S. Afr. I. Zool. 1995,30(3) 91

The tortoises (Testudinidae) and terrapins (Pelomedusidae) of southern Africa: their diversity, distribution and conservation

WR. Branch- Port Elizabeth Museum, P.O. Box 13147, Humewood, 6013 Republic of Soulh Africa GA Benn and A.T, Lombard Percy FitzPatrick Inslilute of African Ornilhology, University of Cape Town, Rondebosch, 7700 Republic of South Africa

Received 15 February 1995; accepted 21 April 1995

Southern Africa has the richesl diversity of land tortoises in the world, as well as an imporlant radiation of pelomedusid terrapins. Total species richness has two epicentres, including the Transvaallowveld and adjacent KwaZuluJNatal (owing to the prevalence of pelomedusid terrapins) and the Eastern and south-western Cape (owing to small testudinids). The area encompassing Lesolho, Transkei and adjacent regions, lacks testudinids for unknown reasons. Archaeological data indicates that this gap is natural, and not the result of man-induced extinctions. Endemic species are clustered in the Cape, whilst the few threatened species are more widely dis­ tributed. The majority of species is well protected in existing reserves. The small number of chelonian species in southern Africa and their relatively well-known distributions, test the efficacy of an iterative reserve selection algorithm. The presence of many allopatric (or nearly so) congeneric species leads to the selection of ~erative reserves that protect peripheral populations. To avoid this, marginal records and isolated, peripheral populations should be excluded from the analysis.

Suidelike Afrika het die rykste diversiteit van grondlewende skilpaaie ter wereld, sowel as 'n belangrike uitstra­ ling van pelomedusiede varswaterskilpaaie. Die totale spesierykheid het twee episentrums, naamlik die Trans­ vaalse laeveld en aangrensende KwaZuluJNatal (te danke aan die oorwig van pelomedusiede varswaterskil­ paaie) en die 005- en suid-westelike Kaap (te danke aan die klein testudiniede soorte). Die gebied wat Lesotho, die Transkei en aangrensende streke insluit, het om onbekende redes geen testudiniede nie. Argeologiese data wys dat hierdie leemte natuurlik is, en nie die resultaat van mens-veroorsaakte uitstelWings. En-demiese spe­ sies is in die Kaap gekonsentreer, terv.tyl die enkele bedreigde spesies wyer versprei is. Die meeste spesies word binne bestaande bewaringsgebiede goed beskerm. Die klein hoeveelheid cheloniese spesies in suidelike Afrika en hulle betreklik bekende verspreidings, het die doeltreffendheid van die herhalende bewaringsgebied­ seleksieprosedure getoets. Die teenwoordigheid van baie allopatriese (of byna) gelyksoortige spesies het veroorsaak dat bewaringsgebiede wat omliggende aHopatriese populasies beskerm, geselekteer word. Om dit . ) te voorkom, behoort marginale aantekenings en ge'isoleerde omringende populasies van die analise uitgesluit te 9

0 word. 0 2

d *To whom correspondence should be addressed e t a d (

r Introduction & Hillis 1990) can be expected to intensify the recognition of e h taxa with limited geographical range, resulting in new s Lombard (1995) has noted the likely South African contrac­ i l descriptions as well as the revival of taxa previously syno- b tual obligations to the Convention on Biological Diversity. u

P This will require the country to develop national strategies, e plans or programmes for the conservation and sustainable use h t RATE OF DESCRIPTION OF TAXA of biological diversity. The lack of a national database that y HERPETOFAUNA b

inventories the distribution and diversity of the fauna and 350,---- d e nora of South Africa threatens this commitment. This is par­ t n ticularly true of South African herpetology, which is presently 300 a r g undergoing somewhat of a renaissance, with burgeoning 250 e

c interest in the diversity of reptiles and amphibians. It has n .8 200 e resulted in the recent discovery of numerous new taxa, includ­ c § i l ing obvious novelties (e.g. Bourquin 1991; Mouton & van z 150 r e Wyk 1994; Broadley 1994). The rate of description of new 100 d

n southern African herpetological taxa shows no evidence of u decline since Smith's (1838-49) founding studies (Figure 1), y a and new reptiles and amphibians are currently being discov­ w e ered or described at a rate of about one new species per t a month. Despite our incomplete knowledge of herpetofaunal G

- Amphibians ...... Che\onians .. __ .. Snakes Uzards t diversity in the subcontinent, it is already evident that reptiles e n

i are exceptionally speciose and form a major component of Figure 1 Rate of description of currently recognized southern Afri­ b

a southern African vertebrate diversity, particularly if only can herpetofauna. Note that the rate of deSCription of new chelonians S endemic srecies are considered. It should be noted that is in the plateau phase indicating that local diversity is relatively y b

increasing acceptance of evolutionary species concepts (Frost well-known. d e c u d o r p e R 92 S. Afr. Tydskr. Dierk. 1995,30(3)

nymized under wide-ranging species (c.g. Broadley 1993). (i) Where are the centres of diversity for With the exception of crocodilians (23 species) and rhyn­ (a) tortoises and terrapins cocephalians (2 species), chelonians are the least speciose of (b) endemic species, and extant reptiles, and two main suborders survive. Pleurodirans (c) threatened species? are primitive aquatic chelonians that withdraw the head side­ (ii) Are southern African tortoises and terrapins adequately ways. They have a Gondwanaland distribution and are repre­ protected in existing reserves? sented in Africa only by the Pelomedusidae. Cryptodiran (iii) Where are the 'ideal' reserves for protecting southern chelonians withdraw the head straight backwards, and include African chelonians? all land tortoises (family Testudinidae), sea turtles, and (iv) What new reserves are required to fully protect southern aquatic New World terrapins including soft-shelled triony­ African chelonians? chids. Excluding the Amazonian genera Podocnemis (six spe­ An analysis of habitat association and zoogeography of cies) and Peltocephalus (one species), pelomedusid southern African chelonians will be presented elsewhere plcurodirans are restricted to Madagascar (Erymnochelys, one (Branch & Benn in prep.) species) and Africa (Pelusios, 14 species; Peiomedusa, one species). Terrestrial testudinids have a long evolutionary his­ Methods tory, with fossils dating from thc Middle Eocene (Auffenberg Taxonomy 1974), and an almost cosmopolitan continental distribution in Chelonian diversity for the region was compiled from the temperate and tropical habitats, with a limited colonization of reviews of Branch (1989), Boycott (1989), Broadley (1989, oceanic islands (e.g. the Aldabra Atoll, Indian Ocean, and the 1993), and Iverson (1992). Analysis was restricted to the fam­ Galapagos Islands, eastern Pacific Ocean), but excluding ilies Testudinidae and Pelomedusidae. Owing to their marine Australia. However, although numerous fossil species are or marginal occurrence, sea turtles and fresh-water trionychid known (Auffenberg 1974), the extant fauna is relatively de­ terrapins have been excluded. The latter have only two repre­ pauperate and currently only 42 spccies, in 11 genera, are rec­ sentatives in southern Africa. The Nile soft-shell terrapin (Tri­ ognized (Iverson 1992; Broadley 1993). The surviving spe­ onyxs triunguis) occurs only in the lower Cunene River, cies are also patchily distributed and sympatry between even downstream of the Ruacana Falls, whilst the Zambezi soft­ two species is rare. shelled terrapin (Cycloderma Jrenatum) is known from the It has been noted on many occasions (e.g. Greig & Burdett Sava River in Mozambique and the Sabi-Lindi river conflu­ 1976; Branch 1988b, 1989) that southern Africa has a sur­ ence in eastern Zimbabwe (Branch 1988a). Even though five prisingly rich chelonian fauna. Although comprising only species of sea turtle have been recorded from South African 0,8% of the world's total land surface, South Africa is home

. coastal waters (Branch 1988a), most are considered vagrant )

9 to 13 species of tortoise (31 % of living species), whilst or marginal vistors. In contrast to the global situation, where 0

0 another (Homopus sp.) occurs in adjacent Namibia and may most sea turtles are threatened by exploitation or indirect 2 yet be recorded within the Richtersveld area. Generic diver­

d mortality (Frazier 1992), they are considered well protccted e t sity is even higher, with five (45,5%) of the eleven recognized in South Africa. Only two species, the loggerhead turtle a d genera present. The diversity of other chelonians in the sub­ (Caretta caretta) and leatherback turtle (Dermochelys cori­ (

r continent is less remarkable, but by no means insignificant. It caea), are known to utilize Maputaland beaches for breeding e

h includes two soft-shelled trionychids, six side-necked pleu­

s (Hughes 1974a, b). These fall in well-protected areas and the i l rodirans, and five sea turtles (Branch 1988a; Broadley 1993). breeding populations of both are rising. Isolated records of b u Although part of this diversity has been summarized in recent nesting leatherbacks on Eastern Cape beaches (Mullins 1984) P popular reviews (e.g. Branch 1988a; .Boycott & Bourquin e may reflect expansion of the Maputaland colony, as this spe­ h

t 1988), the detailed diversity, distribution and conservation

cies is known to show the least site fidelity to its natal y status of chelonians in the region has not previously been

b beaches.

d reviewed. Subspecies have been described for only a few southern e t Before an objective analysis of the conservation status of African chelonians, although trinomials are required for some n a

r any group can be undertaken it is necessary to have a sound species with extralimital races: Kinixys belliana nogueyi is g understanding of its taxonomic diversity and distribution. restricted to West Africa, whilst the typical race is found in e c These criteria are not easily met for southern African reptiles. East Africa and the subcontinent (Broadley 1993); Pelus;us n e c Lombard, Nicholls & August (1995) applied an iterative subniger parietalis and Pelusios castanoides intergularis are i l reserve selection algorithm to South African snakes, using the r restricted to the Seychelles, with the typical races of both spe­ e

d base maps of Broadley (1990). They concluded that most of cies occurring in the subcontinent (Bour 1983). Chelonians n the snake species in South Africa were adequately protected. with subspecific differentiation within the subcontinent u

y However, knowledge of the current composition and distribu­ include the following. a

w tion of other southern African squamates, particularly lizards, Marsh terrapin (Pelomedusa subrufa) e t is so poor that it is impossible to objectively assess their In a footnote, Bour (1986) provisionally recognized two a

G diversity or conservation status. Fortunately, chelonians are races within southern Africa. P. s. nigra, is considered to be t e better known with relatively well-documented distributions. restricted to KwaZulu-Natal, Free State and adjacent East­ n i The rate of description of new taxa has stabilized, relative to ern Capc (Iverson 1992), whilst typical P. s. subrufa b a other groups in the region (Figure 1), and they thcrcforc offer extends from Somalia and Ghana through the central and S

y a suitable reptilian study group. Our studies, detailed below, western regions of the subcontinent, to Cape Town. A fuller b address the following questions: analysis of the situation has not been published, and the d e c u d o r p e R S. Afr. J. Zool. 1995.30(3) 93

matter requires further attention. many of their chelonian rccords have been incorporated Mountain tortoise (Geochelolle pardaJis) from the published literature. The lack of the remainmg Broadley (1989) continues to recognize a dwarf northern material is not believed to have seriously limited the follow­ race (G, p. babcocki) extending throughout most of eastern ing analysis. and southern Africa; typical G. p. pardalis is considered 10 A number of unusual complications may attend tortoise he restricted to the Eastern Cape, with relict populations on records in museum collections. Tortoise shells formed a com­ the southern Namihian escarpment. This arrangement has mon cultural artefact of indigenous peoples (e.g. 'huccu' been questioned by Boycott & Bourquin (1'1~~) and Bauer. pouche~), and as such were often acquired hy early European Branch & Haake (1993). rhe latter considering thc large explorers and suhsequently deposited in museum collections, size of the 'typical' race to he an ccotypic response to the The accompanying locality data may be very vague (e.g. the relatively moist and temperate climate of the Eastern Cape type localities of: Psarnmobates oculifer. Cape; G. pardalis. regIOn. Cape of Good Hope) or retlect the locality of the individual Speckled padloper (Homopus signatus) from which it was acquired. Another common complication is Boycott (1986) validated the southern race. H. s. peersi. that of human trans1ocations. The popular appeal of tortoises that was later (Bour 1988) placed in the synonymy of H. s. results in them often being illegally collected and subse­ cafer. Although both races appear well defined morpholog­ quently released in localities distant from their natural distri­ ically. there is a large area of intcrgradation in the ealvinia­ bution. The distrihution maps of Greig & Burdett (1976) LOCTicsfontcin region, at the junction of Western Mountain renect this with the frequent use of "?' symhols. and such a Karoo and Succulent Karoo vegetation types. translocation misled Pooley (1'165) into including the angu­ Karoo tent tortoise (Psammohates tentorius) late tortoise (Chersina angulata) as part of the Natal herpeto­ Following their monographic revision. Loveridgc & Wil­ fauna. Many questionable records were deleted from the liams (1957) reduccd all 16 taxa dcscrihed by Hewitt (I '133. analysis to avoid possihle skewing of reserve selection. 1934) within this complex to synonymy, and recognized The presence of tortoises and terrapins in protected areas only three races (P. t. tentorius, P. t. trimelli and P. t. ver­ was determined from published and unpublished sources (see roxii), all of which are restricted to the Cape provinces and references in Table 1). Threatened species were taken to be adjacent Namibia. Nonetheless the races remain poorly those listed in (he most recent South African Red Data Book defined, with massive areas of intergradation (see Greig & (RDB) - Reptiles and Amphihians (Branch 1988h). Burdett 1'176). The complex is in urgent need of a modcrn Data analysis for ilerative reserve (lR) selection, congru­ taxonomic revision, and may yet he found 10 contain cryp­ ence with existing protected areas, and hotspot analy~is were tic taxa. performed with the geographic information system (GIS) . )

9 ARCJINFO vcr. 0.1.1 (Environmen[al Systems Research Insti­

0 Distribution data base

0 [ute. Redlands. California) using the methodology detailed in 2 Although for many specimens point locality data were availa­ Lombard (1995). On somc occasions, analyses were rc-run d e

t hIe, a signiticant numher of early records had less detailed with the removal of marginal records or those of questionable a

d provenance. Analysis of distribution was therefore limited to status. (

r quarter-degree grid squares (QDS). The data base was derived e h from a variety of sources, including the following. Results and Discussion s i l (a) Puhlished litera[ure b Centres of diversity in southern Africa u A number of recent provincial and national herpetofaunal P

All tortoises and terrapills

e surveys have puhlished up-dated point locality maps of the h t chelonian fauna (e.g. Transvaal - Jacohscn 198'1; Orange Generalized range maps for all southern African chelonians

y Free State - Bates 1'1'12; Swaziland - Boycott 1'1'13; Bot­ have been included in several recent popular guides to the b

d swana - Auerbach 1987). These have been compiled into subcontinent's reptiles (Branch 1988a; Boycott & Bourquin e t the current data hase. Recent taxonomic revisions (e.g. IYHH). However. only one previous attempt has been made to n a Pelusi{)s, Broadley 1'181; and Kinixys. Broadley 1'1'13) were prepare detailed point locality maps for the region's cheloni­ r g also incorporated. ans (Greig & Burdett 1976). This. however. was restricted [0 e c (b) Museum collections land tortoises and did not include freshwater terrapins (triony­ n e c The extensive Cape tortoise collections made hy Greig & chids or pelomedusids) or sea turLles. i l Burdett (1'176) have heen consolidatcd into the South Afri­ The tinal data base for pelomedusids and testudinids com­ r e YY7 d can Museum (SAM) and Port Elizabeth Museum (PEM). prises a total of I records. Individual species maps will be n both of which are fully computerized and readily accessi­ presented elsewhere in a fuller analysis of hahitat association u

y hIe, and already contained significant chelonian collections. and zoogeography of southern African chelonians (Branch & a

w In addition. the tortoise holdings of the Albany Museum. Benn in prep.). Geographic coverage of ehelonians in south­ e t including the historically important collections of Hewitt ern Africa is not uniform, and large areas have few or no che­ a

G and Duerden. have also been incorporated into the PEM lonian records (Figure 2). For the central Kalahari regions and t

e collection. Owing to the paucity of Cape terrapin records, Namihia this renects, in part, poor collecting. although it is n i additional records from the Transvaal Museum were also also ohvious that chelonian diversity is low in these regions. b a included. Although a numher of major collections were not The paucity of records from Zimhahwe is an artefact as the S

y funy surveyed (e.g. the Transvaal Museum, Pretoria, and large herpetological collections amassed by Dr Broadley in b the Natural History Museum of Zimbabwe. Bulawayo), Zimbahwe and housed in Bulawayo were not incorporated. d e c u d o r p e R 94 S. Afr. Tydskr. Dierk. 1995,30(3)

Table 1 Chelonians recorded within major reserves in 52, Happy Rest NR; 53. Langjan NR; 54, Messina NR; 55, Nylsvley NR; 56, southern Africa Percy Fyfe NR; 57, Abe Bailey NR; 5H. Barberspan NR; 59, BlocmhofDam NR; 60, Boskop Dam NR; 61, Hartebeestpoort Dam NR; 62. Marievale NR; Species Reserves Reference 63, Roodeplaat Dam NR; 64, Rustenburg NR; 65, Rust-der-Winter Dam NR; Ge(lchei(me pardalis 1,2,3,4,6,10,11, 12, 13, 14 a, b, c, d, 66, S,A. Lombard NR; 67, Vaalkop Dam NR; 68, Wolwespruit NR Free Stale Resen'es. 69. Sandveld NR. 15,26,27,28,36,37,38,41,42, f, i, j, k, Botswana Reser~'es. 70. Moremi WR; 71. Chobe NP. 44,45,48,49,50,51,52,53,54, I. p, q. v, Swaziland Reserves. 72, Hiane GR; 73. Mlawula NR; 74, Ndzindza NR; 75, Mbulu:li GR. 55,56,58,59,61,63,64,65,66, W, x Zimbabwe Reserves. 76. Gonarezhou NP; 77, Chizariri NP; 78, Lake Kyle 67,72,73,74,76,77,78,81,83, RP; 79, Mcllwaine RP; 80, Matobo NP; 81, Matusadona NP; 82, Ngezi RP; 84, 85, 86 83, Victoria Falls - Zamhezi NP; 84, Hwange NP; 85, Mana Pools NP. Namibia Reserves. 86, Ethosha NP; 87, Sperrgebiet Diamond Area I. Chersina anJ.:ulata 2,3,5,8,26,27,28,29,30,31, h, c, e, h, Reference legends 36,87 k,l,m,n, a, Pienaar, Haacke & Jacobsen 1983; b. Branch & Braack 1989; c. Branch &

0, y Braack 19H7; d, Grobler & Bronkhorst 1981; e. Braack 1981; f. Haacke

Humopus areola/us 1,5,8,26,30,31,32,33,34,35 c, e, h, n, 0 1984; g, Branch, unpubL obs.; h, G. Thomsen, unpubL obs.; i, Bourquin 1990; j, Burger 1993; k, Burger unpubl obs.; I, Burger & Branch 1994; m, Homopus bouienKeri 2, 36 h, k Branch. unpub!. obs; n. Hensley, pers. comm.; 0, Baard 1993a; p. Branch, HOf1!opuJ si,;natuJ 7,29 g,m unpub!. obs.; q, Jacobsen, Newberry & Petersen 1986; r, Broadley 1993; s, Honu,pus fem(m.1lis 2,4,26 (?) b, d, j Bates 1992; t, Auerbach 19H7; u, Broadley In I, v, Boycott 1993; w, Broad­ Hotn0puJ sp. ley & Blake 1979; x. Hoffmann 1989; y, Branch 1994; z, Haagner & Eis Kinixys belliarw 9,10, II, 12, 13,16, 17,24 r, Z 1986. Kinix)'s speW I, 10, 37, 40, 42, 43, 44, 50, 72, a, q, r 71,76,77,81,82,84,85 However, only two testudinids (Killixys spekii and C. parda­ Kinixys [ohmsiana 41,44,55 lis) are widespread in the country, although a number of Kinix)!s natalensis 1,14,15,45,73,74,75 pelomedusids (e_g, Pelomedusa subrufa and Pelusios sillua­ Pmmmobafes fenforiu,\' 2,26,29,36 b,j, m. p tus) are also common. P,wmmohates Despite these collection artefacts, it is evident that one Neometricus :I I, 32, 33, 34, 35 large, well-collected region is depauperate in chelonians, It ° extends through the eastern regions of the Eastern Cape Psammobutes (lcul!fer 6,53,55, 5S, 59, 66, 69 g, S (Transkei), Lesotho, the KwaZulu-Natal midlands, north­

. Peiomedusa subrufa I, 2, ], 4, 5, 6, 8, 11, 12. 13, 14, a, b, c, d. ) eastern Free State and the North West Province. The only spe­ 9 15, IS, 19,20,21,27. 2X, 36, :n, e, f, h, i, 0 cies occurring in this large region, albeit known from only a 0

2 18,40,41,43,44,46,47,48,49, k, I, p, q, few isolated records, is the Cape terrapin (Pelomedusa sub­ d e 50,51,53,55,56,57,58,59,60, v, w, x rufa), The absence of major river systems precludes other ter­ t a 61,62, 6J, 65, 66, 67, 68, 72. 76, rapins (Pelusios spp,), but it is not obvious why land tortoises d ( should be absent. Although a large part of the region is mon­ r 83,84,86 e tane grassland, the mountain tortoise (G. pardalis) occupies h 9, 10, 11, 12, 13, 14,24,37,45, a, i, II, s Pelusios sinuatus I. q. i l similar habitat in the southern Free State, and its absence in

b 49, 50, 73. 76, 77, 7H. 79, 81, 82, w, z

u the lowlands of the old Transkei region is inexplicable. Local

P 83, 84, H5 extinction, via human agency, is possible as mountain tortoise e h Pelllsio.\' subniKer 1,76,77,81,84,85 n, W are known to have been readily eaten by early man. Although t

y Pelllsios castwwides 9,10,17,24 r,w Greig & Burdell (1976) considered tortoises to be '",now b extinct in the Transkei, and nearly so in Lesotho', there is no d Pe/usios rhodesianus 9,22,23,24,25,79 i,w e t Peiusios he(-huani('us 70,71,83 t,w supportive evidence that they were either present throughout n a the region, or that any extinctions resulted directly, or even r Resenre Legends g

indirectly, from anthropomorphic factors. e National Parks, I, Kruger NP; 2, Karoo NP; 3, Addo Elephant NP; 4, Moun­ c tain Zebra NP; 5, Bontebok NP; 6, Kalahari Gem~bok NP; 7, Richter~veld Insight into the topic may be gained from studies at archae­ n e

c NP; 8. Langebaan NP. ological sites. Although tortoise remains are often noted in i l Natal Reserves. 9, St Lucia GR; 10, Ndumu GR; II, Mkuzi GR; 12, Hluh­ reports of faunal assemblages from such sites, Ihey are rarely r e luwe GR; 13, Umfolozi GR; 14, ItalaGR; 15, Weenen NR; 16, Kosi Bay NR;

d determined to species level. In a noteable exception, during n 17, Sodwana Bay NP; 18, Tugela Drift NR; 19, UmLamvuna NR; 20, Oribi study of two sites (Edgehill and Welgeluk) on the Konaap u

Gorge NR; 21, KranLkloof NR; 22, Bluff NR; 23, Charters Creek NR; 24, y Mapelane NR; 25 Umlalazi NR. River near Adelaide in the Eastern Cape, Hall (1990) a w Cape Resen'e,~, 26 Anysberg NR; 27. Andries Vosloo Kudu and Srun Knoll recorded four taxa (G. pardalis, Pelomedusa subruja, C. e t NR; 28, Thomas Baines NR; 29, Goegap NR; 30. De Hoop NR; 31, Elands· angulata and Homopus areolatus) in levels from 6000 years a

G berg PNR; 32, Voelvlei NR; 33, J.N. Briers-Louw NR; 34, Harmony Flals BP to the present. Tortoises were present in numbers

t NR; 35. Riverlands NR; 36, Karoo NR. e inversely related to their size (which may be a reflection of n i Transvaal Reserves_ 37, Blydesrivierpoort NR; 38, Bronkhorstspruit Drun the ease of transport andlor the amount of meat present on a b NR; 39. Cynthia Letty FR; 40, Jerico Dam NR; 41. Loskop Dam NR; 42, Lil­ a carcass). All four species are still common in the region. Fur­ S lie FR; 43, Nooitgedacht Dam NR; 44. Ohrigstad Dam NR; 45. Pongola NR;

y 46, Suikerbosrand NR; 47, Vaal Dam NR; 48, Doomdraai Dam NR; 49, ther north, on the border of the area presently lacking tor­ b Fanie Botha Dam NR: 50, Hans Merensky NR; 51. Hans Strydom Dam DR; toises, WeI bourne (in Derricourt 1977) analysed faunal d e c u d o r p e R S. AfT. J. ZooL 1995,30(3) 95

.~: •

• •

D 1 -2 .3-4 • 5 . ) 9 0

0 Figure 2 Density map showing areas of land tortoise and pelomedu-sid terrapin species richness in southern Africa. 2 d e t a

d remains from a late Stone Age site at Oakleigh, near Queens­ ment (Yates, pers, comm). (

r town, recording Testudo sp. (probably Geochelone) shell frag­ The only Lesotho site from which tortoise records are too e h ments only from recent strata (levels 3 and 4, 500-400 years numerous to be artefacts is Ntloana Tsoana in north-western s i l BP). The paucity and even absence of chelonian remains in Lesotho (29" 19'5; 27"49'E) where 5H unidentified tortoise b u archaeological sites in the tortoise-free area is confirmed by fragments were collected from four levels dated to between P

e other studies. At the Sehonghong Rockshelter (29·"44'S; HnO±HO BP and 1211O±120 BP (Mitchell 1993b). Their h t 29"47'E, Qacha's Nek District, western Lesotho) only a single identity was undetermined, however, and they may be refera­ y tortoise fragment was noted in a layer dating from 1400±50

b ble to the terrapin Pelomedusa subrufa, which is known from

d BP. No tortoise remains occurred in older layers, c. 7000 to the area, In the eastern Free State. Plug & Engela (1992) e t 32000 BP At two other large sites in eastern Lesotho no tor­ recorded no chelonians in their detailed study of the Rose n a toises were recovered from levels broadly contemporary with Cottage Stone Age site near Ladybrand, whilst Esterhuysen, r g the upper Sehonghong spcCimcn (Carter, Mitchell & Vinni­ Behrens & Harper (1994) recorded only a few tortoise pieces e c combe 1988). Similarly, only two tortoise fragments were from the Leliehoek Shelter, another Holocene sequence from n e

c found from a layer datcd to 6140±IOO BP at Tloutle (29"28' the eastern Free State. Unfortunately, the latter are not specifi­ i l S; 27"46'E) in north-western Lesotho (Mitchell 1993a), and cally identified. Together these archaeological records sup­ r e only a single fragment from 260 to 330 BP was found at port the historical absence of tortoises in the region until at d n Hololo Crossing (28" 44'S; 28"27'E) (Mitchell, Parkington & least 32000 BP, and do not support Greig & Burdett's (1976) u

y Yates 1994). In the samc paper, a fairly large faunal assem­ contention that tortoises in the region were exterminated. a blage from the site Bolahla (30"04'S; 28"24'E; at the conflu­ Studies on archaelological sites in lowland regions of the w e t ence of the Bolahla River and the SenquJOrange River) Transkei, however, are required to confirm the situation a & G contained no tortoises at all. The few tortoise remains in all throughout the region. Although Greig Burdett (1976) also t

e these sites may be San hunter-gatherer artefacts as tortoise considered the mountain tortoise to ' ... have been extermi­ n i carapaces were used as bowls and a fragment from such an nated in the western Cape coastal areas, largely for food', this b a artefact would be indistinguishable from an animal eaten and again remains conjecture. Cheniina angulata remains are S

y discarded. Artefacts could be traded in from quite a distance common in archaeological sites from the western Cape, but b and thus not be representative of the immediate site catch- G. pardalis has not been identified (Klein, pers. comm.), and d e c u d o r p e R 96 S. Afr. Tydskr. Dierk. 1995,30(3)

there is therefore no confirmation that the species was previ­ K. spekii, and Pelusios spp.). Their exclusion lcavcs only two ously present. endemic testudinids (Kinixys lobatsiana and K. natalensis), Within the subcontinent two general epicentres of chelo­ which have mutually exclusive ranges in the west and east of nian diversity are evident (Figure 2). The first, centred in the the Transvaal, respectively. The two species arc not sister taxa low veld regions of the Eastern and Northern Transvaal and within Kinixys; K. lobatsiana shows affinities with K. belli­ northern Maputaland, includes both tortoises and terrapins. ana and the Central and West African forest species K. emsa Previous analysis (Greig & Burdctt 1976) considered the and K. homeana, whilst K. natalensis is more similar to K. region relatively depauperate in testudinid diversity, but the spekii (Broadley 1993). reccnt study by Broadley (1993) recognizes four species of Within the Cape provinces, the ranges of the three Psam­ southern African hingeback tortoises whereas only a single mobates species are mainly exclusive, with P geometricus species was thought to bc present by Greig & Burdett (1976). restricted to the south-western Cape, and P ocullfer occurring Both the Kruger National Park and Northern Transvaal locali­ north of the Orange River. The geometric tortoise lives in ties with five taxa include the two common pelomedusids well-established sympatry with two other tortoises (c. angu­ (Pelomedusa subrufa and Pelusios sinuatus). Only a single lata and H. areolatus) in a number of sites in the southern peJomedusid enters the Cape provinces, and the high tortoise Wcstern Cape (Figure 3; Beard 1990). The only well-estab­ diversity in this second epicentre results from a dramatic radi­ lished region of sympatry within Psammobates occurs ation in testudinids. Of these, only the mountain tortoise is between P telltorius and P oculifer in West Griqualand not endemic to the subcontinent. Numerous QDSs record the (Power 1932). As Grieg & Burdett (1976) noted, the record of presence of five taxa, but the Cape terrapin is present in all P. oculifer from the Aus region (Mertens 1955), where it with varying arrangements of four testudinids. These are dis­ would occur in sympatry with P t. verroxii, should be treated cussed in fuller detail in the following section. with circumspection, as no recent records confirm the pres­ ence of this species in the area. Similarly, a number of other Endemic species species also reputed to have been collected at Aus (Mertens Chelonian endemicity is mainly restricted to the Cape region 1955), e.g. H. signatus and C. angu[ata, have not been subsc­ (Figure 3), where the testudinid genera Chersina, Homopus quently collected in the region. and Psammobates have their evolutionary centre. Along with The angulate tortoise (c. angu[ata) is distributed in a wide Madagascan Pyxis, these are considered the most derived tes­ belt through coastal habitats from East London to the Orange tudinid lineages (Gaffney & Meylan 1988). The rich chelo­ River, with scattered inland populations. These probably rep­ nian diversity in the Kruger National Park region comprises resent relict populations surviving in moister, more vegetated wide-ranging species with tropical affinities (e.g. G. pardalis, habitats associated with localized high rainfall in the escarp- . ) 9 0 0 2 d e t a d ( r e h s i l b u P e h t y b d e t n a r g e c n e c i l r e d n u y a w o e t a 112 G t e n 113 i b a S y b Figure 3 Density map showing areas of endemic land tortoise and pelomedusid terrapin species richness in South Africa. d e c u d o r p e R S. Afr. J. Zool. 1995, 30(3) 97

ment mountains (Branch 1990). The species occurs in high chelonian fauna including six species (Branch & Braack densities in many Cape coastal regions (Branch 1984), and is 1989). These are identical to those found around Cradock, commonly translocated a, a pet (e.g. Pooley 1965). Many iso­ with the exception that H. areloatus does not occur as a lated northern records in Namibia (Greig & Burdett 1976) mesic-adapted relict, but is replaced on the dolerite outcrops probably arise in similar fashion, and the most northerly con­ by Hornopus boulengeri. The presence of another relict popu­ firmed record is within the northern part of the Obib dune sea lation of H. areolatus in the poorly surveyed region on the in the southern Sperrgebiet (Branch 1994). The species has Roggeveld Mountains around Middelpos (Greig & Burdett also been introduced and subsequently formed large popula­ 1976), may reflect another minor centre of diversity. tions on a number of offshore islands, e.g. Dyer and Dassen Two specimens of H. boulengeri (PEM specimens 3976 & islands (Branch 1991). 3981) from the Hantam Mountains, north-west of Calvinia With the exception of Griqualand Psammobates, the only (3119BC) represent new distribution records for the species other regions of sympatry between congeneric tortoises and are the first records of sympatry between this species and involve padlopers (Homopus spp.). Greig & Burdett (1976) the speckled padloper (H. signa/us). Greig & Burdett (1976) note the extension of the range of the parrot-beaked tortoise discuss confusion over a H. boulengeri specimen doubtfully (H. areolatus) inland into the Cradock region, where it over­ recorded from Piketberg (3218DC), whilst Mertens' (1955) laps the range of the greater padloper (H. femoralis). How­ report of both species from the Aus area in southern Namibia ever, these species do not occur syntopically, H. femoralis can also be discounted. The records of H. boufengeri are now being restricted to montane grassland on the summit plateaus known to be referable to a new species (Branch 1992), whilst whilst H. areolatus occurs in relict populations inhabiting the the H. signatus specimen appears to have been obtained from elsewhere. Reported sympatry between H. signatus and H. well-vegetated lower slopes. The presence of C. angulata. boulengeri in the Calvina region is of importance, as these Psammobates tentorius. C. pardalis and Pelornedusa subrufa sister species are sometimes placed in a separate subgenus in the region means that it is theoretically possible to find up Chersobius (Cooper & Broadley 1990), and have similar to six chelonians in a single QDS, although only five have as ecologies. The presence of C. angulata and P. tentorius in the yet been recorded. A similar mixture of relict, montane, and Hantam Mountains indicates another minor centre of diver­ mesic- and xeric-adapted species occurs in the Karoo sity in the western escarpment region. National Park, Beaufort West. The park straddles the arid southern plain of the Great Karoo, and the foothills and sum­ Threatened species mit plateau of the Nuweveld Mountains. The mosaic of habi­ tats generated by edaphic and elevational variation, including A map of the distributions of RDB species (Figure 4) shows a

. localized high rainfall on the summit plateau, supports a rich single 'hotspot' in Natal. This results from the inclusion of ) 9 0 0 2 d e t a d ( r e h s i l b u P e h t y b d e t n a r g e c n e c i l r e d n u y a w e t a G t e n i b a S y b Figure 4 Density map showing areas of threatened land tortoise and pelomedusid terrapin species richness in South Africa. d e c u d o r p e R 98 S. Afr. Tydskr. Dierk. 1995,30(3)

two terrapins (PeJusios casfal10ides and P rhodesiallus). restricted to the the Pieketberg-Klawer region (Boycott] 986). However, both are listed only as 'Peripheral' in the RDB and The conservation of the geometril: tortoises has been well have large distributions in Mozambique, and Mozambique addressed (Beard 1990, 1993a, b), and it is well protected in and northern Zimbabwe, respectively, throughout most of all possible existing reserves in the south-western Cape. which they are unthreatened. The remaining species. Psam­ mobates geometricus (Endangered), K. natalensis (Rare) and Southern African tortoises and terrapins in existing Homopus signatus cafer (Restricted), have non-overlapping reserves ranges. Analysis of the distributions of all species, including The results of a survey of tortoises and terrapins in existing RDB species, was performed at the species level resulting in protel:ted areas in southern Africa is summarized in Table], the full range of H. signatus being illustrated. In fact, the face Relatively few protel:ted areas have prepared faunal invento­ H. signatus cafer has a much more restricted range and is ries of the species they are designed to protect. Siegfried (1989) notes the presence of 582 statutory nature reserves in South Afril:a alone, but we were able to lind reports (pub­ Table 2 The eight iterative reserves that protect all lished and unpublished) of chelonians in only 87 southern Southern African land tortoises and pelomedusid terra- African reserves (i.e. less than 15%). Despite this poor docu­ pins mentation, however, it is evident that most spel:ies are already Number QDGS Species protected recorded from at least one exisiting protected area (Table I). Ideally, for maintenance of genetic diversity and protection 2!LUAD Kinixys bellwna (belliana), Pelusios rhodesianus, Pelu.\"/{!J cmranoides, Pelusil!s simlatus from localized disasters (natural and man-induced), all spe­ cies should be protected in a number of separated reserves 2 2716DD Cher.~ina anJ.:ulma, p.wmmobates lentOrlU,f (verroxii), (populations). Tortoises and terrapins recorded from only a Homopu,\' sp. nov ('berKeri') few reserves (number shown in brackets) include: H. boul­ 3 3319CH Homopus arco/muJ, P,mmmobare.\' J.:eomelrica.I', engeri (2): H. femoralis (3); Homopus sp. (0); K. lobatsiana Pe/omedusu .mbrutil (3); Psammobafes fenforius, Pelusius caslanoides (4) and P. 4 1823DA Pe/usios subn;Ker. Pelusios bechuanicus bechuanicus (3). 5 24~I])A KinixYJ spe/w, Gcochelone purduh,\' (bubcock;), Kinuys natalen.fi,f Iterative reserves 6 3119HC Homopus boulenKeri, llomopu.f ,~jJ:natu.i A total of eight iterative reserves (IRs) were determined to 7 2425DH Psammobule,f (!{'u/!ter, Kinix)'s lobarsianu represent all southern African non-trionychid chelonians at 8 2823CC 1l(lmupusfemoralis . lea... t oncc. These, along with the species they 'protect', arc )

9 Subspecies shown in hrackets detailed in Table 2, and their distribution shown in Figure 5. 0 0 2 d e t a d ( r e h s i l b u P

e ! h t y b d e t ) n a r g e c n e c i l r e d n u y a w e t a G t e n i b a S y b Figure 5 Distribution of Ilerative Reserves conserving all land tortoises and pelomedusid terrapins in southern Afri(;3, d e c u d o r p e R S. Air. J. Zoo!. 1995, 30(3) 99

Table 3 Existing reserves falling within selected Iterative IR 3 (3319CB) Reserves Although targeted for three species (Le. H. areolatus, Reserve E"isling Reserve Owner Psammobates gevmetricus, and Pelomedusa subrufa), the proposed TR is now only marginal for the endangered geo· I (2H~2A(» Mhlalul.e State Forest N:lIn! Parks Board metric tortoise. Although three existing protected areas fall Mihobi Nature Res State Forest (K waZulu) within the QDS it contains none of the four viable reserves Hell's Gate MIlitary Area National Defence Force lor this species listed by Baard (1993a). It does, however, Cape Vidal Stat!! Forest Natal Parb Hoard contain viable populations for the remaining target species, Dukuduku State Forest Natal Parks Hoard as well as the angulate tortoise (c. angulata). Mnpelane Nat. Res. Natal Parks Board IR 4 (1823DA) The IR is extralimital hut borders protected areas in north· Nyalazi State Fore~t Natal Parks Hoard ern Botswana, i.e. Moremi WR and Chobe NP, from which SI Lucia Game Reserve Natal Parks Board one of the targeted terrapins (Pelusios bechuanicus) is 2 (2716()U) No reserves recorded. Viable popUlations of the other target species, 3 (J319CB) Matroosbcrg State Forest Western Cape Nat. Cons Pelusios subniger, are protected in numerous Zimbahwe Karon Nm. Botanical Garden Nat. Botanical Gardens reserves (Tahle I). -+ (I K2JDA) No rc ..... erves IR 5 (2431 DA) 5 (243IDA) Kruger NP National Parks Board Fortunately the QDS contains three well-established reserves (e.g. Kruger NP) from which all three target spe­ ManyeJcti Game Rescf'I.'e N.TvJ Environment and cies (i.e. K. ,pekii, C. purda!i.,· (babcocki), and K. natalen­ Tourism sis) have been recorded. However, the IR is marginal for K. Sahie-Sand Game Reserve spekii and protects only the northern race (C. p. babocki) of n (3119RC) Akkcrendam Nat Res. Local authority: Calvinia the mountain tortoise. 7 (2425DH) No reserves IR 6 (3119BC)

8 (2823CC) No reservt'~ The only existing reserve (Akkerendam NR) in the QDS belongs to a local authority. The IR is marginal for both tar­ get species (i.e. H. boulengeri and H. signatus) and, more· Existing protected arcas falling within the same QDS as an IR over, protects only an intergrade popUlation between the are listed in Table 3. two races of H. signatus (Boycott 1986) It is desirable that reserve selection analysis should be IR 7 (2425DB) .

) restricted to localities taken from within the centre of a spc· No existing protected areas occur in the QDS. The IR is 9

0 cies' range. This was not done in the present study. Isolated or well situated for conserving the Kalahari tent tortoises 0

2 peripheral records may not represent viable popUlations, and (Psammobates ocuJifer), but is unsuitahle for K. Joba/siana

d are more likely to represent translocated specimens or inaccu­

e as it occurs at the western extreme of its range. t a racies in distribution data. In addition, marginal populations IR 8 (2823CC) d (

may not protect maximum genetic diversity for the species No existing protected areas occur in the QDS, and the TR is r e concerned. also marginal for the greater padloper (H. femoralis), occur­ h s i With these caveats in mind, how well do the determined ring at the western limit of northern populations. l b IRs (Table 2) protect the species under discussion? u P

IR I (2832AD) What new reserves are required to fully protect e

h southern African chelonians?

t No less than eight protccted areas fall within the IR, and

y the four species targeted for protection by the IRs (i.e. If marginal locality records are not excluded from the iterative b

d Kinix:rs belliwIll (helliana). Pelusios rhodesianus, P casta· reserve selection analysis, the analysis is often biased. This is e t noides, and P sinuatus) are all recorded from at least one of particularly prohlematic with tortoises that show little sympa­ n a these reserves (Table I). The conservation status of these try among congeneric species. As noted earlier, the ranges of r g species in the area is considered good. the three species of Psammobates, four hingeback tortoises e c IR 2 (2716DD) (Kinix}'s spp.), and five Homopus species arc mainly all opaL­ n e

c As the IR is not in South Africa. no statutory protected ric, with only marginal overlap with that of congeners. How­ i l reserves are listed in the data hase as occurring in the QDS. ever, it is these areas (as illustrated by the selection of IRs 5 r e However, the region horders the southern Sperrgebict, a and 6) that arc preferentially selected by the algorithm in its d n wilderness region currently inaccessible owing to mining search to minimize the numher of IRs required to fully protect u

y activites, but that is under consideration for proclamation as members of the group. a Of the eight IRs selected to conserve the 20 species of w a Namibian reserve. Of the three species that the IR is tar­ e t geted to protect [i.e. C. angulata, Psammobates tentorius southern African land tortoises and pclomedusid terrapins, a only the first (QDS 2832AD) can he considered ideally situ­ G (\'erroxii), and Homopus sp. nov ('hergeri')], the population t

e of C. angulata is at least marginal and may even be a trans­ ated to protect all four target species. The remaining seven n i location (Branch 1994). The IR is also marginal for the new IRs only poorly protect their target species. either including b a Homopus sp. (whose main distribution is centred around peripheral popUlations or conserving only limited genetic S

y Aus), and protects only the northern race (P. t. verroxii) for diversity within the species (particularly in polytypic species). b the tent tortoise. In addition, the IRs 2, 4, 7 and 8 have no existing protected d e c u d o r p e R 100 S. Afr. Tydskr. Dierk_ 1995,30(3)

Table 4 Existing major reserves protecting the majority single record is known from the latter, where its presence is of southern African land tortoises and pelomedusid ter­ also marginal. rapins In order to identify a more appropriate set of reserves, we recommend that, (i) all species currently protected, and (ii) all Reserve Species protected (M = marginal) marginal locality records, should be excluded from the data­ Homopusfemoruiis, H. boulengeri, Karoa National Park base. A re-run of the iterative algorithm would then provide Geocheione pardalis, Psammobates t. more appropriate recommendations for the protection of cur­ tenwrius. Pe/omeduS£l .fUbrufu, Chersina anRulaw (M) rently unprotected species. Elandsberg PN R * Psamm()butes xeomefricu,s. Homopus Final comments areo/mus, Cher.tinu unxu[ata St Lucia Game Reserve Kinix.vs bellianu. Pelusios custunoides, P This study has shown that the rich chelonian fauna of south­ rhode.tianus, P .dnuatus ern Africa is relatively well protected in existing reserves. The small number of chelonian species in southern Africa and Kruger National Park Kinixys nata/ensis, K. srek;i (M), Geochelone pardalis, Pelusios sinuarus, P subniger, their relatively well-known distributions test the efficacy of Pe{omedusa subrufa the iterative algorithm developed by Rebelo & Siegfried (1992). Owing to the presence of many allopatric (or nearly Goegap Nature Reserve Humopus .t. signarus, Psammobates tentorius trimen! so) congeneric testudinid species, many of the selected reserves protect only peripheral populations. Seven of the Kalahari Gemsbok NP P.w.mmobates oculirer eight iterative reserves selected by the algorithm were consid­ Moremi Wildlife Reserve Petusius bechuunicus, P subniger ered unsuitable to protect the target species. By their nature *Private nature reserve, but the largest and only viable population of the these may represent relict, non-viable populations in decline, endangered geometric tortoise. owing to either natural or man-induced environmental changes. They may also more readily represent translocations reserves within their borders. Analysis of Table 1 reveals that or mistakes in the recording of distributional data. Peripheral most species are already adequately protected in existing populations may also partially protect genetic diversity within reserves, of which only seven are required to conserve 17 a species, particularly where polytypic species are involved. (B5%) of the region's tortoises and terrapins (Table 4). The This aspect was compounded in the iterative reserve analysis outstanding species, not protected in these reserves, are given which was perfonned only at the species level. We recom­ mend that peripheral populations of all species should be below. excluded from iterative reserve selection procedures. .

) The study revealed that several tortoise species are inade­ Kinirys lobatsiana 9

0 quately protected. These include two Cape endemic padloper

0 The species is currently recorded from only three reserves,

2 species (H. signatus and H. boulengeri) which are amongst

d including Loskop Dam NR, Nylsvley NR, and Ohrigstad NR the smallest extant testudinids and which are subject to e t (Table 1). Twenty-seven of the other QDS recorded for this

a increasing illegal collecting for the international pet trade. It d species include (at least in part) a number of statutory pro­ ( is recommended that consideration be given to proclaiming

r tected areas, of which the most important (over 10000 hal e further reserves in the succulent Karoo biome. h are: Kransberg National Park (2427CB, National Parks s i l Board, NPB) and Blyde River Canyon NR (2430DA, Eastern b Acknowledgements u Transvaal Nature Conservation). The latter is at the extreme P

We thank Dr O. Bourquin (Natal Parks Board), Marius

e east of the species range. h Burger (Eastern Cape Conservation), and Jenni Hensley t

y (Western Cape Conservation) for information relating to che­

b Homopus sp. lonians in reserves in their regions. Wulf Haacke supplied d e A new species restricted to the highlands around Aus, and t locality data for Pelomedusa in the Transvaal Museum collec­ n isolated rocky mountains in the Namib Desert near Luderitz,

a tion, and Gerald Haagner willingly analysed data in the PEM r where the species is currently afforded protection in the g herpetological database. Denise Drinkrow extracted data from

e northern parts of the Sperrgebiet. Proclamation of a reserve in c the SAM database. The SAM collection incorporates the n the Aus vicinity would safely protect this species, as well as e former herpetological collection of Cape Nature Conserva­ c i (P. l the northern race t. verroxii) of the tent tortoise.

tion, which is gratefully acknowledged. Professor R. Klein r e (Stanford), Dr R. Yates (University of Cape Town) and Dr

d Within the Cape, the following two endemic species are

n Simon Hall (University of the Witwatersrand) quickly relatively poorly protected. u answered queries relating to tortoise remains in archaeologi­ y a cal sites. ATL and GAB acknowledge financial support from w Homopus boulengeri e

t the Department of Environmental Affairs and Tourism, and a Currently recorded only from the Karoo NP (NPB) and Karoo the Foundation for Research Development. G NR (Eastern Cape Nature Conservation). t e n

i References b Homopus signatus a AUERBACH, R. D. 1987. The Amphibians and Reptiles of S Currently recorded from the Goegap NR (Northern Cape Botswana, Mokwepa Consultants (privately published), y b

Nature Conservation) and Richtersveld NP (NPB). Only a Gaborone, 295 pp. d e c u d o r p e R S. Afr. 1. ZooJ. 1995,30(3) 101

AUFFENBERG, W. 1974. Checklist of fossil land tortoises South Affica. Ann. Cape Provo Mu,\·. 18: 205-225. (Testudinidae). Bull. Fla St. Mus. 18: 120--251. BRANCH, W.R. 1992. Homopus 'bergeri' - A wrong name for a BAARD. E. H. W. 1990. Biological aspects and conservation status new tortoise from southern Namibia (absL). In: Proc. 2nd HAA of the geometric tortoise. Psammohales geometricus (Linaells, conference. (Eds.) W.R. Branch, G.Y. Haagner and R.C. Boycott. 1758) (Cryptodira: Testudinidae). Ph.D. thesis, University of 1. Herpetol. Assoc. Afr. 40: 11. Stellenbosch. BRANCH, W.R. 1994. Herpetofauna of the Sperrgebeit region of BAARD, E.H.W. 1993a. A conservation strategy for the geometric southern Namibia. Herpetol. Na,t. Hist. 219: i-ii. tortoise, Psammobates geometricu.~, in the south-western Cape BRANCH, WR. & BRAACK, H.H. 1987. The reptiles and Province, South Africa. Internal report 11, Cape Nature amphibians of the Addo Elephant National Park. Koedoe 30: Conservation, Stcllcnbosch. 64 pp. 61-112. BAARD, E.H. W. 1993b. Distribution and status of the geometric BRANCH, WR. & BRAACK. H.H. 1989. Reptiles and Amphibians tortoise, P,wmmobates geometricu.~, in South Africa. BioI. in the Karoo National Park: A surpri sing diversity. In: Proceedings Conse~. 63: 235-239. of thc First H.A.A. Conference, (ed.) W.R. Branch. 1. Herpetof. BATES, M.E 1992. The herpetofauna of the Orange Free State, with ASJ. Afr. 36: 26-35. special emphasis on biogeographical patterning. M.Sc. thesis, BROADLEY, D.G. 1981. A rcvicw of the of the genus Pelusio,~ Univ. Natal, 420 pp. Waglcr in southern Africa (Pleurodira: Pelomedusidae). Occ. pap. BAUER, A.M., BRANCH, W.R. & HAAKE, WD. 1993. The Natl. Mus. Rhod. B. Nat. sci. 6: 633-686. herpetofauna of Kamanjab and adjacent Damaraland, Namibia. BROADLEY, D.G. 1989. Species Status reports (Geochelone Madoqua 18: 117-145. pardalis, pp. 43-44; KinixyJ sp., pp. 49-61). In: The BOUR, R.H. 1983. Trois populations endcmiques du genre PeLusios Conservation Biology of Tortoise, (cds.) 1.R. Swing1and and (Reptilia, Chelonii, Pelomedusidae) auxiles Seychelles; relations M.W. Klemens, Occ. Pap. IUCN Species Survival Commission avec les especes africaines et malagaches. Bull. Mu,~. natn. Hi,~t. (SSC) 5. nat. Paris 4 ser. 5(AI): 34.1-382. BROADLEY, D. G. 1990. ritzSimons' Snakes of southern Africa. BOUR, R.H. 1986. Note sur Pelusios adansonii (Schweigger, 1812) rev. ed. Delta Books, Johanncsburg, 376 pp. + 10 pp. addendum. et sur une nouvelle espece affine du Kenya (Chelonii, BROADLEY, D.G. 1993. A review of the southern African species Pelomedusidae). Stud. Palaeochelon. 2: 23-54. of Kinixy.\· Be1i (Reptilia: Testudinidae). Ann. Transvaal Mus. 36: BOUR, R.H. 1988. Taxonomic and nomenclatural status of 41-52. Homopus signalus (Gmclin 1789): Reptilia: Chelonii. 1. Herpetol. BROADLEY, D.G. 1994. The genus Seelores Fitzinger (Reptilia: Auoc. Afr. 35: 1-6. Scincidae) in Mozambique, Swaziland and Natal. South Africa. BOURQUIN, O. 1990. Presence, distribution, population sizes, Ann. Natal Mu,," 35: 237- 259. survival requirements, threats to and importance values of BROADLEY, D.G. & BLAKE, D.K. 1979. A check list of the chclonids in Natal. Intcrnal report, Natal Parks Board. 50 pp. reptiles of the national parks and other conservation areas of BOURQUIN, O. 1991. A new genus and species of snake from the Zimbabwe Rhodesia. Arnoldia Rhod. 8(35): 1-15. .

) Natal Drakensberg, South Africa. Ann. TransvaaL Mus. 35: BURGER, M. 1993. The herpetofauna of Anysberg Nature Reserve, 9

0 199~203. Cape Province, South Africa. 1. Herpelol. Assoc. Afr. 42: 1-12. 0

2 BOYCOTT, R.C. 1986. A review of Homopus signatus (Schoepff) BURGER, M. & BRANCH, WR. 1994. Tortoise mortality caused

d with notes on related species (Cryptodira: Testudinidae). 1. by electritied fences in the Thomas Baines Nature Reserve, e t Herpetof. Assoc. Afr:. 32: 10-16. Eastern Cape. S. Afr. 1. Wildl. Res. 24: 32-37. a d BOYCOTT, R.c. 1989. Species status reports (Homopw' boulengeri, CARTER, P.L., MITCHELL, P.J. & VINNICOMBE, P. 1988. (

r H. signatus, and Psammobates oeubler). pp 78-79, 82-84, 88-90. Sehonghong: The Middle and Later Stone Age industrial sequence e

h In: The Conservation Biology of Tortoise, (eds.) I.R. Swingland at a Lesotho rock-shelter. BAR International Series 406, Oxford s i l and M.W. Klemens, Occ. Pap. IUCN Species Survival COOPER, M.R. & BROADLEY, D.G. 1990. A new species of fossil b

u Commission (SSC) 5. Homopu.\· (Cryptodira: Testudinidae) from South Africa. Stud. P

BOYCarr, R.C. 1993. A herpetofaunal survey of Swaziland. M.Sc. PaLaeoehelon. 2: 41-55. e

h thesis, Urnv. Natal, 175 pp., 43 figs. DERRICOURT, R.M. 1977. Prehistoric Man in the Ciskei and t

y BOYCarr, R.C. & BOURQUIN, O. 1988. The South African Transkei, C. Struik, Cape Town. 284 pp. b

Tortoise book. A guide to South African tortoises, terrapins and ESTERHUYSEN, A. B., BEHRENS, J. & HARPER, P. 1994. d e turtles. Southern Book Pub!., Johannesburg, 148 pp. Leliehoek Shelter: a Holocene sequence from the eastern Orange t n BRAACK, H.H. 1981. Lower vertebrates of the Bontebok National rree State. S. Afr:. Archaeol. Bull. 49: 73-78. a r Park. Koedoe 24: 67-77. FRAZIER, N.B. 1992. Sca turtlc conservation and halfway g

e BRANCH, W. R. 1984. Preliminary observations on the ecology of technology. Consem Bioi. 6: 1791-84. c FROST, D.R. & HILLIS, D.M. 1990. Species in concept and n the angulate tortoise (Chen'ina angulata) in the Eastern Cape e

c Province, South Africa. Amphibia-Reptilia 5: 43-55. practise: Herpetological applications. Herpetologica 46: R7-104. i l BRANCH, W.R. 1988a. Field guide to the snakes and other reptiles GA~'~NEY, E.S. & MEYLAN, P.A. 1988. A phylogeny of turtles. r e of southern Africa, Struik Publishers, Cape Town, 326 pp. pp. 157-219. In: The Phylogeny and Classification of the d n BRANCH, W.R. (Ed.). 1988b. South African Red Data Book­ Tetrapods. Vol. I., (Ed) M.J. Benton, Clarendon Press, Oxford. u Reptiles and Amphibians. S. Afr. Nat. Sci. Prog. Rep. 151. 241 pp. GREIG, J. C. & BURDETT, P.D. 1976. Patterns in the distribution y a BRANCH, W.R. 1989. Species Status reports (Chersina angulala, of southern African terrestrial tortoises. Zool. Afr. 11: 249-273. w e Homopus areolatus, H. bergeri, H. femora lis, and Psammobates GROBLER,1. H. & BRONKHORST, PJ. 1981. A preliminary t a tentorius). pp 68-77, 91-93. In: The Conservation Biology of checklist of the reptiles and amphibians in the Mountain Zebra G

Tortoise, (cds.) I. R. Swingland and M.W. Klemens, Occ. Pap. National Park. Koedoe 24: 193-197. t e IUCN Species Survival Commission (SSC) 5. HAACKE, WD. 1984. The herpetology of the southern Kalahari n i domain. Koedoe (suppl). 1984. 171 ~ 186. b BRANCH, W.R. 1990. Herpetofauna of the Cape Province, South a Africa: New distribution records and zoogeography. 1. Herpetol. HAAGNER, G.Y. & ELS, R.E. 1986. The reptiles of Mapelane. S

y Ass. Afr. 37: 17-44. Lammergeyer 37: 14-21. b BRANCH, W.R. 1991. Thc herpetofauna of the offshore islands of HALL, S. 1990. Hunter-gatherer-fishers of the Fish River Basin: A d e c u d o r p e R 102 s. Afr. Tydskr. Dierk. 1995, 30(3)

contribution to the Holocene prehir>tory of the Eastern Cape. Ph.D. 77-132. thesis (unpuhL), University of Stellenbosch. MITCHELL. P.J. 1991b. Archaeological investigations at two HEWITT, J. t 933. On the Cape species and subspecies of the genus Lesotho rock-shelters: Terminal Pleistocene! early Holocene Chersinella, Gray. Part I. Ann. Natal Mus. 7: 255-293. assemblages from Ha Makatoko and Ntloana Tsoana. Proc. HEWITT, 1. 1934. On the Cape species and subspecies of the genus Prehis!. Soc. 59: 39-60. Chersinella Gray. Part 2. Ann. Natal Mus. 7: 303-349. MITCHELL, PJ.. PARKINGTON, J.E. & YATES, R. 1994. Recent HOFFMANN, L.A.C 1989. An annotated list of amphibian and Holocene archaeology in western and southern Lesotho. S. Afr. reptile observations from the Etosha National Park. Madoqua 16: Archaeol. Bull. 49: 33-52. 87-92. MOUTON, P. Ie EN. & VAN WYK, 1.H. 1994. Taxonomic status 01 HUGHES, G. R. 1974a. The sea turtles of south-east Africa. I. geographical isolates in the Cordylus minor complex (Reptilia: Status, morphology and distributions. Oceanog. Res. Inst. Invest. Cordylidae).1. Herpetol. AHOc. Afr. 43: 6--18. Rep. 35: 1-144. MULLINS, A. J. 1984. Leatherback turtle (Dermochelys coriacea), HUGHES, G. R. 1974b. The sea turtles of south-cast Africa. II. The breeding on the Bathurst coast. The Naturalist 28: 14. biology of the Tongaiand Loggerhead turtle Caretta caretta L. PIENAAR, U. de Y., HAACKE, W.O. & JACOBSEN, N.H.G. 1983. with comments on the Leatherback turtle Dennochelys coriacea The Reptiles of the Kruger National Park, Nat. Pks. Board, L. and the Green turtle Chelonia mydas L. in the study region. Pretoria. 236 pp. Oceanog. Res. Ins!. Invest. Rep. 36: 1-96. PLUG. I. & ENGELA, R. 1992. The macrofaunal remains from IVERSON, J. 1992. A revised Checklist with Distribution Maps of recent excavations at Rose Cottage Cave, Orange Free State. S. the Turtles of the World, privately pUblished, Richmond. Indiana, AI' Archaeol. Bull. 47: 16-25. 363 pp. POOLEY, A.C. 1965. A preliminary checklist of the reptiles found JACOBSEN, N.H.G. I9!N. A herpetofaunal survey ofthc Transvaal. within the Mkuzi and Ndumu Game Reserves in northern Ph.D. thesis, Univ. Natal. 1621 pp. Zululand. Lamme'Reyer 3: 12-40. JACOBSEN, N.H.G., NEWBERY, R.E. & PETERSEN, WL. 1986. A checklist of the herpetofauna of the Transvaal Provincial Nature POWER, J.H. 1932. Tesludo verreauxii Smith: a study in variation. Rescrves. Transvaal Nature Conservation Division. Pretoria, 38pp. s. Air 1. Sci. 29: 466-472. LOMBARD, AT. 1995. Introduction to an evaluation of the REBELO, A.G. & SIEGFRIED, W.R. 1992. Where should nature protection status of South Africa's vertebrates. S. Afr. 1. Zoot. 10: reserves be located in the Cape floristic region, South Africa? 63-70. Models for the spacial configuration of a reserve network aimed at LOMBARD, AT., NICHOLLS, A.O. & AUGUST, PY 1995. maximising the protection of floral diversity. Conserv. BioI. 6: Where should nature reserves be located in South Africa? A 233-252. snake's perspective. Conserv. Bioi. (in press). SIEGFRIED, W.R. 1989. Preservation of species in southern African LOVERIDGE, A. & WILLIAMS, E.E. 1957. Revision 01 the nature reserves. pp. 186--201. In: Biotic Diversity in southern African tortoises and turtles of the suborder Cryptodira. Bull. Mus. Africa, (ed.) B.J. Huntley, Oxford University Press, Cape Town.

. SMITH, A. 1838-49. Illustrations of the Zoology of South Africa ) compo ZooL. 115: 161-557.

9 MERTENS, R. 1955. Die Amphibien und Reptilien Stidwestafrikas. consisting chiefly of figures and descriptions of the objects of 0

0 Abh senckenb. naturforsch. Ges. 490: 1-172. Natural History collected during an Expedition into the Interior of 2

d MITCHELL, P.J. 1991a. The archaeology ofTloutle rock-shelter. South Africa in the years 1834. 1835 and 1836. Reptilia. London. e t Maseru District, Lesotho. Navor. Nat. Mus. Bloemfontein 9: pis. i-xxviii) pp 1-28. a d ( r e h s i l b u P e h t y b d e t n a r g e c n e c i l r e d n u y a w e t a G t e n i b a S y b d e c u d o r p e R