Amphibian diversity in Serranía de Majé, an isolated mountain range in eastern Panamá

Abstract Eastern Panamá is within the Mesoamerican biodiversity hotspot and supports an understudied amphibian fauna. Here we characterize the amphibian diversity across an elevational gradient in one of the least studied mountain ranges in eastern Panamá, Serranía de Majé. A total of 38 species were found, which represent 17% of all species reported for Panamá. Based on expected richness function and individual-based rarefaction curves, it is estimated that this is an underestimate and that at least 44 amphibian species occur in this area. Members of all three amphibian orders were encountered, represented by ten families and 22 genera, including five species endemic to Central America. Estimated species richness decreased with elevation, and the mid-elevation site supported both lowland and highland species. Our study provides a baseline for understanding the distribution pattern of amphibians in Panamá, for conservation efforts, and for determining disease-induced changes in amphibian communities.


Introduction
Mesoamerica is a global biodiversity hotspot (Johnson et al. 2015). Within this region, Panamá has the second greatest number of reptile and amphibian species, containing 26% of all amphibian species reported for Mesoamerica (Jaramillo et al. 2010). However, a substantial portion of eastern Panamá has been understudied. Geographically, eastern Panamá comprises the northernmost part of the Chocó biogeographical region (Duque-Caro 1990), and it is part of the Tumbes-Chocó-Magdalena global biodiversity hotspot (Mittermeier et al. 1999). This region includes a number of relatively low mountain ranges, including the Serranía de San Blas + Serranía del Darién on Caribbean side, the inland Serranía de Pirre + Altura de Nique + Altos de Quía, Serranía de Majé, and the Serranía de Sapo + Serranía de Jingurudó + Altos de Aspavé + Cordillera de Juradó along the Pacific Ocean (Duque-Caro 1990; Batista et al. 2016). Whitfield et al. (2016) analyzed regional trends and reported that eastern Panamá has a very small number of recognized species in relation to its geographic area, which reflects the limited number of field surveys in the area. A sharp increase in the number of field surveys during the last decade has led to the discovery of several new amphibian species with restricted distribution ranges (e.g., Ibáñez and Crawford 2004;Crawford et al. 20104a;Batista et al. 2014aBatista et al. , 2014bBatista et al. , 2016 supporting the hypothesis that eastern Panamá is a region with a high endemic amphibian diversity. This is in contrast to the claim that it was mainly a dispersal route during the Great American Biotic Interchange (Webb 2006), and was colonized by species groups from the north and South America (Vanzolini and Heyer 1985;Pinto-Sánchez et al. 2012).
One reason to establish baseline estimates of amphibians is to assess changes following loss caused by disease epidemics. The pathogenic fungus Batrachochytrium dendrobatidis (Bd) causes population declines and extinctions of many amphibian species worldwide, particularly in the Neotropics (James et al. 2015, Lips 2016. Bd has caused dramatic declines of amphibian communities in the highlands of western and central Panamá (Lips 1999;Lips et al. 2006;Crawford et al. 2010b). Importantly, to our knowledge, at the time of sampling there were no published data reporting the presence of Bd in the region -though the amphibian species present at this region can either represent the original community or a subset as a consequence of an undetected Bd epidemic. Here, we describe the results from field surveys to characterize α and β diversity along an altitudinal gradient in the isolated Serranía de Majé of eastern Panamá.

Study sites
During the wet season, from June 23 to July 2 2007, we conducted field surveys at three study sites located at a low, middle, and high elevations in the Serranía de Majé. This mountain range is located on the Pacific coast, previously known as Serranía de Cañazas (Myers 1969), and is isolated from others mountainous areas by the Chepo and Chucunaque Rivers (Figure 1; Angehr and Christian 2000). Its highest point, Cerro Chucantí (1,489 m), stands on the eastern end of the mountain range, at the boundary of the Panamá and Darién provinces (Angher and Christian 2000).
The three study sites were located in Lowland Wet/Moist Forest (LWM) below 600 m, and Premontane Rain Forest/Wet Forest (PRW) above 600 m (Holdridge 1967). The sites were: a low elevation site, Centro Cristo Misionero (8. 96N, 78.457W) at 120-150 m elevation; a mid-elevation site, located within the Reserva Natural Privada Cerro Chucantí (8.79N, 78.451W) at 797 m elevation; and, the high elevation site, also located in the Cerro Chucantí private natural reserve (8.80N, 78.462W), near the top of the Cerro Chucantí, at 1,240-1,365 m elevation. The approximate airline distances between the study sites were 19, 18, and 2 km for lowland-mid-elevation, lowland-highland and mid-elevation-highland sites, respectively.

Data collection
The surveys were conducted using the sampling technique "free and unrestricted search", which is considered to be one of the most efficient methods to record a high number of species in a relatively short amount of time (Rueda et al. 2006). Different types of habitat such as forest, streams, ponds, and open areas with grass were surveyed during the day and night. Species identification and individual counts were performed using the techniques 'visual encounter survey' (VES) and 'acoustic encounter survey' (AES). In addition, the search effort invested (in person-hours) at each sampling site was calculated by multiplying the search time by the number of observers, and the catch per unit of search effort for each site was calculated by dividing the number of post-metamorphic amphibians encountered by the search effort at the respective site as estimated by Kilburn et al. (2011).
A few specimens of each species were collected as voucher specimens (Suppl. material 1: Table S1), photographed, and deposited in the reference Collection of Herpetology (specimen tags CH and AJC) at the Smithsonian Tropical Research Institute, and in the Museo de Vertebrados de la Universidad de Panamá (tags MVUP). Amphibians to be preserved were first euthanized using Orajel (benzocaine 20%) or occasionally 10% ethanol. Before fixation, liver samples were taken from each specimen and preserved for future phylogenetic and phylogeographic analyses (Seutin et al. 1991). Vouchers were then fixed in 10% formalin in a position that facilitates examination. To verify the identification of specimens we used all relevant literature available on the amphibians of Panamá (e.g., Ibáñez et al. 1999a), and compared specimens with those in the CH reference collection. The identification of anuran advertisement calls was facilitated by audio recordings of Panamanian frogs (Ibáñez et al. 1999b).

Data analyses
We calculated α diversity based on all post-metamorphic amphibians captured at each site (Hortal et al. 2006), using the software EstimateS 8.0.0 (Colwell 2006). We calcu- lated Mao Tau (Colwell et al. 2004) and plotted sample-based rarefaction curves with 95% confidence intervals.
To determine β diversity for assessing the variation in species composition across sites, we also used all post-metamorphic amphibians captured at each site, and conducted a cluster analysis based on Jaccard dissimilarity measures estimated with the R function vegdist from the vegan package (Oksanen et al. 2017). In order to identify clusters, we built a dendrogram using the unweighted pair-group method based on arithmetic averages (UPGMA), using function hclust from the default R package stats. This analysis was completed in the R version 3.3.3 (R Core Team, 2017).

Results
Our team conducted 280 person-hours of surveys (lowland site: 125; mid-elevation site: 96; highland site: 59) and identified 38 amphibian species from all three amphibian orders, ten families, and 22 genera ( Table 1). The total number of species for the surveyed area within the Serranía de Majé was estimated as 44 species based on the upper 95% confidence interval of the Mao Tau function (Table 2). Table 1. List of species and number of post-metamorphic individuals found at the three surveyed sites across the elevational gradient in the Serranía de Majé. The letter 'L' refers to a species that was recorded by its larvae and 'V' by its vocalizations. The IUCN conservation status is based on the IUCN (2018). 'E' represents a species that is endemic to Central America (CA) based on Johnson et al. (2015). The greatest number of species was found at the lowland site (24 spp., Table 2; individuals catch per unit of search effort: 1.04), where the search effort was the highest, and where multiple aquatic habitats were available (i.e. ponds and forest streams). The most abundant species at this site were Diasporus aff. diastema (sensu Batista et al. 2016), Craugastor raniformis and Engystomops pustulosus (Table 1). Espadarana prosoblepon and Agalychnis callidryas were detected at this site with larval surveys. The mid-elevation site had fewer species than the lowland site (22 spp., Table 2; individuals catch per unit of search effort: 1.03); however, despite lower search effort at this site, the upper 95% confidence intervals of the Mao Tau function estimated very similar species number (i.e., 25 spp. at the lowland site and 26 spp. at the mid-elevation site). The most abundant species at the mid-elevation site were Dendrobates auratus, Silverstoneia aff. nubicola and Smilisca sila (Table 1). In addition, at this site two species were detected only by their vocalizations: Pristimantis taeniatus and Leptodactylus savagei. The lowest richness was observed at the highland site (13 spp., Table 2; frog catch per unit of search effort: 2.81), because of limited searching effort, fewer habitats, and the lower diversity of the upland area. The estimated number of species for this site based on the upper 95% confidence intervals of the Mao Tau function was 17 spp., and the most abundant species at this site were Rhaebo haematiticus, Pristimantis caryophyllaceus, P. cruentus, and P. moro (Table 1). Moreover, the glassfrog, Espadarana prosoblepon, was detected at this site only by its vocalization.
The individual-based rarefaction curves for the total area surveyed ( Figure 2A) and at the site level ( Figure 2B) showed a substantial decrease in the slope as the number of individuals increased with search effort. Thus, while the upper 95% confidence interval of the Mao Tau function suggests that not all species present in the area were observed, the amphibian community determined in these surveys might be representative of the extant community in Serranía de Majé.
Based on the Jaccard dissimilarity coefficients calculated, the community composition was more similar between the low and mid-elevation sites relative to the high elevation site (Table 3, Figure 3). As expected, the sites that most differed were the low versus high elevation sites. However, six species were consistently present across all three elevation sites: Rhinella alata, Espadarana prosoblepon, Craugastor crassidigitus, Colostethus aff. pratti, and Silverstoneia aff. nubicola.  Table 3. Number of species shared between pairs of sites along an elevational transect of the Serranía de Majé (below the diagonal); total number of species per site including the species registered by post-metamorphic stages, vocalization, or larval stage (diagonal); and Jaccard similarity coefficients (1 -dissimilarity estimate) for each pair of sites (above the diagonal).

Discussion
The present study represents the first attempt to characterize the composition and altitudinal diversity pattern of the amphibian community from the isolated Serranía de Majé of eastern Panamá. We determined that the composition of the species community across the altitudinal gradient was comprised by species from both Mesoamerican and South American groups, and that taxonomic genera from South America dominated the composition of the community (South American genera: 82%; Mesoamerican genera: 18%). In addition, the observed proportion in the composition of genera is consistent with the diversity pattern determined by Savage (2002) for eastern Panamá, where genera from South American groups represented over 50% of the genera comprising the amphibian assemblage.
The species found during this study represent 17% of the native amphibian species of Panamá (AmphibiaWeb 2018). However, the estimated total number of species based on the rarefaction analysis suggests that the richness of the study area is slightly higher than what we observed. In addition, the recent discovery of two new amphibian species from the Serranía de Majé, Bolitoglossa chucantiensis and Diasporus majeensis (see Batista et al. 2014bBatista et al. , 2016, suggest that this region might be high in endemism, as previously suggested for eastern Panamá (Crawford et al. 2010a).
Amphibians, occurring in Central America, have their highest species richness at intermediate elevations (Savage 2002, Wiens et al. 2006, Whitfield et al. 2016). This general altitudinal diversity pattern might also apply to the Serranía de Majé considering that, despite the relatively lower search effort and lower number of individuals encountered at the mid-elevation site, we observed and estimated a species richness simi- lar to that of the lowland site (i.e., site with the highest observed richness). In addition, we determined similar estimates of individuals catch per unit of search effort (lowland site: 1.04 vs. mid-elevation site:1.03) between the lowland and mid-elevation sites. Hence, these results suggest that an increase in sampling effort at the mid-elevation site will potentially increase the number of species detected. Lastly, the observed altitudinal pattern of species richness could have been influenced by the variation across sites in the area covered during the surveys and the habitat types present at the sampling sites. In particular, the number of observed species at the highland site was potentially affected by the absence of streams and ponds, and the reduced patch size of the cloud forest.
In terms of β diversity, the higher similarity in the community composition between the mid-elevation and highland sites compared to that between the lowland and highland sites, suggests that the composition at intermediate elevations in Serranía de Majé might result, in part, by an overlap in the altitudinal distribution of the species associated with higher and lower altitudes; a pattern previously observed for the anuran communities from the Panamá Canal watershed (Ibáñez et al. 2002). In addition, despite the mid-elevation and highland sites being closer to each other (i.e., ~2 km apart) than to the lowland site, the community composition between the mid-elevation and highland sites was less similar than the composition between the mid-elevation and lowland sites. The higher similarity in the community composition between the midelevation and lowland sites compared to that with the highland site suggests that the highland site might be comprised by species with restricted altitudinal distributions. For instance, the dissimilarity associated with the highland site in our study was potentially influenced by the observation of species with restricted altitudinal ranges, such as: Pristimantis aff. latidiscus, P. caryophyllaceus, P. moro, P. pardalis and Diasporus majeensis.
The Serranía de Majé is isolated from the other mountain ranges in the region by the valleys of the Chepo and Chucunaque Rivers (Figure 1), which could have represented physical barriers leading to genetic isolation of populations that could have resulted in allopatric speciation (Cadena et al. 2011). Preliminary results from a comparison between the amphibian communities from the Serranía de Majé and Serranía de Piedras-Pacora (Ibáñez et al. 1994, Sosa andGuerrel 2013), located across the valley of the Chepo River (Figure 1), showed a lower species diversity at the Serranía de Majé and a decrease in the similarity of species composition as elevation increases (Figure 4). In addition, the highest elevations studied at these two mountain ranges are about 106 km apart (airline distance), and their dissimilarity is largely due to the disproportionate number of species that are present in Serranía de Piedras-Pacora but potentially absent in Serranía de Majé. Hence, thus far, seems that dispersal limitation has potentially played a major role in shaping the amphibian community at Serranía de Majé; nonetheless, more studies would be necessary to address this. Lastly, the decrease in similarity in species composition as elevation increases is consistent with the general pattern of amphibians endemism observed in Central America, that shows that a substantial portion of endemic species in the region are associated with upland regions (Whitfield et al. 2016).
Central America, while being a hotspot for amphibian diversity, is a region with a high proportion of threatened amphibian species. For instance, 41% of the regional pool of species that have been assessed by the IUCN (International Union for Conservation of Nature) are under one of the following categories of the Red List of Threatened Species: critically endangered, endangered or vulnerable (reviewed in Whitfield et al. 2016). Within this context, the amphibian community of the Serranía de Majé does not seem, at first, to be comprised of species of high conservation concern given that 76% of the species registered in this study are under the category 'least concern' of the IUCN Red List of Threatened Species (IUCN 2018). However, the Serranía de Majé harbors amphibian species that could be regarded as threatened species, as well as poorly known species lacking an evaluation of their conservation status. For example, based on the IUCN criteria, two of the recorded species are considered near threatened (i.e., Pristimantis caryophyllaceus and P. pardalis), and two others are data deficient (i.e., Caecilia isthmica, and Hyalinobatrachium vireovittatum). Importantly, these two species that are considered near threatened and the two data deficient ones are endemic to Central America (Johnson et al. 2015). Notably, in this study we also found four species, which include one species from the genus Pristimantis (P. aff. latidiscus), two dendrobatids (Colostethus aff. pratti and Silerstoneia aff. nubicola) and one species from the Diasporus diastema species group (i.e., Diasporus aff. diastema suggested by Batista et al. 2016), that are potential new species and, together with the recently described Diasporus majeensis, lack an assessment by the IUCN.
Our survey provides baseline information for exploration and conservation efforts by identifying species in the area requiring immediate assessment and conservation action (Table 1). Importantly, this study might also inform the delimitation of protected areas based on species with restricted distribution ranges. This is particularly relevant given the absence of biological reserves within this mountain range that are recognized by the national system of protected areas (Jaramillo et al. 2010), and the increasing deforestation pressure in the region (Parker et al. 2004). Lastly, considering the arrival of Bd to the Serranía de Majé some years after this study (Küng et al. 2014), the Diagram showing a decrease with elevation in the similarities of amphibian species assemblages associated with sites from the Serranía de Piedras-Pacora mountain range and the isolated Serranía de Majé mountain range. The numbers represent the shared species between sites (N), Jaccard similarity coefficients (N) and total number of species at the site level (N). Each color represents an elevation category, where the lowlands (< 400 m) are represented in yellow, mid-elevation sites (400-800 m) in green, and highlands (> 800 m) in blue. NA = no data available. baseline information provided by this inventory could potentially serve to determine Bd-induced changes in the amphibian community. In particular, at mid and high elevations, where disease-induced losses of amphibian diversity have been substantial in Central America, including Panamá (Lips 2016).