African Journal of , 2004 53(1): 35-61.

Original article The of southeast Katanga, an overlooked ‘hot spot’

DONALD G. BROADLEY1 AND F.P.D. COTTERILL2

1Research Associate, Natural History Museum of , Bulawayo Present address: Biodiversity Foundation for , P.O. Box FM 730, Bulawayo, Zimbabwe [email protected]

2Evolutionary Genomics Group, Department of Zoology, Natural Sciences Building, University of Stellenbosch, Private Bag X1, Matieland 7602, and Principal Curator of Vertebrates, Department of Mammalogy Natural History Museum of Zimbabwe, P O Box 240, Bulawayo, Zimbabwe Correspondence: Biodiversity Foundation for Africa, Secretariat: P O Box FM 730, Famona, Bulawayo, Zimbabwe [email protected]

Abstract.—Areview of the reptiles recorded from southeast in the Democratic Republic of Congo, employing the evolutionary concept, indicates that there are at least 119 species in the area (7 chelonians, 37 , 4 amphisbaenians, 69 and 2 crocodilians), 15 of which are endemic (12.6 %). Analysis of zoogeographical affinities shows that this fauna is derived from all directions, with forest forms entering from the north accounting for 14.4 % and savan- na forms entering from the east contributing 22 %.

Key words.—Reptiles, Katanga, Upemba, Kundelungu, , zoogeography.

chouteden and de Witte collected herpeto- All surveys of Katanga’s biodiversity have Slogical material in Katanga during 1924-26 been sporadic and patchy. By far the greatest and 1930-31 (Witte 1933) and de Witte effort has been devoted to Upemba National reviewed the reptiles of Katanga in his 1953 Park and its environs. Nevertheless, despite study of the reptiles of the Upemba National these gaps in knowledge, the overall patterns of Park (P.N.U.), while Laurent (1950, 1955) species richness, faunal affiliations and described several new taxa from the endemism exhibited by the reptiles are suffi- Kundelungu plateau, which is still poorly cient for us to make an up to date assessment of known. No major collections have been made zoogeographic patterns, and to determine the in this area subsequently. Some specimens factors that account for the evolution of this from the P.N.U. (including paratypes of new rich fauna. These patterns are complemented taxa) were transferred to the Umtali Museum by other vertebrates (Cotterill in press a,b; sub- collection (now in the Natural History Museum mitted). Together, this combined knowledge of of Zimbabwe in Bulawayo) in exchanges of Katanga emphasizes the interesting evolution- material during 1965-6 and additional material ary history of the region. We re-emphasize the was borrowed from Brussels and Tervuren for importance and further scope for research and revisionary studies. conservation of the biodiversity in Katanga.

35 AFRICAN JOURNAL OF HERPETOLOGY 53(1) 2004

STUDY AREA varies from riparian and flooded to montane (Witte 1966; Lubini 1994a,b). Centred on these Extent, Physiography, and .—The plateaux, this southeast corner of the Congo approximate limits of the area under considera- basin has an especially high endemism of tion are the Lualaba valley in the west, the savanna plant species (Colyn 1991; Linder in the north, and the 2001). Furthermore, species richness of plants in the east and the Congo- is the highest in the miombo ecoregion; and watershed in the south (Fig. 1). The Katanga lies within a centre of high species geomorphology of the area has been described richness (> 3000 species) of the Afrotropical by Dapper (1981) and the geology mapped by flora (Malaisse 1997; Burgess et al. 2004). The Aderca (1966). The principal formations are evergreen gallery forests (mushitu) confined Precambrian quartzites, metamorphosed sedi- along wetlands often grow within and are bor- ments, and sedimentary Karoo formations. dered by dambos. Mushitu are present across Extensive shallow alluviums represent recent large gradients in altitude, from valley floors to episodes of deposition in shallow lakes, tops of ravines abutting the highest plateaux notably in the Lufira, Lower Luapula and (Witte 1966; Malaisse 1997). Lualaba valleys (Fig. 2). The high flat plateaux are major landforms in Three main areas of relief dominate the phys- Katanga. Their greater extents are covered with iography of Katanga in the southeast Congo dilungu, shallow layers of fine loose sands. basin. The largest are the Bia, Kibara and They have been reworked from Kalahari sands Kundelungu plateaux, bordered by deep of Neogene and Plio-Pleistocene age. The finer ravines and wide depressions eroded by a structure of the landforms within dilungu com- northerly flowing drainage system. The high, prises a varied and extensive micro-relief, flat plateaux, notably Kibara, rise to altitudes which vary across a range of spatial and tem- over 1890 m above the Upemba swamps and poral scales. The largest are of fossil aeolian floodplains. The latter lie at approximately origin (probably Late Pleistocene), and more 680m asl (Witte 1966; Dapper 1981). The hills, recent micro-topography of shallow dune mountains and ravines and steep margins of the ridges and pans. Dilungu support a vegetation plateaux provide rocky habitats for rupicolous of short grasses and fruit-bearing shrubs reptiles, especially along fault scarps. (including Parinari capensis spp. latifolia, and Syzygium guineense spp. huillense (Dapper This high physiographic diversity is expressed 1988). These sandy habitats appear to be in a correspondingly rich flora. The Upemba- important for small vertebrates, including the Kundelungu area lies in the Katanga-Bemba herpetofauna and certain small and region, the northern subregion of the . The most extensive dilingu cover the Zambezian Phytochorion as defined by Kundelungu and Kibara plateaux, with much Malaisse (1997). In addition to its many smaller patches on hilltops in the southwest, endemic plants, the flora is enriched by both where they reach their southern limit at Kalene Guineo-Congolian and Zambezian elements Hill in the Ikelenge Pedicle of northwest (White & Werger 1978). This diverse flora is (Fig. 1). represented in a corresponding variety of vege- tation. It is principally mesic miombo savanna, The variety of wetlands in Katanga equals that but there are also important patches of forest, of terrestrial habitats. This complexity includes savanna woodlands, and alluvial and plateau upland ravines and streams, large rivers, both grasslands. The diversity of forest patches sedate and fast flowing, and huge swamps in

36 BROADLEY & COTTERILL — Reptiles of southeast Katanga

Figure 1. Protected areas and principal land forms in Katanga, Democratic Republic of Congo (DRC). Localities mentioned in the text are shown.

37 AFRICAN JOURNAL OF HERPETOLOGY 53(1) 2004

Figure 2. Map of south- depicting deposits of recent alluvium in the context of modern drainage sys- tems. As explained in detail by Cotterill (2003), the original headwaters of the Palaeo- were the modern Chambeshi River; when this drainage system terminated in Palaeo-Lake Makgadikgadi in the Kalahari basin. The proposed link between the Palaeo-Chambeshi and Upemba was across Palaeo-Lake Lufira through the channel of the Palaeo-Kafila River. The arrowheads depict predatory headwaters of the Luapula, and mid-Zambezi Rivers that broke up the Palaeo-Chambeshi in the vicinity of southeast Bangweulu, Kafue Flats and Victoria Falls. (Modified from Cotterill in press a)

38 BROADLEY & COTTERILL — Reptiles of southeast Katanga and around large shallow lakes. Saline and 1956), decades before protected areas were thermal springs also occur. Dambos are impor- proclaimed. The first to be declared a protected tant, which although individually of small area, area was Upemba National Park (P.N.U.) in collectively cover a vast area along tributaries 1939 (Symoens 1963); and its boundaries were and at their headwaters. Three of Africa’s later extended to include the lakes and associ- largest wetlands have formed in Katanga. ated wetlands along the Lualaba. Kundelungu These include Lake Mweru (and the associated was only proclaimed in 1970, and is also com- Luapula Swamps to its south), Lake Lufira plemented by a large annex to its north and (and its associated mosaic of dambos down- west (Demey & Louette 2001, Fig. 1). The stream) along the upper , and integrity and future of Katanga’s protected Upemba Swamps in the Kamalondo depression areas (especially P.N.U.) has recently precipi- (Malaisse 1997). The latter is a large mosaic of tated grave concern. Large popula- exhoreic deflation lakes situated within exten- tions were decimated by commercial poaching sive swamps of Phragmites and Papyrus, sur- in the 1980s (Wolanksi 1998; Hasson & rounded in turn by floodplains (Banister & Wolanski 1999). Fund raising and support by Bailey 1979). the conservation NGO Nouvelle Approaches has alleviated some problems in Kundelungu Both Mweru and Upemba lie in rift valleys - and Upemba, but much work remains fed by the Luapula and Upper Lualaba rivers (Wolanski 1998; Hasson & Wolanski 1999, respectively (Dapper 1981, Fig. 1). Lake http://www.nouvellesapproches.org). Mweru is shallow (average depth 10 m). Its levels appear to have fluctuated greatly in recent geological times; as indicated by exten- ANNOTATED CHECKLIST OF THE sive recent alluvium to its south that nearly REPTILES OF SOUTHEAST KATANGA equals the area of the modern lake. Mweru’s major tributary, the Luapula River, drains the ORDER TESTUDINES , which in turn are fed by Family PELOMEDUSIDAE the Chambeshi. In contrast to the valleys, cool, fast flowing streams drain the plateaux, frac- Pelomedusa subrufa (Bonnaterre 1789).—The tured by many rapids and waterfalls (Witte Helmeted Terrapin is found throughout the dry 1966; Banister & Bailey 1979). The 342 m savannas of sub-Saharan Africa, but is absent drop of the Loloi Falls, flowing into the Lufira from the Kibara plateau in the east of the depression off the western Kundulungu, is one P.N.U. (Witte 1953). of the highest in Africa (Demey & Louette 2001). The wetlands of Katanga form a distinct Pelusios gabonensis (Duméril 1856).—The ichthyofaunal subprovince within the overall Forest Hinged Terrapin has a wide distribution Congo basin, separated at the Portes d'Enfer in the Congo basin and was recorded from (Gates of Hell) from the lower Nyonga on the western shore of Lake Upemba (Fig. 1). Its ichthyofaunal affinities are a mix- in 1925 (Witte 1933). ture of Congolian and Zambezian species with a significant proportion of endemics (Teugels Pelusios nanus Laurent 1956.—The Dwarf & Guégan 1994; Malaisse 1997). Hinged Terrapin occurs from central along the Congo-Zambezi watershed into Protected Areas and Conservation.—The con- northern Zambia along the Chambeshi River servation significance of Katanga’s landscapes (Broadley 1981). It has been recorded from was noted early in the 20th century (Sharp Kanzenze and (Witte 1953 [as P.

39 AFRICAN JOURNAL OF HERPETOLOGY 53(1) 2004 subniger]; Laurent 1956), but not yet from the ORDER SAURIA P.N.U. Family

Pelusios subniger (Bonnaterre 1789).—The Acanthocercus atricollis (A. Smith 1849).— Pan Hinged Terrapin has a wide range in south- The Southern Tree has a wide range in east and central Africa and is common in the eastern Africa and Madagascar, extending west P.N.U. and environs (Witte 1953). to the Okavango delta and Katanga and north to (Bour 1983). It has been recorded Agama armata Peters 1854.—The Tropical from Kikondja on Lake Kisale and Kando, near Spiny Agama has a wide range in east and cen- Tenke (Witte 1933 [as Sternothaerus der- tral Africa. It occurs throughout Katanga and in bianus]). the P.N.U. it is mainly found on the Kibara plateau at 1300 to 1800 m (Witte 1953 [as A. Pelusios upembae Broadley 1981.—The hispida aculeata]). Upemba Hinged Terrapin is endemic to the swamps in the Kamalondo basin, where it is Family CHAMAELEONIDAE sympatric with P. subniger and P. rhodesianus. Chamaeleo gracilis Hallowell 1842.—This It is the sister species of P. bechuanicus savanna species has an enormous range in west FitzSimons of the Okavango delta, upper and central Africa and is found throughout the Zambezi and central Angola (Broadley 1981). P.N.U. (Witte 1953).

Pelusios rhodesianus Hewitt 1927.—Described Chamaeleo dilepis Leach 1819.—The Flap- from southeast of , the necked Chameleon has a wide range in the Zambian Hinged Terrapin has a wide range in savannas of east and central Africa and is com- central Africa, with relict populations in eastern mon throughout the P.N.U. at lower altitudes South Africa (Broadley 1981; Bour 1983). (Witte 1953, who also listed C. quilensis Some specimens were collected on the Bocage, considering it a distinct species). Kundelungu Plateau. Many specimens from southeast Katanga in the Tervuren collection Chamaeleo anchietae Bocage 1872.—This ter- have been examined, but all the P.N.U. materi- restrial chameleon has a discontinuous distrib- al listed by Witte (1953) as P. subniger needs to ution in montane grasslands from southwestern be sorted, as it includes all four species listed Angola to southwestern . The syn- onym C. a. vinckei Laurent was described from above. the Kundelungu plateau and is common on the Family TESTUDINIDAE Kibara plateau in the P.N.U. (Witte 1953).

Kinixys spekii Gray 1863.—Speke’s Hinged Family GEKKONIDAE Tortoise has a wide range on the central plateau Pachydactylus (Elasmodactylus) tuberculosus from to South Africa and west to (Boulenger 1894).—The Tuberculate Thick- Katanga (Broadley 1993). It appears to be toed ranges through savannas of central replaced by K. belliana Gray further west in the Africa from northeast Tanzania to the lower D.R.C. and Angola. In the P.N.U. it occurs at Congo region. In the P.N.U. and environs it is lower altitudes, but was not found on the common below an altitude of 1000 m (Witte Kibara plateau (Witte 1953 [as K. b. belliana]). 1953). The synonym Elasmodactylus triedus

40 BROADLEY & COTTERILL — Reptiles of southeast Katanga

Boulenger was described from Kikondja in the Albertine Rift to the P.N.U., which is its south- Upemba rift. ern limit (Pasteur 1964). It does not occur on the Kibara plateau (Witte 1953). Pachydactylus katanganus Witte 1953.—This small gecko is endemic to the P.N.U., where it Hemidactylus mabouia (Jonnès 1818).—The occurs from Mabwe (585 m) to Lusinga at Tropical House Gecko seems to have originat- 1810 m on the eastern boundary of the park ed in , but is rapidly expanding its (Witte 1953). It was described as a distribution westwards and southwards due to of P. capensis, then placed as a northern race of accidental translocation by man. It is uncom- P. oshaughnessyi (Broadley 1977b) and recent- mon in the P.N.U. and was not found on the ly recognised as an evolutionary species Kibara plateau (Witte 1953). (Broadley 2003). Family SCINCIDAE Pachydactylus punctatus Peters 1854.— Subfamily SCINCINAE Although the Speckled Thick-toed Gecko Sepsina hemptinnei Witte 1933.—This fossori- ranges from west to , al is widespread in Katanga and has been there are few records north of the Zambezi. It is recorded from near Ujiji on the eastern side of common in the P.N.U. below 1300 m (Witte (Loveridge 1942). The 1953), but this population is well isolated from species was described from the Kundelungu the nearest records in northern and plateau and is found throughout the P.N.U. It southern Zambia (Broadley 2003). was placed as a subspecies of S. tetradactyla Peters by Loveridge (1942), but it is a more Lygodactylus heeneni Witte 1933.—This rare gracile form with more rudimentary digits and arboreal dwarf gecko was described from a bronze rather than bluish tail. Kapiri to the south of the P.N.U., and is almost endemic to southeast Katanga, but it has also Subfamily been recorded from the northern Mwinilunga District in Zambia (Broadley 1991b). In the The Afro-Malagasy , formerly included P.N.U. it only occurs on the Kibara plateau in the New World Mabuya Fitzinger, are between 700 and 1750 m (Witte 1953). This now correctly assigned to the revived genus form was described as a subspecies of L. angu- Trachylepis Fitzinger (Bauer 2003). laris Günther, common on buildings on the Zambian copperbelt, but there is no indication Trachylepis maculilabris (Gray 1845).—The of intergradation in throat pattern and they are Speckle-lipped Skink has a wide range in forest best regarded as sister species. and savanna through west, central and east Africa as far south as central Mozambique. It is Lygodactylus angolensis Bocage 1896.—The common throughout the P.N.U. (Witte 1953). Angolan Dwarf Gecko has a wide range in south-central Africa from Angola to central Trachylepis upembae Witte 1953.—This skink Tanzania. It is common in the P.N.U. except on is endemic to the P.N.U. It was described as a the Kibara plateau (Witte 1953). subspecies of T. perroteti (Duméril & Bibron) of the Sudanese savanna, distinguished by a Lygodactylus gutturalis (Bocage 1873).—This higher number of midbody scale rows (32-36) is the most widely distributed savanna species and colour pattern. In fact T. upembae agrees in the L. picturatus complex, ranging from with T. perroteti in the former character, but east to the and south via the seems to be more closely related to T. plani-

41 AFRICAN JOURNAL OF HERPETOLOGY 53(1) 2004 frons (Peters) of East Africa, which occurs at Eumecia anchietae Bocage 1870.—Anchieta’s Nyamkolo at the southern end of Lake Serpentiform Skink with reduced digits has an Tanganyika (Loveridge 1933, 1957) and has extensive range from southwest Angola 26-32 midbody scale rows. Trachylepis per- through Zambia and the D.R.C. to western roteti is a large robust skink with a relatively Kenya and Tanzania. It occurs throughout the short tail (ca.160 to 180% of snout-vent P.N.U., but is more common on the Kibara length). Trachylepis planifrons is a more slen- plateau (Witte 1953). der species with a long tail (ca 180 to 240% of Leptosiaphos dewittei (Loveridge 1934).— SVL). Trachylepis upembae (two paratypes This slender skink ranges from western Angola examined) is intermediate in build, but agrees to the P.N.U., where it occurs throughout, but is with T. perroteti in tail length; its colour pattern more common in the plateau areas. It is a sister differs from both. species of L. kilimensis Stejneger of Kenya and Tanzania, Witte (1953) regarded it as a syn- Trachylepis pulcherrima Witte 1953.—This onym (Broadley 1989). large rupicolous skink is endemic to the P.N.U. It was described as a subspecies of T. quinque- Panaspis seydeli (Witte 1933).—Seydel’s taeniata (Lichtenstein), but was recognised as a -eyed Skink is only known from four full species by Broadley & Bauer (1998). specimens, which have the prefrontals in broad contact: the type of P. seydeli from Trachylepis megalura (Peters 1878).—The Lubumbashi, the type of Ablepharus moeruen- slender smooth-scaled Grass-top Skink ranges sis (Witte) from Kilwa on Lake Mweru, the from south to central Mozambique type of A. anselli FitzSimons from Kasempa, and west to the Albertine rift. In the P.N.U., it northwest Zambia, and BMNH 1940.1.19.34 occurs on the Kibara plateau above 1300 m from Mwenzo, northeast Zambia. Fuhn (1964, (Witte 1953). 1970) placed A. anselli in the synonymy of Ablepharus tancredii Boulenger of Ethiopia Trachylepis varia (Peters 1878).—The Variable (known only from the type) and consequently Skink ranges from the Sudan to South Africa all three southern taxa were provisionally and west to Namibia. It is common throughout assigned to P. tancredii by Jacobsen & the P.N.U. and environs (Witte 1953). Broadley (2000), but on zoogeographical grounds it seems better to separate them from Trachylepis striata (Peters 1844).—The P. tancredii under the earliest name. There is some variation; the type of A. moeruensis has Striped Skink ranges from Ethiopia to South three supraoculars instead of two and five labi- Africa and west through northern Zambia to the als anterior to the subocular instead of four. The P.N.U., where it is common throughout (Witte Mwenzo specimen agrees with the type of P. 1953). seydeli in having 20 midbody scale rows, the others have 22. Lygosoma afrum (Peters 1854).—The Mozambique Writhing Skink has a wide range Panaspis smithii (Witte 1936).—The Upemba from the Sudan to central Mozambique Snake-eyed Skink is probably endemic to the (Broadley 1966a), and extends west through P.N.U. and environs at low altitudes (Witte northern Zambia to the P.N.U., where it occurs 1953). It is readily distinguished from P. sey- on the edge of the Kibara plateau (Witte 1953 deli by having the prefrontals separated or [as Riopa sundevalli]). meeting at a point and having 24-26 midbody

42 BROADLEY & COTTERILL — Reptiles of southeast Katanga scale rows. It also resembles P. wahlbergii in Family GERRHOSAURIDAE having white lateral stripes. In his list of mate- rial, Witte (1936) included MRAC 6774 from Gerrhosaurus nigrolineatus Hallowell 1857.— Lodja, in northern Kasai province 600 km The taxonomy of the Black-lined Plated northwest of the P.N.U., but in view of the wide is still in a state of flux, but it has a wide range gap, this specimen should be re-examined. in south-central Africa from coast to coast. It is found throughout the P.N.U., but is less com- Panaspis wahlbergii (A. Smith 1849).— mon on the Kibara plateau than at lower alti- Wahlberg’s Snake-eyed Skink is distinguished tudes (Witte 1953). from the previous two by having the fron- toparietals fused. It has a wide range in savan- Gerrhosaurus bulsi Laurent 1954.—This large nas from Ethiopia south to South Africa grey plated lizard ranges from northeast Angola (Jacobsen & Broadley 2000) and reaches its through northwest Zambia (G. nigrolineatus western limits in the P.N.U., where it is com- anselli Broadley is a synonym) to southern mon throughout (Witte 1953). Katanga. In the P.N.U. it is sympatric with G. nigrolineatus at lower altitudes (Witte 1953 [as Subfamily FEYLININAE G. auritus]). This species can be distinguished Feylinia currori Gray 1845.—The Forest from both G. nigrolineatus and G. auritus by its Limbless Skink ranges from strongly spinose proximal supracaudal scales. through the Congo basin to its southeastern It has been erroneously treated as a synonym of limit in the P.N.U. (Brygoo & Roux-Estève the poorly known G. multilineatus Bocage of 1983), where it occurs below 1300 m (Witte western Angola (Broadley 1971d). The status 1953). of the latter is still under investigation.

Family CORDYLIDAE Tetradactylus ellenbergeri (Angel 1922).—The Zambian Whip Lizard ranges from eastern Cordylus tropidosternum (Cope 1869).—The Angola through Zambia and southern Katanga Tropical Girdled Lizard ranges from southeast- to southeastern Tanzania. Tetradactylus ern Kenya south to central Mozambique, west boulengeri Witte (type locality: Kansenia) and through northern Zambia to the P.N.U. T. fitzsimonsi simplex Laurent (type locality: (Broadley & Branch 2002). It is common on the Kibara plateau, but rare at lower altitudes, Kundelungu) are considered synonyms and is usually found in rocky places (Witte (Broadley 1971d). The four specimens from the 1953). P.N.U. were all collected on the Kibara plateau at an altitude of 1750 m (Witte 1953 [as T. Chamaesaura miopropus Boulenger 1894.— fitzsimonsi boulengeri]). The Zambian Snake Lizard inhabits grasslands from Angola east to northern Malawi and Family LACERTIDAE southern Tanzania. The only records from the D.R.C. are from on Lake Mweru and the Adolfus africanus (Boulenger 1906).—The P.N.U., where it is only found on the Kibara Multi-scaled Forest Lizard ranges through cen- plateau (Witte 1953). Although it has previous- tral Africa from Cameroon to northwestern ly been treated as a subspecies of C. Zambia (Broadley 1991b) and Katanga. The macrolepis, which lacks forelimbs (Broadley specimens from the P.N.U. were collected at 1971d), the latter does not occur north of the altitudes between 950 and 1750 m (Witte Chimanimani Mountains in eastern Zimbabwe. 1953).

43 AFRICAN JOURNAL OF HERPETOLOGY 53(1) 2004

Holaspis laevis Werner 1895.—The Eastern east to the Mweru Wantipa in northern Zambia Serrate-toed Tree Lizard inhabits coastal (Broadley & Broadley 1997). In the P.N.U. this forests and miombo woodland from northeast species is very common at low altitudes, but is Tanzania south to central Mozambique not found on the Kibara plateau (Witte 1953 [as (Broadley 2000). There are no records from Amphisbaenia quadrifrons capensis Thominot, Zambia, but it is recorded from Lubumbashi also including A. q. katangae Witte & Laurent, (Bourgeois 1964) and in the P.N.U. at altitudes both are synonyms]). between 890 and 1140 m (Witte 1953 [as H. guentheri]). The sister species H. guentheri Monopeltis remaclei Witte 1933.—This large ranges from through the northern spade-snouted amphisbaenian was described D.R.C. to and south to northeastern from Lukulu, just south of the Lualaba-Luvua Angola (Laurent 1964). confluence. It is otherwise only known from Kabalo, 80 km downstream from the conflu- Ichnotropis bivittata Bocage 1866.—The ence, and Nyunza 120 km east of Kabalo. It has Angolan Rough-scaled Lizard ranges from not yet been found in the P.N.U. Angola east through southern Katanga, north- ern Zambia and northern Malawi to southern Monopeltis adercae Witte 1953.—This large Tanzania. In the P.N.U. it inhabits the Kibara species was described from Bukena on the plateau (Witte 1953). Luapula River just north of the P.N.U. and was subsequently recorded from Mamono, 80 km to Ichnotropis capensis (A. Smith 1838).—The the northeast (Broadley et al. 1976), so it is Cape Rough-scaled Lizard has a wide range in endemic to the area under consideration. south-central Africa. In the P.N.U. it occurs at low altitudes, 585 - 1050 m (Witte 1953 [as I. Monopeltis scalper (Günther 1876).—This longipes Boulenger, a synonym]). moderate-sized species is endemic to Katanga, as the type localities for M. scalper (Kilembo) Family VARANIDAE and its synonym M. gerardi Boulenger Varanus albigularis angolensis Schmidt (Kikondja) are in the Upemba rift, and M. 1933.—This subspecies inhabits Angola, scalper bulsi Laurent (Kamina) on the western northern Zambia and southern Katanga. In the rim (Broadley et al. 1976). The specimens from P.N.U. it is common at low altitudes, but does the P.N.U. were taken between 585 and 890 m not occur on the Kibara plateau (Witte 1953). (Witte 1953).

Varanus niloticus (Linnaeus 1766).—The Nile SUBORDER SERPENTES Monitor occurs throughout savannas of sub- Family Saharan Africa. In the P.N.U. it was not found angolensis (Bocage 1866).— above 1000 m (Witte 1953). The Angolan Blind Snake occurs in both forest SUBORDER AMPHISBAENIA and savanna from Cameroon south through the Family AMPHISBAENIDAE Congo basin to northern Angola, apparently reaching its southeastern limit at Kasenga on Zygaspis quadrifrons (Peters 1862).—The the Luapula River south of Lake Mweru Kalahari Round-headed Worm Lizard has a (Roux-Estève 1974). The Zambian specimens huge range in southern Africa centred on the that she recorded from Abercorn (= Mbala) are Kalahari, reaching the northern limits of its more likely to have been collected by H.J. range along a line extending from the P.N.U. Bredo in the Mweru Wantipa on the border of

44 BROADLEY & COTTERILL — Reptiles of southeast Katanga

Katanga. This species and the next two were Lake Mweru. The genus Cope will transferred from to Rhinotyphlops by be revived for the slender blind species includ- Broadley & Wallach (2000). ed in the Rhinotyphlops Group VI of Roux- Estève (1974) (Broadley & Wallach, in prep.). Rhinotyphlops lineolatus (Jan 1864).—The Lineolate Blind Snake has a wide distribution Rhinotyphlops (Letheobia) kibarae (Witte through the savannas of sub-Saharan Africa 1953).—This species is endemic to the P.N.U., from Senegal east to Kenya and Tanzania and but one of the paratypes from Kaswabilenga south to northern Angola and Katanga (Roux- (type locality) at 700 m is actually a R. gracilis, Estève 1974). The specimens from the P.N.U. establishing sympatry between these two were taken at altitudes between 1140 and 1700 species (Roux-Estève 1974). One paratype is m (Witte 1953 [as T. punctatus (Leach)]). from the Pelenge gorge (1250 m).

Rhinotyphlops schmidti (Laurent 1956).—This Family LEPTOTYPHLOPIDAE species has a wide range in Katanga, extending into northeast Angola (Laurent 1964) and Leptotyphlops kafubi (Boulenger 1919).—The northern Zambia (Broadley 1971d). Some of Katanga Worm Snake was described from the specimens from the P.N.U. assigned to Lubumbashi and has a limited range in south- Typhlops p. punctatus by Witte (1953) are ern Katanga, extending into northwestern referable to R. schmidti (Laurent 1956). Roux- Zambia (Broadley & Broadley 1999). In the Estève (1974) placed R. schmidti (midbody P.N.U. it is common below 1000 m, but does scale rows 22-26) in the synonymy of R. lineo- not occur on the Kibara plateau (Witte 1953 [as latus (m.s.r. 28 - 32, rarely 26); the two species L. emini emini Boulenger of East Africa]). may be sympatric (Laurent 1956). Family PYTHONIDAE Rhinotyphlops mucruso (Peters 1854).—The Python natalensis A. Smith 1840.—The Zambezi Blind Snake is the largest in the world Southern African Python ranges across south- (maximum length 950 mm), ranging from central Africa from Angola southeast to South coastal Kenya south to Zimbabwe and central Africa and north to the southeastern D.R.C., Mozambique and west through the southern Burundi, Tanzania and the Kenya highlands provinces of the D.R.C. to northeast Angola (Broadley 1984). In the P.N.U. it was not (Roux-Estève 1974 [as R. schlegelii dinga]). In encountered on the Kibara plateau (Witte 1953 the P.N.U. it does not seem to occur above an [as P. sebae]), but Laurent (1956) recorded a altitude of 1300 m (Witte 1953). specimen from the Kundelungu plateau at 1750 m [as P. sebae]. This taxon was originally Rhinotyphlops (Letheobia) gracilis (Sternfeld revived as a subspecies of the Northern African 1910).—The Gracile Blind Snake has a Python P. sebae (Gmelin), but was reinstated as restricted range extending from Kitungulu in a full species when it became evident that the southwest Tanzania (type locality) through two forms overlapped in distribution by 900 northern Zambia to Katanga east of the km (Broadley 1999). Luapula River (Roux-Estève 1974). Witte (1953) assigned his material to Typhlops lepto- Family soma Witte 1933, a synonym described from Subfamily CAUSINAE Lukafu at the foot of the Kundelungu plateau, while another synonym, Typhlops katangensis lichtensteinii (Jan 1859).—Although Witte, had been described from Lukonzolwa on the Forest Night Adder has not yet been record-

45 AFRICAN JOURNAL OF HERPETOLOGY 53(1) 2004 ed from the P.N.U., there are specimens from Family ATRACTASPIDIDAE Lubumbashi and Kiambi on the Luvua River (Witte 1953). To the southwest it has been Atractaspis leleupi Laurent 1950.—This stilet- recorded from Ikelenge in northwest Zambia to snake seems to be endemic to the (Broadley 1991b). Kundelungu plateau at 1750 m and is only known from two specimens. These differ from Causus bilineatus Boulenger 1905.—The the western A. congica Peters [Angola to Lined Night Adder inhabits plateau areas of Cameroon, east to western Katanga (below Angola and northwest Zambia, but occurs at 1000 m) and northern Zambia (1300 m at lower altitudes in Katanga. Laurent (1956) Mbala) - Broadley 1971d, 1991b] by having 17 described C. Iineatus from the Kundelungu midbody scale rows instead of 19-21, so they plateau and also recorded it from . do not fit into a cline. Subsequently this was treated as a dwarfed Atractaspis bibronii A. Smith 1849.—Bibron’s subspecies of Causus bilineatus with lower Stiletto Snake has a wide range in savannas of ventral counts (Laurent 1964), but these differ East Africa from to South Africa, west little from those of the type series from to eastern Katanga, Angola and Namibia “Benguela to Bihe”, Angola. (Broadley 1991a). In the P.N.U., which is on the northwestern periphery of the species Causus rhombeatus (Lichtenstein 1823).—The range, this species was not found above 1300 Rhombic Night Adder has an extensive range m (Witte 1953). in moist savannas of south, central and east Africa, extending as far west as eastern (Bocage 1873).—The (Hughes 1978). It is common throughout the Common Purple-glossed Snake ranges from P.N.U. coastal Kenya to South Africa, west to Angola and northern Namibia (Broadley 1971b). The Subfamily VIPERINAE two specimens from the P.N.U. (the northwest- ern limit of the species range in the D.R.C.) Bitis arietans arietans (Merrem 1820).—The came from below 1000 m (Witte 1953). Puff Adder occurs throughout the savannas of sub-Saharan Africa and is found in all parts of Amblyodipsas rodhaini Witte 1930.—This the P.N.U. (Witte 1953 [as B. lachesis lach- species is only known from four specimens, the esis]). only precise localities being Kikondja, 35 km north of Lake Upemba, and Kamina on the Bitis gabonica (Duméril & Bibron).—The western edge of the Upemba rift valley. The Gaboon Viper is found in areas of forest-savan- latter is type locality of Calamelaps ventrimac- na mosaic throughout central and southeast. It ulatus katangensis Witte 1951, which seems to is found throughout the P.N.U. up to an altitude be a juvenile of A. rodhaini (Broadley 1971b). of 1810 m at Lusinga (Witte 1953). This snake may be endemic to Katanga.

Atheris katangensis Witte 1953.—The Upemba Amblyodipsas katangensis Witte & Laurent Bush Viper is endemic to the P.N.U., where it 1942.—This small species is only known from was collected at altitudes between 1250 and five specimens. The type series came from 1480 m (Witte 1953). It appears to be related to N’Gayu at the eastern end of the pedicle south A. rungweensis, which occurs 400 km to the of Lake Bangweulu. Witte collected two more east in northeastern Zambia (Broadley 1998). at Mabwe on the eastern margin of Lake

46 BROADLEY & COTTERILL — Reptiles of southeast Katanga

Upemba (585 m) and there is one Zambian the P.N.U. (Witte 1953 [as A. c. punctatolinea- specimen from 14 km northwest of Ndola tus Boulenger, a synonym]). (Broadley 1971b), so the species is a near- endemic for Katanga. Amblyodipsas katangen- Family sis ionidesi Loveridge of southeast Tanzania Elapsoidea guentheri Bocage 1866.— should probably be regarded as a sister species. Guenther’s Garter Snake ranges from the lower Xenocalamus michelli L. Müller 1911.—This Congo, south to Angola and east to Katanga, large quill-snout is only known from three western Zambia and Zimbabwe (Broadley specimens, one of which came from Bukena in 1971a). In the P.N.U. it was not found above the rift valley just north of the P.N.U. (Witte 1140 m (Witte 1953 [as E. semiannulata semi- annulata]). 1953). Laurent (1956) recorded another from , 240 km to the northeast. Boulengerina annulata stormsi Dollo 1886.— This water cobra was considered endemic to Chilorhinophis gerardi (Boulenger 1913).— Lake Tanganyika, but there is a single record Gerard’s Striped Burrowing Snake is found from Pweto on Lake Mweru (Witte 1953). from Katanga south through Zambia to north- ern Zimbabwe. In the P.N.U. it was not found anchietae Bocage 1879.—Anchieta’s above 1300 m (Witte 1953). Cobra ranges from Angola and northern Namibia to northwestern Zimbabwe and west- Polemon christyi (Boulenger 1903).— ern Zambia, reaching its eastern limit at Lake Christy’s Snake-eater is found from eastern Bangweulu (Broadley & Wüster, in review). D.R.C. and northern Zambia east to Uganda One specimen has been recorded from the and western Tanzania. The P.N.U. specimens Katanga pedicle south of the lake (Witte 1953 were taken between 1300 and 1810 m (Witte [as N. haje anchietae]). 1953 [as Miodon christyi]). Naja melanoleuca Hallowell 1857.—The Hypoptophis wilsoni Boulenger 1908.—The Forest Cobra (or species complex) has a huge rare Wedge-snouted Burrowing Snake occurs range in both forest and savanna through west, from Sankuru in the Congo basin south to central and east Africa. In the P.N.U. it does not Katanga and western Zambia (Broadley seem to occur above 900 m (Witte 1953). 1971b). In the P.N.U. it was only found below 900 m (Witte 1953). Naja crawshayi Günther 1893.—This big spit- ting cobra is a sister species of Naja nigricollis Aparallactus moeruensis Witte & Laurent Reinhardt. It occurs south of the Congo Basin, 1943.—This large centipede-eater is endemic from Congo-Brazzaville, southern D.R.C., to southeast Katanga. The type locality is Angola, western and northern Zambia through Pweto on Lake Mweru and 11 specimens were northern Malawi to Tanzania (Broadley & collected in the P.N.U. below 900 m (Witte Wüster, in prep.). It was described from Lake 1953). Mweru and is found throughout the P.N.U. (Witte 1953). Aparallactus capensis A. Smith 1849.—The Cape Centipede-eater ranges from Tanzania Naja mossambica Peters.—The Mozambique south to South Africa and west through Zambia Spitting Cobra has a wide range from Tanzania and Katanga to Angola. It is found throughout south to South Africa and west to northern

47 AFRICAN JOURNAL OF HERPETOLOGY 53(1) 2004

Namibia. There is a relict population in the the subcaudals in L. mlanjensis. The nearest D.R.C. pedicle southwest of Lake Bangweulu, populations are separated by a gap of 1000 km. represented by a pair of specimens in Tervuren (Laurent 1956, pl. xxviii, fig. 1). Lamprophis fuliginosus (Boie 1827).—The Sooty House Snake is found throughout west Dendroaspis polylepis (Günther 1864).—The and central Africa, extending into the Horn of Black Mamba has an enormous range through- Africa and plateau areas of Kenya, being out the savannas of sub-Saharan Africa. In the replaced on the coast by L. capensis (Duméril P.N.U. it was not found on the Kibara plateau, & Bibron) (Hughes 1997). Witte (1953) record- only below 1000 m (Witte 1953 [as D. angusti- ed it [as Boaedon l. lineatus (Duméril & ceps]). Dendroaspis jamesoni (Traill), a forest Bibron)] throughout the P.N.U. and illustrated a species, occurs at Kamina on the western specimen from Lubumbashi. periphery of the area under consideration, so may eventually be found in the P.N.U. Mehelya poensis (A. Smith 1847).—This forest file snake occurs from east to Uganda, Family south to Angola and the D.R.C., apparently Subfamily LAMPROPHIINAE reaching the southern limit of its range in the Lycodonomorphus upembae Laurent 1954.— P.N.U. at Munoi (Witte 1953). The Upemba Water Snake is endemic to Katanga, and was described from Nyonga on Mehelya nyassae (Günther 1888).—The Nyasa the western side of Lake Upemba. De Witte File Snake inhabits savannas from Kenya to collected 20 specimens there in 1925 (Witte South Africa, west to Burundi, Katanga 1933 [as Boodon lineatus Duméril & Bibron]), (P.N.U.), Zambia, and Namibia. In but he did not get any from the eastern edge of the P.N.U. it was not taken above 1140 m the swamps during 1946-49. It was described (Witte 1953). as a subspecies of L. subtaeniatus Laurent, of the western D.R.C. and northeast Angola, but it Mehelya capensis capensis (A. Smith 1847).— is more robust and has lower ventral and sub- The Southern File Snake has a wide range in caudal counts, while the hemipenis is bifurcate savannas of eastern Africa from northeast distally as in L. rufulus (Lichtenstein). Tanzania to South Africa and west to Katanga, Zambia, Angola and northern Namibia. Lycodonomorphus leleupi Laurent 1950.— Although not yet recorded from the P.N.U., This endemic water snake was described from Laurent (1956) noted that Witte (1953) had the Kundelungu plateau and also occurs on the assigned a specimen from Lubumbashi to M. Kibara plateau in the P.N.U. between 1250 and poensis. 1810 m (Witte 1953). It was described as a sub- species of L. whytei (Boulenger) of northern INCERTAE SEDIS Malawi and adjacent Tanzania, but was later thought to be conspecific with L. rufulus mlan- Lycophidion multimaculatum (Boettger jensis Loveridge of southern Malawi and east- 1888).—The Blotched Wolf Snake ranges from ern Zimbabwe (Broadley 1967). These are now south to Angola east to the southeastern considered to be sister species, distinguished D.R.C., Zambia and adjacent Tanzania by ventral colouration: dark with yellow patch- (Broadley 1996). It is found throughout the es and a dark tail in L. leleupi and yellow with P.N.U. (Witte 1953 [as L. capense capense (A. a few dark spots and a dark median stripe on Smith)]).

48 BROADLEY & COTTERILL — Reptiles of southeast Katanga

Pseudaspis cana (Linnaeus 1758).—The Mole (Broadley 1977a). Although widespread in Snake is found throughout southern Africa, Katanga, it has not yet been recorded from the extending north to Angola, Zambia, Katanga, P.N.U. Rwanda and the Kenya highlands. In the P.N.U. it is found on the Kibara plateau between 1320 acutus (Günther 1888).—The and 1810 m (Witte 1953), and it has been Striped Beaked Snake ranges from Angola recorded from the Kundelungu plateau at 1750 through Katanga and northern Zambia to south- m in Tatera burrows (Laurent 1956 [as P. cana western Tanzania and Burundi (Broadley anchietae (Bocage), a synonym]). 1971c). In the P.N.U. it was only collected on the Kibara plateau (Witte 1953): Laurent Grayia ornata (Bocage 1866).—The Ornate (1956) included this material in his new taxon Water Snake inhabits rivers from Cameroon R. a. wittei, described from the Kundelungu south to Angola and west to Katanga and north- plateau at 1750 m and synonymised by western Zambia (Broadley 1991b). In the Broadley (1971c). It seems that R. a. jappi P.N.U. it was taken at 700 to 800 m (Witte Broadley, described from the Barotse flood 1953). plains in western Zambia, is probably a good evolutionary species. Subfamily PSAMMOPHIINAE Dromophis lineatus (Duméril & Bibron nototaenia (Günther 1864).— 1854).—The Lined Olympic Snake has a wide The Southeastern Bark Snake has a wide range range from Senegal east to Ethiopia, south to in southeast Africa from coastal Kenya to Angola, Zambia and northern Malawi. South Africa, east to Botswana, Zambia and Although not yet recorded from the P.N.U., Katanga (Broadley & Hughes 2000). It is com- there are records from Bukena on the Luapula mon at low altitudes in the P.N.U., but is not River just southwest of the park boundary, found on the Kibara plateau (Witte 1953). Pweto and Kilwa on Lake Mweru, and variabilis Günther 1893.—The Lubumbashi to the south (Witte 1953). Two Grey-bellied Grass Snake has a discontinuous specimens recorded from the Kundelungu distribution on the southern rift valley high- plateau by Laurent (1956) actually belong to lands, but also inhabits flood plains further zambiensis. west (Broadley 1977a). In the P.N.U. it occurs on the Kibara plateau at 1750-1830 m (Witte Psammophis zambiensis Hughes & Wade 1953 [as Cerastes t. tritaeniatus]). Laurent 2002.—The Zambian Whip Snake was recent- (1956) included these specimens in his new ly described from northern Zambia and ranges taxon P. tritaeniatus festivus, described from from northern Mwinilunga District (Broadley the Kundelungu plateau, and subsequently 1991b [as P. sibilans leopardinus]) to the placed in the synonymy of P. v. variabilis Mweru Wantipa in Mporokoso District (Broadley 1977a). The status of the subspecies (Broadley & Pitman 1960 [as P. ? sibilans]). of P. variabilis is being investigated by C.M.R. The “Abercorn” [= Mbala] specimens collected Kelly. by H.J. Bredo probably came from the Mweru Wantipa while on Red Locust control opera- Psammophylax tritaeniatus (Günther 1868).— tions. It was not collected at Mbala by D. The Three-lined Grass Snake has a wide range Vesey-FizGerald, Bredo’s successor. Two spec- in southeastern Africa, extending north to imens from Kundelungu plateau at 1750 m Tanzania and west to southern Angola were assigned by Laurent (1956) to the superfi-

49 AFRICAN JOURNAL OF HERPETOLOGY 53(1) 2004 cially similar Dromophis lineatus (Hughes & western Katanga (Witte 1953 [as Neusterophis Wade 2002). fuliginoides, part]).

Psammophis mossambicus Peters 1882.—The olivacea (Peters 1854).—The taxonomy of the P. sibilans complex is still has a wide range in savan- unresolved. This name was provisionally nas from Guinea, east to the Sudan and revived for eastern material previously Ethiopia, south through southern Somalia to assigned to P. phillipsii Hallowell of West southern Mozambique and west to Gabon and Africa. The Olive Whip Snake ranges from the Angola. In the P.N.U. it was only found below Sudan to South Africa, west through Katanga 1000 m, mainly at Mabwe on Lake Upemba to Angola (Broadley 2002). It is common (Witte 1953 [as Neusterophis o. olivaceus]). throughout the P.N.U. (Witte 1953 [as P. sibi- lans sibilans]). Limnophis bangweolicus (Mertens 1936).— The Bangweulu Water Snake ranges from Psammophis angolensis (Bocage 1872).—The Zambia, west to Katanga and eastern Angola. It Dwarf Whip Snake has a wide range in savan- has not yet been found in the P.N.U., but it has nas of eastern Africa, from Ethiopia to South been recorded from Pweto, Lubumbashi and Africa and Mozambique, west through Zambia near Kanzenze on a tributary of the upper and Katanga, to Angola (Broadley 2002). In the Lualaba (Witte 1953 [as L. bicolor (Günther)]). P.N.U. it was not found on the Kibara plateau (Witte 1953). Subfamily BOIGINAE

Subfamily Boiga blandingii (Hallowell 1844).— Blanding’s Tree Snake inhabits forests and Natriciteres bipostocularis Broadley 1962.— wooded savannas from Guinea, east to western This small marsh snake was described from Kenya, south to Angola, Katanga and northern northeastern Zambia and extends west through Zambia (Mporokoso District). The P.N.U. Katanga to the highlands of central Angola specimen came from Pelenge gorge at 1250 m (Broadley 1966b). The specimens from the (Witte 1953); the other Katangan localities are P.N.U. and environs were collected between Sandoa and Kanzenze, the latter being the 1320 and 1585 m (Witte 1953 [as Neusterophis southernmost record. fuliginoides (Günther)]), but were assigned to N. olivacea uluguruensis (Loveridge) by Dipsadoboa shrevei shrevei (Loveridge Loveridge (1958). This taxon was originally 1932).—This nocturnal tree snake ranges from described as a subspecies of N. olivacea, and the lower Congo and Angola, east through the then placed as a subspecies of N. variegata D.R.C. and northern Zambia to southeast (Peters) of West Africa (Broadley 1966b). It Tanzania (Rasmussen 1986). The only definite now seems better to recognise all the well iso- locality in Katanga is Lubumbashi (Bourgeois lated “races” as evolutionary species (Broadley 1964), but there are numerous records from & Kelly, in prep.). northern Zambia (Broadley 1971d, 1991b).

Natriciteres fuliginoides (Günther 1858).— Dipsadoboa unicolor Günther 1858.—This This forest marsh snake ranges from Guinea slender arboreal snake occurs from Guinea to east to the D.R.C., reaching its southern limit at Cameroon, is apparently replaced by D. weileri Lofoi at the base of the Kundelungu plateau, it (Lindholm) in the Congo basin, and then reap- is also recorded from Kapanga and Sandoa in pears along the Albertine rift (Rasmussen

50 BROADLEY & COTTERILL — Reptiles of southeast Katanga

1993). The southernmost specimens are from to the southern D.R.C. and Angola. It is found the P.N.U. and were taken between 1250 and throughout the P.N.U. (Witte 1953). 1300 m (Witte 1953). Philothamnus heterolepidotus (Günther Crotaphopeltis hotamboeia (Laurenti 1768).— 1863).—The Slender Green Snake ranges from The Herald Snake is common throughout moist Sierra Leone east to Ethiopia, south through the savannas of sub-Saharan Africa, including the D.R.C. to Angola and northern Zambia P.N.U. (Witte 1953). (Hughes 1985). It is common throughout the P.N.U. and environs (Witte 1953). semiannulatus A. Smith 1849.— The Eastern has a wide range in Philothamnus angolensis Bocage 1882.—The eastern Africa from Kenya to South Africa, Angolan Green Snake has a wide range in west to the D.R.C., Angola and Namibia. It is savannas from Cameroon east to the Sudan, found throughout the P.N.U. (Witte 1953). south to South Africa (Hughes 1985). It is found throughout the P.N.U. (Witte 1953) [as P. Subfamily COLUBRINAE irregularis irregularis (Leach) of West Africa].

Scaphiophis albopunctatus Peters 1870.—The Philothamnus semivariegatus (A. Smith Grey Beaked Snake ranges from east to 1847).—The Variegated Bush Snake has the Kenya, south to the D.R.C., northern Zambia widest range of any species in the genus, inhab- and Tanzania (Broadley 1994). In the P.N.U. it iting savannas from Senegal east to Ethiopia, was not found above 900 m (Witte 1953). south to South Africa (Hughes 1985). It is found throughout the P.N.U. (Witte 1953). Prosymna ambigua Bocage 1873.—Bocage’s Shovel-snout has an extensive range around the Dispholidus typus (A. Smith 1829).—The periphery of the Congo basin (Broadley 1980). Boomslang is found throughout the savannas In the P.N.U. it was not taken above 800 m of sub-Saharan Africa, including the whole of (Witte 1953). the P.N.U. (Witte 1953). Laurent (1955) assigned material from northern Angola and Philothamnus carinatus (Andersson 1901).— the southern D.R.C. to a new subspecies D. t. This forest species ranges from Guinea to east- punctatus, which intergrades with his other ern Kenya and Tanzania, south to Katanga new subspecies D. t. kivuensis in northeastern (Hughes 1985). Specimens from the P.N.U. Zambia. were taken at an altitude of 680 - 700 m (Witte 1953 [as P. heterodermus carinatus]). Hughes Thelotornis kirtlandii (Hallowell 1844).—The (1985) recorded sympatry between this species Forest Vine Snake ranges from Guinea-Bissau and P. ruandae Loveridge at M’bala about 90 east through forested areas of West Africa and km further west, but Laurent (1956) suggested the Congo basin to Uganda and the southern that the locality data for the latter specimen is Sudan, south to northern Angola, northwestern probably erroneous. Transposition of specimen Zambia and south-central Tanzania (Broadley labels at Tervuren is easy, as they are not tied to 2001). In the P.N.U. it was not found on the the specimens. Kibara plateau, but was taken between 700 and 1320 m (Witte 1953). Philothamnus hoplogaster (Günther 1863).— The Southeastern Green Snake ranges through Thelotornis capensis oatesii (Günther 1881).— savannas from Kenya to South Africa and west This savanna vine snake has a wide range in

51 AFRICAN JOURNAL OF HERPETOLOGY 53(1) 2004 south central Africa (Broadley 2001). In the Guinean-Congolian Forest species (17).— P.N.U. it was not found on the Kibara plateau Pelusios gabonensis; Feylinia currori; Adolfus (Witte 1953). africanus; Rhinotyphlops angolensis; Causus lichtensteinii; Bitis gabonica; Polemon (Linnaeus 1754).—The christyi; Naja melanoleuca; Mehelya poensis; Common Egg-eater is found throughout the Grayia ornata; Natriciteres fuliginoides; Boiga savannas of sub-Saharan Africa, including the blandingii; Dipsadoboa unicolor; Prosymna P.N.U. (Witte 1953). ambigua; Philothamnus carinatus; Thelotornis kirtlandii; Crocodylus cataphractus. This list is ORDER CROCODYLIA probably incomplete, as some forest species Family CROCODYLIDAE recorded from the Ikelenge pedicle in north- west Zambia can be expected to occur in south- Crocodylus cataphractus Cuvier 1824.—The east Katanga, i.e., Thrasops jacksonii Günther Slender-snouted Crocodile ranges through and Rhamnophis aethiopissa Günther forested areas from Senegal east to Lake (Broadley 1991b; Broadley & Wallach 2002). Tanganyika and south to Angola and Katanga. Dendroaspis jamesoni (Traill) is also likely to occur. It occurs in Lakes Upemba and Mweru and the Luapula River to the south (Witte 1953). Widespread sub-Saharan savanna species (10).—Pelomedusa s. subrufa; Varanus niloti- Crocodylus niloticus Laurenti 1768.—The Nile cus; Bitis arietans; Dendroaspis polylepis; Crocodile occurs in rivers and lakes of sub- Natriciteres olivacea; Crotaphopeltis hotam- Saharan Africa south to eastern South Africa. boeia; Philothamnus semivariegatus; In the P.N.U., it has always been less abundant Dispholidus typus; Dasypeltis scabra; than the previous species, but numbers of both Crocodylus niloticus. crocodiles declined between 1925 and 1947-49 (Witte 1953). Northern Guinean savanna species (8).— Chamaeleo gracilis; Lygodactylus gutturalis; Trachylepis maculilabris; Rhinotyphlops lineo- ZOOGEOGRAPHY latus; Lamprophis fuliginosus; Dromophis lin- eatus; Scaphiophis albopunctatus; Philotham- Analysis of the distribution patterns of the rep- nus heterolepidotus. tiles of Katanga suggests that they can be roughly allocated to 13 range clusters, as fol- South-central African savanna species (20).— Pelusios rhodesianus; Kinixys spekii; lows: Lygodactylus angolensis; Sepsina hemptinnei; Gerrhosaurus nigrolineatus; G. bulsi; Species endemic or near-endemic to Katanga Ichnotropis capensis; Varanus albigularis (15).—Pelusios upembae; Pachydactylus kat- angolensis; Monopeltis remaclei; Rhino- angensis; Trachylepis upembae; T. pulcherri- typhlops schmidti; Leptotyphlops kafubi; ma; Panaspis smithii; Monopeltis adercae; M. Xenocalamus michelli; Hypoptophis wilsoni; scalper; Rhinotyphlops (Letheobia) kibarae; Chilorhinophis gerardi; Elapsoidea guentheri; Atheris katangensis; Atractaspis leleupi; Boulengerina annulata stormsi; Naja craw- Amblyodipsas rodhaini; A. katangensis; shayi; Lycophidion multimaculatum; Philo- Aparallactus moeruensis; Lycodonomorphus thamnus angolensis; Thelotornis capensis upembae; L. leleupi. oatesii.

52 BROADLEY & COTTERILL — Reptiles of southeast Katanga

Southern rim of the Congo Basin species DISCUSSION (15).—Pelusios nanus; Chamaeleo anchietae; Lygodactylus heeneni; Eumecia anchietae; An encompassing explanation for the richness Leptosiaphos dewittei; Panaspis seydeli; of this cosmopolitan herpetofauna needs to Chamaesaura miopropus; Tetradactylus ellen- consider the ecological determinants of its composition. Faulting in the southwest branch bergeri; Ichnotropis bivittata; Rhinotyphlops of the African Rift created high, flat plateaux, (Letheobia) gracilis; Causus bilineatus; whose steep sides surround wide valleys. This Rhamphiophis acutus; Psammophis zambien- diverse geomorphology of Katanga is the tem- sis; Natriciteres bipostocularis; Dipsadoboa s. plate on which the many different habitats have shrevei. evolved. This variety can be roughly classified into: wetland (lowland and highland), savannas Eastern savanna species (27).—Pelusios sub- and grasslands (in valleys, and on relief), aeo- ; Acanthocercus atricollis; Agama arma- lian sediments (dilungu on plateaux), lacustrine ta; Chamaeleo dilepis; Hemidactylus mabouia; alluvium, and forest. The forests and savannas are especially variable in composition, context Trachylepis megalura; T. varia; T. striata; and extent. Lygosoma afrum; Panaspis wahlbergii; Cordylus tropidosternum; Holaspis laevis (+ Equally, elements of neighbouring biomes have coastal forest); Rhinotyphlops mucruso; Python significantly elevated the diversity of Katanga’s natalensis; Causus rhombeatus; Atractaspis herpetofauna. In its position under contempo- bibronii; Amblyodipsas polylepis; Aparallactus rary climatic conditions, Katanga straddles a capensis; Naja mossambica; Mehelya nyassae; broad transition zone of forest and savanna, juxtapositioned between the main Congo forest M. c. capensis; Hemirhagerrhis nototaenia; block and the Zambezian savannas. Climatic Psammophylax tritaeniatus; Psammophis change has been as, if not more, important in mossambicus; P. angolensis; Telescopus semi- changing the extent of habitats. This was espe- annulatus; Philothamnus hoplogaster. cially so through the Plio-Pleistocene, with plu- vials alternating with arid periods. These deter- East African Savanna to Congo Basin (1).— minants have interacted in complex ways, Pachydactylus tuberculosus. which complicates attempts to tease out each of their influences on biogeographical patterns. South-west savanna (2).—Pachydactylus Their relative importance has also varied punctatus; Naja anchietae. across different scales of space and time. The contributions of these determinants are dis- Southeastern savanna (1).—Pseudaspis cana. cussed below, with focus on the evolution of endemics in SE Katanga. Kalahari sand fauna (1).—Zygaspis quadri- DIVERSITY frons. Contemporary Ecological Conditions, and Southern rift valley highlands fauna (1).— Elevated β and γ Diversity.—We can invoke Psammophylax variabilis. contemporary patterns in the ecology of the landscape to explain the richness of reptile Palaeo-Upper Zambezi - Okavango fauna assemblages. The high hetereogeneity of habi- (1).—Limnophis bangweolicus. tats stands out as a major determinant. These

53 AFRICAN JOURNAL OF HERPETOLOGY 53(1) 2004 represent a great range in ecological condi- Higher, and perennial, availability of soil mois- tions: from mesic forest to comparatively arid ture along drainage lines supports gallery dilungu. High heterogeneity is exhibited both forests both on more gentle slopes (in drainage within and among the various habitats. An basins between interfluves), and also in steep example is how altitude augments variety in valleys and ravines. Thus, ecological condi- the miombo and forests, from lowlands to tions in these forested drainage channels feed- plateaux, across a large altitudinal range (400 ing the Lualaba differ markedly from that on m in the Lualaba valley to over 1800 m on the interfluves under the more extensive miom- Kibara and Kundelungu plateaux), and these bo woodlands (White & Werger 1978). Plants gradients are often steep (exemplified across hold a keystone role in structuring these patch- the Upemba horst). This high richness of ter- es of habitat in which other forest biota persist. restrial habitats is matched by (and integrated The overall result has been an enrichment of with) a complementary richness of wetlands the region's biodiversity by forest organisms, as (Dapper 1981). For example, ecological diver- indicated by certain invertebrates and small sity is enhanced by thermal and saline springs vertebrates (Cotterill 2002a,b). where the thermophilic crustacean, Thermo- bathynella adami occurs. This was the fourth CLIMATIC CHANGE parabathynellid ever described - isolated from a hot spring in Upemba (Schminke 1987). Shifting Habitats.—Both the composition and Another interesting vertebrate endemic, the extent of habitats across Africa changed radi- cichlid fish, Oreochromis salinicola (Poll cally through the Tertiary, during which cli- 1948), is known only from Mwashya geother- mates varied across a range of scales in space mal saline springs north of Lake Lufira and time. An important influence on the con- (Trewavas 1983; Malaisse 1997). Overall, this temporary biodiversity of Katanga has been great variety in both physiography and vegeta- alternations between pluvial and xeric regimes. tion (resplendent in the high β and γ diversity These altered ranges and composition of prin- of the flora) is the proximate determinant of the cipal habitats, notably wetland, grassland, for- high species richness of reptiles, and indeed the est and savanna (Colyn 1991). In addition to rich biodiversity of Katanga. mushitu, signatures of these changes through recent geological time are represented by the The Example of Guineo-Congolian Elements.— remarkable sandy habitats - the dilungu - on The 17 species of forest reptiles in the contem- Katanga’s plateaux. These record comparative- porary fauna exemplify the biogeographical ly recent arid episodes in variable climates, consequences where mushitu has persisted where regimes of precipitation and temperature deep within the miombo savannas. This forest- fluctuated in complex ways through the Plio- savanna mosaic is a more apparent aspect of Pleistocene and Holocene (Dapper 1981, 1988; elevated habitat hetereogeneity, and thus land- Colyn 1991). scape structure, in the Katanga region. Mushitu not only elevates β and γ diversity, but is a crit- There are at least two likely reasons why the ical habitat for those species of Guineo- biodiversity in Katanga has been especially Congolian affinity up to 250 kms south of the influenced by these climatic changes. One is its main belt of moist forest. This augments strategic position abutting the great equatorial species richness of reptiles significantly. divide — between the Zambezian and Guineo- Edaphic conditions appear to be the singular Congolian regions (Fig. 2); in this situation the geomorphological control that structures this region has been influenced by both biotic forest-savanna mosaic in the landscape. (notably phytogeographic associations) as well

54 BROADLEY & COTTERILL — Reptiles of southeast Katanga as physiographic events (especially exchanges as well as alluvial and aeolian sediments. In between drainage systems, see below). Second, total, they account for at least 12% of overall patches of both xeric and mesic habitats have diversity. One of these species exhibits forest persisted successfully, with elements of their affinities and likely speciated according to the biodiversity intact, sheltered in the region’s forest-savanna model described above. physiographically diverse landscapes. Con- Fossorial endemics point to vicariant specia- temporary examples are the dilungu on tion in patches of dilungu and alluvium. The Katanga’s plateaux, and mushitu forests in val- apparent dearth of endemic fossorial reptiles in leys, respectively (Malaisse 1997). the dilungu more likely represents a lack of collecting. The isolation of populations of Evolutionary Significance of the Persistence of grassland birds (notably pipits and larks) on Forests.—A subtle connectivity of Katanga’s these sandy habitats on Katanga’s plateaux is gallery forests with moist forests in the Congo attributed to a history of climatic changes in the lowlands is the most plausible explanation for late Pleistocene. It appears that these Kalahari not only why Guineo-Congolian species occur sands were initially deposited in one or more deep within mesic miombo savanna, but also interpluvials and then isolated in wetter periods recent speciation. The congruent pattern is of (Cotterill in press b). species isolated from sister species in the main forest block, with their ranges restricted within Lycodonomorphus leleupi isolated on the the southern Congo basin against the Congo- Kibara and Kundelungu plateaux is hypothe- Zambezi watershed. This is exemplified in the sized to represent a relict link to taxa along the biodiversity of the Ikelenge pedicle on the East African Rift in east Africa and Malawi. southwest border of Katanga. Examples Conversely, L. upembae has been isolated in include three small mammals currently collect- the Upemba depression from its sister species, ed only from the Ikelenge Pedicle, namely: L. subtaeniatus, in the lower Congo basin. The Ansell’s long-footed , Malacomys australis evolutionary history of the other wetland Ansell; Ansell’s musk shrew, Crocidura ansel- endemic, Pelusios upembae, is complicated by lorum Hutterer & Dippenaar, and Sakeji horse- its vicariance from P. bechuanicus. Speciation shoe , Rhinolophus sakejiensis Cotterill of these terrapins has been interwoven with the (Cotterill 2002a,b). It appears that a congruent intricate history of drainage evolution within evolutionary history is exhibited by particular and around Katanga (discussed below). reptile endemics in Katanga, notably Atheris katangensis. This endemism emphasizes the evolutionary significance of different patches (forests, sedi- Interestingly, a parallel example of this land- ments and wetlands) - exemplified by specia- scape pattern (namely a forest-savanna mosaic) tion in mushitu. Vicariant speciation of these occurs in the Amazon Basin. Likewise, gallery endemic reptiles illustrates how these patches forests account for forest species occurring have functioned as source, as opposed to sink, deep into Neotropical savanna; where their per- habitats through evolutionary time. Again, the sistence has sometimes resulted in evolution. region’s complex physiography has governed Gallery forests are not only refugia for forest these complementary histories in forest, sedi- organisms, but have also fostered speciation of ments, and wetland. It would be interesting to these relict populations (Meave et al. 1991). quantify how these suites of assemblages (e.g., forest) differ from extant populations in their Evolution of Endemics.—Katanga’s endemic ancestral source habitats (pertinently the reptiles are represented in forest and wetlands, Congo lowlands). Phylogeographic studies of

55 AFRICAN JOURNAL OF HERPETOLOGY 53(1) 2004 these geographically dispersed lineages should influence on speciation of vertebrates, notably help to elucidate their evolutionary histories. reptiles, within and around Katanga.

GEOLOGICAL HISTORY A model of drainage evolution proposed for vicariant speciation of the Upemba lechwe Tectonic Events in the East African Rift (Cotterill submitted) in the Kamalondo depres- System.—Ultimately, the singular determinant sion centres on past links between the Upper of the diverse topography of Katanga has been Lualaba-Lufira drainage and the Upper activity in the East African Rift system during Zambezi through a trans-Katanga drainage sys- and after the Miocene. Associated with Africa’s tem. The first event entailed dispersal by aquat- rift valley system, tectonic activity in the ic organisms along a connection, the Palaeo- Pliocene across the south-central African Kafila; which was then a major river connected plateau reactivated erosion. Significantly, this to the Palaeo-Chambeshi drainage system. This was when Katanga’s plateaux were uplifted to was when, or soon after, the latter was the prin- accentuate the Kamalondo depression. These cipal tributary of Palaeo-Lake Makgadigadi geologically-recent events account for the var- through the Plio-Pleistocene; such that the ied topography ranging from plateaus and Palaeo-Chambeshi flowed southwest across the ravines to deep and wide valleys (Dapper 1981; Lukanga depression into Palaeo-Lake Sebagenzi & Kaputo 2002). Makgadigadi. Large, shallow endoreic lakes played a key role in these dynamics; and espe- Landscape Evolution in and Around Katanga, cially significant was Palaeo-Lake Lufira, and Vicariant Speciation of Aquatic whose recent existence is represented by sig- Vertbrates.—Drainage across Katanga has nificant alluvium in the Lufira depression undergone interesting changes and this has (Cotterill in press a, Fig. 2). influenced biotic evolution profoundly. This history has been intimately interlinked with Alternatively, it can be argued that vertebrates dispersed between the Upper-Zambezi and drainage evolution across the south-central Upper Lualaba across the Congo-Zambezi Africa plateau overall, especially the Upper watershed prior to speciation of Pelusios upem- Zambezi and neighbouring systems. Signifi- bae; so that a link between the Palaeo- cant events involved shrinkage and expansion Chambeshi and the Kamalondo graben via a of wetlands, with changes in their connections. trans-Katanga drainage system is thus unlikely. The most interesting of these events were most The close proximity of the modern Upper likely Plio-Pleistocene. This complex history is Zambezi and Upper Lualaba headwaters is still incompletely understood, and its deserving likely geologically recent; and fieldwork by treatment is beyond the scope of this paper. Bell-Cross (1965) discounted the likelihood of Nevertheless, sufficient biotic and geological fishes dispersing across this watershed. evidence points to past connections of wetlands in Katanga with Bangweulu, and with the This favoured explanation for evolution of the Upper Zambezi and Okavango systems further endemic terrapin Pelusios upembae is support- south. A model of this recent drainage evolu- ed by interesting congruence in two complexes tion explains vicariant speciation of lechwe of wetland birds. The one, its distribution cen- antelopes in south-central Africa - most perti- tred on Katanga, entailed vicariant speciation nently the Upemba lechwe (Cotterill in press a, of masked weaver birds (the Ploceus reichardi submitted). A synopsis of this drainage evolu- group). These are entirely swamp-dwelling, tion, presented below, illustrates this important and dependent on large permanent wetlands,

56 BROADLEY & COTTERILL — Reptiles of southeast Katanga especially for nesting. This species complex catchment (Teugels & Guégan 1994; Malaisse comprises three endemics: with P. ruweti 1997). Louette & Benson restricted to the Lufira depression; P. upembae (Verheyen) in Upemba; Congruent patterns of speciation in all these and P. katangae (Verheyen) occurring along the mammals, fishes, birds and reptiles (pertinent- Luapula River and around Lakes Bangweulu ly Pelusios upembae) is most plausibly and Mweru. The other species, P. reichardi accounted for by recent drainage evolution. Reichenow is isolated in southeast Tanzania Thus, their vicariant speciation followed on around Lake Rukwa and its drainage system; fragmentation of the trans-Katanga drainage while the recently discovered P. burnieri Baker system that isolated the Kamalondo depression & Baker is isolated in the Kilombero valley. A from the Palaeo-Chambeshi, and thus Palaeo- congruent pattern is shown by the distribution Lake Makgadikgadi. The history of this huge of Swamp Flycatchers of the Muscicapa lual- drainage system, the Palaeo-Chambeshi, is abae (Chapin) complex (Cotterill in press a). being investigated in detail, with special focus on its former connections with Katanga, to Speciation of endemic fishes in Katanga has account for past links and fissions. This new been especially prolific. There is an interesting model which emphasizes past connections congruence between endemic fishes in the between the Congo and Zambezi drainage sys- Kamalondo with the water snake Lycodon- tems obviously endorses a radical reconsidera- omorphus upembae and Upemba terrapin, tion of landscape evolution in south-central Pelusios upembae. The zoogeography of these Africa (Cotterill 2003, in press a, submitted; F. fish faunas also exhibits a significant comple- P. D. Cotterill unpublished data). mentation of Upper Zambezi species. Poll (1963) originally contrasted the aquatic ecolo- gy of the lower Congo, which drains and flows CONCLUSIONS under Guineo-Congolian moist forests, with the Upper Lualaba-Lufira, whose tributaries 1. The above explanation has focused on the drain a patchwork mosaic of dambos, riparian more obvious patterns of endemism, faunal forests and savanna. These headwaters espe- mixing and species persistence exemplified by cially differ in the predominance of cool, fast indicator species in Katanga. The processes flowing mountain streams (Banister & Bailey responsible for the rich reptile fauna have acted 1979; Teugels & Guégan 1994). Thus, over 600 through complimentary, interlinked, and thus species of lower Congo fishes do not occur complex mechanisms. Proximally, the very upstream of the Portes d’Enfer; but the fish diverse physiography and vegetation are com- fauna of the Upper Congo has been augmented bined determinants of high β and γ diversity by Upper Zambezian species. The relatively across Katanga’s landscapes. Ultimately, evo- high endemism in three ichthyofaunal subre- lution of the region’s biodiversity has been dri- gions of Katanga (Upper Lualaba-Lufira, ven by a complex history of climatic and geo- Kamalondo Basin, and Luapula) is especially logical change. interesting (Malaisse 1997). This endemicity is illustrated by comparisons of fishes (total 2. It is unfortunate that knowledge of the biodi- endemics/total species richness) of the Kasai- versity of this fascinating region (Katanga and Bushimaie-Lulua (2/77) with the Kamalondo northeast Zambia) is so incomplete. The princi- basin (47/183), Upper Lualaba-Lufira (12/87) pal herpetological surveys have been very and Luapula (39/146). The latter also includes patchy in their focus on protected areas (espe- Lake Bangweulu and the Upper Chambeshi cially Upemba) and near human settlements

57 AFRICAN JOURNAL OF HERPETOLOGY 53(1) 2004

(e.g., Lubumbashi). In particular, we single out ACKNOWLEDGMENTS Lake Bangweulu, from which very few speci- mens of any organisms have been collected. We thank Drs Michel Hasson and Jo Only more thorough faunal surveys of Katanga Thompson, Nouvelle Approaches, who provid- ed valuable information to complete this paper. and surrounding landscapes will provide the Cotterill gratefully acknowledges the support necessary data for fascinating phylogeographic of a Biodiversity Leadership Award from the studies, and above all quantify zoogeographic Bay Foundation, and the Josephine Bay Paul patterns more thoroughly. and C. Michael Paul Foundations, New York City. Part of the results was presented at the 3. Opportunities for research on Katanga’s rep- Third Meeting of the Southern African Society tiles, as for all its biodiversity, are practically for Systematic Biology in July 2003, Pretoria, unlimited. This synthesis provides a framework South Africa. for further research on this biodiversity hotspot to understand how tectonics, drainage evolu- LITERATURE CITED tion, and climatic change have influenced evo- lution. This needs to employ phylogeographic ADERCA, B. 1966. Notice Géologique. Pp. 21-37. In studies, with palaeoecological reconstructions Witte, G.-F. de Introduction. Explor. Parc natn. Upemba, Miss. G.-F. de Witte. 1. that include dating of lake sediments. BANISTER, K.E. & R.G. BAILEY. 1979. Fishes collected Phylogeographic comparisons of forest, savan- by the Zaire River Expedition, 1974-75. Zool. J. na, wetland and fossorial species, respectively Linn. Soc. 66: 205-249. BAUER, A.M. 2003. On the identity of Lacerta punctata and in comparison, promise to yield fascinating Linnaeus 1758, the type species of the genus insights into late Tertiary history of the land- Euprepis Wagler 1830, and the generic assignment scape. of Afro-Malagasy skinks. Afr. J. Herpetol. 52: 1-7. BELL-CROSS, G. 1965. Movement of fish across the Congo-Zambezi watershed in the Mwinilunga dis- 4. Notwithstanding deficiencies in data, the dis- trict of . Pp. 415-424. In tributions exemplified by our current knowl- Snowball, J. G. (Ed.) Proc. Med. and Sci. Congr., edge outline a remarkable history of complex Lusaka 1963. Pergamon Press, Oxford. BOUR, R. 1983. Trois populations endémiques du genre faunal dynamics and speciation. The excep- Pelusios (Reptilia, Chelonii, Pelomedusidae) aux tional species richness and endemism of the îles Seychelles; relations avec les espèces africaines reptile fauna of southeast Katanga constitutes a et malagaches. Bull. Mus. natn. Hist. nat. Paris (4) 5 (A): 343-382. strong argument for the region’s conservation. BOURGEOIS, M. 1964. Les reptiles de la Kasapa Brooks et al. (2001) identified “Upemba Plain” (Elizabethville). Publ. Univ. Off. Congo 7: 65-89. among the top 10 sites for priority conservation BROADLEY, D.G. 1966a. A review of the Riopa sunde- valli group (Sauria: Scincidae) in southern Africa. in Africa, based on continental patterns of Arnoldia (Rhod.) 2, No. 34: 1-7. species richness and endemism of amphibians BROADLEY, D.G. 1966b. A review of the genus and snakes; but they did not specify the indica- Natriciteres Loveridge (Serpentes: Colubridae). Arnoldia (Rhod.) 2, No. 35: 1-11. tor taxa. The global significance of this hotspot BROADLEY, D.G. 1967. A review of the genus of biodiversity is strengthened even more when Lycodonomorphus Fitzinger (Serpentes: Colubr- one acknowledges complementary patterns of idae) in southeastern Africa, with a key to the genus. Arnoldia (Rhod.) 3, No.16: 1-9. other aquatic and terrestrial vertebrates and BROADLEY, D.G. 1971a. A revision of the African snake plants (Cotterill, in press a,b, submitted). genus Elapsoidea Bocage (Elapidae). Occ. Pap.

58 BROADLEY & COTTERILL — Reptiles of southeast Katanga

natn. Mus. Rhod. 4 (32B): 577-626. Python natalensis A. Smith 1840, is a valid species. BROADLEY, D.G. 1971b. A revision of the African snake Afr. Herp. News 28: 31-32. genera Amblyodipsas and Xenocalamus (Colubr- BROADLEY, D.G. 2000. Geographical Distribution: idae). Occ. Pap. natn. Mus. Rhod. 4 (33B): 629- Reptilia, Sauria, Lacertidae, Holaspis laevis 697. (Werner 1895), Eastern Serrate-toed Tree Lizard. BROADLEY, D.G. 1971c. A review of Rhamphiophis Afr. Herp News 31: 13-14. acutus (Günther), with the description of a new BROADLEY, D.G. 2001. A review of the genus subspecies from Zambia. Arnoldia (Rhod.) 5, No. Thelotornis A. Smith in eastern Africa, with the 8: 1-8. description of a new species from the Usambara BROADLEY, D.G. 1971d. The reptiles and amphibians of Mountains (Serpentes: Colubridae: Dispholidini). Zambia. Puku No. 6: 1-143. Afr. J. Herpetol. 50: 53-70. BROADLEY, D.G. 1977a. A revision of the African BROADLEY, D.G. 2002. A review of the species of snakes of the genus Psammophylax Fitzinger Psammophis Boie found south of Latitude 12°S (Colubridae). Occ. Pap. natn. Mus. Rhod. (B) 6: 1- (Serpentes: Psammophiinae). Afr. J. Herpetol 51: 44. 83-119. BROADLEY, D.G. 1977b. A review of the northeastern BROADLEY, D.G. 2003. Pachydactylus katanganus de forms of the Pachydactylus capensis complex Witte 1953, a species endemic to the Upemba (Sauria: Gekkonidae). Arnoldia (Rhod.) 8, No. 18: National Park (Sauria: Gekkonidae). Afr. J. 1-19. Herpetol. 52: 69-70. BROADLEY, D.G. 1980. A revision of the African snake BROADLEY, D.G. & A.M. BAUER. 1998. A review of the genus Prosymna Gray (Colubridae). Occ. Pap. Mabuya quinquetaeniata complex in East Africa natn. Mus. Rhod. (B) 6: 481-556. (Sauria: Scincidae). Afr. J. Herpetol. 47: 43-58. BROADLEY, D.G. 1981. A review of the genus Pelusios BROADLEY, D.G. & W.R. BRANCH. 2002. A review of the Wagler in southern Africa (Pleurodira: small east African Cordylus (Sauria: Cordylidae), Pelomedusidae). Occ. Pap. natn. Mus. Rhod. (B) 6: with the description of a new species. Afr. J. 633-686. Herpetol. 51: 9-34. BROADLEY, D.G. 1984. A review of geographical varia- BROADLEY, D.G. & S. BROADLEY. 1997. A review of the tion in the African Python, Python sebae (Gmelin). African genus Zygaspis Cope (Reptilia: Brit. J. Herpetol. 6: 359-367. Amphisbaenia). Syntarsus No. 4: 1-23. BROADLEY, D.G. 1989. A reappraisal of the genus BROADLEY, D.G. & S. BROADLEY. 1999. A review of the Panaspis Cope, with the description of a new African worm snakes from south of Latitude 12° S species of Leptosiaphos (Reptilia: Scincidae) from (Serpentes: Leptotyphlopidae). Syntarsus No. 5: 1- Tanzania. Arnoldia (Zimbabwe) 9 (32): 439-449. 36. BROADLEY, D.G. 1991a. A review of the southern BROADLEY, D.G., C. GANS & J.D. VISSER. 1976. Studies African Stiletto Snakes of the genus Atractaspis A. on amphisbaenians (Amphisbaenia, Reptilia) 6. Smith (Serpentes: Atractaspididae). Arnoldia The genera Monopeltis and Dalophia in southern (Zimbabwe) 9 (36): 495-517. Africa. Bull. Amer. Mus. nat. Hist. 157: 311-486. BROADLEY, D.G. 1991b. The herpetofauna of northern BROADLEY, D.G. & B. HUGHES. 2000. A revision of the Mwinilunga District, northwestern Zambia. African genus Hemirhagerrhis Boettger 1893 Arnoldia (Zimbabwe) 9 (37): 519-538. (Serpentes: Colubridae). Syntarsus No. 6: 1-17. BROADLEY, D.G. 1993. A review of the southern African BROADLEY, D.G. & C.R.S. PITMAN. 1960. On a collec- species of Kinixys Bell (Reptilia: Testudinidae). tion of snakes made in Northern Rhodesia by Ann. Transv. Mus. 36: 41-52. Monsieur H.J. Bredo. Occ. Pap. natn. Mus. S. BROADLEY, D.G. 1994. A revision of the African genus Rhod. 24B: 437-451. Scaphiophis Peters (Serpentes: Colubridae). BROADLEY, D.G. & V. WALLACH. 2000. A new blind Herpetol. J. 4: 1-10. snake (Serpentes: Typhlopidae) from montane BROADLEY, D.G. 1996. A revision of the genus forests of the Eastern Arc Mountains in Tanzania. Lycophidion Fitzinger (Serpentes: Colubridae) in Afr. J. Herpetol. 49: 165-168. Africa south of the Equator. Syntarsus No. 3: 1-33. BROADLEY, D.G. & V. WALLACH. 2002. Review of the BROADLEY, D.G. 1998. A review of the genus Atheris Dispholidini, with the description of a new genus Cope (Serpentes: Viperidae), with the description and species from Tanzania (Serpentes, Colubridae). of a new species from Uganda. Herpetol. J. 8: 117- Bull. nat. Hist. Mus. Lond. (Zool.) 68: 57-74. 135. BROADLEY, D.G. & W. WÜSTER. in review. A review of BROADLEY, D.G. 1999. The Southern African Python, the southern African ‘non-spitting’ cobras

59 AFRICAN JOURNAL OF HERPETOLOGY 53(1) 2004

(Serpentes: Elapidae: Naja). Afr. J. Herpetol. ered plateaux in southern Shaba, Zaire. Pp. 115- BROOKS, T., A. BALMFORD, N. BURGESS, J. FJELDSA, L. 135. In G.F. Dardis & B.P. Moon. Geomorpholog- A. HANSEN, J. MOORE, C. RAHBEK & P. WILLIAMS. ical Studies in Southern Africa. Balkema, 2001. Toward a blueprint for conservation in Rotterdam. Africa. Bioscience 51: 613-624. DEMEY, R. & M. LOUETTE. 2001. Democratic Republic BRYGOO, E.R. & R. ROUX-ESTÈVE. 1983. Feylinia, of Congo. Pp. 199-218. In: Important Areas in genre de Lézards africains de la famille des Africa and its Islands: priority sites for conserva- Scincidae, sous-famille Feylininae. Bull. Mus. tion. L.D.C. Fishpool & M.I. Evans (Eds), . Birdlife natn. Hist. nat. Paris (4) 5 (A): 307-341. International, Cambridge. BURGESS, N., W. KUPER, J. MUTKE, J. BROWN, S. FUHN, I.E. 1964. Ablepharus anselli FitzSimons 1955, WESTAWAY, S. TURPIE, C. MESHACK, J. TAPLIN, C. synonyme d'Ablepharus tancredii Boulenger 1909 MCCLEAN & J. C. LOVETT 2004. in press. Major (Reptilia: Sauria). Doriana 3, No. 150: 1-3. gaps in the distribution of protected areas for FUHN, I.E. 1970. Contribution à la systématique des threatened and narrow range Afrotropical plants. Lygosomines Africains (Reptilia, Scincidae). I. Les Biodiv. Cons. espèces attribuées au genre Ablepharus. Rev. COLYN, M. 1991. L'importance zoogeographique du Roum. Biol., Sér. Zool. 15: 61-71. basin du fleuve Zaire pour la speciation: la cas des HASSON, M. & E. WOLANSKI. 1999. Upemba National primates simiens. Ann. Mus. r. Afr. centr., Zool. Park shall not die! Swara. East African Wildlife 264: 1-250. Society 22: 48-49. COTTERILL, F. P. D. 2002a. Notes on mammal collec- HUGHES, B. 1978. Latitudinal clines and ecogeography tions and biodiversity conservation in the Ikelenge of the West African Night Adder, Pedicle, Mwinilunga District, northwest Zambia. (Hallowell, 1842), Serpentes, Viperidae. Bull. Inst. Occ. Publ. in Biodiver. 10: 1-18. Bulawayo: Fond. Afr. Noire 39 (A): 356-384. Biodiversity Foundation for Africa. HUGHES, B. 1985. Progress on a taxonomic revision of COTTERILL, F. P. D. 2002b. A new species of horseshoe the African Green Tree Snakes (Philothamnus bat (Microchiroptera: Rhinolophidae) from south- spp.). Proc. Intern. Symp. Afr. Vertebr. Bonn: 511- central Africa: with comments on its affinities and 530. evolution, and the characterization of rhinolophid HUGHES, B.1997. Dasypeltis scabra and Lamprophis species. J. Zool., Lond. 256: 165-179. fuliginosus — two pan-African snakes in the Horn COTTERILL, F. P. D. 2003. Geomorphological influences of Africa: a tribute to Don Broadley. Afr. J. on vicariant evolution in some African mammals in Herpetol. 46:68-77. the Zambezi Basin: Some lessons for conservation. HUGHES, B.& E. WADE. 2002. On the African leopard Proceedings of the Ecology and Conservation of whip snake, Psammophis leopardinus Bocage, Mini-antelope: An International Symposium on 1887 (Serpentes: Colubridae), with the description Duiker and Dwarf Antelope in Africa. A. Plowman of a new species from Zambia. Bull. nat. Hist. Mus. (Ed.). Filander Verlag: Füürth. pp. 11-58. Lond. 68:75-81. COTTERILL, F. P. D. in press a. Drainage evolution in JACOBSEN, N.H.G. & D.G. BROADLEY. 2000. A new south-central Africa and vicariant speciation in the species of Panaspis Cope (Reptilia: Scincidae) Swamp-dwelling weaver birds and Swamp fly- from southern Africa. Afr. J. Herpetol.49: 61-71. catchers. The Honeyguide. LAURENT, R.F. 1950. Reptiles nouveaux des COTTERILL, F. P. D. in press b. Taxonomy and conser- Kundelungu. Rev. Zool. Bot. Afr. 43: 349-352. vation importance of particular birds occurring in LAURENT, R.F. 1954. Reptiles et Batraciens de la region Katanga (southern Congo basin) and its environs. de Dundo (Angola) (Deuxième note). Publçoes The Honeyguide. cult. Co. Diam. Angola 23: 35-84. COTTERILL, F. P. D. submitted. The Upemba lechwe: an LAURENT, R.F. 1955. Diagnoses préliminaires de antelope new to science which emphasizes the con- quelques Serpents venimeux. Rev. Zool. Bot. Afr. servation importance of the Upemba-Kundelungu 51: 127-139. ecosystem in the southern Congo Basin. Anim. LAURENT, R.F. 1956. Contribution à l'Herpétologie de la Cons. Région des Grands Lacs de l'Afrique centrale. I. DAPPER, M. DE. 1981. Geomorphologische studie van Généralités - II. Chéloniens - III. Ophidiens. Ann. het plateau complex rond (Shaba - Zaire). Mus. roy. Congo Belge, Sér. 8vo, Sci. Zool. 48: 1- Verh. Kon. Acad. Wet. Lett. Sch. K. Belg. Kl. Wet.: 390, pl. i-xxxi. 43: 11-204. LAURENT, R.F. 1964. Reptiles et Amphibiens de DAPPER, M. DE. 1988. Geomorphology of the sand-cov- l'Angola. (Troisième contribution). Publçoes cult.

60 BROADLEY & COTTERILL — Reptiles of southeast Katanga

Co. Diam. Angola 67: 11-165. genetic analysis of the genus (Serpentes, LINDER, H. P. 2001. Plant diversity and endemism in Dipsadidae, Boiginae). Steenstrupia 19: 129-196. sub-Saharan tropical Africa. J. Biogeog. 28: 169- ROUX-ESTÈVE, R. 1974. Révision systématique des 182. Typhlopidae d'Afrique: Reptilia - Serpentes. Mém. LOVERIDGE, A.1933. Reports on the scientific results of Mus. natn. Hist. nat. Paris, Sér. A, 87: 1-313. an expedition to the southeastern highlands of SCHMINKE, H. K. 1987. The genus Thermobathynella Tanganyika Territory. VII Herpetology. Bull. Mus. Capart 1951, Bathynellacea malacostraca and its comp. Zool. Harv. 74: 197-416, pl. i-iii. phylogenetic relationships. Revue d'Hydrobiologie LOVERIDGE, A.1942. Scientific results of a fourth expe- Tropicale 20: 107-112. dition to forested areas in east and central Africa. SEBAGENZI, M. N. & K. KAPUTO. 2002. Geophysical IV Reptiles. Bull. Mus. comp. Zool. Harv. 91: 237- evidence of continental breakup in the southeast of 373, pl. i-vi. the Democratic Republic of Congo and Zambia LOVERIDGE, A.1956 (1957). On a third collection of rep- (Central Africa). EGU Stephan Mueller Spec. Publ. tiles taken in Tanganyika by C.J. Ionides, Esq. Series 2: 193-206. Tanganyika Notes & Rec., No. 43: 1-19. SHARP, R. R. 1956. Early Days in Katanga. Privately LOVERIDGE, A.1958. Revision of five African snake Printed Memoir, Hillside, Bulawayo. genera. Bull. Mus. comp. Zool. Harv. 119: 1-198. SYMOENS, J.-J. 1963. Le Parc National de L'Upemba. LUBINI, C.A. 1994a. Kundelungu. Zaire. Pp. 188-189. Son histoires on intérêt. Pp. 6:43-55. In Colloque In Centres of Plant Diversity. A Guide and Strategy sur les Problèmes Biogéographiques du Parc for their Conservation. Volume 1. WWF & IUCN, National de L'Upemba. Publications de l'Universite IUCN Publications Unit, Cambridge. d'Elisabethville, Elizabethville. LUBINI, C.A. 1994b. Upemba National Park. Zaire. Pp. TEUGELS, G.G. & J-F. GUÉGAN.1994. Diversité 190-191. In Centres of Plant Diversity. A Guide and biologique des poissons d'eaux douces de la Basse- Strategy for their Conservation. Volume 1. WWF & Guinée et de l'Afrique Centrale. Ann. Mus. r. Afr. IUCN, IUCN Publications Unit, Cambridge. Centr., Zool. 275: 67-85. MALAISSE, F. 1997. Se Nourrir en Forêt Claire TREWAVAS, E., 1983. Tilapiine fishes of the genera Africaine: Approache écologique etnutrionnelle. Sarotherodon, Oreochromis and Danakilia. British Presses Agronomiques de Gembloux, Gembloux. Museum of Natural History, London, UK. MEAVE, J., M. KELLMAN, A. MCDOUGALL & J. ROSALES. WHITE, F. 1983. The vegetation of Africa. 356 pp. + 3 1991. Riparian habitats as tropical forest refugia. maps. UNESCO, Paris. Global. Ecol. Biogeog. Letters 1: 69-76. WHITE, F. & M. J. A. WERGER. 1978. The Guineo- PASTEUR, G. 1964. Recherches sur l'evolution de Congolian transition to Southern Africa. Pp. 599- Lygodactyles, lézards afro-malagaches actuels. 620. In M. J. A. Werger (Ed.). Biogeography and Trav. Inst. sci. Chérifien, Sér. zool., No. 29 1-132, Ecology of Southern Africa. W H Junk, The Hague. pl. i-xii. WITTE, G.-F. DE 1933. Reptiles récoltés au Congo Belge POLL, M. 1963. Zoogéographie ichthyologique du cours par le Dr H. Schoutenden et par M. G.-F. de Witte. supérieur du Lualaba. Colloque sur les Problèmes Ann. Mus. Congo Zool. (1) 3: 55-100, pl. i-v. Biogéographiques du Parc National de L'Upemba. WITTE, G.-F. DE 1936. Description d'une espèce nou- Publications de l'Universite d'Elisabethville 6: 95- velle du genre Ablepharus provenant du Congo 106. Belge. Rev. Zool. Bot. Afr. 29: 29-33. RASMUSSEN, J.B. 1986. On the taxonomic status of WITTE, G.-F. DE 1953. Reptiles. Explor. Parc natn. Dipsadoboa werneri (Boulenger), D. shrevei Upemba, Miss. G.-F. de Witte. 6: 1-322, pl. i-xli. (Loveridge), and Crotaphopeltis hotamboeia WITTE, G.-F. DE 1966. Introduction. Pp. 1-122. In kageleri Uthmöller (Boiginae, Serpentes). Exploration du Pare National de I'Upemba, Mission Amphib.-Reptil. 7: 51-73. G.F. De Witte 1946-1949. Fasc.1. Institut des Parcs RASMUSSEN, J.B. 1993. A taxonomic review of the Nationaux du Congo, Bruxelles. Dipsadoboa unicolor complex, including a phylo- WOLANSKI, E. 1998. Upemba dying. Nat. Parks J. 42: 8.

Received: 23 September 2003; Final acceptance: 1 March 2004.

61

View publication stats