In the past decade, a steady stream of taxonomic discoveries have come out of the Chortís Highlands of Mesoamerica, a biogeographic region found to the south and east of the tectonic boundary between the Chortís and Mayan Blocks and north of the Nicaraguan Depression (Townsend 2011, Townsend et al. 2011). Fifty new species of amphibians and reptiles have been described from the region’s montane forests since 2000 (Cadle 2012, Cadle and Savage 2012, McCranie and Hedges 2012, Rovito et al. 2012), with literally dozens more awaiting description (Townsend 2011).

Our knowledge of the taxonomic diversity of Mesoamerican pitvipers has also greatly increased since the turn of the century (e.g. Campbell and Flores-Villela 2008, Jadin et al. 2011). Three species of endemic pitvipers have been described from the Chortís Highlands since 2000: Atropoides indomitus Smith & Ferrari-Castro 2008, Bothriechis thalassinus Campbell & Smith 2000, and Cerrophidion wilsoni Jadin, Townsend, Castoe & Campbell 2012. Two of these three taxa, Bothriechis thalassinus and Cerrophidion wilsoni, had previously been concealed within more widespread taxa only to be revealed by more focused sampling and phylogenetic analyses.

The green palm-pitvipers (genus Bothriechis) of Mesoamerica have long been a source of taxonomic uncertainty and confusion (Campbell and Lamar 2004). Ambiguities among type specimens and localities, the imprecise provenance of many available specimens, disjunct distributions limited to fragmented highland forests, and misleading external morphology have all contributed to a lack of taxonomic resolution among populations currently assigned to two species from the Chortís Highlands; Bothriechis marchi (Barbour and Loveridge 1929) and Bothriechis thalassinus. As currently understood, these two taxa inhabit a number of disjunct localities in the Chortís Highlands (Campbell and Lamar 2004, McCranie 2011a). Available molecular data for these two taxa are limited to a single sample assigned to each nominal form, which indicate the two are sister species nested within a Nuclear Central American clade (also including Bothriechis aurifer, Bothriechis bicolor, and Bothriechis rowleyi) that is, in turn, sister to the highland Bothriechis (i.e. Bothriechis lateralis and Bothriechis nigroviridis) found in lower Central America (Taggart et al. 2001, Castoe and Parkinson 2006, Castoe et al. 2009).

Bothriechis marchi sensu lato is known from localities in the Cordillera Nombre de Dios, Cordillera de Merendón, and Sierra de Sulaco of Honduras and adjacent areas of Guatemala, with Bothriechis thalassinus being found in the Cordillera de Merendón of Guatemala and Honduras, Cerro Santa Bárbara and nearby highland forests in Honduras, Cerro del Mono in eastern Guatemala, and the highlands around the El Salvador-Guatemala-Honduras border area (Campbell and Lamar 2004; McCranie 2011a). Both of these taxa include a number of allopatric highland populations that have not been assessed using phylogenetic methods. Of interest here are populations from the Cordillera Nombre de Dios, found within and around Refugio de Vida Silvestre Texíguat and Parque Nacional Pico Bonito. These two cloud forest reserves are each recognized for their diverse endemic herpetofauna (McCranie and Castañeda 2005; Townsend et al. 2012) and are taxonomically and biogeographically distinctive from that of the northern Cordillera de Merendón, which includes the vicinity of the type locality of Bothriechis marchi in the Sierra de Omoa (Townsend and Wilson 2008).

Two expeditions in 2010 provided the first herpetofaunal inventory of the extensively forested windward portions of Refugio de Vida Silvestre Texíguat, one of the most endemism-rich highland forests in Mesoamerica (Townsend et al. 2012). During two visits to the windward side of Refugio de Vida Silvestre Texíguat in June and July 2010, we collected a series of arboreal pitvipers representative of those assigned to Bothriechis marchi. Phylogenetic analyses revealed that the population from Refugio de Vida Silvestre Texíguat is not conspecific with the nominal taxon Bothriechis marchi, nor are they part of the Nuclear Central American clade containing Bothriechis marchi and Bothriechis thalassinus. Remarkably, this population is shown to represent a relict northern lineage that is most closely related to Bothriechis lateralis from Costa Rica and western Panama. We herein describe the Refugio de Vida Silvestre Texíguat population of Bothriechis as a new taxon, and discuss the implications for systematics, biogeography, and conservation.

Bayesian and Maximum Likelihood phylogenetic analyses produced congruent topological results. Our phylogeny (Fig. 1) is generally congruent with those of Castoe et al. (2009) and Daza et al. (2010), recovering two clades of nominal Bothriechis schlegelii (one Mesoamerican, one from Ecuador) rendered paraphyletic with respect to Bothriechis supraciliaris, and showing strong support for a Bothriechis marchi–Bothriechis thalassinus clade and a Bothriechis aurifer–Bothriechis rowleyi clade, together forming a Bothriechis aurifer–Bothriechis bicolor–Bothriechis marchi–Bothriechis rowleyi–Bothriechis thalassinus clade that geographically corresponds to Nuclear Central America (Fig. 1). Both of our analyses recovered a weakly supported clade that includes the Costa Rica/Panama taxa Bothriechis lateralis and Bothriechis nigroviridis, along with the Bothriechis population from Refugio de Vida Silvestre Texíguat, Honduras. Within this primarily southern clade, nine samples from Refugio de Vida Silvestre Texíguat show virtually no genetic divergence from one another, and form a monophyletic group with a well-supported (PP = 1.0; bs = 100) sister clade to Bothriechis lateralis from Costa Rica and western Panama (Fig. 1).

Based on the phylogenetic results, we examined morphological variation among populations of Bothriechis marchi sensu lato, which confirm the evolutionary and taxonomic distinctiveness of the Texíguat population as well as its apparent morphological affinity with a specimen from Parque Nacional Pico Bonito, approximately 75 km to the east of Refugio de Vida Silvestre Texíguat. We present the following description of this relict northern lineage as a new species.

Conservation status of Bothriechis guifarroi. With the description of Bothriechis guifarroi, there are now at least three species of palm pitvipers endemic to the Chortís Highlands, including Bothriechis marchi and Bothriechis thalassinus, with the potential for additional undescribed taxonomic diversity pending phylogenetic evaluation of allopatric populations in central Honduras. Based on the IUCN Red List criteria (2012), Bothriechis guifarroi should be classified as Critically Endangered (B1ab[iii]+2ab[iii]) due to its limited known area of occurrence and the potential for anthropogenic damage to its habitat. According to the algorithm developed by Wilson and McCranie (2004a), we calculated the Environmental Vulnerability Score (EVS) for Bothriechis guifarroi as 5+8+5=18, allocating it to the category of a high vulnerability species. Given this conservation status, Bothriechis guifarroi becomes the 48^th member of the critically endangered endemic component of the Honduran herpetofauna (Wilson et al. 2012), and the tenth snake species and the first viperid species so designated. This species also warrants immediate consideration for protection under CITES, given its potential for exploitation in the pet trade.

The vicinity of the type locality of Bothriechis guifarroi is part of the relatively large and intact premontane rainforests and cloud forests of Refugio de Vida Silvestre Texíguat, one of the most important areas of herpetofaunal endemism in Mesoamerica (Townsend et al. 2012). While deforestation in the leeward portion of Refugio de Vida Silvestre Texíguat has been documented since at least the early 1990’s (see summary in Townsend et al. 2010), Townsend et al. (2012) reported that the windward portion of the reserve contained a large intact expanse of virtually undisturbed forest. In late 2012, a plot was cleared in the upper Río Jilamito watershed to the south of La Liberación, marking the first time farmers from adjacent Yoro had crossed into the Río Jilamito watershed and illegally cleared land (L. Herrera-B., pers. comm.). This is an ominous development, particularly in light of the recent drastic reduction in financial support for conservation efforts in Refugio de Vida Silvestre Texíguat, which had funded the training and employment of a team of local park guards during 2010–2012.

Herpetofaunal Endemism in the Chortís Highlands. Whereas Nuclear Central America has long been accepted as a region of high biodiversity and endemicity, some observers have further recognized the western and eastern portions of this highland block as distinct biogeographic entities (Johnson 1989; Campbell 1999; Townsend 2006, 2009). Eastern Nuclear Central America has been shown to have a distinctive component of endemic biodiversity, particularly in amphibians and reptiles (Wilson and Johnson 2010); however, molecular characterization of evolutionary diversification patterns in this region has been limited to a few studies of a restricted taxonomic breadth and broader geographic focus (e.g. Castoe et al. 2009). This region is geographically analogous to the Chortís Block, an allochthonous geological formation that currently forms the only modern continental portion of the Caribbean Tectonic Plate and the largest terrestrial segment of the contemporary Central American land bridge (Rogers 2003; Marshall 2007). The Chortís Block has a challengingly complex history, and recently has been the subject of increased focus, and sometimes contentious debate, within the geological research community (James 2007; Mann et al. 2007; Ortega-Gutiérrez et al. 2007; Silva-Romo 2008; Morán-Zenteno et al. 2009).

The majority of the geographical extent of the Chortís Highlands is found within the political boundaries of Honduras, the country with the highest degree of herpetofaunal endemism of any Central American nation (Wilson and Johnson 2010). Townsend and Wilson (2010) reported 91 endemic species (47 amphibians and 44 reptiles) from Honduras. Since that work went to press, an additional ten endemic species have been described from Honduras, including three new plethodontid salamanders (Nototriton picucha, Townsend et al. 2011; Nototriton tomamorum, Townsend et al. 2010; Oedipina petiola, McCranie and Townsend 2011), a new black iguana (Ctenosaura praeocularis, Hasbún and Köhler 2009), a new skink (Marisora roatanae; Hedges and Conn 2012), two new dwarf geckos (Sphaerodactylus guanajae and Sphaerodactylus leonardovaldesi; McCranie and Hedges 2012), and three new colubrid snakes (Omoadiphas cannula, McCranie and Cruz-Díaz 2011; Tantilla psittaca, McCranie 2011b; Tantilla olympia, Townsend et al. 2013). With these species included, the total stands at 101 species, making Bothriechis guifarroi the 102nd described herpetofaunal species endemic to Honduras.

Bothriechismarchi and the status of populations from Yoro. We recognize Bothriechis marchi sensu stricto as occurring in localities in the Cordillera de Merendón in the Honduran departments of Cortés and Santa Bárbara along the border with Guatemala, as well as for at least one isolated locality in eastern Guatemala (Fig. 7). This Guatemalan locality, Cerro del Mono in Departamento de Zacapa, previously was the source of the only sequenced sample assigned to Bothriechis marchi (UTA R-52959; Castoe and Parkinson 2006). Although the population of Bothriechis from Cerro del Mono does not agree morphologically with the typical form of Bothriechis marchi (E.N. Smith, pers. comm.; see Plates 422–424 in Campbell and Lamar 2004, as Bothriechis thalassinus), the sequence data attributed to UTA R-52959 are not notably divergent from a sample of typical Bothriechis marchi (MVZ 263604) collected from the Sierra de Omoa in northern Honduras (Fig. 1).

The type localities of Bothriechis marchi and Bothriechis thalassinus are both in the Sierra de Caral, within approximately 20 km of one another on opposite sides of the Guatemala/Honduras border (Campbell and Lamar, 2004) suggesting that the two species occur either in parapatry or sympatry in the limited forest remaining in that mountain range (the type locality of Bothriechis marchi is not precise; Wilson and McCranie [1992] restricted it to the forested hills above El Oro, Departamento de Santa Bárbara). Individuals of both taxa in the Sierra de Caral exhibit a primarily green dorsal coloration, with some scattered bluish middorsal blotches (Campbell and Smith, 2000; Campbell and Lamar, 2004: plate 425). While the presence of Bothriechis thalassinus has not been confirmed by vouchered specimens from the Sierra de Omoa, the proximity of the Sierra de Caral and the Sierra de Omoa in northwestern Honduras and the similarity in coloration exhibited between these nominal taxa in that vicinity suggest the possibility that Bothriechis thalassinus may have gone unnoticed in the Sierra de Omoa. We have evaluated photographs of over a dozen individuals of green Bothriechis from the Sierra de Omoa encountered as part of an expedition-tourism operation in that area over the past five years, and have noted considerable variation in head scalation in the photographs. Unfortunately, none of the photographed individuals were collected nor were genetic samples taken to allow for more detailed evaluation of the Bothriechis of the Sierra de Omoa. In addition to the nominal form of Bothriechis marchi, it is likely that more than one species of Bothriechis occurs in sympatry or parapatry in the Sierra de Omoa, possibly including Bothriechis thalassinus and/or an unidentified sister taxon of Bothriechis guifarroi. Focused sampling and phylogenetic analysis of Bothriechis from the Sierra de Omoa is needed to better characterize the taxonomic diversity present in that mountain range.

Paraphyly in Bothriechis marchi sensu lato in terms of populations from the Cordillera de Merendón and the Cordillera Nombre de Dios, the latter now known to represent Bothriechis guifarroi, calls into question the taxonomic status of populations from isolated localities in the Sierra de Sulaco in Departmento de Yoro (Fig. 7). These populations are represented in collections by one specimen from Cerro de Pajarillos (USNM 561085), three specimens from the Montaña de Mataderos (FMNH 21777, MCZ R-38785–86), 14 specimens from “Portillo Grande” near Montaña Macuzal (FMNH 34732–35, 35895, 35999–601, 37217, 38542, 41621; MCZ R-38789–91), and two specimens from “Subirana Valley” (FMNH 21892, MCZ R-38788). Four of these specimens examined by us for this paper (MCZ R-38785–86, 38790–91) differ from Bothriechis guifarroi in having 21–19–15 dorsal scale rows (versus 19–19–15) and having varying conditions of fusion of the prelacunal and second supralabial (fused on both sides in MCZ R-38790, separate on both sides in MCZ R-38786, and fused on one side and separate on the other in MCZ R-38785 an R-38791). Given the considerable phylogenetic diversification presented by analysis of Bothriechis guifarroi and Bothriechis marchi s. s., we cannot justify assignment of the Yoro populations to either taxon in the absence of molecular characterization of those populations. Therefore, we tentatively refer to the Yoro populations as Bothriechis species inquirenda pending collection of fresh material and phylogenetic characterization. Efforts to secure this material are currently underway.

Evolutionary and biogeographic implications. The phylogenetic position of Bothriechis guifarroi has significant implications for our understanding of the biogeography and evolution of palm-pitvipers. Crother et al. (1992) first presented a phylogenetic hypothesis for the genus Bothriechis supporting an eyelash pitviper clade (Bothriechis schlegelii and Bothriechis supraciliaris) that is sister to a clade containing the remaining species of Bothriechis, all of which have highland-associated distributions (typically <1, 000 m elevation; Fig. 7). Castoe et al. (2009) and Daza et al. (2010) provided the most recent phylogenetic estimates for nine species of Bothriechis, as part of a comparative phylogeographic study of three co-distributed, predominantly Mesoamerican genera of pitvipers. Their phylogenetic hypotheses supported the contention of Castoe and Parkinson (2006) that Bothriechis lateralis is the sister group to the Nuclear Central American highland species of the genus (Bothriechis aurifer, Bothriechis bicolor, Bothriechis marchi, Bothriechis rowleyi, and Bothriechis thalassinus) and rejected the Crother et al. (1992) and Taggart et al. (2001) estimates that Bothriechis lateralis is the sister clade to Bothriechis bicolor. Our phylogenetic hypothesis also supports a Nuclear Central American clade, with Bothriechis marchi and Bothriechis thalassinus in the Chortís Highlands comprising a group that is sister to Bothriechis aurifer, Bothriechis bicolor, and Bothriechis rowleyi from the Nuclear Central American highlands to the west.

The Cordillera Nombre de Dios, which includesRefugio de Vida Silvestre Texíguat and Parque Nacional Pico Bonito, is home to a distinct endemic biota that, in many cases, bears little resemblance to the endemic communities found in other nearby cloud forests (Wilson and McCranie 2004b). While published phylogenetic datasets of taxa endemic to the Cordillera Nombre de Dios are limited, evidence is beginning to accumulate that this region, and Refugio de Vida Silvestre Texíguat in particular, represents a paleo-refugium where relict lineages have accumulated and persisted while disappearing elsewhere in northern Central America. Bothriechis guifarroi provides the best evidence to date in support of the paleo-refugium hypothesis, given its phylogenetic relationship with Bothriechis lateralis and Bothriechis nigroviridis from southern Central America and lack of close evolutionary relationships with other taxa from Nuclear Central America. Another recently described species from Refugio de Vida Silvestre Texíguat, the moss salamander Nototriton tomamorum, is morphologically and phylogenetically more closely associated with the Costa Rican taxon Nototriton richardi than it is with members of the northern Nototriton barbouri group, which includes an undescribed taxon that occurs in sympatry with Nototriton tomamorum (Townsend et al. 2010). Isthmohyla insolita, another Texíguat endemic, is one of only two Isthmohyla in northern Central America (the other being Isthmohyla melacaena from the Sierra de Omoa; neither species has been subjected to phylogenetic analysis), with the remaining 13 species restricted to highlands of Costa Rica and Panama (Köhler 2011).

Castoe et al. (2009) estimated that Lower Central American and Nuclear Central American Bothriechis diverged at the Nicaraguan Depression between 5.73–9.87 mybp, based on the hypothesis that Bothriechis lateralis represented the sister lineage to the Nuclear Central American clade. Given the discovery of Bothriechis guifarroi, the timing of divergence across the Nicaraguan Depression could be much more recent, similar to that seen between Cerrophidion sasai and Cerrophidion wilsoni (3.06–6.03 mybp; Castoe et al. 2009). Alternatively, it is possible that a Lower Central American Bothriechis (i.e., Bothriechis guifarroi, Bothriechis lateralis, Bothriechis nigroviridis) may have split from Nuclear Central American taxa via the Nicaraguan Depression earlier than suggested by Castoe et al. (2009) and Daza et al. (2010). Following additional sampling and the evaluation of populations presently referred to as Bothriechis sp. inquirenda, a reevaluation of Mesoamerican pitviper biogeography is warranted, and may shed further light on the complicated biogeographic relationship between northern and southern Central America.