Diversifying into the branches: Species boundaries in African green and bush snakes, Philothamnus (Serpentes: Colubridae)

Highlights

• The colubrid Green and Bush Snakes of the genus Philothamnus are monophyletic.

• Improved genetic and geographical sampling (14 of the 21 nominal taxa).

• Up to seven candidate species were identified.

• Philothamnus natalensis occidentalis is raised to a full species.

Abstract

The African green and bush snakes of the genus Philothamnus currently comprises 21 species and three subspecies and occurs throughout sub-Saharan Africa. The genus has been the subject of previous taxonomic revisions based on traditional morphological characters and limited genetic assessment, and may not reflect their evolutionary history. Indeed, previous findings based on phylogenetics show discordant results of interspecific relationships and question the monophyly of the genus, although taxon sampling has been limited to date. We investigated phylogenetic affinities within Philothamnus with more inclusive genetic and geographical sampling, with the aim of better understanding their evolutionary history, so that future taxonomic revision of Philothamnus can be better informed. Species relationships were examined within a phylogenetic context and sampling included 133 ingroup samples from 16 taxa. Phylogenies were constructed in Bayesian and likelihood frameworks using three mitochondrial (16S, cyt b and ND4) and two nuclear (c-mos and RAG1) markers. Competing hypotheses relating to the monophyly of the genus were tested with a Shimodaira-Hasegawa test. To examine species boundaries, Bayesian General Mixed Yule-Coalescent Model and multi-rate Poisson Tree Processes analyses were conducted. In addition, a barcoding approach was used to further clarify species-level relationships by comparing frequency distributions between intra- and interspecific sequence divergence. The genus was recovered as monophyletic; however, species-delimitation results suggest that the current taxonomy does not reflect the evolutionary history of this group. For example, Philothamnus s. semivariegatus is paraphyletic, with at least four distinct clades. Philothamnus carinatus consists of two cryptic (sister) lineages from Central and West Africa that are deeply divergent, suggesting a long history of isolation between those regions. Furthermore, the subspecies P. n. natalensis and P. n. occidentalis show strong support for species-level divergence, which reflects their morphological and ecological differences. Accordingly, we elevate P. occidentalis nov. comb. to a full species. A fully informed taxonomic revision of these taxa will require additional morphological and ecological data for corroboration, but it seems that the morphological characters (e.g. scalation, dentition) used to describe these species to date are labile within and between species. This most likely has clouded our understanding of the species boundaries within the genus. Our phylogeny and species-delimitation analyses should provide a sounder framework for taxonomy, but may also prove useful toward understanding the morphological adaptations of these species to their respective habitats.

Introduction

Colubridae, with 1922 species (Uetz et al., 2018), is the most speciose snake family globally and is found on every continent, except Antarctica (Cogger et al., 1998). The monophyly of the family and respective subfamilies, as well as intergeneric relationships within subfamilies however, has been called into question (e.g. Chen et al., 2013, Lawson et al., 2005, Pyron et al., 2011, Pyron et al., 2013, Pyron et al., 2014, Vidal et al., 2007, Zheng and Wiens, 2016). Characteristically, traditional taxonomy is dominated by the Morphological Species Concept, even though numerous other species concepts exist. Over the last decades, molecular phylogenetic methods have proven to be practical for identifying independent evolutionary lineages. This is particularly relevant under the General Lineage Concept (GLC), where species are defined as independently evolving metapopulation lineages (de Queiroz, 1998, de Queiroz, 2007). Multiple criteria are favorable when defining species under the GLC (e.g. pre- and post-zygotic isolation, monophyly, ecology and morphology), but differences across all these features are not prerequisite. Indeed, metapopulation segments can differ for some of these traits. One approach to using the GLC is to apply the Evolutionary Species Concept (ESC) (Simpson, 1951; Wiley, 1980), which often relies on molecular phylogenetics to identify species-level divergences and linking these to other criteria such as ecological and morphological differences. Several studies provide evidence of hidden genetic diversity that are linked to taxonomic discrepancies for snake taxa, and highlight the necessity of comprehensive molecular phylogenetic studies to validate species and further uncover cryptic species, thus targeting more holistic approaches for taxonomic revisions (e.g. Bryson et al., 2007, Glaw et al., 2013, Mezzasalma et al., 2015).

The African green and bush snakes of the genus Philothamnus are widely distributed across the entire sub-Saharan region. This genus is comprised of 21 species, of which three contain at least one subspecies (Chippaux, 2006, Hughes, 1985, Trape and Baldé, 2014, Trape and Roux-Esteve, 1990, Uetz et al., 2018, Wallach et al., 2014). Members of this genus are primarily arboreal, slender, with elongated and distinct heads, proportionally large eyes, and they feed mainly on frogs and lizards. One-third of the species are considered habitat specialists, occurring predominantly in either forest or savanna habitats, and most have relatively small distributions (e.g. P. carinatus, P. girardi, P. natalenis natalensis, P. nitidus nitidus, P. nitidus loveridgei, P. ornatus, P. ruandae, and P. thomensis). The remainder of species are generalists occurring in both forest and savanna biomes and have relatively wide distributions (e.g. P. angolensis, P. dorsalis, P. hoplogaster, P. natalensis occidentalis, P. semivariegatus semivariegatus; Wallach et al., 2014).

Species boundaries within green and bush snakes have historically been based on a number of key morphological features, including lepidosis, dentition, ventral keeling, and pigmentation. However, these traditional characters often display considerable overlap between species, leading to taxonomic confusion. In addition, ecologically adaptive morphological features, particularly ventral and subcaudal keeling, may provide little insight on their evolutionary history. For example, Loveridge (1958) recognised Chlorophis Hallowell 1857 (C. angolensis - as P. irregularis, C. macrops, C. heterodermus heterodermus, C. h. carinatus, C. h. ruandae, C. hoplogaster, C. irregularis irregularis, C. i. battersbyi, C. i. ornatus, C. natalensis, C. heterolepidotus) as a subgenus based on minor features of subcaudal keeling and tail length, with little discussion of how habitat selection may influence the expression of these characters. In fact, ventral keeling is thought to indicate more arboreal habits within this genus of snakes (Hughes, 1985, Loveridge, 1958). Moreover, a few species show intraspecific variation for the degree of ventral keeling, for example, P. angolensis, P. battersbyi, P. hoplogaster, and P. ornatus (Hughes, 1985). Interestingly, the intraspecific variation in ventral keeling for at least one of the species (P. hoplogaster) seems to be regional and could indicate climbing adaptations in different habitat types. Similarly, the regional variation in the number of tooth sockets (e.g. P. carinatus and P. semivariegatus; Hughes, 1985) could indicate adaptation to certain prey types in different habitats. Thus, other morphological characters used for taxonomic diagnoses of Philothamnus species may be labile and present ecomorphological adaptations irrespective of evolutionary relationships.

Few Philothamnus species have been included in higher-level phylogenies, and most of the species have been from islands rather than mainland Africa. There has been only a single study focussed on Philothamnus, but this included species found only near the Gulf of Guinea (Jesus et al., 2009). These studies showed discordant results for interspecific relationships and the monophyly of Philothamnus (Jesus et al., 2009, Pyron et al., 2013; Tonini et al., 2015; Zheng and Wiens, 2016). The phylogenetic placement of Hapsidophrys as the sister taxon of Philothamnus, as opposed to Hapsidophrys rendering Philothamnus paraphyletic, has been contested (Pyron et al., 2013, Zheng and Wiens, 2016) and resolution requires more inclusive taxon and intraspecific sampling for Philothamnus. Herein, we investigated phylogenetic relationships within Philothamnus with improved taxon sampling at the species level, as well as multiple samples within species with widespread distributions. The aim was to investigate if the present morphologically defined species are supported in a phylogenetic context, and whether cryptic species are present in the genus. We hypothesise that Philothamnus is monophyletic with respect to Hapsidophrys. Moreover, we hypothesise that currently recognised species of Philothamnus are valid under coalescent-based species delimitation. However, we expect that the most widespread generalist taxa, P. angolensis and P. s. semivariegatus, comprise multiple cryptic evolutionary lineages, because similar morphological features such as lepidosis, dentition, ventral keeling, and pigmentation may not correlate with the evolutionary history of species within this genus, but instead represent convergent ecomorphological adaptations across similar habitat types.