PEDRO VAZ PINTO THESIS

“Sable subspecies

Contrasting with roan, sable taxonomy suggests a much better defined structure, making the published classifications, at least until recently, relatively unambiguous (Ansell 1972; Groves 1983; Estes & Kingdon 2013) (Fig. 1.18b). Following the discovery and description of H. niger HARRIS 1838 from South Africa, the first split led to the description of H. n. kirkii GRAY 1872, named after the famous explorer Sir John Kirk, who while exploring the Batoka hills was struck by how different sable looked on both sides of the Zambezi river (Harper 1940; Ansell 1974). Unlike the very dark females characteristic of the typical race, to the north of the Zambezi the females are reddish to chestnut-brown, while males tend to have darker faces (Harper 1940; Ansell 1972; Groves & Grubb 2011). Subsequently, a third subspecies was described from Kenya, H. n. roosevelti HELLER 1910, corresponding to the smallest body-sized sable, also carrying noticeably shorter horns, with a broader white muzzle line, and females being light to chestnut-brown (Groves & Grubb 2011). In 1912 another subspecies was described from the Caprivi Strip, Namibia, H. n. kaufmanni MATSCHIE 1912, characterized by relatively lighter-coloured females, but it has remained obscure and was eventually synonymized with H. n. niger following observations of individuals with intergrading features (Harper 1945; Ansell 1972, 1974).

Fig. 1.18 – Subspecies boundaries for (a) roan (Hippotragus equinus) following Ansell (1972) adapted to the species distribution range from IUCN (2008), and for (b) sable (H. niger), following Ansell (1972) and Ansell & Dowsett (1988), adapted to the species distribution range from IUCN (2008).

The year of 1916 marks the formal description of the giant sable H. n. variani in Angola, characterized by much larger horns and corresponding skull measurements, and by complete obliteration of white muzzle line in bulls (Thomas 1916; Blaine 1922). Female giant sables are golden to chocolate-brown, often with dark faces. Possibly even more than Roosvelt’s sable and on morphological grounds alone, the giant sable may be the most distinctive sable subspecies, as reflected by significant larger size not only of its enormous horns but also skull, and with teeth that often compare to roan measurements (Thomas 1916; Blaine 1922; Klein 1974; Groves & Grubb 2011). As result, not only Thomas (1916) hesitated, but actually Blaine (1922) proposed elevating the giant sable to specific status. However Blaine’s suggestion was not followed by most subsequent authors (but see Harper 1945), and a four sable subspecies classification would remain undisputed for almost 70 years. In 1983, a comparative morphological study on museum specimens resulted on the description of a fifth subspecies, H. n. anselli, corresponding to sable from Malawi and eastern Zambia (Groves 1983). Reportedly, this race was of average body size but with narrow skulls, with females of red-golden colour, and having a distinctive facial mask with very broad white muzzle line (Groves 1983). This study was based on a small sample size, and often H. n. anselli has been ignored or overlooked by subsequent authors, but it is nevertheless a geographically coherent population and is here considered (Fig. 1.18b).

If some subspecies boundaries seemed to have clear-cut geographical definition, others warranted clarification because of under-sampling of critical locations, leading to several “grey areas” of uncertain correspondence or possible intergradation (Ansell 1972, 1974; Estes & Kingdon 2013). Specifically, with H. n. variani isolated in an Angolan “island”, H. n. niger has been the only race claimed for southern Africa south of the Zambezi and H. n. roosevelti for Kenya, while H. n. kirkii occupies most of Zambia and extends into the Democratic Republic of Congo (Ansell 1972, 1978; Estes & Estes 1974; Estes 2013; Butynski et al. 2015).

Conversely sable populations from east Zambia and Malawi, from west Tanzania, east Tanzania and northern Mozambique, have been historically difficult to ascribe to subspecies. Regarding the geographic region of east Zambia and Malawi between the dry corridor of Muchinga/ Luangwa valley (Groves & Grubb 1999) and the southern extension of the EARS, Sweeney (1959) separated those sable from the rest of Zambia and attributed them to the typical race, but this statement was considered doubtful by Ansell (1972) who temporarily ascribed them to H. n. kirkii as in the rest of Zambia, before they were later proposed as H. n. anselli (Ansell 1972, 1974; Groves 1983; Ansell & Dowsett 1988; Groves & Grubb 2011). The status of Tanzanian sables has often been confusing, and more frequently referred to as H. n. kirkii or H. n. roosevelti (Swynnerton & Hayman 1951). However there seems to be two distinct populations in Tanzania, separated by the Somali arid corridor that runs along the rain shadow of the EAM (Ansell 1972; Grubb 1999). The west Tanzanian sables have been relatively less studied and are poorly represented in museum collections, and were suggested to be H. n. kirkii simply by exclusion of other races (Ansell 1972). Eastern Tanzania sable have been more often classified as H. n. roosevelti, while northern Mozambican sable were occasionally assigned to H. n. kirkii or H. n. niger, but without clear justification. One key factor is determining the southern extension of H. n. roosevelti, and although it has been suggested a southern limit defined by river Rufiji (Estes 2013), mounting evidence suggest a continuum distribution and phenotypical gradation of H. n. roosevelti along the coastal eastern regions southwards and into northern Mozambique (Siege & Baldus 1999; Booth 2002), although an intergradation with H. n. anselli has also been suggested (Groves 1983).

Sample collection, DNA extraction and library preparation

A total of 233 modern and 33 museum samples were collected from 22 populations covering the whole geographic range of H. niger in Africa (Fig. 1 and supplementary Tables S1, S2). Four Hippotragus equinus were included to outgroup H. niger in data analysis.

The full set of 262 H. niger and four H. equinus complete mitochondrial genomes were aligned using MUSCLE v3.8.31 (Edgar 2004) and Mafft v.7.017b (Katoh & Standley 2013).”