High Andean environments of central Chile (32°–38°S) are inhabited by several endemic species of the genus Alsodes. Two of them, A. pehuenche and A. hugoi, have geographic distributions restricted to their type locality and surroundings. The Chilean government classifies A. pehuenche as Critically Endangered (like the IUCN) and A. hugoi as Vulnerable. In this study we report 16 new localities of Alsodes, corresponding to first order streams, located in the Andes of Chile between 35°58’ and 36°32’S (1800–2470 m). In some of these sites, adults and juveniles morphologically similar to A. pehuenche and A. hugoi were observed, as well as specimens of Alsodes that could not be identified by their external morphology. A Bayesian phylogenetic analysis with mitochondrial sequences (cytochrome b) was performed to identify the new populations to species level. All populations around 36°S belong to A. pehuenche, while most of those located south of that area would be A. hugoi. The exception is Cajón de Plaza (36°23’S), where specimens with sequences of A. hugoi or A. pehuenche coexist, whose taxonomic status could not be determined. These findings imply not only a westward range extension of A. pehuenche in Chile of about 14.5 km and of A. hugoi about 100 km southward, but also that practically all the first order streams of the Andes of central Chile would be inhabited by populations of Alsodes. Both results demonstrate the scarce biogeographic knowledge of the genus in the Andes, which has important implications for its conservation at local and species levels.

Alsodes pehuenche , Alsodes hugoi , Chilean Andes, distribution extensions, first order streams, conservation, Critically Endangered, Vulnerable

The amphibian fauna of the Andes of southern South America (Chile and Argentina) is relatively poor in comparison to tropical high-altitude environments (Duellman 1979, 1999). In fact, only three anuran genera, Alsodes Bell, 1843, Rhinella Fitzinger, 1826 and Pleurodema Tschudi, 1838, are represented in the Andes south of 30°S above 1500 m. Alsodes is the most diversified of these genera, with five species (A. montanus (Lataste, 1902), A. tumultuosus Veloso, Iturra & Galleguillos, 1979, A. hugoi Cuevas & Formas, 2001, A. pehuenche Cei, 1976 and A. nodosus (Duméril & Bibron, 1841)) present in the western slopes (Chile) between 33° and 36°S (Charrier et al. 2017). Four of these species are endemic to Chile, since A. pehuenche is also present in Argentina in a small area around the border between these countries (Valle Pehuenche, its type locality, 36°S) (Corbalán et al. 2010, Correa et al. 2013).

There are few published records of Andean populations of Alsodes in Chile and most of these were reported in the last decade. For instance, several reports since 2008 increased significantly the distribution ranges and number of known localities of A. montanus and A. tumultuosus from 33°20’ to 35°S (Araya and Riveros 2008, Correa et al. 2008, Mora et al. 2015, Ramírez 2015, Correa 2017). Likewise, new records of A. hugoi (Araya and Cisternas 2008) and A. pehuenche (Corbalán et al. 2010, Correa et al. 2013) extended their distributions, but in both cases within a small area around their type localities. These last two studies are particularly relevant because they extended the distribution of A. pehuenche to Chile, a species previously known just from its type locality in Argentina. The fifth species, A. nodosus, also inhabits the western slopes of Andes, but it is not clear what altitude it reaches (e.g. Cuevas 2013). Despite this significant increase in the number of known localities of Alsodes in the Andes, there are still substantial distribution gaps. For example, there is a stretch of 380 linear kilometers where only A. hugoi and A. pehuenche are known, between the Tinguiririca River basin (34°55’S), the southern end of the distributions of A. montanus and A. tumultuosus, and San Ignacio de Pemehue, type locality of A. vittatus (Philippi, 1902) (38°05’S) (Fig. 1).

Figure 1. 

New and literature records of Alsodes from the Andes Range between 34°50’ and 38°05’S. Yellow circles represent the new localities reported in this study; squares represent type localities. A. Andean localities of Alsodes of the literature between 34°50’ and 38°05’S: the southernmost localities of A. montanus and A. tumultuosus, the type locality of A. hugoi, localities of A. pehuenche (within red box B, see map B) and the type locality of A. vittatus. There is a record of a putative new species related to A. nodosus in Pemehue (Alsodes sp. 1 of Blotto et al. 2013; not included in the map), presumably the same type locality of A. vittatus. Red boxes correspond to the three explored areas described in this study (maps B, C and D). B. Area explored during the first field campaign (Paso Pehuenche, Laguna del Maule and surroundings). All colored symbols correspond to localities of A. pehuenche: yellow circles correspond to new records; green circles and the square are all previously known localities of the species. The white circle is the place where no amphibian was found. C. Area and sites explored during the third field campaign (tributaries of the Guaiquivilo River). D. Area and sites explored during the second field campaign (surroundings of Laguna El Dial). See details of the localities and the populations discovered in Table 1. Orange lines represent the boundary between Chile and Argentina; thinner yellow lines indicate the boundaries of the administrative regions of Chile.

The distribution ranges of A. pehuenche and A. hugoi are extremely restricted. Alsodes pehuenche was first reported as Telmatobius montanus (now A. montanus) by Cei and Roig (1965) at two very close localities located on each side of the border of Chile and Argentina (36°S). The Argentinean locality (Valle Pehuenche) became the type locality when this species was formally described by Cei (1976), but the Chilean locality was subsequently ignored (see comment in Correa et al. 2013). For 45 years it was assumed that A. pehuenche inhabited only in its type locality, located in a drainage basin on the eastern slopes of the Andes (that ultimately flows into the Atlantic Ocean), until Corbalán et al. (2010) extended its distribution about 4 km to the southwest in Chile, in a contiguous drainage basin, but that drains to the west into the Pacific Ocean. Corbalán et al. (2010) also specified that on the Argentinian side the species only occupies five very close streams and indicated that attempts to find individuals in other nearby streams failed. Later, Correa et al. (2013) reported a second locality in Chile, which further extended the distribution approximately 3.4 km west. Similarly, Alsodes hugoi was known since its description only in its type locality (Altos de Vilches, 35°33’S, 900 m, Cuevas and Formas 2001) until Araya and Cisternas (2008) reported its presence in several sites around this locality. Although these new records extended the altitudinal range up to 2115 m, all sites are hydrologically connected to the Lircay River and the maximum distance between them does not exceed 10 km.

The known distribution ranges of these species have been fundamental in establishing their national and international conservation categories. The Chilean government, through its legal instrument Reglamento de Clasificación de Especies Silvestres (RCE, which applies similar criteria to those of IUCN), classifies A. pehuenche as Critically Endangered and A. hugoi as Vulnerable. These categories rest on the application of criterion B, which exclusively considers the extent of the geographical distribution. On the other hand, A. pehuenche is Critically Endangered and A. hugoi is Data Deficient according to the IUCN (2017). This last categorization of A. pehuenche is fundamentally based on estimates of the extent of occurrence (about 9 km²) and area of occupancy (about 5 km²) (IUCN SSC Amphibian Specialist Group 2013).

In this study we report new Andean localities of Alsodes situated in Chile between 35°58’ and 36°32’S. These localities were found during three field campaigns aimed to locate new populations of A. pehuenche, one in the surroundings of the known localities in Chile (around 36°S) and the other two south of the known distribution range of the species. We performed a phylogenetic analysis with mitochondrial cytochrome b sequences to identify most of the populations discovered to species level. We discuss the implications of these findings for the biogeography and conservation of the genus in Chilean Andes.

Considering the three field campaigns we found tadpoles of Alsodes in 18 of the 19 sites explored (Fig. 1, Table 1). In 16 of those 18 sites we also observed adults and/or juveniles (see their tentative identifications in Table 1). The only site where no specimen was observed was a relatively short and narrow stream located north of Laguna del Maule (Stream km 143.5, Fig. 1), which at the sampling date (January) had an intermittent flow. Most of the other streams had a continuous flow, except Vegas de Miranda and Vega Sin Nombre. In both sites we observed large tadpoles in isolated remnant pools located in dry sectors of the streambeds. Tadpoles could not be identified to species by their external characteristics, so their identification relied exclusively on the phylogenetic analysis. Some of the adults and juveniles were tentatively identified as A. pehuenche or A. hugoi by their external characters (Table 1), but most specimens from Vegas de Miranda could not be identified to species level (Fig. 2D).

Figure 2. 

Adult males of Alsodes from the new discovered localities. In parentheses the specific identification according to the phylogenetic analysis (Fig. 3), the snout-vent length (SVL) and code of the respective buccal mucosa sample are indicated. A. Vegas de Cisternas (A. hugoi, SVL = 68.3 mm, VC1m). B. Cajón de Plaza (undetermined, SVL = 54.8 mm, CP5m). C. Vegas de Andrade (A. hugoi, SVL = 57.2 mm, VA1m). D. Vegas de Miranda (A. hugoi, SVL = 56.3 mm, VM6m). E. Cajón Lagunitas (site 3) (A. hugoi, SVL = 71.7 mm, CLP3-5m). F. Lo Aguirre Chico (A. pehuenche, SVL = 52.0 mm, AgCh4m). Some populations are characterized by well-developed interdigital webbing in the hind feet (shown in the insets).

We obtained an alignment of the cytochrome b of 933 nucleotide sites, although this matrix included a few sequences of Blotto et al. (2013) of 318 base pairs. We did not observe effects of saturation or stop codons in the cytb sequences. The phylogenetic analysis shows that all the new populations are related to A. pehuenche or A. hugoi (most) (Fig. 3). All specimens from Laguna del Maule and surroundings comprise one well-supported clade together with the reference samples of A. pehuenche (from the type locality, Valle Pehuenche, Blotto et al. 2013, and from Paso Pehuenche and Stream km 154.7, Correa et al. 2013). The geographic distribution of this lineage implies a range extension of this species of 14.5 km to the west (Lo Aguirre Chico and Lo Aguirre Grande, Fig. 1B). Within this clade also four specimens from Cajón de Plaza (Figs 1 and 2) are included (see comment below). The genetic divergence among the specimens within this group is extremely low, including some that have identical sequences coming from localities separated by more than 60 km (Lo Aguirre Grande and Cajón de Plaza, Fig. 1). On the other hand, almost all samples coming from the ten southern sites (field campaigns two and three, Figs 1C and 1D), which are genetically and phenotypically more variable (Fig. 2), comprise two well-supported sister clades in the Bayesian consensus tree (Fig. 3). One of these clades includes the representative of the type locality of A. hugoi (Altos de Vilches). These clades do not exactly match the two groups of geographically closest localities, since specimens of some localities are distributed into the two clades, so we consider them as conspecific with that species. Although the specimens that constitute these two groups come from localities separated by up to 100 km they have very low genetic distances, which reinforces the hypothesis of their conspecificity. The population of Cajón de Plaza (Fig. 1D), represents a special case, since there specimens with sequences of the A. pehuenche and A. hugoi type are mixed, but they all resemble A. pehuenche phenotypically (Fig. 2B).

Figure 3. 

Bayesian consensus tree (50% majority-rule) showing the relationships of the new Andean populations of Alsodes. Representatives of the new populations are labeled in green (related to A. pehuenche) and red (related to A. hugoi) (see details of the new localities in Table 1). Note that specimens from Cajón de Plaza are distributed in both the red and green clades. Reference sequences of the type localities of A. pehuenche and A. hugoi are in bold. Numbers next to the nodes correspond to posterior probabilities (only values ≥ 0.95 of the more internal nodes are shown). The scale bar in the lower left corner represents the expected substitutions per site along the branches.

In summary, the phylogenetic analysis indicates that all sampled individuals belong to only two previously known species in the Andes Range between 35°30’ and 36°S, A. pehuenche or A. hugoi. These results imply the extension of the distribution range of A. pehuenche 14.5 km to the west in Chile, and of A. hugoi of 100 km to the south. Also, the altitudinal range of A. hugoi was slightly increased since 2115 m (Araya and Cisternas 2008) to 2157 m (Table 1). We do not recognize Cajón de Plaza as a new population of A. pehuenche because even though all the specimens we observed resemble this species externally (Fig. 1B), some of them have mitochondrial sequences of the A. hugoi type. More phenotypic and genetic data are needed to determine the taxonomic status of this population.

This study reports new localities of two species of Alsodes endemic to the Andes Range, whose geographic distributions were considered extremely restricted. These findings add to a series of Andean populations of Alsodes reported in the last decade between 33°25’ and 36°S in Chile (Araya and Cisternas 2008, Araya and Riveros 2008, Correa et al. 2008, Corbalán et al. 2010, Correa et al. 2013, Mora et al. 2015, Ramírez 2015, Correa 2017), which have expanded considerably our knowledge about the geographic distribution of the genus. Below we discuss the implications of these discoveries for the biogeography and conservation of the species involved and the genus.

The localities of A. pehuenche reported here constitute the third and widest distribution extension since the species was described (Cei 1976), more than doubling the known distribution until 2013 (Correa et al. 2013). Moreover, the occurrence of this species is extended to another hydric system in Chile (Lo Aguirre Grande and Lo Aguirre Chico), not connected directly to Laguna del Maule (Fig. 1). However, the area of occupancy is still very small as this species is restricted to watercourses (see comment below). Although all known sites in Chile are found within the hydrographic basin of the Maule River, it is possible that currently populations from Lo Aguirre Grande and Lo Aguirre Chico are demographically disconnected from those streams that discharge directly into the Laguna del Maule. The new localities also allow us to discard the presence of A. montanus around Laguna del Maule mentioned by Cei and Roig (1965). We could not confirm the presence of A. pehuenche south of 36°S. Although in Cajón de Plaza we found specimens phenotypically similar to A. pehuenche, even with mitochondrial sequences identical to those found in the type locality, they coexist with individuals with sequences of A. hugoi. Currently, we have no additional data to identify the possible causes of this mixture of mitochondrial sequences, so we consider the taxonomic status of this population as uncertain. The distribution extension of A. hugoi reported here is much more significant, reaching the northeastern end of the Ñuble River basin about 100 km south of the type locality (which is in the Maule river basin). Moreover, this species presents high levels of geographical variation in attributes such as body coloration and development of interdigital webbing, which had been described so far only in another species of the genus, A. gargola (Blotto et al. 2013).

These findings have important implications for the biogeography of the genus at different spatial scales. The new localities discovered south of 36°S are in a stretch of the western slopes of the Andes of about 250 km where no populations of Alsodes were known. However, an apparent discontinuity in the distribution of the genus of approximately 180 km still persists, covering almost completely the mountainous zones of the Administrative Region of Biobío (Fig. 1A). The relief, altitude, climatic conditions and types of vegetation in this area of the Andes are similar to those found where the new populations of A. hugoi were discovered, so this distribution gap is probably due to a lack of exploration. The new localities of Alsodes recently reported in the Andes Range (see references above), several of them located in areas of difficult access, support this idea. One of the issues that remains to be clarified is the reciprocal distribution limits between A. hugoi and A. pehuenche, since A. hugoi is to the south and north of the Maule River, whereas A. pehuenche is restricted to the head of that river at an intermediate latitude. Although there seems to be an altitudinal segregation between them (A. hugoi is between 900 and 2157 m, A. pehuenche between 2215 and 2463 m; Table 1), the data are very scarce and fragmentary to establish if one species replaces the other along the water systems. The mixture of mitochondrial sequences at Cajón de Plaza is compatible with the syntopy of both species, but this issue must be assessed with more phenotypic and genetic data.

At a smaller spatial scale, the effectiveness of our searches, together with previous antecedents, suggest that the genus is present in practically all first order watercourses of the western slopes of the Andes. We found tadpoles in 18 of the 19 sites explored (adults and/or juveniles in 16), with a sampling effort of up to three hours per day per site. Similar patterns have been reported for species of the same genus in the Andes. For example, Araya and Cisternas (2008) found reproductive specimens of A. hugoi in the 10 sites that they explored, expanding the altitudinal range and the type of environments where this species inhabits (although they did not specify the time spent in the searches). Further north, at Potrero Grande (Metropolitan Region), Correa (2017) reported three localities of A. montanus and A. tumultuosus discovered in a three-day field campaign, two of them first order streams. Moreover, the relative ease of finding individuals in various stages of development in most localities reported here (for example, Paso Pehuenche, Lo Aguirre Grande and Lo Aguirre Chico; see Table 1) suggests high local abundance. It is important to note that these observations were made at night, when the activity of these species is greater. In many cases the tadpoles were observed immediately at the beginning of the surveys (at sunset) and the first adults or juveniles in less than half an hour. The presence of numerous tadpoles, and juveniles in some cases, indicates that first order streams are reproduction and recruitment sites that could harbor abundant populations. Although we did not quantify exactly the number of larvae, our estimates are compatible with the number of specimens and densities reported for the type locality of A. pehuenche (Corbalán et al. 2008, 2014). This last study showed that A. pehuenche has a prolonged larval development that lasts at least four years, a strategy that could also occur in other Andean species of the genus (see references in Corbalán et al. 2014). This development strategy would require hydric systems with permanent flow, so our observations of tadpoles in two intermittent watercourses suggest that adverse environmental conditions, such as a drought (Corbalán et al. 2014), could affect the recruitment.

The distribution range extensions of A. pehuenche and A. hugoi have the potential to change their conservation categories, as happened with A. montanus and A. tumultuosus, which recently were down-listed from Critically Endangered to Vulnerable due to significant range extensions (IUCN 2017). However, the known range of A. pehuenche remains extremely restricted and is not represented in any protected area. Moreover, it is unclear whether more than one subpopulation (as defined by the IUCN) can be considered to exist. This species, due to its aquatic habits, only occupies streams and contiguous swamps (Corbalán et al. 2008, 2010; C. Correa, D. Vásquez, personal observations), so its area of occupancy (AOO) is probably much less than its extent of occurrence (EOO). The estimates of the IUCN (2017) reflect these observations, considering an EOO of 9 km² and an AOO of about 5 km² (although this last figure would be an overestimation as it is based on lower resolution satellite images). In addition, it remains to be assessed whether the localities of the western Andean slopes (Chile) are genetically and/or demographically separated from those located on the eastern slopes (Argentina), since they belong to different drainage basins. However, the limit of the drainage basins, corresponding to the border crossing between both countries, does currently not seem to be an important geographical barrier since the nearest streams of both basins are less than 50 m apart. Preliminary mitochondrial data are consistent with the absence of a barrier, suggesting that the localities from both sides of the border can be considered as inhabited by one population (Correa et al. 2013; unpublished data). Some of the threats identified by Corbalán et al. (2010) (habitat alteration, contamination, aquatic invasive species, livestock, natural floods, climate change) and Ghirardi et al. (2014) (chytrid fungus) also might affect the new described localities. Until now we have only been able to verify the effects of livestock and the presence of garbage, which are particularly intense at sites close to the border crossing (here called Paso Pehuenche), but all sites may be potentially threatened by environmental factors such as natural floods and climate change. In the case of A. hugoi, its distribution is currently composed of three highly disconnected areas separated by 60 and 40 km approximately, so they surely represent at least three populations, only one of which is present in a protected area (the type locality). This species is threatened by forest fires and tourism in its type locality (IUCN 2017), but we detected the effects of transhumant livestock activities on all discovered sites.

The observations made in this study add to a series of antecedents that show that high Andean environments where Alsodes species inhabit are threatened by natural and anthropogenic causes (Araya and Riveros 2008, Corbalán et al. 2010, Charrier et al. 2017). In this context, the geographic data obtained in this study acquire a special relevance since they suggest that the genus is widely distributed in the Andes, so any type of anthropic intervention (for example, mining projects, hydroelectric plants, mountain recreational centers, intensive livestock) has the potential to affect some population. Therefore, although the knowledge of the biogeography of Andean Alsodes species has improved, several aspects must be considered to reassess their conservation status. First, the few known populations seem to be threatened by several natural and anthropogenic factors. Second, their known distribution ranges are highly fragmented so their estimated AOOs would be extremely small. And third, there is no information on population connectivity, abundance or demographic trends, so we recommend maintaining their categories until more population and distribution data are obtained.

We are very grateful to José Miguel Quintana and Bernardino Camousseigt both from Enel Generación Chile, the company that funded this study. We also acknowledge the invaluable help of Kora Menegoz (guide) and Miguel Almuna (muleteer) from Andeaë (Andean Expeditions), and Juan Pablo Donoso and David Espinoza (field assistants) with whom we made the two horseback expeditions to the mountain range. We also thank Víctor Ardiles, José Luis Fajardo and Marjorie Correa for their field assistance in Laguna del Maule, and to René Loyola and Carmen Andrade from El Médano (San Clemente). We greatly appreciate the comments of Carmen Úbeda and an anonymous reviewer who contributed to improve this work. Claudio Correa thanks the financing of the project Fondecyt PAI 79130032.