African snake-eaters of the genus Polemon are cryptic, fossorial snakes that mainly inhabit the forests of central, eastern, and western Africa. Molecular results from a previous study demonstrated that Polemon christyi is not monophyletic—two distinct lineages were recovered from Uganda (the type locality) and southeastern Democratic Republic of the Congo (DRC). Genetic data indicated differences in sequence divergence and encoded amino acids between these lineages. Based on these molecular differences and diagnostic differences in morphology, we describe the lineage from southeastern DRC as a new species. Literature records indicate that it likely occurs in adjacent Tanzania and Zambia. It is the first species of Polemon to be described in over 70 years.
The 13 species of currently recognized snakes of the genus Polemon are cryptic, fossorial inhabitants of forests and woodland/savanna mosaic habitats throughout central, eastern, and western Africa (Chippaux, 2006; Uetz et al., 2018). Snakes, particularly typhlopids, form their main diet, hence the common name ‘Snake Eaters.’ Most species of the genus rarely exceed 80 cm total length, but they can be voracious predators that consume snakes of equal size (Pitman, 1974; Kusamba et al., 2013; Spawls et al., 2018). Like many other lamprophiids, these fossorial and secretive snakes are poorly known, both in terms of natural history and taxonomy (i.e., low numbers of specimens available in museums), but most species have grooved rear fangs and are ophiophagous, nocturnal, and oviparous (Spawls et al., 2018). Many species also have prominent yellow or orange neck bands (Underwood and Kochva, 1993; Spawls and Branch, 1995; Spawls et al., 2018).
Based on hemipenial, dentition, and osteological characters, Bogert (1940) gave the first definitive arrangement of aparallactines, which included the genera Amblyodipsas, Aparallactus, Brachyophis, Chilorhinophis, Elapotinus, Hypoptophis, Macrelaps, Micrelaps, Poecilopholis, Polemon, and Xenocalamus. De Witte and Laurent (1943, 1947) revised aparallactines into three groups, with Elapocalamus (Boulenger, 1911), Chilorhinophis (Werner, 1907), Polemon (Jan, 1858), Miodon (Duméril, 1859), Cynodontophis (Werner, 1902), and Melanocalamus (de Witte, 1941) comprising their “Deuxième Groupe,” which was characterized by the presence of a maxillary-ectopterygoid foramen. Currently, only two genera from the “Deuxième Group” are recognized, with all genera except Chilorhinophis placed in the synonymy of Polemon (Laurent, 1956a; Hughes and Barry, 1969; Resetar and Marx, 1981). In a recent, major phylogenetic analysis of aparallactines, these two fossorial genera were recovered as sister taxa with strong support (Portillo et al., 2018: fig. 2).
Relationships within the genus Polemon have historically been poorly understood and reliant solely on morphological data. Only recently have relationships within the genus been studied with molecular data (Fig. 1; e.g., Pyron et al., 2013; Figueroa et al., 2016; Portillo et al., 2018). Because many of the species share morphological characters that often overlap, taxonomic classification of species of Polemon has been challenging. Many of the 13 currently recognized species were historically considered to be geographic variants or synonyms of fewer species (Loveridge, 1942, 1944; de Witte and Laurent, 1947; Pitman, 1974; Spawls et al., 2018). For example, the poorly known species P. christyi was considered to be a race of either P. collaris or P. gabonensis at different times (Loveridge, 1942, 1944, 1957; Pitman, 1974). This is unsurprising, because the three taxa overlap considerably in morphological characters, including ventral and subcaudal scale counts (de Witte and Laurent, 1947; Pitman, 1974). However, de Witte and Laurent (1947) and Laurent (1956a) considered P. christyi to be a distinct species based on dorsal coloration, because it has a grayish black dorsum and lacks a neck band, whereas P. collaris and P. gabonensis both have a yellowish band on the neck. Polemon christyi is also partially sympatric with another superficially similar member of the genus, P. graueri, which can be grayish black dorsally, but with a more slender build and a larger number of ventral scales (de Witte and Laurent, 1947; Pitman, 1974). Currently, P. christyi is considered to occur throughout much of the Albertine Rift region (Spawls et al., 2018), including eastern Democratic Republic of Congo (DRC; absent from the western lowlands, sensu Broadley, 1998), South Sudan (Wallach et al., 2014), and western Kenya (Lötters et al., 2007), to northern Zambia (Broadley et al., 2003) and Tanzania (Loveridge, 1944; Caro et al., 2011; Spawls et al., 2018), with a single record from northern Malawi (Mercurio, 2007). Recent molecular results demonstrated that P. christyi is not monophyletic—topotypic P. christyi is found in eastern Africa and is sister to central African P. robustus, whereas a second lineage occurs in southeastern DRC and is sister to the widespread species P. collaris (Figs. 1, 2; Portillo et al., 2018). Herein, we examine the taxonomic status of these lineages of P. christyi in more detail with morphological data and additional molecular analyses.
Fig. 1
Phylogenetic tree depicting relationships of African Polemon, based on Portillo et al. (2018). Branch support values are Bayesian inference posterior probabilities/maximum likelihood bootstrap support values. See Data Accessibility for tree file.
MATERIALS AND METHODS
Molecular analyses.—To understand molecular variation within the genus Polemon, we used data from three mitochondrial (16S, cyt b, and ND4) and two nuclear genes (c-mos and RAG1) that were sequenced and analyzed in a phylogenetic context in the study of Portillo et al. (2018). Levels of sequence divergence between haplotypes were inferred using uncorrected p-distances calculated from MEGA version 7.0 (Kumar et al., 2016).We also analyzed differences in amino acid translation of the protein-coding genes ND4 and cyt b for samples of Polemon christyi (UTEP 21618) and P. cf. christyi (PEM R20734 and PEM R17452). No nuclear data were available for P. christyi.
Morphological analyses.—Specimens examined for this study (Material Examined) were collected from multiple localities throughout sub-Saharan Africa, and are housed in the collections of the Port Elizabeth Museum, the University of Texas at El Paso Biodiversity Collections, and the Royal Belgian Institute of Natural Sciences. Additional aparallactine specimens, including relevant type specimens, were cited or examined from a diversity of collections and institutions as listed in the Material Examined.
Specimens were examined under a Zeiss Stemi 2000-C stereo microscope, and photographs were taken with a Canon Rebel T3i DSLR camera and Canon 6D full frame DSLR camera. Characters were chosen from previous taxonomic studies of advanced snakes (LaDuc and Johnson, 2003; Devitt et al., 2008; Moyer and Jackson, 2011; Anderson and Greenbaum, 2012; Greenbaum et al., 2015). Snout–vent length (SVL) was measured with a metric ruler and rounded to the nearest 1.0 mm. All other measurements were taken from the right side of the body with digital calipers under a dissecting microscope and rounded to the nearest 0.1 mm.
Morphological data consisted of 18 mensural and 14 meristic characters. Mensural data included: snout–vent length (SVL); tail length (TL); head width (HW)—measured at widest point of head; head length (HL)—measured at angle of jaw, from posterior edge of mandible to tip of snout; naso-ocular distance (NOD; taken at the anterior border of the eye to posterior border of nare); eye to snout length (ES; taken at the anterior border of the eye); interocular distance (ID); eye diameter (EW; anterior–posterior); internasal scale width (INAS); rostral scale height (RH); rostral scale width (RW); frontal scale length (FL); frontal scale width (FW); chin shield length (CSL); chin shield width (CSW); prefrontal scale length (PFL); prefrontal scale width (PFW); and mouth gape (MG). Meristic data included: number of ventral scales (VENT)—following the standard and Dowling methods (Dowling, 1951); subcaudals (SCDL); dorsal scale rows at the neck (DSRN; one head length posterior to jaw rictus); dorsal scale rows at midbody (DSRM); dorsal scale rows one head length anterior to the cloaca (DSRC); prefrontals (PFRO; size and number); internasals (INAS; size and number); number of supralabials (SUPRA); number of supralabials in contact with the eye (SUIE); number of infralabials (INFRA); number of supralabials in contact with chin shields (LLC); preoculars (PRE); postoculars (PO); and temporal arrangement (T).
Mean, standard deviation, and range of mensural and meristic characters were calculated for each group indicated by the phylogeny. To eliminate the effect of size, analyses of covariance (ANCOVA) were conducted with snout–vent length as the covariate (Packard and Boardman, 1999). Principal components analyses (PCA) of mensural data were conducted in Minitab 15 (Minitab Statistical Software, State College, PA) and used to identify patterns of variation in the data. All analyses used the covariance matrix. PCA analyses were conducted using log-transformed mensural data (measurements pertaining to the head). Residual data obtained from the ANCOVA analyses were used for PCA analyses.
RESULTS
Molecular analyses.—Cytochrome b distances ranging from 2.5% to 5.3% were observed between Polemon cf. christyi (hereafter referred to as P. ater, new species) and its sister taxon P. collaris, and a genetic divergence of 14.0% was recovered between P. ater, new species, and P. christyi (Supplementary Table 1; see Data Accessibility). Genetic distances of ND4 ranged from 2.4% to 3.6% between P. ater, new species, and P. collaris, and a genetic divergence of 14.1% was recovered between P. ater, new species, and P. christyi (Supplementary Table 1; see Data Accessibility). Divergences recovered from the 16S gene and nuclear DNA data sets were noticeably lower; uncorrected p-distances calculated from 16S, c-mos, and RAG1 ranged between 0% to 0.8% (P. ater, new species, and P. collaris) and 0.4% (P. ater, new species, and P. christyi; Supplementary Table 1; see Data Accessibility). Cytochrome b and ND4 divergences between P. ater, new species, and other species of Polemon are shown in Supplementary Table 1 (see Data Accessibility).
Twenty-five major differences in amino acid translation were noted in ND4 and cyt b data from lineages of Polemon from DRC and Uganda. For ND4, amino acid codon positions 43, 47, 116, 130, 134, 164, 169, 189, 211, 213, 217, and 223 translated to threonine, alanine, isoleucine, threonine, leucine, threonine, leucine, methionine, methionine, methionine, methionine, and proline for P. ater, new species, whereas the same positions translated to alanine, serine, threonine, alanine, isoleucine, valine, phenylalanine, threonine, leucine, threonine, alanine, and serine in P. christyi. For cyt b, amino acid codon positions 57, 58, 63, 69, 76, 108, 113, 149, 151, 153, 162, 171, 175, 180, and 185 translated to methionine, threonine, cysteine, isoleucine, threonine, threonine, isoleucine, threonine, leucine, valine, asparagine, alanine, isoleucine, threonine, and leucine for P. ater, new species, whereas the same positions translated to isoleucine, methionine, tyrosine, threonine, isoleucine, valine, threonine, alanine, isoleucine, isoleucine, aspartic acid, threonine, leucine, isoleucine, and serine in P. christyi.
Morphological analyses.—Morphometric data for examined specimens of Polemon (P. christyi, P. ater, new species, and P. collaris) are presented in Tables 1 and 2. The principal components analysis (PCA) with head morphometric data regressed against SVL is shown in Figure 3 and Table 1. The first two PC axes accounted for most of the variation in the data (87.5%; Fig. 3, Table 1). The first PC axis was an indicator of general head size relative to body size; specimens towards the right of the graph were considered to have larger heads relative to body length. The second PC axis loaded negatively for head length; negative values on this axis were correlated with shorter heads relative to body length (Fig. 3). The PC analysis showed a clear separation between P. ater, new species, and samples of P. collaris, but there was some overlap between P. ater, new species, and samples of P. christyi (Fig. 3).
We found a large amount of overlap in morphological characters (size and scalation) between several taxa of Polemon (P. christyi [n ¼ 5], P. ater, new species, [n ¼ 2], P. collaris [n ¼ 8], and P. gabonensis [n ¼ 8]; Table 2; Chippaux, 2006; Chirio and LeBreton, 2007; Pauwels and Vande weghe, 2008). This is common within the genus because species are morphologically conserved. The easiest way to distinguish these lineages of Polemon is by a combination of characters including shape of the postocular, shape of the nasal scales, dorsal coloration, ventral coloration, and presence/absence of an orange or yellow neck band.
Conclusion.—The two populations, P. christyi and P. ater, new species, are distinguished by subtle differences in morphology and substantial mitochondrial molecular differences. Moreover, P. ater, new species, is known only from grassland/ miombo woodland, whereas P. christyi occurs mostly in forest, although it enters grassland and woodland in northeastern DRC (de Witte, 1966; Pitman, 1974). We consider these differences to indicate that these lineages are specifically distinct, and therefore describe the population from southeastern DRC as a new species.
Polemon ater, new species
Black Snake-eater
urn:lsid:zoobank.org:act:6256E503-4C74-4033-BABE-CEE8 C42A067E
Figures 4, 5, 6; Table 2
Miodon gabonensis christyi, Loveridge (1944; part): 170, 178– 180.
Miodon christyi, Laurent (1947; part): 10.
Miodon christyi, de Witte and Laurent (1947; part): 8, 60, 73– 75, figs. 67–69.
Miodon christyi, de Witte (1953): 264–265, fig. 91.
Miodon christyi, Laurent (1955; part): 293.
Miodon christyi, Laurent (1956b): 252.
Miodon collaris christyi, Loveridge (1957; part): 283.
Miodon christyi, Broadley and Pitman (1960): 437, 447.
Miodon christyi, Bourgeois (1968: part): 179, 284.
Polemon christyi, Broadley (1971): 26, 76.
Miodon christyi, Pitman (1974: part): 135, 165–168, 205, colour plate M, fig. 3, plate XII.
Polemon christyi, Welch (1982; part): 142.
Polemon christyi, Hughes (1983; part): 316, appendix A.
Polemon christyi, Chifundera (1990; part): table 1.
Polemon christyi, Broadley and Howell (1991; part): 29, 35, 62.
Polemon christyi, Broadley (1998; part): xxx.
Polemon christyi, Behangana and Goodman (2002; part): 64.
Polemon christyi, Spawls et al. (2002; part): 426.
Polemon christyi, Broadley et al. (2003): 95–96, fig. 17.
Polemon christyi, Broadley and Cotterill (2004; part): 47, 52.
Polemon christyi, Spawls et al. (2004; part): 426.
Polemon christyi, Chirio and Ineich (2006; part): 58.
Polemon christyi, Lötters et al. (2007; part): 98–99, plate 12.
Polemon christyi, Caro et al. (2011; part): 561.
Polemon christyi, Wallach et al. (2014; part): 561, table 1.
Polemon christyi, Tilbury and Branch (2014): 36–38, figs. 1 (two figures labeled fig. 1).
Polemon christyi, Spawls et al. (2018; part): 461, unnumbered figure.
Holotype.— PEM R20734, subadult male, Democratic Republic of the Congo, Lualaba Province, Fungurume, 10.53388S, 26.33758E, 1189 m, C. Tilbury, 12 February 2014 (Tilbury and Branch, 2014).
Paratype.— PEM R17452, adult female, Democratic Republic of the Congo, Lualaba Province, Kalakundi, 10.65508S, 25.93258E, 1472 m, W. R. Branch, 25 January 2008.
Referred material.— Given the morphological similarities (scale counts and coloration) between southern populations previously referred to P. christyi, we provisionally assign records from southeastern DRC (de Witte and Laurent, 1943, 1947; de Witte, 1953; Laurent, 1956b), Zambia (Broadley, 1971; Broadley et al., 2003), and west-central Tanzania (Loveridge, 1944; Caro et al., 2011; Spawls et al., 2018) to P. ater.
Diagnosis.— Polemon ater is a medium to large aparallactine. The dorsum and venter are uniformly grayish black or black, with ventrals and subcaudals each edged posteriorly in silver white, lacking any lighter tones or shades anywhere on the dorsum and lacking a distinct collar; the preocular scale is irregular in shape (somewhat triangular with a rounded top); dorsally the head narrows towards the snout. Cytochrome b and ND4 pairwise sequence divergence rates between P. ater and its closest relative (P. collaris) ranged between 2.5% to 5.3%.
Comparisons.— (Figs. 4, 5, 6; Table 2) Based on examined material and published details (Boulenger, 1903; de Witte, 1941, 1953, 1962, 1966; de Witte and Laurent, 1943, 1947; Laurent, 1956a, 1956b, 1960; Pitman, 1974; Broadley and Howell, 1991; Meirte, 1992; Broadley et al., 2003; Chippaux, 2006; Trape and Mané, 2006; Chirio and LeBreton, 2007; Pauwels and Vande weghe, 2008), P. ater differs from P. acanthias by dorsal coloration (grayish black or black vs. whitish or pale reddish with five black stripes in P. acanthias), having a divided cloacal plate (entire in P. acanthias), and ventral coloration (grayish black or black with silver-white edging on ventral and subcaudal scales vs. white in P. acanthias); from P. barthii by the number of postocular scales (two vs. one in P. barthii), the shape of the preocular scale (irregular vs. trapezoidal in P. barthii), dorsal coloration (grayish black or black vs. olive in P. barthii), having a divided cloacal plate (entire in P. barthii), and ventral coloration (grayish black or black with silver-white edging on ventral and subcaudal scales vs. yellowish white in P. barthii); from P. bocourti by the shape of the preocular scale (irregular vs. triangular in P. bocourti), and lacking a distinct collar (distinct creamy yellow collar in P. bocourti), having a divided cloacal plate (entire in P. bocourti), and having a narrower snout; from P. fulvicollis by the number of ventral scales (202–242 vs. 247–267 in P. fulvicollis), body shape (stout vs. slender and long in P. fulvicollis), ventral coloration (grayish black or black with silver-white edging on ventral and subcaudal scales vs. white in P. fulvicollis), and lacking a distinct collar (yellowish or orange collar present in P. fulvicollis); from P. gracilis by the number of infralabials (seven vs. six in P. gracilis), the number of ventral scales (202–242 vs. 246–284 in P. gracilis), ventral coloration (grayish black or black with silver-white edging on ventral and subcaudal scales vs. white or cream in P. gracilis), and absence of a collar (yellowish collar present in P. gracilis); from P. graueri by the number of ventral scales (202–242 vs. 222–262 in P. graueri), ventral coloration (grayish black or black with silver-white edging on ventral and subcaudal scales vs. cream or white in P. graueri), and shape of the preocular (irregular vs. triangular in P. graueri); from P. griseiceps by the number of ventral scales (202–242 vs. 177–200 in P. griseiceps) and ventral coloration (grayish black or black with silver-white edging on ventral and subcaudal scales vs. cream or white in P. griseiceps); from P. neuwiedi by dorsal coloration and pattern (grayish black or black vs. pale brown with three black stripes in P. neuwiedi) and ventral coloration (grayish black or black with silverwhite edging on ventral and subcaudal scales vs. white in P. neuwiedi); from P. notatus by dorsal coloration (grayish black or black vs. pale brown with two series of round black spots in P. notatus), number of ventral scales (202–242 vs. 181– 200 in P. notatus), ventral coloration (grayish black or black with silver-white edging on ventral and subcaudal scales vs. white in P. notatus), and number of postocular scales (two vs. one or two in P. notatus); from P. robustus by the shape of the preocular scale (irregular vs. rectangular and long vertically in P. robustus), lack of a distinct collar (yellowish orange collar present in P. robustus), shape of the snout laterally (narrow vs. wide in P. robustus), and number of ventral scales (202–242 vs. 163–189 in P. robustus); from P. christyi, to which it is morphologically most similar, by the shape of the postocular scales (upper postocular scale is noticeably larger than the lower postocular scale vs. equalsized postocular scales in P. christyi) and shape of the nasal scales (square-like vs. irregular shaped in P. christyi); from P. collaris by lacking a distinct collar (tan or yellow collar present in P. collaris), the shape of the postocular scales (top postocular scale is noticeably larger than the bottom postocular scale vs. equal-sized postocular scales in P. collaris), shape of the nasal scales (square-like vs. irregular shaped in P. collaris), ventral coloration (grayish black or black with silver-white edging on ventral and subcaudal scales vs. white or cream in P. collaris), and a narrower head; and also from P. gabonensis by the shape of its preocular scale (irregular vs. elongated and triangular in P. gabonensis), shape of the postocular scales (top postocular scale is noticeably larger than the bottom postocular scale vs. equal-sized postocular scales in P. gabonensis), shape of the nasal scales (square-like vs. irregular shaped in P. gabonensis), a less robust snout, lack of a distinct collar (yellowish light gray collar present in P. gabonensis), and ventral coloration (grayish black or black with silver-white edging on ventral and subcaudal scales vs. creamy yellow lower labials and venter in P. gabonensis).
Description of the holotype.— (Figs. 4, 5, 6; Table 2) Subadult male 254 mm SVL; interocular distance 3.1 mm, pupil round, eye diameter 0.9 mm; no loreal; body cylindrical; tail short (17.0 mm, 6.69% of SVL); body stout; head slightly distinct from neck; dorsally, head slightly wider than neck and progressively narrower towards tip of snout; laterally, head narrow, widest point at back of head and narrower at nostrils; nostrils visible from above; scales smooth and glossy. Supralabials 7 (left)/7 (right), 3rd–4th/3rd–4th contacting orbit; infralabials 7/7, 1st on each side in contact behind mental, 1st–4th/1st–4th contacting anterior chin shields; 1/1 preocular; 2/2 postoculars; temporals 1+1/1+1; two internasals; nasal divided; frontal is longer (2.5 mm) than wide (1.6 mm); dorsal scales 15 one head length posterior to jaw rictus, 15 at midbody, and 15 one head length anterior to cloaca; ventrals 211 (Dowling count: 208); cloacal plate divided; all paired subcaudals 20. Maxillary dentition—two small anterior teeth, followed by a very large, deeply grooved fang positioned anterior to eye, followed posteriorly by 12 smaller teeth on each side. These data are nearly identical to those reported by Tilbury and Branch (2014).
Coloration of the holotype in life.— (Fig. 4) Dorsum and venter uniform glossy grayish black, with ventrals and subcaudals each edged posteriorly in silver white (Tilbury and Branch, 2014). The anterior forked portion of the tongue is silver white, which transitions to grayish black posteriorly.
Coloration of the holotype in preservative.— Dorsum and venter uniform grayish black; slightly lighter in color than found in life.
Variation.— Mensural and meristic variation between the two examined specimens of Polemon ater are shown in Table 2. The paratype (PEM R17452) was a badly damaged adult female. There were no differences between the two specimens in terms of coloration in preservative, as both were uniform grayish black dorsally and ventrally. The female is larger (640 mm SVL), has fewer subcaudals (15), and has a proportionately shorter tail (4.35% of SVL). The largest known specimen (806 mm SVL) is from Solwezi, Zambia (Broadley et al., 2003). Ventrals were not counted for the paratype because it was badly damaged. Literature records of specimens from southeastern DRC and Zambia report ventral ranges of 202–242 and subcaudal ranges of 15–24 (de Witte, 1953; Laurent, 1956b; Broadley and Pitman, 1960; Broadley et al., 2003). Broadley and Pitman (1960) and Broadley et al. (2003) noted that specimens from Zambia may have one or two postoculars, and temporal formulas were either 1+1 or 0+1+1. De Witte (1953) reported that a specimen from Upemba National Park in southeastern DRC contained one postocular on the left side, and two postoculars on the right side. Specimens from southeastern DRC and Zambia are reported to be uniformly grayish black, bluish black, or black, both dorsally and ventrally (de Witte, 1953; Broadley et al., 2003), but Broadley et al. (2003: 95) noted ventral coloration “may have varying degrees of white on the neck or belly.” Laurent (1956b) noted that a young male specimen from Dilolo (Lualaba Province, DRC) still had a distinct grayish collar in preservative, suggesting that juvenile or subadult P. ater might have a distinct collar. Morphometric and meristic data for examined specimens of Polemon christyi, P. collaris, and P. ater are shown in Table 2.
Habitat.— Specimens of P. ater were collected from localities in or near Brachystegia (i.e., miombo) woodlands of Lualaba Province, DRC (Fig. 7). The paratype was found dead in a pit in Kalakundi Copper Mine, where it had been killed by mine workers. Specimens from Upemba National Park in southeastern DRC were found in grassland-miombo woodland habitat near tributaries. Specimens from Zambia were found in miombo woodland and in some cases, there was gallery forest in the vicinity, although none of the specimens were found in gallery forests (Broadley et al., 2003). Specimens for this study were found in elevations ranging from 1189–1472 m. In Upemba National Park, de Witte (1953) found a male specimen as high as 1810 m.
Natural history.— Very little is known about the ecology and natural history of this species. Upon discovery at about 20:00 hrs, the behavior of the holotype was described as “atractaspoid,” but it did not produce the neck flexure posturing that is typical for Atractaspis snakes that are in a defensive mode. However, “it did thrash and jerk, freeze with body dorsoventrally flattened, and occasionally display a small degree of neck flexion” (Tilbury and Branch, 2014: 36). The holotype was kept in captivity for some time after capture, during which it burrowed into soil of its container, but it preferred to shelter under pieces of bark at the surface of the soil. The animal refused offerings of food including earthworms, grasshoppers, newly-metamorphosed toadlets, and geckos (Hemidactylus mabouia and Lygodactylus gutturalis), but eventually it ate one L. gutturalis gecko (Tilbury and Branch, 2014).
Based on the natural history of other species of Polemon, P. ater is likely nocturnal and fossorial, although Hinkel and Fischer (1988) noted that P. christyi in Rwanda can be diurnal or nocturnal. The new species is known to consume snakes that are relatively large. The paratype (PEM R17452) was found with a very large (480 mm SVL) Afrotyphlops schmidti (PEM R17440) in its gut. The Afrotyphlops schmidti was about halfway consumed, but the thickness of this prey item (14.8 mm) exceeded the thickness of the specimen of P. ater (9.8 mm). Broadley et al. (2003) reported a 806 mm P. ater (as P. christyi) that consumed a 600 mm Crotaphopeltis hotamboeia and a 430 mmP. ater that consumed a 305 mmC. hotamboeia. Broadley et al. (2003) reported that Zambian specimens were usually seen at night after heavy rainfall. Spawls et al. (2018) noted this species may be found in leaf litter or below the surface, and it emerges from underground during the rainy season. Additionally, Spawls et al. (2018) stated that the species is known to consume Afrotyphlops, Leptotyphlops, and C. hotamboeia. Polemon ater is thought to lay eggs, but no clutch details are known (Hinkel and Fischer, 1988; Spawls et al., 2018).
Distribution.— The new species most likely occurs in southeastern DRC, Zambia, west-central Tanzania, and possibly as far north as Burundi (Broadley and Howell, 1991; Caro et al., 2011; Spawls et al., 2002, 2004, 2018; Tilbury and Branch, 2014). Specimens noted from Rwanda and Malawi (some of which were found in elevations above 1995 m; de Witte, 1941; Laurent, 1956a; Hinkel and Fischer, 1988; Mercurio, 2007) may be attributable to P. christyi, P. ater, or an unknown species. Specimens noted from northeastern DRC, Uganda, South Sudan, and western Kenya are attributable to P. christyi (Tilbury and Branch, 2014; Wallach et al., 2014; Spawls et al., 2018), but some populations (e.g., Virunga National Park, DRC) require additional study to confirm their identification.
Etymology.— Derived from the Latin atrum in reference to the grayish black or black dorsal and ventral coloration that is present in all known specimens of P. ater.
Table 1
Principal components analysis (PCA) comparing Polemon ater, new species, with P. christyi and P. collaris, with natural logtransformed morphometric data regressed against SVL. Eigenvalues, percent variance, cumulative variance, and loadings are shown for the first three principal components. See Materials and Methods for abbreviations.
Table 2
Morphometric data (in mm) and meristic scale counts for examined specimens of Polemon christyi, P. collaris, and P. ater, new species. For character abbreviations, see Materials and Methods. Data are shown as mean6standard deviation with range in parentheses. Asterisks next to species names indicates data included from type specimens.
Fig. 3
Scatter plots of PC1 and PC2 scores for the analysis with morphometric data regressed against SVL for examined specimens of Polemon.
Fig. 4
Photographs of the holotype of Polemon ater, PEM R20734 (254 mm SVL), subadult male from Fungurume, Lualaba Province, Democratic Republic of the Congo, in life (photos: CRT). (A) Closeup of head; (B) anterior body and tongue; (C) entire body.
DISCUSSION
Polemon ater is one of only a few lamprophiid species to be described from Central Africa in recent years (e.g., Greenbaum et al., 2015; Trape and Mediannikov, 2016) and the first species described from the genus in over 70 years. Micrelaps tchernovi was described in 2006, but in recent phylogenetic analyses, the genus was recovered outside the subfamily Aparallactinae (Figueroa et al., 2016; Portillo et al., 2018). As is the case with most species of aparallactines, morphological conservatism is common within Polemon. Meristic characters for several species (e.g., P. collaris, P. christyi, and P. ater) display considerable overlap, making specimens difficult to distinguish (Table 2). These species are most easily distinguished by coloration (presence of an orange or yellow collar, presence of dorsal stripes), head shape, and head scalation shape (Figs. 4–6). ]
Morphologically, the most similar species to the newly described P. ater is P. christyi, but the latter species is not sister to the former one, and was recovered in a well-supported clade with P. robustus (Fig. 1; Portillo et al., 2018). The latter study lacked genetic samples of P. gabonensis, but morphologically, P. gabonensis is readily distinguished by its large, broad snout (Figs. 5, 6; de Witte and Laurent, 1947). Moreover, P. gabonensis also has a distinct, creamy yellow venter that easily distinguishes it from P. ater, and the former species is only known from lowland rainforest, a habitat that is distinct from the miombo woodland/savanna habitat of P. ater (Broadley and Howell, 1991; Broadley et al., 2003; Chippaux, 2006; Chirio and LeBreton, 2007; Pauwels and Vande weghe, 2008; Portillo et al., 2018; Spawls et al., 2018).
Polemon ater is genetically most similar to P. collaris, and the sister relationship between the two species was strongly supported in maximum likelihood (RAxML) and Bayesian inference (MrBayes and BEAST) analyses (Fig. 1; Portillo et al., 2018). Morphologically, the two species can usually be distinguished by the grayish black or black dorsal and ventral coloration of P. ater. This contrasts with the vibrant yellow or cream collar (that may fade with age) and contrasting, creamy white venter that seems ubiquitous in many adult specimens of P. collaris (de Witte and Laurent, 1947; FP, pers. obs.). Polemon collaris was described from “Macange” (¼Malanje, Malanje Prov., N Angola, 098330S, 168200E) and characterized (in part, and as its name implies) by a pale collar. An anomalous Angolan sample (PEM R19893), with a very faded grayish collar, grayish black dorsum, and grayish black, white-edged ventrals, showed substantial genetic differentiation (i.e., long branch length) from P. ater and other samples of P. collaris from DRC (Fig. 1). Its relationship to other Angolan populations possessing a pale collar (e.g., from Malanje, Cazengo ,and Pungo-Andongo) awaits further study (de Witte and Laurent, 1947; FP, pers. obs.). A surprising result from the phylogenetic analyses of Portillo et al. (2018) was the placement of P. christyi, which was recovered as sister to P. robustus. Polemon christyi, P. collaris, P. gabonensis, and P. ater all have similar ranges of ventral scale counts, yet P. christyi was found to be more closely related to P. robustus (Portillo et al., 2018), which is stockier in build and has substantially fewer ventral scales relative to most congeners (de Witte and Laurent, 1947).
Fig. 7
Photograph of miombo woodland habitat of the paratype of Polemon ater in Kalakundi, Lualaba Province, Democratic Republic of the Congo (photo: WRB).
Polemon christyi has been recorded from Uganda, western Kenya, Virunga National Park (eastern DRC), Upemba National Park (southeastern DRC), Garamba National Park (northeastern DRC), Lualaba Province (DRC), Central African Republic (record considered doubtful sensu Chirio and Ineich, 2006), Rwanda, Burundi, west-central Tanzania, Zambia, South Sudan, and northeastern Malawi (Boulenger, 1903, 1911, 1915; deWitte, 1941, 1953, 1955, 1975; deWitte and Laurent, 1943, 1947; Loveridge, 1944; Laurent, 1955, 1956a, 1956b, 1960; Broadley, 1971; Pitman, 1974; Spawls, 1978; Hinkel and Fischer, 1988; Joger, 1990; Broadley and Howell, 1991; Meirte, 1992; Vonesh, 2001; Behangana and Goodman, 2002; Broadley et al., 2003; Chippaux, 2006; Mercurio, 2007; Caro et al., 2011;Wallach et al., 2014; Spawls et al., 2018). These records encompass a large geographic area with multiple habitats in different elevations, and in some cases, it is not clear whether the specimens are referable to P. ater, P. christyi, or an unknown species.
Schmidt (1923) described Miodon unicolor (later placed in the synonymy of P. christyi by de Witte and Laurent, 1947) based on a single male specimen from lowland rainforest in Poko (Ituri rainforest), northeastern DRC. This specimen has 202 ventral scales and a uniformly dark bluish gray dorsum, with ventral scales edged with white. Based on these features and its locality, the specimen is likely attributable to P. christyi rather than P. ater. The shape of the nasal and postocular scales (based on the original description) of Miodon unicolor also closely matches that of P. christyi (Schmidt, 1923). De Witte (1941) described Melanocalamus leopoldi based on a female specimen from montane forest in Rwankeri, Rwanda (2200 m) with 245 ventral scales and fused preocular and prefrontal scales. This specimen contains more ventral scales than the two female specimens of P. christyi examined herein (but within range of Ugandan specimens, sensu Pitman, 1974), and also fused preocular and prefrontal scales, a trait that is not exhibited by either P. ater or Ugandan P. christyi. Laurent (1956a) placed M. leopoldi in the synonymy of P. christyi, but this action was seemingly rejected by de Witte (1962), and based on the absence of a preocular (because of fusion with the prefrontal), Meirte (1992) continued to recognize the former taxon as a valid species, and he retained Melanocalamus as a subgenus of Polemon.Wagner et al. (2014) also recognized Polemon leopoldi as a distinct species. Although Laurent's (1956a) action has been accepted by most authorities (e.g., Wallach et al., 2014; Spawls et al., 2018; Uetz et al., 2018), further examination of Rwandan populations is needed to determine with certainty whether M. leopoldi is conspecific with topotypic P. christyi.
Loveridge (1944) noted a specimen of Miodon gabonensis christyi (¼Polemon christyi) from 4600 feet (1402 m) at Ilolo, located in present-day Ruaha National Park, Tanzania. No distinctive morphological features were noted by Loveridge (1944) for this specimen (MCZ R30401), and because the park contains both miombo woodland and evergreen forest (Mtui et al., 2016), further study is needed to determine the taxonomic status of this population. Several illustrations of the head of a specimen of P. christyi (RGMC 9809) from Usumbura (¼Bujumbura), Burundi were shown by de Witte and Laurent (1947: figs. 67–69), and Laurent (1960) provided additional records from Uvira and nearby Makobola (DRC)— these records are intriguing because they are in a floodplain near the shore of Lake Tanganyika (EG, pers. obs.), although it is possible that some gallery forest was intact at the time of collection. Polemon christyi has also been recorded from Garamba National Park in northeastern DRC, but the specific locality and habitat where the specimen was found were not noted by deWitte (1966). Although the park is dominated by grasslands and woodlands, it contains some gallery forest (Hillman Smith et al., 2014). Six specimens of P. christyi reported by de Witte (1955, 1975) from Virunga National Park included Mutsora (savanna habitat, 1200 m) and Indray (“spiny” savanna and euphorbia habitat, 900 m), which are relatively xeric habitats in the park. Mercurio (2007) recorded a specimen of P. christyi from montane forest (1995 m) in the Wilindi Forest Reserve, northeastern Malawi. This specimen has six supralabial scales, unlike P. christyi and P. ater, which both have seven supralabials (Boulenger, 1903; de Witte, 1941, 1953; de Witte and Laurent, 1943, 1947; Laurent, 1956a, 1960; Pitman, 1974; Meirte, 1992; Broadley et al., 2003; Chippaux, 2006; Mercurio, 2007), but otherwise it has similar morphology. Further genetic and morphological examination of the Malawi population is needed to determine whether it represents P. christyi, P. ater, or an unknown species. The montane forest records of P. christyi from Rwanda (de Witte, 1941) and Malawi (Mercurio, 2007) were both found at higher elevations (2200 m and 1995 m, respectively) than the known elevational range of P. ater (1189–1810 m) or P. christyi (600–1760 m; de Witte, 1941, 1953; Pitman, 1974; Broadley et al., 2003; Mercurio, 2007; Spawls et al., 2018). We thus restrict the known range of P. christyi to forests, and possibly grasslands, woodlands, and savannas, of northeastern DRC, Uganda, South Sudan, and western Kenya. Additional work is needed on the species complex, but because specimens are rare and fieldwork in DRC is problematic for many reasons (Greenbaum, 2017), it will likely be many years before all of these populations can be examined with molecular data.
The BEAST results from Portillo et al. (2018) suggested that P. ater and P. christyi last shared a common ancestor during the early to mid-Miocene (around 16 mya), which coincided with a climactic optimum (Couvreur et al., 2008; Feakins and Demenocal, 2010). These results also indicated that P. ater diverged from its sister taxon, P. collaris, during the late Miocene (ca. 6 mya), when increasingly cool and arid conditions throughout central and eastern Africa likely fragmented populations of many squamates, eventually leading to their speciation (Greenbaum et al., 2018). Other Central African snake groups with similar dates of divergence between sister taxa include the lamprophiid genus Boaedon and viperid genus Atheris (Menegon et al., 2014; Greenbaum et al., 2015).
Interestingly, the species of Polemon that are morphologically similar to P. ater (P. christyi, P. collaris, and P. gabonensis) are mainly inhabitants of rainforests. Specifically, Polemon christyi is known from rainforests and associated forest relicts in Uganda, but it might occur in grasslands and woodlands in Garamba National Park, DRC (de Witte, 1966; Pitman, 1974). Polemon ater inhabits southeastern Lualaba, Haut-Katanga, and Haut-Lomami provinces of DRC, and Zambia, which are dominated by grasslands and miombo woodlands (deWitte, 1953; Broadley et al., 2003).Within Lualaba, Haut-Katanga, and Haut-Lomami provinces, plant species richness was highest within the miombo ecoregion (Broadley and Cotterill, 2004), and several unique species of reptiles are known from the region (e.g., Greenbaum et al., 2012; Medina et al., 2016). Polemon ater might have adapted to the miombo woodlands and savannas when arid climates in Africa increased after 9.6 mya (Feakins and DeMenocal, 2010).
Many non-forest habitats in southeastern DRC that are potential habitats of P. ater and other aparallactines (FP, unpubl. data) are exposed to degradation because of poor farming management, uncontrolled fires, mining, and other environmental degradation linked to human population growth. Because of these factors, these habitats are constantly at risk, especially unprotected regions in southeastern DRC miombo woodlands and savannas (Sodhi et al., 2007; Herrmann and Branch, 2013). Additionally, southeastern DRC is known to harbor high species diversity of plants, amphibians, reptiles, and birds (Broadley and Cotterill, 2004; Greenbaum et al., 2012; Larson et al., 2016; Medina et al., 2016). Given the results herein and from Portillo et al. (2018), it is likely that P. ater is endemic to the grasslands, miombo woodlands, and possibly forests of southeastern DRC, Zambia, and west-central Tanzania. Possible populations in Rwanda, Burundi, and Malawi require further study. Given the proximity of the Dilolo locality (Laurent, 1956b) to the border of DRC, P. ater is likely to be found in neighboring Angola.
MATERIAL EXAMINED
* Indicates type specimens. Institutional abbreviations follow Sabaj (2016), with the Royal Museum for Central Africa, Tervuren listed as RMCA R and the Royal Belgian Institute of Natural Sciences listed as RBINS.
Polemon acanthias: ZMB 51389, West Africa.
Polemon bocourti: MNHN-RA 1896.0553–54*, Vallee de l'Ogoone, Congo.
Polemon christyi: BMNH 1946.1.8.88*, Uganda; CAS 111863, Mabira Forest, Uganda; CAS 147905, Kakamega Forest, Kenya; CAS 204334, road between Kibale National Park and Fort Portal, near Kibale National Park turnoff, Uganda; UTEP 21618, road to Budongo Central Forest Reserve, Western Region, Uganda, 01.653578N, 31.328068E, 1084 m.
Polemon collaris: PEM R19893, Lunda Norte, Angola, 9.399228S, 20.413238E; UAC 62 (RBINS 18544), Yoko, DRC, 1.93008N, 25.25258E; UTEP 21612, Byonga, South Kivu, DRC, 03.336948S, 28.124198E, 710 m; UTEP 21613, Fizi, South Kivu, DRC, 04.274708S, 28.929308E, 1268 m; UTEP 21614, Salonga River, Watsi Kengo, Tshuapa, DRC, 00.910068S, 20.620768E, 332 m; ZMB 10045*, Kwango River, Angola; ZMB 20308, Cameroon.
Polemon collaris longior: RMCA R 1629*, Medje, DRC; RMCA R 15843*, Lutunguru, DRC.
Polemon fulvicollis: MNHN-RA 1886.0211*, Congo Brazzaville, Franceville.
Polemon fulvicollis laurenti: RMCA R 4771*, Tongo, DRC; RMCA R 15430, Stanleyville (i.e., Kisangani), DRC; RMCA R 18246, Katche, Kivu, DRC; RMCA R 21569, Kivu, DRC; UTEP 21615, Bombole village, Bas-Uele, DRC, 02.278058N, 25.146798E, 467 m.
Polemon gabonensis: BMNH 1946.1.3.4, 5*, Cameroon Mountains, Cameroon; RMCA R 16087, Ibembo, DRC; RMCA R 16545, Bokoro, DRC; RMCA R 21030, Ikela, DRC; ZMB 21142, Cameroon.
Polemon gabonensis brachyurus: RMCA R 20326*, Idjwi Sud, DRC.
Polemon gabonensis schmidti: RMCA R 8008*, Stanleyville (i.e., Kisangani), DRC; RMCA R 10545*, Karawa, Ubangi, DRC.
Polemon graueri: CRT 4007 (RBINS 18543), Bomane, DRC, 1.2708N, 23.7328E; UTEP 21611, Rwenzori Mountains National Park, near Nyakalengisa entrance, Uganda, 00.362478N, 29.998638E, 2075 m; UTEP 21650, Idjwi Island, Bugarula, South Kivu, DRC, 02.058158S, 029.057918E, 1541 m; ZMB 20721*, Entebbe, Uganda.
Polemon griseiceps: BMNH 1946.1.21.90*, Bitye, South Cameroon.
Polemon notatus: ZMB 10271*, West Africa.
Polemon robustus: RMCA R 6803, 6851, 8761, 11839*, Kunungu, DRC; UTEP 21616, Lake Mai-Ndombe, Isongo, Mai-Ndombe, DRC, 01.342228S, 18.237748E, 309 m; UTEP 21617, Salonga River, Itala Village, Equateur, DRC, 00.626158S, 20.208968E, 322 m.
ACKNOWLEDGMENTS
Fieldwork by the last author in DRC was funded by the Percy Sladen Memorial Fund, an IUCN/SSC Amphibian Specialist Group Seed Grant, K. Reed, M.D., research funds from the Department of Biology at Villanova University, a National Geographic Research and Exploration Grant (no. 8556-08), UTEP, and the US National Science Foundation (DEB-1145459); E. Greenbaum, C. Kusamba, W. M. Muninga, and M. M. Aristote thank their field companions M. Zigabe, A. M. Marcel, M. Luhumyo, J. and F. Akuku, F. I. Alonda, and the late A. M'Mema. The Centre de Recherche en Sciences Naturelles and Institut Congolais pour la Conservation de la Nature provided project support and permits. Fieldwork by Z. T. Nagy in DRC was supported by the Belgian National Focal Point to the Global Taxonomy Initiative. D. F. Hughes, E. Greenbaum, and M. Behangana thank the Uganda Wildlife Authority of Kampala for necessary permits to work in Uganda. W. R. Branch thanks the National Research Foundation of South Africa (NRF) for funding, and Thomas Branch for collecting important specimens in northeastern Angola during mine exploration in the region. We thank W. Conradie of Port Elizabeth Museum for facilitating loans of specimens and providing measurement data for a prey item of the paratype of P. ater. F. Portillo and E. Greenbaum thank Danny Meirte of the Royal Museum for Central Africa for facilitating access to specimens and correcting crucial details of this paper. It is with great sadness that we note that our friend and colleague Prof. William Roy Branch did not live to see the publication of this manuscript. His passing is a huge loss to the many colleagues who treasured his friendship, advice, and counsel. We salute him as a true giant of African herpetology.
