Based on morphology, biogeography, biology, and the results of a molecular phylogenetic analysis, a new species of Dasypeltis is described from Harenna Forest in the southern Somalian plateau in Ethiopia. It is a medium-sized member of the genus (females up to 92 cm total length, males up to 58 cm), usually uniformly black dorsally but occasionally light brown with a dense and somewhat irregular pattern of narrow brown saddles and vertically elongate lateral markings, sometimes forming bands across the back, and small white spots and streaks, or reddish-brown with indistinct darker dorsal and lateral markings. Interstitial skin is light grey, appearing bluish in adults. Infralabials, chin shields, gulars, and 20–62 anterior ventral plates (medially) are mostly white, the other plates being ash-grey or black in colour. Scales in the first row of dorsals on either side of the body are smooth or almost smooth; those of the second row are weakly keeled; other dorsal scales are moderately keeled with two apical pits each. Pairwise genetic distance between the new species and its closest congener, D. atra, is similar to or greater than the distances between some long-recognised species in the genus. The new species is also geographically and ecologically isolated from closely related congeners.

Anhand morphologischer, biogeographischer und biologischer Merkmale sowie der Ergebnisse einer molekularphylogenetischen Analyse wird eine neue Art der Gattung Dasypeltis vom südlichen somalischen Hochplateau in Äthiopien beschrieben. Es handelt sich um einen mittelgroßen Vertreter der Gattung (Weibchen erreichen bis zu 92 cm Gesamtlänge, Männchen bis zu 58 cm), welcher in der Regel einheitlich schwarz am Rücken gefärbt ist. Die dorsale Zeichnung ist gelegentlich hellbraun mit einem dichten und etwas unregelmäßigen Muster aus schmalen braunen Sattelflecken und vertikalen, länglichen Seitenzeichnungen, die manchmal Bänder über den Rücken bilden, sowie kleinen weißen Flecken und Streifen. Gelegentlich sind diese Zeichnungselemente rotbraun mit undeutlich angedeuteten Wirbel- und Seitenzeichnungen. Die Zwischenhaut ist hellgrau und erscheint bei adulten Tieren bläulich. Infralabiale, Kinnschilde, Kehlschilde und 20–62 vordere Bauchplatten sind in der Mitte weiß, danach aschgrau oder schwarz. Die Schuppen der ersten Reihe auf beiden Körperseiten sind glatt oder fast glatt. Die der zweiten Reihe sind nur schwach gekielt. Die übrigen Rückenschuppen sind mäßig gekielt und weisen jeweils zwei apikale Gruben auf. Die genetische Distanz zwischen der neuen Art und ihrer nächsten Verwandten, D. atra, ist vergleichbar mit oder größer als die Distanzen zwischen seit langem bekannten Arten der Gattung. Die neue Art ist geografisch und ökologisch von nah verwandten Arten der Gattung isoliert.

Dasypeltis albigularis sp. nov., Dasypeltis atra, Dasypeltis fasciata, Dasypeltis gansi, egg-eater, Harenna Forest, Horn of Africa, Serpentes, Somalia, taxonomy

Dasypeltis albigularis sp. nov., Dasypeltis atra, Dasypeltis fasciata, Dasypeltis gansi, Eierschlange, Harenna-Wald, Afrikanischer Horn, Schlangen, Somalia, Taxonomie

As recently as the mid-2000s, only six species of egg-eating snakes of the genus Dasypeltis Wagler, 1830 were recognised, namely D. atra Sternfeld, 1912, D. fasciata Smith, 1849, D. inornata Smith, 1849, D. medici Bianconi, 1859, D. palmarum (Leach, 1818), and D. scabra (Linnaeus, 1758). At that time, populations in the Horn of Africa were treated as either D. scabra only (Largen and Rasmussen 1993) or D. scabra and D. atra (Hughes 1997). Trape and Mané (2006), while describing three new species from western Africa, resurrected the Ethiopian endemic D. abyssina (Duméril, Bibron & Duméril, 1854) from the synonymy of D. scabra. However, this act was only acknowledged but not followed by Largen and Spawls (2010), who again included only D. scabra and D. atra in their book on the herpetofauna of Ethiopia, Eritrea, and Djibouti, noting that D. abyssina needed to be “redefined”. This was addressed eight years later by Bates and Broadley (2018), who provided a comparative morphological analysis and a redescription of D. abyssina. They also described three new species in the genus, of which two – D. crucifera Bates, 2018 and D. taylori Bates & Broadley, 2018 – are from the Horn of Africa. These species, like the majority of Dasypeltis, were described based on external morphology (including colour pattern) and geographical distribution.

It was only in 2012 that, for the first time, a molecular phylogenetic analysis was used to support the description of a new species, i.e. D. parascabra (Trape et al. 2012). For several species, especially those in western and central Africa, this study and similar investigations conducted in subsequent years (see below), confirmed the overall correctness of species delimitation based on colour. Using a phylogeny generated from 16S rRNA sequences, Trape et al. (2012) showed that the colour pattern described for species in western Africa indeed reflects the phylogeny of Dasypeltis. In a paper describing D. congolensis, Trape et al. (2021) used additional DNA sequences and included more species. Unfortunately, this phylogeny was again inferred using only the 16S rRNA marker, and therefore it may not reflect true evolutionary relationships in the genus. See, for example, criticism by Hillis et al. (2021) and Wüster (2025) regarding the use of only mitochondrial markers for inferring phylogeny. Nonetheless, overall, Trape et al. (2012, 2021) largely confirmed the morphology-based species delimitation hypotheses that had been proposed by previous workers. Although the phylogram presented by Trape et al. (2021) contains many nodes with extremely weak support (posterior probabilities of 16–46%), the haplotypes it revealed match those species that had been recognised in the past largely on the basis of colour pattern.

Since meristic morphological characters in Dasypeltis are of limited value for distinguishing between species, identification is often aided by colour pattern and distribution. The use of dorsal colour pattern as an aid to identification was introduced by Gans (1959), who proposed a classification of colour patterns which he applied in a revision of the genus. However, Gans did not imply that colour pattern alone was a satisfactory criterion for identification. This ‘colour pattern’ approach was widely accepted and used by subsequent authors. It has proved especially useful for species with conspicuous or vivid markings on the body. For example, D. confusa Trape & Mané, 2006 can usually be distinguished from most congeners by its distinct ‘5L’ pattern of broad dark bands across the back (saddles and lateral bars are confluent), although it is now inseparable from D. congolensis Trape, Mediannikov & Chirio, 2021, which has the same pattern. It should, however, be noted that some variation does occur with regard to supposedly diagnostic colour patterns, an example being D. confusa, in which some Nigerian and Ghanaian individuals have a ‘5L’-like pattern but with several or most saddles and lateral bars unlinked, and sometimes the pattern is even ‘5N’-like (as in D. scabra), at least on some parts of the body (Bates 2013).

Gans (1959) applied the same coding to plain-coloured Dasypeltis. For example, he used the code ‘2B’ for specimens he identified as D. inornata and for the brown phases of a few other species, including D. atra (which he had subsumed within D. scabra). Plain or sparsely marked specimens occur in some species of Dasypeltis (e.g. D. scabra sensu lato and D. gansi Trape & Mané, 2006) that usually have distinct colour patterns (e.g. Barends et al. 2024). In such cases, information about geographic origin is helpful, as such morphs are not known to be sympatric with other species that are normally plain-coloured (e.g. D. inornata and D. atra). Dasypeltis inornata is a southern African endemic with a plain olive-greyish, sometimes reddish, dorsum. In D. atra, patterned individuals sometimes occur in plain-coloured populations (e.g. fig. 319 in Largen and Spawls 2010: 493) and thus are the exception rather than the rule. Their body pattern is also quite different from that of other Dasypeltis, especially those with which they are sympatric. Brown or red-brown plain-coloured individuals also occur, but generally, D. atra is referred to as the ‘black’ egg-eater, and even its specific epithet ‘atra’ indicates this (although this name nearly lost its priority status to the name D. lineolata Peters, 1878, which refers to a patterned dorsum – see Bates 2018). As black individuals are not known in other populations of egg-eaters, black Dasypeltis are always referred to as D. atra.

Dasypeltis atra (Fig. 1A) is a largely silvicolous species that occurs at elevations of 1,500–2,500 m asl. The black morph is typical of montane populations at high elevations, hence the vernacular name ‘Montane Egg-eater’. The type locality of D. atra is “jungle behind the border mountains on the north-west shore of Lake Tanganyika”, and the syntypes were all black (Sternfeld 1912: 272). Largen and Rasmussen (1993: 340) considered it likely that D. atra was present in the forested south-western provinces of Ethiopia, adjacent to South Sudan. They suggested that “any uniformly dark Dasypeltis that has a single postocular shield and is found in a forest habitat is potentially D. atra, but any such determination should be regarded as tentative”. Black or red-brown egg-eaters, as well as some greyish and brownish patterned specimens, reported from localities in Ethiopia as far north as Lake Tana and as far east as Ogaden, and not only from humid montane forests but also from dry woodland and savannah at low elevations, have also been treated as D. atra (e.g. Largen and Spawls 2010; Bates and Broadley 2018; Spawls et al. 2023).

Figure 1. 

A. Dasypeltis atra from Chuka, Kenya (ZSM 328/2023); B. Dasypeltis sp. from the Harenna Forest, Ethiopia (ZSM 94/2019).

Broadley and Bates (2009) noted considerable geographic variation regarding the occurrence of the various colour phases of D. atra. Nevertheless, they indicated that the dorsally black ‘2M’ form is most common in forest, the brown ‘2B’ form is most common in grassland, and the patterned ‘2Bx’ form is most common in savannah. They also suggested that D. atra persisted in savannahs after the retreat of forests as a result of climatic changes and then evolved the ‘mottled’ dorsal pattern for better camouflage. Bates and Broadley (2018) later named several localities where black and brown, black and patterned, brown and patterned, and all three morphs occur in sympatry. Today, in the context of molecular phylogenetics, such hypotheses may need to be revised because at least some such populations may represent separate species.

In 2019, a population of ‘black’ Dasypeltis was discovered by the first author (AT) on the south-eastern slope of the Somalian Plateau, in the Harenna Forest, one of the last remaining natural forests in the Horn of Africa. To the best of current knowledge, it was the first vouchered record of this genus from the Harenna Forest. Before that, “Dasypeltis scabra” from “S of Shisha River” had been mentioned among “Amphibians and reptiles recorded from the Bale Mountains” (Largen and Spawls 2011), but, as the voucher could not be located, this report has to be treated as anecdotal. The first specimen was a dorsally black juvenile female received from a local collaborator on 30 May 2019. Based solely on its colour, this animal was initially identified as D. atra (Fig. 1). Its dorsal scales were weakly keeled, those of the first lateral row on either side were smooth, the interstitial skin was bluish, the gular scales and anterior ventral plates were white, and the head was decorated with scattered small white dots and speckles. Moreover, the specimen had a loreal scale (which is normally absent in Dasypeltis) on the left side of the head and a small scale between the internasal, prefrontal and nasal scales on both sides of the head (Fig. 2A).

Figure 2. 

Small scales intercalated between the nasal, prefrontal and internasal in A. Dasypeltis sp. from the Harenna Forest, Ethiopia, ZSM 94/2019, which had the scale on both sides of the head; and B. Female offspring of ZSM 94/2019 (hybrid with Dasypeltis atra), with the scale present on the right side of the head only. Note also the ‘loreal’ scale in A.

To determine whether the unusual scales described above were inherited and not merely the result of a mutation, and to experimentally test the degree of relationship between this population and D. atra, AT mated this female, after it had reached sexual maturity, with a male D. atra from Chuka, Mount Kenya. Copulation occurred on 15 July 2022 and lasted for approximately 15 h, the usual duration for this genus. On 10 March 2023, after a gestation of almost seven months, which is unusually long for Dasypeltis, the female deposited 19 eggs, of which 17 were of a colour, shape, and size that suggested that they were fertile. However, all but one egg perished a few days after deposition. In clutches produced by other Dasypeltis in captivity, usually 80–100% of fertile eggs are viable. The remaining egg hatched on 15 June 2023, after 100 days of incubation at a constant temperature of 23 °C. The hatchling, weighing 2.55 g, was a female (Fig. 3). Therefore, the possibility of parthenogenesis was excluded.

Figure 3. 

F1 hybrid between a male Dasypeltis atra from Chuka, Kenya, and a female Dasypeltis sp. (ZSM 94/2019) from the Harenna Forest, Ethiopia, with mixed phenotypes.

The hatchling exhibited attributes of both the mother and father. It had a small scale between the nasal, prefrontal and internasal scales, as in the mother, but only on the right side of the head (Fig. 2B). The interstitial skin was bluish but darker than that of the mother. The colour of the infralabials, gulars, and first seven ventral scales was a mixture of black and light grey. However, it lacked the ‘loreal’ scale that was present in the mother (Fig. 2), and the lateral scales were all distinctly keeled, as in D. atra (Fig. 3). This F1 hybrid female grew rapidly and reached sexual maturity at the age of approximately 8 months, when it produced an infertile clutch of 18 eggs on 24 January 2024. Usually, sexual maturity in females of this genus is reached after 1.5–2 years. Thereafter, it mated successfully with a D. atra male, and on 12 April 2024 it deposited, after a remarkably short gestation of less than 3 months, 11 fertile eggs (maximum length 27 mm, maximum width 16 mm), which were a little smaller than the eggs of D. atra (21–33 × 14–19 mm). Its next fertile clutch of 12 eggs followed on 29 July 2024, i.e. only 3.5 months after the first. The F2 hybrid hatchlings (hatched on 21–23 August 2024) were smaller than D. atra hatchlings (less than 3 g compared with usually over 3 g) but had the phenotype of the latter, i.e. none had the same type of extranumerary scales on the snout, blue skin, or white gulars as did the grandmother from the Harenna Forest. Subsequently, the F1 hybrid female produced more offspring that were, overall, superficially indistinct from D. atra. The D. atra male made no distinction between the hybrid female and a female D. atra, but no longer showed any interest in the female from the Harenna Forest with which it had mated with earlier, and all attempts to mate them again were unsuccessful. In addition, the Harenna Forest female no longer produced fertile clutches (i.e. after the clutch in which only one egg hatched). This seemed quite unusual because females of other Dasypeltis species lay up to three clutches after a single mating. Instead, this female continued laying infertile eggs until it later mated with a male from Harenna Forest (see below).

The unusual morphological traits observed in the first Harenna Forest egg-eater and the insights gained from the hybridisation experiment stimulated further investigations, the results of which are reported below.

Examination of 16 specimens of Dasypeltis sp. from the Harenna Forest confirmed the consistency of the morphological traits mentioned above that distinguish the proposed new species from D. atra (Table 1). All individuals that hatched from 25 February to 2 March 2026 from the eggs deposited by a female on 20 November 2025 (after mating with a conspecific male from the same locality on 5 July 2025) were phenotypically similar to their parents, as were all hatchlings that emerged on 4–8 June 2026 from the next clutch. This further confirmed that the traits were consistent and inherited. In the Harenna Forest snakes, the white under the head and on the anterior parts of the belly is distinctive, and it can be noted that ‘2M’ D. atra always have a black dorsum with black or dark grey underparts (see, e.g., specimens listed in Suppl. material 4).

The phylogenetic analyses recovered a close evolutionary relationship between the Harenna Forest population and D. atra (based on specimens from Kenya and Uganda). According to phylogenetic inferences in both Bayesian (MrBayes v3.2.7a) and ML (IQ-TREE 2) frameworks (Fig. 5), the Harenna Forest population represents a sister species to D. atra, and together they share a common ancestor with D. gansi. These branches are well supported in both the phylograms constructed for single-gene markers (16S, cyt b; Suppl. material 2) and the concatenated multigene phylogeny, combining mitochondrial (cyt b, 16S) and nuclear (RAG2, c-mos) markers. When D. fasciata is included, it is sister to D. gansi, and both species constitute a sister clade to D. atra and the Harenna Forest population (e.g. with a posterior probability of 0.78, as in the 16S-based phylogeny shown in Suppl. material 2).

Figure 4. 

A patterned specimen (HFE2025-023) of Dasypeltis albigularis sp. nov. from the Harenna Forest, Ethiopia.

Figure 5. 

Phylogenetic relationships in Dasypeltis from the Horn of Africa, inferred from concatenated cyt b, c-mos, 16S, and RAG2 data sets using both Bayesian inference in MrBayes v3.2.7a and ML inference in IQ-TREE 2. Node labels indicate posterior probabilities, ultrafast bootstrap support, and approximate likelihood ratio support. A single specimen of Ahaetulla prasina (CHS836) was used as the outgroup; for c-mos, the sequence for this specimen was substituted with that of another specimen of the same species (ROM 30961). The outgroup was excluded from the final tree visualisation.

The pairwise distances between the Harenna Forest Dasypeltis and D. atra are of a similar magnitude to those between D. atra and D. gansi and are similar to or even larger than those between some more distantly related congeners, e.g. D. confusa and D. atra (Suppl. material 3). With regard to the RAG2 nuclear gene, the distance between D. atra from Kenya and D. confusa from Nigeria is 0.8%, whereas it is 1.5% between the Harenna Forest population and D. atra from Kenya (Suppl. material 3: table S3-4). Considering the nuclear c-mos locus, the interspecific difference in this genus remains below 1%, revealing the low phylogenetic signal in this marker. This is further evident in the haplotype network (Fig. 6), which recovered a shared allele for D. atra and D. cf. scabra. Despite the low phylogenetic signal, the Harenna Forest population displays a private allele, separated from D. atra by one mutational step. The intraspecific p-distance for 16S between the populations of D. atra from Uganda and Kenya is 0.4%, whereas it increases to 2.6% between the Ugandan D. atra and the Harenna Forest Dasypeltis and to 3.1% between the latter and the Kenyan D. atra (Suppl. material 3: table S3-3).

Figure 6. 

Haplotype network for Dasypeltis from the Horn of Africa generated using a nuclear gene (c-mos, 370 bp) data set. Each tick represents one nucleotide substitution between alleles. Allele sharing is indicated by multiple specimens (or species) sharing the same node.

Somewhat surprisingly, the Harenna Forest population is rather distantly related to the parapatric population of reddish Dasypeltis (D. sp. ‘Borena’) found in an arid woodland area of Borena approximately 20 km to the south-east at elevations below 1,500 m asl. As noted above, red-brown egg-eaters occurring in Ethiopia have been referred to D. atra (e.g. Largen and Spawls 2010; Bates and Broadley 2018; Spawls et al. 2023), even though their dorsal pattern is more or less (i.e. indistinctly) of type ‘5L’ and not typically ‘2Bx’. Another population of Dasypeltis in an arid bushland and woodland area at 1,400–1,500 m asl, approximately 300 km north of the Harenna Forest, that has also been treated as D. atra, clustered with D. confusa and another potential new species (D. sp. ‘Somalia’) with 100% ultrafast bootstrap support in the cyt b phylogeny (see cyt b phylogram in Suppl. material 2). This haplotype is labelled “19 km S of Sheikh Hussein” and is represented by an almost uniformly dark-yellow specimen (TMHC 2022.09.1012) that had initially been misidentified as D. atra. The evolutionary separation of the Harenna Forest population from D. atra and D. gansi, as well as from geographically proximate populations of potentially new species, such as the above-mentioned population from Borena and D. sp. ‘Somalia’, is even more clearly visualised by the c-mos haplotype network (Fig. 6).

The population of Dasypeltis in the Harenna Forest is separated by a straight-line distance of 300–400 km from its putative closest relatives in the western Ethiopian highlands that are currently treated as ‘D. atra’ (e.g. Spawls et al. 2023) but may belong to a different, undescribed species. The geographical distance between the Harenna Forest and the D. atra samples from Kenya and Uganda used in our analysis is even greater, i.e. 600–800 km (Fig. 7). Isolation of the Dasypeltis population on the forested slope of the southern Somalian plateau from the latter area may be explained by a lack of suitable habitat. Both D. atra and the Harenna Forest Dasypeltis are associated with (although perhaps not restricted to) forests at moderate elevations. Such habitats and suitable climatic conditions are absent in the lowland savannahs, swamps, bushland, woodland, and desert located between the Harenna Forest in the Ethiopian highlands and the confirmed distribution of D. atra in the eastern Congo Forest and Kenyan highlands (Fig. 7).

Figure 7. 

Map showing collection localities of samples used for DNA sequencing in this study: Dasypeltis albigularis sp. nov. – triangles, D. atra – circle, D. sp. ‘Borena’ – diamonds, D. sp. ‘Somalia’ – squares. The red lines indicate apparent biogeographical barriers of lowland savannahs, swamps, and deserts where no species of the D. atra group have been recorded: between the Ethiopian highlands and the Kenyan highlands, as well as between the western and eastern Ethiopian plateaus.

Considering its morphological differences, geographical isolation, and phylogenetic distinctness, the Harenna Forest population is here proposed as a distinct lineage that is isolated from its closest relatives, D. atra, D. fasciata, and D. gansi, and is evolving as a separate species (see De Queiroz 1999, 2007). Here, this lineage is described as a new species.

Taxonomic account

Dasypeltis albigularis sp. nov.

https://zoobank.org/732539B6-49DD-4CF1-A11F-DF034671D43B

Figs 1B, 2A, 4, 8, 10, 11, 13

Common name.

White-throated Egg-eater.

Figure 8. 

Holotype ZSM 629/2025 of Dasypeltis albigularis sp. nov. from approximately 2 km south of Welmel river bridge, south of Hawo village, Harenna Forest, Ethiopia.

Figure 9. 

Habitats in the Harenna Forest, Ethiopia, at collection localities of Dasypeltis albigularis sp. nov.: A. Holotype ZSM 629/2025, approximately 2 km south of Welmel River bridge, in the vicinity of Hawo village (6°35'44"N, 39°35'43"E; 1,659 m asl.); B. Paratype ZSM 94/2019, Manyate village (6°29'00"N, 39°44'52"E; 1,554 m asl.); C. Paratype ZSM 630/2025 and live juvenile specimens HFE2025-031–033 (6°31'34"N, 39°38'36"E; 1,663 m asl.). Note the olive baboons, Papio anubis, in the background, feeding in the glade.

Figure 10. 

Paratype ZSM 630/2025 of Dasypeltis albigularis sp. nov. from a glade south of Hawo village, Harenna Forest, Ethiopia.

Figure 11. 

Male Dasypeltis albigularis sp. nov. (HFE2025-002) from Manyate village, Harenna Forest, Ethiopia. Note the bluish interstitial skin.

Figure 12. 

Incomplete ventral plates in specimen ZSM 254/2026 (top) and single subcaudals mixed with paired subcaudals in specimen ZSM 255/2026 (bottom), of Dasypeltis albigularis sp. nov.

Figure 13. 

A brown morph of Dasypeltis albigularis sp. nov. photographed in Manyate village, Harenna Forest, Ethiopia (ca. 6°29'N, 39°44'E; 1,550 m asl.). Photograph by A. M. Hussein.

Diagnosis.

A medium-sized Dasypeltis (females approximately 90 cm in total length, males less than 60 cm). Dorsally usually uniformly dark-grey or blackish, seldom reddish-brown with faint dark streaks; or light yellow-brown with a dense and irregular pattern of small brown saddles and narrow brown bars on the flanks, and numerous small white markings. Interstitial skin light grey, sometimes bluish, especially in adults. Anterior infralabials, gulars, chin shields and 20–62 anterior ventral plates mostly white, other ventrals gradually transition to ash-grey. Dorsal scales arranged around midbody in 20–23 rows. Scales of the first row of dorsals on either side of the body very weakly keeld or smooth and much larger than other dorsal scales. Scales of the second row on either side of body weakly keeled. Other dorsal scales each with a moderate median keel and a pair of apical pits distally. Head with fine white spots and speckles. Supralabials 7, 3^rd and 4^th bordering orbit. Postocular single. Frontal shield pitted in only one row around edges. Inter-prefrontal sulcus moderately marked, not deep. Prefrontal scales longer and larger overall than internasals. Rostral shield visible from above head.

Holotype.

Ethiopia • ♀; approximately 2 km south of Welmel river bridge, south of Hawo village, Harenna Forest; 6°35'44"N, 39°35'43"E; alt. 1659 m; 28 May 2025; A. Tiutenko leg.; killed by local villagers on a road; ZSM 629/2025.

Paratypes.

(Six specimens) Ethiopia • 1 ♀; glade south of Hawo village, Harenna Forest; 6°31'26"N, 39°38'31"E; alt. 1608 m; 31 May 2025; A. Tiutenko leg.; ZSM 630/2025 • 1 juv.; Manyate village, Harenna Forest; 6°35'N, 39°35'E; alt. 1550 m; 28 Feb.–2 Mar. 2026; A. Tiutenko leg.; hatched in captivity; ZSM 251/2026 • 1 juv.; same data as for preceding; ZSM 252/2026 • 1 juv.; same data as for preceding; ZSM 253/2026 • 1 juv.; same data as for preceding; ZSM 254/2026 • 1 juv.; same data as for preceding; ZSM 255/2026.

Additional referred material.

Ethiopia • 1 juv.; Manyate village, Harenna Forest; 6°35'N, 39°35'E; alt. 1550 m; 5 Mar. 2026; A. Tiutenko leg.; failed to hatch from an egg deposited in captivity; ZSM 256/2026.

Live specimens: Ethiopia • 1 ♀; Manyate village, Harenna Forest; 6°29'07"N, 39°44'44"E; alt. 1554 m; 30 May 2019; A. Tiutenko leg.; ZSM 94/2019 • 1 ♂︎; Manyate village, Harenna Forest; 6°35'N, 39°35'E; alt. 1550 m; 26 May 2025; A. Tiutenko leg.; HFE2025-002 (field number) • 1 ♂︎; same data as for preceding; HFE2025-003 (field number) • 1 ♀; same data as for preceding; 2 June 2025; A. Tiutenko leg.; HFE2025-050 (field number) • 1 ♀; south of Hawo village, Harenna Forest, in a glade in the forest while crossing a road after sunset; 6°31'34"N, 39°38'36"E; alt. 1663 m; 30 May 2025; A. Tiutenko leg.; HFE2025-023 (field number) • 1 ♀; same data as for preceding; 31 May 2025; A. Tiutenko leg.; HFE2025-031 (field number); • 1 ♂︎; south of Hawo village, Harenna Forest, in a glade in the forest while crossing a road after sunset; 6°31'26"N, 39°38'31"E; alt. 1608 m; 31 May 2025; A. Tiutenko leg.; HFE2025-032 (field number); • 1 ♀; same data as for preceding; 31 May 2025; A. Tiutenko leg.; HFE2025-33 (field number).

Description of the holotype.

ZSM 629/2025, adult female (Fig. 8). Body elongated, cylindrical. Dorsal scales on the body ovate and obtusely pointed, moderately unicarinate, those of the 3^rd to 5^th rows on either side reduced in size and obliquely positioned, with serrated keels; scales of the first row on either side are more-or-less rhomboid, very feebly carinate, and approximately one-and-a-half-times larger than those of the second row on either side which are of a similar shape but feebly carinate. Head small (HL 15.24 mm), HL/SVL 0.03 (head damaged posteriorly). SVL 508 mm, TL 74 mm, MDR 23, V 236, SC 52. Fourth and fifth ventral plates merged in the middle, plate 235 incomplete. Preoculars 1; postoculars 2; nasal scale divided between nostril and first supralabial. Temporals 1+2+3, scales of 2^nd and 3^rd transverse rows are keeled and equal in size and shape to dorsals. Supralabials 7, 3^rd and 4^th entering orbit; sixth supralabial the largest, approximately double the size of 3^rd and 4^th. Infralabials 7, 1^st to 3^rd in contact with the anterior chin shield, 4^th in contact with posterior chin shield. Colouration (soon after death): uniformly black above, ventrally uniformly black or ash-grey except for anterior infralabials, chin shields, gulars and anterior 31 ventral plates which are white or partly white. The first 21 plates are white medially and gradually merge into grey on the sides. Some white speckles are visible up to ventral plate 31. After over a year in ethanol, white parts now cream-yellow, otherwise colouration unchanged.

Variation.

There is minor variation in the scale counts of paratypes and additional referred material (Table 1).

The juvenile female paratype ZSM 630/2025 (Fig. 10) has temporals arranged as 2+2+3. Temporals in the first two rows are, unlike in the holotype, smooth, elongate, and larger than the dorsals.

In ZSM 255/2026, the 4^th to 6^th subcaudals are entire, and the posteriormost ventral plate is half the size of the preceding one; the white on the underside reaches the 20^th ventral plate. Plates 21 and 22 are moderately speckled with white, and plate 23 has only a few white spots. From plate 24 onwards, the ventrum is uniformly black. ZSM 251/2026 has particularly extensive white on the ventral surface: Ventral plates 1–62 are white, and white speckles are visible up to and including plate 82.

All specimens of the new species had some aberrant ventrals and subcaudals. This is not known to occur with such regularity in other Dasypeltis species. The wild-caught specimens of D. albigularis sp. nov. had at least one such scale, whereas all hatchlings of the first clutch obtained in captivity had several (e.g. Fig. 12). The hatchlings of the second clutch produced by the same female had few or no such aberrations. It thus remains to be investigated whether abnormal ventral and subcaudal scales are usual for this species or are an occasional developmental artefact resulting from some individual condition of the parents or from environmental parameters of incubation. However, this phenomenon was not observed in D. atra hatched from eggs incubated under the same conditions.

The female ZSM 94/2019 (Fig. 2) is currently the largest known specimen of this species, with a SVL of 80 cm and a TL of 12 cm. This snake hatched from an egg laid in 2018. Body size did not increase significantly after it reached the age of 2 years. Therefore, its current total length of 92 cm at the age of approximately 8 years is considered likely to be the maximum length for this species, although adult size in the wild is expected to be smaller. As noted above, this specimen has a small scale located between the nasal, prefrontal and internasal scales on either side of the head (Fig. 2B). In addition, on the left side of the head, it has a scale between the nasal and the preocular that would normally be identified as a loreal (generally absent in Dasypeltis).

HFE2025-023 (Fig. 4), a young female, is pale brown with a dense pattern of dark brown markings consisting of small, mostly narrow and transversely slightly to distinctly enlarged saddles in the vertebral region, starting at the nape and extending to the tip of the tail, and somewhat narrow, often irregular bars of similar colour on the flanks. The saddles are sometimes blotch-like or in the form of two attached spots side by side or one in front of the other, and in parts, several such confluent spots form a zigzag dorsal stripe. Scales that constitute the saddles and lateral bars usually have dark brown edges. In the case of the bars, these darker edges are often confluent, resulting in undulating or slightly irregular dark streaks. Saddles and bars are mostly not confluent but occasionally form narrow bands across the back. The dorsum is also liberally decorated with small white markings (on the edges of dorsal scales), usually slightly elongated (anteroposteriorly); the largest of these are present between the saddles, but they occur all over the body, becoming less common and prominent on the tail. The pattern of dark markings is not visible in the exuvia. Bluish interstitial skin is most prominent posteriorly, such that the posterior half of the body has a bluish tint. As in black specimens of this species, the underside of the head and the anterior 25 ventral plates are white. The head is a combination of pale brown and grey (especially greyish on the sides), with scattered, small, dark brown or white spots; there is a conspicuous row of 4–5 white, dash-like markings medially on the back of the head; the eye is orange-brown with small dark markings, and the pupil is vertically elliptic. Underparts are mostly pale olive green with numerous scattered small white spots, but laterally the ventral plates are cream-brown, similar in colour to the first two rows of dorsals on either side. The edges of the scales of the latter two rows are, in parts, dark brown, forming a series of three broken stripes along the flanks.

All other additional specimens are dorsally black and ventrally mostly dark grey. Infralabials, chin shields, gulars, and up to 25 anterior ventral plates are white. However, all specimens (not only those hatched in captivity but also wild-caught ones) have some malformed ventrals and subcaudals at different positions, where ventrals extend to half or less of the body width and undivided subcaudals are mixed with paired ones (e.g. Fig. 12).

Comparisons with related species.

Dasypeltis albigularis sp. nov., with a currently known maximum total length of 92 cm, is considerably smaller than its closest relatives – D. atra and D. gansi. For comparison, the largest known female D. atra (ZSM 328/2023) is 108 cm in total length, and the largest specimen of D. gansi (ZSM 326/2023) recorded so far measures 112 cm. The previously reported maximum size for D. gansi was 110 cm (Bates and Ineich 2012). The hatchlings of D. albigularis sp. nov., with weights of 2.17–2.34 g (n = 23), are also much lighter than hatchlings of D. atra, which are often up to 1 g heavier.

The new species has smooth or almost smooth scales in the first row of dorsals on either side of the body, and feebly keeled scales in the second row of dorsal scales on either side, whereas these scales are normally more strongly keeled in other species studied (Fig. 14). However, scales in the first row were found to be smooth in some specimens in museum collections that are currently identified as D. atra (see Suppl. material 4).

Figure 14. 

Comparison of mid-dorsal scales in eight species/lineages of Dasypeltis: A. D. albigularis sp. nov. (Harenna Forest, Ethiopia), B. D. atra (Chuka, Kenya), C. D. gansi (Togo), D. D. fasciata (Cameroon), E. D. sp. ‘Borena’ (Ethiopia), F. D. sp. ‘Somalia’ (Somaliland), G. D. confusa (Nigeria), H. D. taylori (Somaliland). E–G. Parts of photographs by T. Mazuch.

The scales in rows 3–5 that are used by the snakes of this genus for stridulation, i.e. creating a rasping sound by rubbing together the sides of the body, are smoother and flatter than in other species we compared them to. Therefore, the new species does not appear to be as well adapted for stridulation as, for example, D. confusa, D. gansi, and D. scabra, in which these scales are strongly serrated and acutely angled (Fig. 14). None of the specimens of D. albigularis sp. nov. observed produced sounds in this way. Similarly, D. atra and D. fasciata, whose scales in rows 3–5 are relatively moderately serrated, seem not to stridulate. However, the latter three species, when highly irritated, coil their bodies in the same way as congeners do when stridulating.

The new species differs from all other species of Dasypeltis, except for D. atra, by its usually uniformly black dorsal colour. However, in live black specimens, the dorsal and ventral colouration is lighter than in D. atra (Fig. 1). As in D. atra, a brown phase is known. A photographic record of such a specimen found at the type locality was received from a villager (Fig. 13). However, such snakes are not completely plain but have a feeble pattern of dark streaks dorsally. From D. atra, the new species also differs by having pale grey or bluish (versus dark grey) interstitial skin and white or mostly white infralabials, chin shields, gulars, and anterior ventral scales (versus black or grey underparts, although the throat may occasionally be pale, see Suppl. material 4). The white ‘throat’ is also evident in fully patterned specimens (Fig. 4). The colour pattern described above for patterned D. albigularis sp. nov. is unique in that the dorsal saddles are smaller, narrower, often blotch-like, and even confluent, forming zigzag patterns in parts, and the lateral markings are narrower and less regular. In D. atra, similarly patterned individuals (such as those illustrated in fig. 17G, H in Bates and Broadley 2018) have bolder dorsal markings.

The colour pattern in D. albigularis sp. nov. also differs from that of the parapatric, yet-to-be-named Dasypeltis sp. ‘Borena’, which has reddish dorsal colouration with a feeble, ‘5L’-type pattern of brown markings (after Gans 1959). Moreover, the dorsal scales of D. sp. ‘Borena’ lack apical pits (Fig. 14E), whereas these are present in D. atra and D. albigularis sp. nov. (Fig. 14A, B), as well as in their close relatives D. gansi and D. fasciata (Fig. 14C, D). In addition, in contrast to D. sp. ‘Borena’, D. albigularis sp. nov. has a single (versus 2) postocular scale.

The other species of this genus found in the Horn of Africa – D. abyssina, D. crucifera, D. scabra sensu lato, and D. taylori – belong to clades that are more distantly related to D. albigularis sp. nov. and, therefore, are not immediately comparable.

Etymology.

The specific Latin epithet ‘albigularis’ means ‘white-throated’ and refers to the pale ventral colouration on the lower parts of the head and anterior parts of the body of the new species.

Distribution, habitat, and ecology.

Dasypeltis albigularis sp. nov. is known only from the Harenna Forest (Fig. 9) at elevations of 1,500–1,700 m asl. This vegetation zone was classified by Friis (1986) as Podocarpus‒Filicium forest situated between 1,450 and 1,900 m asl. Bussmann (1997) described it as a submontane zone with a “Podocarpo falcati–Coffeetum arabicae” plant community. The new species may have a wider range in the forested areas of the southern Somalian Plateau, where similar habitat is available.

The habitat of D. albigularis sp. nov. at the type locality is dense deciduous forest with tall trees such as Afrocarpus, Celtis, Cassipourea, Croton, Ekebergia, Filicium, Syzygium, and Warburgia. The undergrowth is, to a large extent, composed of wild and semi-wild Coffea arabica. Dasypeltis albigularis sp. nov. was also found in villages and plantations, as well as at the margins of glades within the forest (Fig. 9). Day temperatures in the rainy season usually do not exceed 25 °C. In the dry season, hot, sunny days are frequent. The usual night temperatures throughout the year are 13–17 °C, but in the dry season these can drop to 10 °C.

Natural history.

Dasypeltis albigularis sp. nov. is a scansorial snake displaying a degree of arboreality similar to that of D. atra, D. fasciata, and D. gansi, i.e. adults occupy trees and descend to the ground only when searching for prey or a mating partner, as well as for egg deposition. Despite their arboreal way of life, and similarly to other species of this genus, D. albigularis sp. nov. are slow tree-climbers. When in danger, instead of climbing higher up on branches, they deliberately fall to the ground, where they attempt to escape from intruders.

The diet comprises bird eggs of suitable size, but relatively soft-shelled eggs of tree-nesting bird species are preferred. Large females consume prey up to the size of a small chicken egg. Once the snake has found a nest, it visits it several times until all eggs are eaten. Eggs are consumed at all developmental stages, including those containing an almost fully developed embryo. The annual activity cycles of this species is synchronised with the breeding cycles of birds and with the local climate. In the dry season, starting in December and usually ending in late March, D. albigularis sp. nov., like all other reptiles in this habitat, remains hidden.

The mating season is in summer, at which time males chase females and usually bite them to prevent them from moving and to allow time for alignment of their cloacae. Such behaviour was recently reported in D. atra (Tiutenko 2024) and is also known to occur in D. gansi and D. inornata according to unpublished reports by snake breeders. As in these species, the copulating male and female remain attached to each other for 11–18 h. Gestation lasts 3–4 months. Females first deposit eggs in October–November in wet soil, usually under a wooden log or similar object. To find a suitable oviposition site, females crawl on the ground for about 7–10 days before oviposition. The number of eggs in a single clutch is similar to that of D. atra and varies from 7 to 22, depending on the condition of the female and the number of clutches already produced after mating. The first clutch of the season is usually the largest and the last one the smallest. The size and shape of eggs in a single clutch varies quite substantially (Fig. 15). The interval between clutches is about 3 months, i.e. a female that has just laid eggs, if well fed, usually becomes gravid again soon after oviposition. Eggs hatch after about 3 months of incubation at temperatures of 22–23 °C. Dasypeltis atra eggs hatch in January–February, i.e. approximately 2 or 3 months earlier than those of D. albigularis sp. nov., which hatch at the beginning of the rainy season in April and May. This can be explained by the nesting season of birds, which begins earlier in equatorial East Africa (Stevenson and Fanshawe 2002). In D. atra, incubation at the same temperature takes about 20 days longer.

Figure 15. 

Size and shape variation in a single clutch of eggs of Dasypeltis albigularis sp. nov. deposited by ZSM 94/2019: 1) 4.58 g, 26.64 × 17.10 mm; 2) 4.42 g, 26.67 × 16.91 mm; 3) 2.72 g, 27.58 × 12.69 mm; 4) 4.86 g, 32.95 × 15.49 mm; 5) 3.39 g, 27.24 × 13.38 mm; 6) 3.07 g, 24.94 × 14.47 mm; 7) 4.50 g, 30.32 × 15.69 mm; 8) 4.00 g, 29.07 × 15.60 mm; 9) 3.49 g, 22.13 × 16.21 mm; 10) 3.16 g, 26.64 × 13.81 mm; 11) 4.54 g, 34.02 × 15.07 mm; 12) 1.97 g, 24.16 × 14.69 mm; 13) 4.18 g, 26.94 × 13.81 mm; 14) 2.43 g, 21.95 × 13.84 mm; 15) 3.38 g, 26.18 × 14.83 mm; 16) 3.73 g, 23.78 × 15.64 mm. All eggs except 3, 9, 10, and 11 hatched after 86–90 days of incubation. Egg 11 contained twins that died when they were about to hatch.

As usual in Dasypeltis, the hatching process takes several days. After cutting open the egg shell, the young snake remains in the eggs for up to a day and usually emerges at night. Hatchlings, often several individuals coiled together, hide on the ground in grass tussocks, under pieces of bark or logs, in leaf litter, or bury themselves in soil until their first ecdysis, i.e. for up to 2 weeks after hatching. After their first shedding, they move around, actively searching for bird eggs. Sexual maturity is reached at the age of about 3 years in males and at or before 2 years in females.

As in all Dasypeltis, the new species is crepuscular and nocturnal. In captivity, these snakes were observed to be most active from 17.00–18.00 until midnight. Sometimes, they appeared again at 06.00–08.00. In their natural habitat, nights may be too cold for them to remain active.

In May and early June, in the season of ‘small rains’, D. albigularis sp. nov. of all ages are the most frequently encountered snakes in the Harenna Forest. However, this species was not encountered in the main rainy season between July and December, although sympatric snakes of the genera Philothamnus and Pseudoboodon were common at such times. The reasons for the reduced activity of egg-eaters at such times may be related to the end of the breeding season for birds and the snakes’ habit of remaining high up in the branches of trees. In this season, D. albigularis sp. nov. spends most of its time in elevated hideouts. Captive snakes generally feed throughout the year, but adult males usually refuse food in the dry season (December–April).

This study showed that a uniform dorsal colour in this genus does not automatically imply a taxonomic singularity, i.e. that a plain-coloured Dasypeltis is not always of the same species as other such snakes with the same colour pattern. Given the generally weak meristic morphological characters in Dasypeltis, additional new species of dorsally plain-coloured egg-eaters can be expected to be discovered when molecular genetic methods are used. This may apply to the population in the western Ethiopian highlands (see Largen and Spawls 2010; Bates and Broadley 2018; Spawls et al. 2023) that is isolated by ecological and geomorphological barriers both from populations of D. atra and from D. albigularis sp. nov. (Fig. 7).

Gans (1959: 70) found that black D. atra – which he called “montane scabra” – in Uganda and Burundi were most common above 1,800 m asl. In D. atra, it has been assumed that colour pattern correlates with elevation, with black morphs being more common at higher elevations (e.g. Hughes 1997). All of the more than 70 D. atra (from parents from Chuka, Kenya) that hatched in captivity between January 2020 and March 2026 were dorsally black. Such dependence on high elevation was not observed in the new species, which was never collected above 1,700 by AT in 14 years of expeditions to the Harenna Forest. This elevation that is quite low compared with that of D. atra.

At lower elevations, outside the Harenna Forest, D. albigularis sp. nov. appears to be replaced by D. sp. ‘Borena’. The status of a population of Dasypeltis in Borena, Ethiopia (specimens from a locality “14 km south-east of Negele” – T. Mazuch pers. com.) that was previously treated as D. atra (e.g. Bates and Broadley 2018) requires investigation. This population may represent a distinct species of Dasypeltis, although more data on its distribution and evolutionary relationships are required.

In the cyt b phylogeny (Suppl. material 2), specimens TMHC 2023.01.1022 and TMHC 2015.09.650 from the area around Filtu, Ethiopia (34 km SE and 25 km W), as well as TMHC 2022.09.1012 (19 km south of Sheikh Hussein, Ethiopia), clustered with D. sp. ‘Somalia’ TMHC 2023.01.1023 (W of Dilla, Somaliland) and TMHC 2023.01.1019 (Daallo Forest, Erigavo, Somaliland). The p-distances between the Ethiopian and Somalian specimens (for example, a maximum of 0.85% in cyt b – see Suppl. material 3: table S3-2) are small despite their strong superficial dissimilarity. They may represent a single species conspecific with D. taylori of Somaliland and Djibouti (see Bates and Broadley 2018).

Several ‘scabra-like’ Dasypeltis populations in eastern and south-eastern Africa may not be conspecific with D. scabra sensu stricto. The analyses recovered two such lineages that are not immediately related to each other and may be recognised as separate new species in future research. One of these is represented in the data set by the two patterned snakes from Burundi with voucher numbers BEV 109891 and BEV 109892, sequenced for the paper by Trape et al. (2012), who identified them as “D. cf. lineolata”. Bates and Broadley (2018) suggested that they were either D. scabra sensu lato or D. atra, but more probably the latter. Unfortunately, only 16S sequences are available. In the 16S phylogeny (Suppl. material 2), these specimens clustered with a D. cf. scabra sequence (SNHM N 39422) from Kruger National Park in South Africa (Fig. 5).

Although the D. cf. lineolata lineage is not closely related to D. atra, it is nested in the same clade as D. atra, D. fasciata, D. gansi, and D. albigularis sp. nov. This latter clade is treated as the ‘D. atra group’. Its members are characterised by a high degree of arboreality, large body size, and broad and relatively smooth dorsal scales with apical pits (Fig. 14). In the concatenated phylogeny, combining nuclear and mitochondrial loci, the D. atra group is sister to the clade that includes D. confusa, D. sp. ‘Borena’, and D. sp. ‘Somalia’, represented by snakes with strongly keeled and sharply pointed dorsal scales (Fig. 14) that occupy drier habitats.

The authors thank Tomáš Mazuch for providing tissue samples from Ethiopia and Somaliland and photographs of specimens housed in his collection. TLH is grateful to Oliver Hawlitschek and the laboratory staff at the Leibniz Institute for Biodiversity Change (LIB) in Hamburg, Germany, for generous funding and access to infrastructure for laboratory procedures and DNA sequencing. MFB’s visit to the Royal Museum of Central Africa in Tervuren, Belgium, in October 2012 was facilitated by an African Biodiversity Information Centre scholarship, with guidance from Danny Meirte; he also thanks Nicolas Vidal and Ivan Ineich for their assistance during a visit to the Muséum national d’Histoire Naturelle in Paris, France, in September–October 2015; Silke Schweiger for her assistance during a visit to the Naturhistorisches Museum in Vienna, Austria, in September 2017; and Jeffrey Streicher and Patrick Campbell for their assistance during a visit to the Natural History Museum in London, United Kingdom, in August–September 2025. The curators of the Zoologisches Forschungsmuseum Alexander Koenig in Bonn, Germany, are thanked for the loan of D. atra specimens. AT thanks Abdulkadir M. Hussein for logistical support and guidance during fieldwork in Ethiopia, as well as for providing observation data. The authors thank Jean-François Trape for reviewing the manuscript.