Huallaga valley may constitute a boundary of biogeographic areas and act as a dispersal barrier that significantly influenced the evolution of Pristimantis, leading to speciation through vicariance (Cracraft 1985; Smith et al. 2014). It is easily imaginable that the lower river valley hinders gene flow for montane forest species which are adapted to moister and cooler environments, compared to their relatives in the lowlands. This hypothesis is supported by another species-pair in the Pristimantis conspicillatus species group, namely P. nebulosus, which shares its type locality with P. loeslein in the southern Cordillera Azul and its sister species P. symptosus occurs west of the Huallaga river valley in the Cordillera de Carpish (Köhler et al. 2022).

We are indebted to the Servicio Nacional Forestal y de Fauna Silvestre (SERFOR) for dealing with our research proposals and issuing research and collection permits (RGD 071-2020-MINAGRI-SERFOR-DGGSPFFS and D000067-2021-MINAGRI-SERFOR-DGGSPFFS). We thank Juan C. Chaparro for information on populations from southern Peru. William E. Duellman and Linda Trueb provided access to the collection under their care, and José M. Padial shared his photos of several type specimens of Pristimantis. Juan M. Guayasamin, Edgar Lehr and an anonymous reviewer provided comments that improved the manuscript. Financial support was provided to the first author by the BIOPAT initiative (www.biopat.de).